Predicting the Haemodynamic Response to Prone Positioning: A Novel and Simultaneous Analysis of the Guyton and Rahn Diagrams

Predicting the Haemodynamic Response to Prone Positioning: A Novel and Simultaneous Analysis of the Guyton and Rahn Diagrams

Jon-Emile S. Kenny
Department of Medicine, Icahn School of Medicine at Mount Sinai, New York City
Department of Learning, Informatics, Management and Ethics (LIME), Karolinska Institutet, Stockholm, Sweden
Email: jon-emile@heart-lung.org

Full-text PDF

AbstractFull-TextReference ListCitationVideo

Graphical models of physiology are heuristically appealing as they facilitate qualitative conclusions at the bedside of the critically ill. Historically, the Rahn diagram has portrayed the physiology of the lungs, chest wall and respiratory system, while the Guyton diagram has illustrated cardiovascular physiology. As contemporary methods of haemodynamic monitoring, and their predecessors, are inexorably influenced by the interaction between the respiratory and cardiovascular systems, a qualitative graphical model for mechanical heart-lung interaction holds clinical utility. This short report will present an illustrative clinical case, briefly review the physiological underpinnings of the Rahn and Guyton Diagrams and then introduce a novel combination thereof. It is expected that this new diagram will clarify the case at hand, as well as facilitate the transfer of cardio-respiratory theory to clinical practice.

Keywords:  Prone position; cardio-respiratory interactions; Rahn diagram; Guyton diagram


When Arthur Guyton combined the venous return and cardiac function curves onto a single graph, he created a compelling qualitative construct for physiologists and clinicians alike,. His representation has aided our understanding of cardiovascular physiology for decades but largely neglects the intimate relationship between the cardiovascular system and the thoracic pump. Whereas Guyton correctly predicted that the cardiac function and venous return curves would be altered by intra-thoracic pressure (ITP), his diagram is typically presented independently of the effects of changing ITP and lung volume.

Appreciation for the mechanical relationship between the heart and lungs has been especially prominent since the introduction of the flow-directed pulmonary artery catheter (PAC) in the early 1970s. Following widespread adoption of the PAC, it became apparent that its measured variables are significantly affected by mechanical heart-lung interaction and that the physiological linkage between the respiratory and cardiovascular systems can undermine both acquisition and interpretation of haemodynamic data,. While attempts were made to correct these deficiencies,, the use of the PAC has waned significantly to the point where it was recently eulogized.

Newer, non-invasive methods of haemodynamic monitoring have largely taken the place of the PAC in modern intensive care units (ICUs). Simultaneously, functional haemodynamic monitoring (FHM) is gaining traction amongst intensivists. FHM assays the functional state of the cardiovascular system by measuring a haemodynamic response (e.g. SVC calibre change) to a defined stressor (e.g. a change in ITP). Nevertheless, these contemporary means of haemodynamic assessment still obey fundamental principles of cardiorespiratory physiology. Thus, comprehension of the immediate interdependence of the heart and lungs within the confines of the thorax continues to be of paramount importance.

Illustrative Case

A 70-year-old woman with a body mass index (BMI) of 39, severe chronic bronchitis and profound aspiration pneumonia complicated by acute respiratory distress syndrome (ARDS), is becoming progressively more hypoxaemic. Her partial pressure of oxygen to fraction of inspired oxygen (P/F) ratio is less than 100 mmHg. She is heavily sedated and pharmacologically paralyzed. Her plateau pressure (Pplat) is 37 cm H20 and her oesophageal pressure (Poes) is 17 cm H2O. Her mean arterial pressure (MAP) is 60 mmHg and her urine output has decreased during the last hour. Transoesophageal echocardiography (TOE) is carried out to help predict fluid responsiveness. In the supine position, her superior vena cava (SVC) is noted to be engorged and unvarying with mechanical inspiration. A decision is made to place her in the prone position. Following this, her Pplat remains 37 cm H2O, her Poes falls to 15 cm H2O and P/F ratio rises to 150 mmHg. Importantly, her MAP increases slightly and her SVC now collapses with mechanical inspiration. Noting the change in SVC calibre, 500 ml of balanced crystalloid is infused; the patient’s MAP increases further and urine output normalizes.

Clinical Overview

Turning a paralysed patient into the prone position may change both the pulmonary and chest wall compliances. The combined compliance of the lungs and the chest wall, the respiratory system compliance (Crs), determines the end-inspiratory alveolar pressure, commonly assessed on the ventilator as the plateau pressure. Assuming that the plateau pressure approximates end-inspiratory alveolar pressure (PAlv) and that the oesophageal pressure estimates intra-thoracic (or pleural) pressure, the pressure across the alveolus, called the trans-pulmonary pressure (Ptp), can be assessed at the bedside as the difference between these i.e. Pplat – Poes. Consequently, application of the prone position in this patient resulted in an apparent fall in intra-thoracic (pleural) pressure (decreased Poes) with a contemporaneous rise in the trans-pulmonary pressure as the difference between Pplat and Poes increased.

Ptp (supine) = Pplat – Poes = [37-17] = 20 cm H2O

Ptp (prone) = Pplat – Poes = [37-15] = 22 cm H2O

However, it is important to note that the oesophageal pressure in the supine position tends to over-estimate the intrathoracic (pleural) pressure due to mediastinal compression of the oesophagus. In healthy individuals, the compression-induced rise in oesophageal pressure may be as much as 40%. It is also likely that in the critically ill ARDS patient the dense, oedematous lung and cardiomegaly from volume infusion may compress and increase the oesophageal pressure by an even greater degree. Lastly, it has been known for decades that the oesophageal pressure in the upright, lateral and prone positions are all similar in value, reinforcing the impact of the mediastinum on the oesophageal pressure in the supine position. It is, therefore, reasonable to correct the oesophageal pressure measured in the supine position to more accurately reflect intra-thoracic pressure. A reduction in the measured supine Poes of approximately 40%, based on the data described above, will give a new ’corrected’ supine oesophageal pressure in this patient of 10 cm H2O i.e. an intra-thoracic (pleural) pressure (Ppl) of 10 cm H2O. Using this corrected intra-thoracic (pleural) pressure we can insert the ’new’ value for supine trans-pulmonary pressure (Ptp) as follows:

Ptp (supine) = Pplat – Poes = [37-10] = 27 cm H2O

Ptp (prone) = Pplat – Poes = [37-15] = 22 cm H2O

Taking the aforementioned corrections into consideration, on assuming the prone position the respective haemodynamic impacts upon the right ventricle (RV) are to:

  1. diminish its venous return – as the pressure within the thorax rises relative to the body (i.e. the intra-thoracic pressure Ppl increased from 10 cm H2O to 15 cm H2O),
  2. augment RV forward flow – as the fall in trans-pulmonary pressure (from 27 cm H2O to 22 cm H2O) reduces RV afterload.

In aggregate, therefore, the cardiac output is slightly increased and the RV becomes more preload responsive.

The Physiology and Novel Qualitative Diagram

The Rahn Diagram

A qualitative analysis of respiratory mechanics can be gleaned from the Rahn diagram, which simultaneously depicts the compliances of the passive chest wall, lungs and respiratory system (see Figure 1).

Figure 1. The Rahn diagram. The x-axis is airway pressure and the y-axis is the lung volume. The pulmonary and chest wall compliance curves (Cpulm and Ccw respectively) are depicted in dashed navy lines. The Crs curve in sky blue represents the summative compliance of the respiratory system.

Importantly, in the patient fully passive with the ventilator, the intra-thoracic pressure (or pleural pressure) follows the chest wall compliance curve as this pressure is the surface pressure generated between the lungs and the effortless chest wall. Clinically, this pressure may be estimated by an oesophageal balloon (i.e. the Poes).

Figure 2. The patient in the supine position. The patient’s obesity has caused shift of the chest wall compliance curve to the right. ARDS decreases pulmonary compliance reducing the slope of the Cpulm curve. The combined effect is a reduced Crs. For a given ventilator volume, the plateau pressure Pplat can be estimated from the Crs curve and the pleural pressure can be estimated from the Ccw curve, allowing the trans-pulmonary pressure (red dotted line) to be calculated (27 cm H2O).

Considering the patient while supine (see Figure 2), a given ventilator-delivered volume on the y-axis may be parsed into the plateau pressure (from the respiratory system compliance curve), and the pleural pressure (from the obese chest wall compliance curve); the lateral distance between the plateau pressure (Pplat) and the pleural pressure (Ppl) qualitatively approximates the transpulmonary pressure (Ptp) (dashed red line). When a patient is moved from the supine position to the prone position, the following changes are expected on the Rahn diagram (see Figure 3).

Figure 3. The patient in prone position. Chest wall compliance has fallen further resulting in reduced slope of the Ccw curve. Due to lung recruitment, the slope of the Cpulm has risen. Overall the Crs curve has not changed significantly. For the same ventilator volume, the Pplat (37 cm H2O) remains the same but the pleural pressure Ppl has risen to 15 cm H2O. The trans-pulmonary pressure (dashed red line) has therefore fallen and is now 22 cm H2O.

Firstly, dorsal lung spaces are recruited and apparent pulmonary compliance increases (i.e. an increased slope of the Cpulm curve). Secondly, the chest wall is stiffened on moving to the prone position so its slope falls. Since pulmonary compliance rises and chest wall compliance falls, the combined respiratory system compliance does not vary (compare Figures 2 & 3). However, the trans-pulmonary pressure falls because the pleural pressure rises relative to the plateau pressure. As elaborated below, the haemodynamic consequence of an increase in pleural pressure relative to plateau pressure is a reduction in both RV preload and afterload.

The Guyton Diagram

The venous return curve models blood flow into the thorax. The x-intercept of the venous return curve depicts the mean systemic filling pressure (Pms) which is the equilibrium pressure the circulatory system assumes when blood flow ceases. The mean systemic filling pressure is, essentially, determined by venous blood volume (i.e. ‘volume status’) and venous tone; it is the upstream pressure for blood flow towards the thorax. When the venous return curve is superimposed on cardiac function curve, the Guyton diagram is formed. Importantly, the relationship between the RV cardiac function curve and the SVC venous return curve illustrates the concepts of SVC collapse and fluid responsiveness. An inspiratory augmentation of pleural pressure shifts the cardiac function curve rightwards with respect to the venous return curve, because the pressure within the thorax rises relative to the extra-thoracic venous pressure. When the venous return curve intersects the flat portion of the RV cardiac function curve, that is, when a patient is not fluid responsive, the lateral distance between right atrial pressure (Pra, an estimation of the SVC pressure) and the intra-thoracic pressure (the x-intercept of the cardiac function curve) shrinks minimally. The absence of SVC collapse indicates that the patient is on the flat portion of her SVC transmural pressure–volume compliance curve; a marker that the patient is not fluid responsive (see Figure 4).

Figure 4. The Guyton diagram for a patient who is not fluid-responsive. The venous return curve intersects the plateau of the cardiac function curve. Cardiac output is on the y-axis and right atrial pressure (or SVC pressure) is on the x-axis. The intra-thoracic pressure (or pleural pressure), is the x-intercept of the cardiac function curve and shifts rightwards in response to a passive mechanical breath. The SVC transmural pressure is indicated at end-inspiration. The SVC  transmural pressure is the pressure within the SVC less the intra-thoracic pressure (x-intercept of the cardiac function curve). This graphic is analogous to the patient in the supine position, that is, a mechanical breath results in a relatively small increase in Poes, but a large increase in Ptp and RV afterload. Pms is mean systemic pressure and the x-intercept of the venous return curve.

Conversely, when the venous return curve intersects the ascending portion of the RV cardiac function curve, the patient is fluid responsive and an inspiratory augmentation of intra-thoracic pressure causes the distending pressure of the SVC to fall greatly. This collapse is an echocardiographic marker of fluid responsiveness (see Figure 5).

Figure 5. The Guyton diagram for a fluid-responsive patient. The venous return curve intersects the ascending portion of the cardiac function curve, such that provision of fluids will raise cardiac output. The transmural SVC pressure is much smaller, meaning that its inspiratory fall will lead to collapse. This graphic is analogous to the patient in the prone position, that is, a mechanical breath results in a relatively large increase in Poes, but a decrease in Ptp and, therefore, RV afterload. Again, this graphic represents end-inspiration.

Simultaneous Analysis

How does the Guyton analysis relate to the Rahn analysis and how does this inform the aforementioned clinical scenario? The Rahn diagram, described above, may be altered such that ventilator volume is moved to the z-axis (effectively directed ’into the page’); pressure remains on the x-axis. The superimposed Guyton diagram may share pressure on the x-axis while blood flow (cardiac output or venous return) remains on the y-axis. This composite diagram allows for simultaneous analysis of the ventilator-applied volume, Ppl, Pplat, Ptp, cardiac output and SVC transmural pressure.

Figure 6. The simultaneous Rahn and Guyton analysis of the supine patient. This graphic represents the system at end-inspiration. The Rahn diagram in the z-axis matches Figure 2. Cpulm is omitted for simplicity. Pplat, Ppl and Ptp (dashed red) are depicted. The RV function curve x-intercept shifts in-step with the Ppl along the Ccw and parallels Figure 4. Because there is a large Ptp (RV afterload) without much fall in preload, the SVC trans-mural pressure remains high and the RV is fluid intolerant.

Figure 6 represents the patient while supine, at end inspiration. The primary insult is the excessive stress and strain placed upon the ‘baby lung’ (see also Figure 2) such that Pplat increases much more than Ppl; Ptp therefore rises. Because RV afterload is partly determined by the distending pressure across the alveolus , a large Ptp retards RV ejection and the slope of the RV function curve falls. Additionally, the rightward shift of the cardiac function curve relative to the venous return curve follows the rise in pleural pressure3. While supine, there is a comparatively small inspiratory augmentation of Ppl as compared to the stiffened chest wall when prone. In totality, when the patient is supine, mechanical inspiration drives high RV afterload with minimal reduction in RV preload such that the SVC is non-collapsible, the venous return curve continues to intersect the plateau of the RV function curve and the patient is fluid intolerant.

Figure 7. The simultaneous Rahn and Guyton Analysis of the prone patient. This graphic represents the system at end-inspiration. The Rahn diagram in the z-axis matches Figure 3. Cpulm (omitted for simplicity) rises, while Ccw falls, such that the overall respiratory compliance (Crs) curve remains unchanged. At the same ventilator volume, the Ppl increases to a greater degree and the Ptp (dashed red line) is reduced. The reduction in Ptp raises the slope of the RV cardiac function curve. Simultaneously, the RV function curve x-intercept rises in-step with the Poes such that preload is significantly diminished; the Guyton diagram parallels Figure 5. Because both RV afterload and preload are reduced, cardiac output, on the y-axis, is essentially unchanged or may even rise. The SVC transmural pressure falls and the patient is now fluid responsive.

Figure 7, by contrast, depicts the changes that occur in the prone position. Firstly, the chest wall compliance falls, while the apparent pulmonary compliance improves . Accordingly, the Pplat – derived from the respiratory system compliance curve – changes very little to the extent that the fall in the chest wall compliance is counterbalanced by the increase in lung recruitment. In contrast to the supine position, the Ppl rises notably relative to the Pplat and so the Ptp shrinks. A fall in the Ptp diminishes RV afterload and therefore steepens the slope of the RV cardiac function curve .

Additionally, the fall in chest wall compliance augments Ppl which prominently shifts the RV cardiac function curve relative to the SVC venous return curve. This magnification of end-inspiratory pressure within the thorax – relative to the extra-thoracic veins – diminishes venous return and potentially impairs stroke volume and hence cardiac output. However, this is coupled with a fall in RV afterload, and this fall may be sufficient to offset the impact of reduced venous return, and for stroke volume and cardiac output (i.e. the y-axis) to rise. Notably, the intersection of the cardiac function and venous return curves at their sloped portions shrinks the SVC transmural pressure. Therefore, on TOE, there is inspiratory collapse and provision of fluid will raise cardiac output further. The impact of these changes in summarised in Figure 8.

Figure 8. Summary of the respiratory and circulatory effects of the prone position.

Pathophysiology of the Case

How does the prone position alter respiratory mechanics in this patient?

Assuming the prone position caused the patient’s chest wall compliance to fall (i.e. her chest wall became ’stiffer’). This increased the intra-thoracic (pleural) pressure present for any given ventilator-delivered inspiratory volume. However, if lung is recruited in the prone position then lung compliance will rise (i.e. the lungs will become less stiff). Because the plateau pressure on the ventilator is the summation of the lung and chest wall in series, the plateau pressure may not change if the increase in lung compliance offsets the fall in chest wall compliance. If the plateau pressure stays the same, but the intra-thoracic pressure rises in the prone position then the net effect is a fall in trans-pulmonary pressure.


Importantly, this analysis neglects (for diagrammatic simplicity) changes in the venous return curve in response to both prone position and cyclical mechanical ventilation. In the prone position, pressurization of the abdomen may increase the mean systemic pressure (Pms). This is the pressure head for venous return to the RV and is the x-intercept of the venous return curve. Additionally, the patient’s underlying volume status may alter resistance to venous return. Increased intra-thoracic pressure favours collapse of the great veins on entering the thorax, that is,  vascular waterfall physiology. As demonstrated in the IVC, hypervolaemia retards great-vein collapse while the converse is true when the abdomen’s venous beds are relatively under-filled.


In conclusion, the sedated and paralysed patient placed into the prone position will have changes in chest wall and pulmonary compliance. These changes favour increased intra-thoracic pressure and diminished trans-pulmonary pressure. Consequent reduction in both RV preload and afterload renders the RV fluid-responsive and maintains the cardiac output respectively. Finally, the  aforementioned principles are illustrated by a novel combination of two classic, graphical analyses, emphasising the mechanical linkage of the cardiovascular and respiratory systems for the clinical


I would like to thank Drs. Carla Canepa and Lina Miyakawa for their revisions of this manuscript.


  1. Guyton AC. Determination of cardiac output by equating venous return curves with cardiac response curves. Physiol Rev. 1955;35(1):123–9.
  2. Magder S. Bench-to-bedside review: An approach to hemodynamic monitoring–Guyton at the bedside. Crit Care. 2012;16(5):236. doi:10.1186/cc11395.
  3. Feihl F, Broccard AF. Interactions between respiration and systemic hemodynamics. Part I: basic concepts. Intensive Care Med. 2009;35(1):45–54. doi:10.1007/s00134-008-1297-z.
  4. Tuman KJ, Carroll GC, Ivankovich AD. Pitfalls in interpretation of pulmonary artery catheter data. J Cardiothorac Anesth. 1989;3(5):625–41.
  5. Pinsky MR, Vincent JL. Let us use the pulmonary artery catheter correctly and only when we need it. Crit Care Med. 2005;33(5):1119–22. doi:10.1097/01.CCM.0000163238.64905.56.
  6. Hoyt JD, Leatherman JW. Interpretation of the pulmonary artery occlusion pressure in mechanically ventilated patients with large respiratory excursions in intrathoracic pressure. Intensive Care Med. 1997;23(11):1125–31.
  7. Teboul JL, Pinsky MR, Mercat A, Anguel N, Bernardin G, Achard JM, et al. Estimating cardiac filling pressure in mechanically ventilated patients with hyperinflation. Crit Care Med. 2000;28(11): 631–6.
  8. Marik PE. Obituary: pulmonary artery catheter 1970 to 2013. Ann Intensive Care. 2013;3(1):38. doi:10.1186/2110-5820-3-38.
  9. Marik PE, Monnet X, Teboul JL. Hemodynamic parameters to guide fluid therapy. Ann Intensive Care. 2011;1(1):1. doi:10.1186/2110-5820-1-1.
  10. Pinsky MR, Payen D. Functional hemodynamic monitoring. Crit Care. 2005;9(6):566–72. doi:10.1186/cc3927.
  11. Pinsky MR. Functional haemodynamic monitoring. Curr Opin Crit Care. 2014;20(3):288–93. doi:10.1097/MCC.0000000000000090.
  12. Washko GR, O’Donnell CR, Loring SH. Volume-related and volume-independent effects of posture on esophageal and transpulmonary pressures in healthy subjects. J Appl Physiol.  2006;100(3):753–8. doi:10.1152/japplphysiol.00697.2005.
  13. Talmor DS, Fessler HE. Are esophageal pressure measurements important in clinical decision-making in mechanically ventilated patients? Respir Care. 2010;55(2):162–72; discussion 172–4.
  14. Ferris BG, Mead J, Frank NR. Effect of body position on esophageal pressure and measurement of pulmonary compliance. J Appl Physiol. 1959;14(4).
  15. Rahn H, Otis AB. The pressure-volume diagram of the thorax and lung. Am J Physiol. 1946;146(2):161–78.
  16. Gattinoni L, Chiumello D, Carlesso E, Valenza F. Bench-to-bedside review: chest wall elastance in acute lung injury/acute respiratory distress syndrome patients. Crit Care. 2004;8(5):350–5. doi:10.1186/cc2854.
  17. Talmor D, Sarge T, Malhotra A, O’Donnell CR, Ritz R, Lisbon A, et al. Mechanical ventilation guided by esophageal pressure in acute lung injury. N Engl J Med. 2008;359(20):2095–104. doi:10.1056/NEJMoa0708638.
  18. Behazin N, Jones SB, Cohen RI, Loring SH. Respiratory restriction and elevated pleural and esophageal pressures in morbid obesity. J Appl Physiol. 2010;108(1):212–8.  doi:10.1152/japplphysiol.91356.2008.
  19. Jansen JRC, Maas JJ, Pinsky MR. Bedside assessment of mean systemic filling pressure. Curr Opin Crit Care. 2010;16(3):231–6. doi:10.1097/MCC.0b013e3283378185.
  20. Bodson L, Vieillard-Baron A. Respiratory variation in inferior vena cava diameter: surrogate of central venous pressure or parameter of fluid responsiveness? Let the physiology reply. Crit Care. 2012;16(6):181. doi:10.1186/cc11824.
  21. Vieillard-Baron A, Chergui K, Rabiller A, Peyrouset O, Page B, Beauchet A, et al. Superior vena caval collapsibility as a gauge of volume status in ventilated septic patients. Intensive Care Med. 2004;30(9):1734–9. doi:10.1007/s00134-004-2361-y.
  22. Gattinoni L, Pesenti A. The concept of “baby lung”. Intensive Care Med. 2005;31(6):776–784. doi:10.1007/s00134-005-2627-z.
  23. Vieillard-Baron A, Matthay M, Teboul JL, Bein T, Schultz M, Magder S, et al. Experts’ opinion on management of hemodynamics in ARDS patients: focus on the effects of mechanical ventilation. Intensive Care Med. 2016;42(5):739–49. doi:10.1007/s00134-016-4326-3.
  24. Gattinoni L, Taccone P, Carlesso E, Marini JJ. Prone position in acute respiratory distress syndrome. Rationale, indications, and limits. Am J Respir Crit Care Med. 2013;188(11):1286–93. doi:10.1164/rccm.201308-1532CI.
  25. Repessé X, Charron C, Vieillard-Baron A. Acute respiratory distress syndrome: the heart side of the moon. Curr Opin Crit Care. 2016;22(1):38–44. doi:10.1097/MCC.0000000000000267.
  26. Pinsky MR. Determinants of pulmonary arterial flow variation during respiration. J Appl Physiol. 1984;56(5):1237–45.
  27. Jozwiak M, Teboul JL, Anguel N, Persichini R, Silva S, Chemla D, et al. Beneficial Hemodynamic Effects of Prone Positioning in Patients with Acute Respiratory Distress Syndrome. Am J Respir Crit Care Med. 2013;188(12):1428–1433. doi:10.1164/rccm.201303-0593OC.
  28. Fessler. Effects of CPAP on venous return. J Sleep Res. 1995;4(S1):44–49. doi:10.1111/j.1365-2869.1995.tb00185.x.
  29. Robotham, Takata. Mechanical abdomino/heart/lung interaction. J Sleep Res. 1995;4(S1):50–52. doi:10.1111/j.1365-2869.1995.tb00186.x

Cite this article as follows:

Kenny J-E. Predicting the Haemodynamic Response to Prone Positioning: A Novel and Simultaneous Analysis of the Guyton and Rahn Diagrams. Critical Care Horizons 2017:1-7.

Icatibant and its Therapeutic Indications

Icatibant and its Therapeutic Indications

Christian Longley
Emergency Registrar, St Vincent’s Emergency Department, Melbourne, Australia
Email: christian.longley@googlemail.com

Full-Text PDF

AbstractFull-TextReference ListCitation

Angioedema is a localised dermal and subcutaneous swelling. There are multiple causes for angioedema, including allergic reactions, angiotensin-converting enzyme inhibitor (ACE inhibitor) use and hereditary angioedema (HAE). ACE inhibitor angioedema and HAE are both mediated by an excess of bradykinin, while angioedema secondary to an allergic reaction is mediated by histamine. In these instances, oropharyngeal swelling may occur unpredictably, requiring immediate treatment to prevent airway obstruction. Recent randomised controlled trial evidence suggests early use of 30 mg subcutaneous icatibant, a synthetic bradykinin B2 receptor antagonist, is an effective treatment in HAE. Its potential use in ACE inhibitor angioedema is recognised, but, as yet, less well evidenced. Additionally, its superiority to other therapies is also unclear. This article examines the use of icatibant for hereditary angioedema and ACE inhibitor angioedema.

Keywords: Icatibant; angiotensin-converting enzyme inhibitors; hereditary angioedema; angioedema


In recent years icatibant (Firazyr®) has been increasingly recognised as a therapeutic option for HAE. Interest has also developed in its use for ACE inhibitor angioedema, as it shares a common pathway mediated by bradykinin. The financial cost of this medication is high and its efficacy, like other drugs in this area, is not well studied. Initial trials and case series were encouraging, leading to its licensing for use in HAE attacks in adults over eighteen in the United States and European Union. In the emergency setting it is hoped it may be an excellent addition to the clinician’s armament when faced with cases of angioedema. Other benefits in this setting may also include reduced duration of symptoms and reduced hospital length of stay (LOS). With various novel treatments becoming available, and difficult management decisions facing clinicians who deal with angioedema, it remains a challenging condition to treat. This article will examine icatibant within the larger context of angioedema.


Angioedema is a localised and self-limiting oedema involving deeper layers of the dermis, including the subcutaneous tissue and occasionally the bowel wall. It occurs due to increased vascular permeability caused by release of vasoactive mediators. The swelling affects a variety of areas of the body and presents the most serious threat to life when oedema occurs in the oropharynx (lips, tongue) and larynx. It may occur as a distinct entity, or in addition to urticaria, which only involves the superficial portion of the dermis, and is usually seen in histaminergic mediated angioedema. The episode is considered acute if it lasts less than six weeks, and can occur at any time of life.

The phenotype of this condition is largely dictated by its underlying cause, with common pathways at play (Table 1).

FeaturesAssociated urticaria
Swelling typically resolves in 24-48 hours
No urticaria
Swelling onset is varied but typically quick (hours)
May have a family history or previous bouts of abdominal pain
Related to menstruation
Resolves in 2-7 days
Poor/no response to traditional therapy aimed at histamine-mediated aetiology (steroids/antihitamine)
MechanismImmunoglobulin-mediated IgE antibodiesBradykinin excess due to decreased metabolism (ACE inhibitor)
Abnormal C1-inhibitor gene results in overproduction of bradykinin.
Triggers for attacks include surgery and stress
CausesAllergic – drugs, food, latex, insect bites
HAE (type I and II)
ACE inhibitors
Acquired C1 esterase inhibitor deficiency
Invasive airway management
H1 and H2 antagonists
Invasive airway management
Discontinue ACE inhibitor
Options – Fresh frozen plasma, ecallantide, plasma-derived C1-Inhibitor
concentrate, Icatibant, Recombinant C1-Inhibitor
Table 1. Comparison of of histamine- and bradykinin-mediated angioedema (Adapted from Moellman et al [3])


In 2014, Moellman et al. produced a consensus document for management in the emergency department, classifying angioedema as anaphylaxis, histaminergic angioedema without anaphylaxis and non-histaminergic angioedema. The best-known mediators are histamine and bradykinin. This simple distinction in pathogenesis is of key clinical importance when treating acute attacks of angioedema. Clinical endpoints are similar, but emergency treatment takes one of two distinct pathways. Angioedema can be idiopathic, but recognised causes are multiple, including: allergy and anaphylaxis, HAE, acquired C1 esterase inhibitor deficiency, adverse reaction to ACE inhibitors and some viral, bacterial and parasitic infections (especially in children).

If the cause is known, treatment should be tailored to the individual mechanism. For instance, in anaphylaxis where the underlying cause for angioedema is histamine release, immediate intramuscular or intravenous adrenaline is the mainstay of treatment. If an immune-mediated IgE response triggers a lesser clinical manifestation, steroids and antihistamines can be given. However, if HAE or ACE inhibitors are known to be responsible, or adrenaline fails to improve the situation in angioedema of unknown cause, excess circulating bradykinin is likely to be the causative factor and the medication icatibant is amongst treatment options. Several consensus guidelines suggest adrenaline, corticosteroids and antihistamines have no effect on HAE or ACE inhibitor angioedema,,. Several drugs have been considered in the past decade for the treatment of acute HAE, including ecallantide (a kallikrein inhibitor), icatibant and C1-inhibitor concentrate. Due to its similar bradykinin pathway, interest has also developed in whether similar medications will work in ACE inhibitor angioedema.

Bradykinin Pathway

Bradykinin is a vasoactive peptide formed at the endpoint of kallikrein-kinin system, and acts as the primary mediator in non-allergic angioedema.   Factor XIIa from the coagulation cascade converts prekallikrein to kallikrein.  Subsequently, kallikrein converts high-molecular-weight kininogen to produce bradykinin.  Usually bradykinin concentration is very low, with a half-life less than thirty seconds.  Genes for two bradykinin receptors (B1, B2) are found on chromosome 14.  When bradykinin activates endothelial cells via B2 receptors, vascular permeability increases and oedema results.

Hereditary Angioedema

Hereditary angioedema is a rare autosomal dominant condition usually due to a quantitative or functional deficiency of C1 esterase inhibitor (C1-INH).  Low levels of C1-INH result in uninhibited action of vasoactive mediators, especially bradykinin, causing vasodilation, smooth muscle contraction, submucosal and subcutaneous oedema without urticaria. The underlying cause is a mutation of the gene encoding the C1-INH, inherited in an autosomal pattern with high penetrance. Recent estimates place its prevalence at 1:50,000. Three subtypes are defined:

  • type I, due to low levels of C1-INH (85%),
  • type II, due to normal C1-INH levels, but abnormal C1-INH function (15%), and
  • type III, an extremely rare, poorly understood HAE, involving normal C1-INH levels.

Due to its autosomal dominant nature, a family history is expected; however, de novo mutations are found in 25% cases. Clinical expression may differ significantly, even in members of the same family. Attacks can result in gastrointestinal, subcutaneous or oropharyngeal swelling. Gastrointestinal attacks range from mild to severe and usually result in recurrent pain, generally resolving without serious complication. Diagnosis can be difficult in this instance and patients may undergo multiple medical investigations including spectroscopy before diagnosis. Cutaneous attacks are clinically more apparent but also not usually associated with serious complication. However, when the attack involves the larynx, it poses a definite risk of asphyxiation and death. Fortunately, most attacks present with mild symptoms, which develop over twenty-four hours or so and resolve over two to five days. Signs and symptoms may include tingling followed by a non-itchy macular rash, abdominal pain, diarrhoea, vomiting and swelling of the limbs, face, or genitals. If cases of HAE are suspected, blood testing may confirm diagnosis. Occasionally, genetic testing may be required.

Treatment of acute attacks depends on clinical presentation. There has been significant recent evolution of the pharmacological treatment for acute attacks and longer-term prophylaxis (Table 2).

Medication (trade name)RouteDoseMechanismApproval statusCost/dose (US$)
C1-Inhibitor concentrate (plasma-derived)
(Berinert®, Cinryze®)
Intravenous20 units/kg or
1000 units for Cinryze®
C1-INH protein replacementBerinert® is approved in Europe and USA for all HAE attacks in adults and adolescents.
Cinryze® is approved for all attacks in Europe.
Berinert® $6,377
Cinryze® $4,906
Ecallantide (Kalbitor®)Subcutaneous30 mgPlasma-kallikrein inhibitorApproved in USA for all acute facial and abdominal attacks in patients >16 years old.
Not approved in Europe
Icatibant (Firazyr®)Subcutaneous30 mgBradykinin 2-receptor antagonistApproved in Europe and USA for all acute attacks in adults$8,004
Recombinant C1-Inhibitor (Ruconest®)Intravenous50 units/kgC1-INH protein replacementApproved in Europe for all acute attacks in adults.
Approved in USA for acute attacks of HAE in adult and adolescent patients.
Table 2. Current Treatment Options for Angioedema [15]


Therapies aim to provide C1 inhibitor protein replacement or to antagonise bradykinin or kallikrein (Figure 1).


Figure 1.Reprinted with permission from EB Medicine, publisher of Emergency Medicine Practice, from: R. Gentry Wilkerson, MD, FACEP, FAAEM. Angioedema in the Emergency Department: An Evidence-Based Review. Emergency Medicine Practice. 2012;14(11):3. © 2012. EB Medicine. www.ebmedicine.net

Depending on a specific patient’s pattern of attacks, they may be able to initiate treatment themselves at home.  However, patients who suffer a laryngeal attack are always advised to seek immediate medical assistance.  In cases of airway compromise, endotracheal intubation, or even a surgical airway, may be required.  It is hoped that novel agents may prevent this critical intervention in some circumstances and reduce symptom duration.  Symptoms of dysphonia, dysphagia, stridor and globus pharyngeus must be taken very seriously as airway obstruction has been reported as early as twenty minutes from the onset of symptoms.

Angiotensin-Converting Enzyme Inhibitor Angioedema 

ACE inhibitors are commonly prescribed medications used to treat congestive cardiac failure, diabetic nephropathy and hypertension. ACE inhibitors prevent the conversion of angiotensin I to angiotensin II. The angiotensin-converting enzyme is also the principal enzyme responsible for the breakdown of bradykinin. Several side-effects of ACE inhibitors are recognised, including cough and rash. Angioedema is a serious complication of their use, usually occurring within the first month of treatment but may occur years later,. Why some people are affected in this way is not fully understood, but may be related to genetic variations producing altered function in other enzymes that also help metabolise bradykinin. ACE inhibitor use results in excess circulating bradykinin levels. Patients with decreased ability to metabolise bradykinin, via one of the other redundant mechanisms, are at increased risk of developing angioedema. Angioedema is estimated to occur in 0.1-0.7% of patients taking ACE inhibitors, classically presenting with facial, tongue or lip swelling, in the absence of urticaria,. ACE inhibitor angioedema can be severe, and with limited treatment options, fatal cases have been reported. As ACE inhibitors are such a widely prescribed medication in cardiovascular and renal disease, ACE inhibitor angioedema is a relatively common cause of angioedema presenting the emergency department. In a retrospective review at a large community hospital over a five-year period, 11% of patients who presented with ACE inhibitor angioedema required definitive airway intervention (5/45), with 2% requiring a surgical airway (1/45). Another study found 30% (15/50) required ventilation, with 4% requiring a surgical airway (2/50). With angioedema being so unpredictable, the decision of when and how to secure the airway in the cases of ACE inhibitor angioedema is challenging.

The standard treatment of ACE inhibitor angioedema is to secure the airway when clinically indicated and provide supportive management whilst stopping the causative medication.  Unfortunately, there are no approved medical treatments for ACE inhibitor angioedema and no widely accepted consensus algorithm exists. Again, cases will not respond to steroids, antihistamines or adrenaline, as ACE inhibitor angioedema is bradykinin mediated, as opposed to histaminergic.  This bradykinin pathway, shared with HAE, has stimulated interest in the use of icatibant.  Evidence currently stems from isolated case reports and case series,,,,,.  Similar interest has been shown in the potential of ecallantide and CI-INH concentrates.  Improvement of symptoms following the cessation of the ACE inhibitor is usually enough to confirm diagnosis.

Icatibant (Firazyr ®)

Icatibant (Firazyr®, Shire) is a synthetic decapeptide similar in structure to bradykinin and acts as a competitive antagonist at the bradykinin (B2) receptor (Figure 2). 

Figure 2. Mechanism of Icatibant blockage of bradykinin pathway (adaptation of image courtesy of Professor Simon Carley)

Figure 2. Mechanism of Icatibant blockage of bradykinin pathway (adaptation of image courtesy of Professor Simon Carley)

In Europe, Icatibant was approved by the Committee for Medicinal Products for Human Use (CHMP) in 2008 for angioedema in adults linked to C1 esterase inhibitor deficiency. The Therapeutic Goods Association (TGA) approved its use for symptomatic attacks of HAE in adults in Australia in 2010 and it has been FDA (U.S. Food and Drug Administration) approved in the United States since 2011 for treatment of acute attacks of HAE in people aged eighteen years and older.

Product information and characteristics of icatibant are detailed in material produced by the pharmaceutical company Shire. Icatibant is available as a fixed dose preparation, in a 3 ml (30 mg) pre-filled syringe. It is administered as a slow subcutaneous injection into the abdominal area. A single injection is currently thought to be sufficient to reduce oedema, but if needed, a second and third injection can be given at six-hour intervals to a maximum of three doses in a twenty-four hour period. The most commonly reported adverse effects include local injection site skin reactions and pain. Its bioavailability is 97% and time to maximum concentration is approximately 30 minutes. Ninety percent is metabolised by enzymes to inactive metabolites and excreted via the kidneys whilst 10% is excreted unchanged. It is not metabolised by the cytochrome P450 system and no dose adjustment is needed in hepatic or renal impairment. There is relatively little experience worldwide with this medication; therefore caution is advised in concomitant acute coronary syndrome, stroke and pregnancy. It is not recommended in children. It is not known whether icatibant is excreted in breast milk. Currently, it is recommended that breastfeeding should be delayed by a minimum of 12 hours after icatibant administration.

Its use is recommended in known cases of HAE. It may also be used in suspected cases of ACE inhibitor angioedema. It can be considered for empiric use in cases of angioedema when the cause is unknown, and a response has not been seen with adrenaline. Icatibant is not universally available and remains an expensive drug, therefore its use in the emergency setting is suggested when swelling is seen in the tongue, neck, oropharynx or larynx (to avoid an airway obstruction). It should not be given for minor peripheral swelling, where simple observation may be appropriate. Whilst expensive, it may preserve the airway from life threatening compromise, avoiding costly admissions to the intensive care unit. It is easily administered and repeatable if needed.

Self-administration of icatibant for HAE attacks was assessed in a small case series of fifteen patients. Whilst limited by small sample size and no comparator group, the authors were encouraged by the ease of use and low side effect profile patients’ experienced with icatibant. No patients required hospitalisation following use of icatibant.

Current Evidence

Icatibant in acute HAE attacks

In 2007, icatibant was given intravenously or subcutaneously in an open label pilot study involving fifteen patients with HAE attacks. There was an improved median time to symptom relief in the subcutaneous group compared to the intravenous group (30 minutes vs. 1 hour); therefore, further research used the subcutaneous route. When compared to untreated attacks, time to resolution of symptoms was quicker (reducing mean time to onset of symptom relief by 97%).

This phase II study paved the way for subsequent phase III trials, sponsored by the original manufacturer, Jerini. Three randomised, double blind, control trials (FAST-1, FAST-2, FAST-3 – “For Angioedema Subcutaneous Treatment”) have been published. FAST-1 and FAST-3 were placebo-controlled trials, while in FAST-2 icatibant was compared to oral tranexamic acid. The primary endpoint in all three FAST trials was assessed using a three-point visual analogue scale (VAS) assessment of symptoms – skin swelling, skin pain and abdominal pain. To reach the primary endpoint, a 30% reduction in symptoms was needed on three consecutive scores in the FAST-1 and FAST-2 trials and a 50% reduction in the FAST-3 trial.

FAST-1 compared icatibant with placebo in patients with moderate to severe cutaneous, laryngeal or abdominal attacks of HAE. This showed a non-statistically significant improvement in the time to reduction of symptoms. The FAST-2 trial reported a ten-hour quicker recovery time with icatibant, compared to oral tranexamic acid at a dose of 3 g orally daily for three days. The results of FAST-1 and FAST-2 were published together. The FAST-3 trial compared icatibant with placebo in moderate to severe attacks of HAE and showed the strongest statistically significant improvement.

To improve the safety of these trials, laryngeal attacks were not included in the FAST 1 and 2 trials. Instead, patients with laryngeal symptoms were included in an open label study. In FAST-3, patients with mild laryngeal symptoms were included, whereas patients with severe symptoms were treated in an open label study, resulting in a total of 40 unblinded patients who received icatibant. This provides further limited information on how icatibant fares when dealing with a threatened airway. It is unlikely studies in this area will proceed due to obvious ethical barriers. The unblinded data for laryngeal attacks showed median times for 50% reduction in symptoms were 2.5 and 3.2 hours for icatibant and placebo, respectively. These results emphasise that, in the emergency setting of a laryngeal attack of HAE, it may be appropriate to trial a dose of icatibant, although administration should never supersede the proactive securing of a patient’s airway.
As a result of these trials, icatibant gained approval status in Europe and the U.S. The European CMHP used data from FAST-1 and FAST-2 to decide that “Firazyr’s benefits are greater than its risk”. The FDA cited data from all the FAST trials to provide approval in 2011 for adults over eighteen with HAE.

Prior to the publication of the FAST-3 trial, an international working group met to assess the evidence base for current HAE treatments. The FAST-1 and -2 trials were reviewed favourably, suggesting the available evidence shortened attacks of HAE, with approximately 10% requiring a second dose and 1% requiring a third dose. There were no major safety concerns with the medication, with the main side effect registered being local site irritation. Icatibant compares well with the other angioedema therapies in terms of safety. Allergy, or possible anaphylaxis, has been reported in the other available drugs – ecallantide, recombinant human C1 inhibitor and plasma-derived C1 inhibitors, but not with icatibant. In conclusion, the group reported patients should have access to one of the medications available to treat HAE (which includes icatibant), but do not assign superiority to any of these agents.

The major breakthrough of the FAST trials is that they managed to collate international data on the use of icatibant in a rare condition. Whilst FAST-1 did not show statistical improvement on its primary endpoint, FAST-2 and FAST-3 showed benefit when compared to tranexamic acid and placebo, respectively. Multiple end-points suggest consistent and beneficial results (Tables 3 and 4).

Trial (Year)Study GroupsPrimary EndpointResult
FAST-1 (2010)Icatibant (27) vs. placebo (29)Time to 30% decrease in VAS score2.5 hours vs. 4.6 hours (p=0.14)
FAST-2 (2010Icatibant (36) vs. tranexamic acid (38)Time to 30% decrease in VAS score2 hours vs. 12 hours (p<0.001)
FAST-3 (2011)Icatibant (43) vs. placebo (45)Time to 50% decrease in VAS score2 hours vs. 19.8 hours (p<0.001)
Table 3. Primary endpoint results in FAST trials (n=number of patients in group)


Trial (Year)Study Groups (n)Results
FAST-1 (2010)Icatibant (27) vs. placebo (29)0.8 hours vs. 16.9 hours (p<0.001)
FAST-2 (2010)Icatibant (36) vs. tranexamic acid (38)0.8 hours vs. 7.9 hours (p<0.001)
FAST-3 (2011)Icatibant (43) vs. placebo (45)1.5 hours vs. 18.5 hours (p<0.001)
Table 4. FAST trial outcomes using time to onset of primary symptom relief as endpoint


Although the FAST trials have resulted in licensing in the US and EU, certain weakness and unanswered questions remain. The study populations were small. FAST-2 compared icatibant to tranexamic acid, a medication already considered to be of very limited benefit in HAE. No head-to-head trials are currently available comparing icatibant to other treatments such as C1 inhibitor concentrate, ecallantide or fresh frozen plasma (FFP). Clinicians therefore have a range of options when faced with HAE presentations, with little evidence to support superiority of any treatment. The 2012 international working group reviewed all phase 3 trials for acute treatment of HAE, comparing mainly time for improvement of symptoms and complete resolution of symptoms.[37] Whilst no specific medication was definitively superior, the group made a number of consensus statements regarding an acute attack, suggesting all HAE patients should have access to one of the available medications and that all attacks are eligible to administration of one of the medications either by the patient or healthcare professional as soon as symptoms are recognised. Based on this consensus, it appears the clinician may be justified in giving a novel therapy, such as icatibant, in any patient with HAE, not only one presenting with a threatened airway.

No other published major phase 3 trial data is currently available. There is no data on the efficacy and safety in children (0-18 years old) or use in pregnancy. Currently, there is a multi centre, open label study running involving thirty paediatric subjects with HAE using subcutaneous icatibant 30 mg in the U.S. The results are expected in 2017.

Icatibant in ACE Inhibitor angioedema

The use of icatibant for ACE inhibitor angioedema is, as yet, neither well tested nor established. Most evidence is available in the form of case reports or small series.

In 2010, Bas et al. published the first small case series of eight patients with ACE inhibitor angioedema who were treated with icatibant27. The mean time to initial improvement of symptoms was 50.6 minutes, and to resolution of symptoms, 4.4 hours. No patient required intubation or tracheostomy. In comparison, their historic control group showed an average of 33 hours for resolution of symptoms in patients treated with steroid and antihistamine.

Bas et al. subsequently published a randomised phase two trial in The New England Journal of Medicine in January 2015. This study included 27 patients with ACE inhibitor angioedema of the upper aerodigestive tract treated with either icatibant 30 mg subcutaneously or intravenous prednisolone 500 mg and clemastine 2 mg. Patients presented within ten hours from symptom onset. Icatibant shortened the time to both improvement of oedema (2 hours vs. 11.7 hours) and resolution of symptoms (8 hours vs. 27.1 hours). Three patients in the steroid and antihistamine group needed rescue treatment with icatibant and one required tracheostomy. This small study seems to support the anecdotal evidence so far, that icatibant is useful in ACE inhibitor angioedema. However, it has drawn criticism for describing prednisolone and clemastine as “standard therapy”. It is generally recognised that steroids and antihistamine actually have little or no effect on ACE inhibitor angioedema, and therefore comparing icatibant to this therapy may be doing little more than comparing icatibant to placebo.

Another case series from Italy reported the treatment of thirteen patients with ACE inhibitor angioedema with icatibant. All patients had shown no response to treatment with adrenaline, steroids and antihistamines. After treatment with icatibant 30 mg subcutaneously, the median time to symptom relief was thirty minutes and time to complete resolution was five hours. No patient required a definitive airway and all were discharged within 24 hours of presentation.

Results from the first phase III, double blind, placebo- controlled trial examining icatibant in cases of ACE inhibitor angioedema are expected shortly (CAMEO trial). One hundred and eighteen adult patients have been enrolled in a 1:1 ratio to examine the safety and efficacy of icatibant in patients presenting within twelve hours of the development of suspected ACE inhibitor angioedema. Two primary outcomes are under examination – time to meeting discharge criteria and safety and tolerability of icatibant. In addition, a number of secondary outcomes, including time to onset of symptom relief and occurrence of airway intervention, will be considered.

Discussion Points

Clinicians currently face difficult decisions when managing angioedema, regardless of its aetiology. Firstly, if possible, they must establish what the likely cause may be, and subsequently what treatment may be most effective. Once decided, the type of intervention required is often a judgement decision. One of the most important questions is when to secure the airway in the presence of oropharyngeal angioedema. This remains a complex clinical decision and one taken on a case-by-case basis. It remains unclear at what time point expensive drugs, such as icatibant, should be considered and administered. As clinician awareness of these medications increases, there is a risk that clinical usage patterns will change, in the absence of robust evidence of benefit from large clinical trials. In a condition which is not particularly conducive to guideline-based management, there is concern this medication could be used at lower and lower thresholds. Angioedema patients who may have been reasonably managed by observation could now be treated on a “just in case” basis.

Ishoo proposed a simple staging system based on eighty patients with angioedema requiring admission to ICU over a ten year period (Table 5). This may help clinicians to risk-stratify patients and identify those most at risk of requiring an active airway intervention.

StageClinical FindingsDispositionAirway Interventions (%)
IFacial rash, facial oedema, lip oedemaHome or admission0
IISoft palate oedemaHome or admission0
IIILingual oedemaIntensive Care Unit7
IVLaryngeal oedemaIntensive Care Unit24
Table 5. Ishoo Classification for Monitoring Severity of Upper Airway Oedema [44]


It has also been suggested patients presenting to the emergency department with head and neck oedema may benefit from flexible fibreoptic laryngoscopy to identify the extent of oedema. Other suggested therapies for treating ACE inhibitor angioedema, such as FFP and C1 esterase concentrate, have limited study data. Again, the main body of evidence is currently from case reports. No randomised control trials have been published comparing icatibant to these therapies. FFP contains multiple enzymes that degrade bradykinin, yet no prospective trials have looked at its use in ACE inhibitor angioedema. Successful use has been reported in several case reviews,,,,. Whilst FFP has limitations, including viral transmission, fluid overload and is unsuitable as an out-of-hospital treatment, its relative availability and overall safety profile suggests it could be a therapeutic option.


Icatibant is now available as an easily used treatment in certain forms of angioedema. Its use in severe HAE and ACE inhibitor angioedema appears to be supported by the small-scale data available. It remains expensive and high quality randomised control trials comparing its use to other therapies are absent.

If available, icatibant 30 mg subcutaneously may be a viable treatment in attacks of HAE, especially in the emergency setting when a patient’s airway is threatened. Current evidence suggests more rapid improvement than with oral tranexamic acid, however, it is unclear whether it is superior to other available treatments, such as C1 esterase inhibitor concentrate, ecallantide or human plasma.

Its use in ACE inhibitor angioedema is also theoretically strong but evidence is limited. Case series and a single phase II randomised trial suggest efficacy in comparison to steroids and antihistamines, but no published trial evidence is available comparing it to FFP or other novel medications. The results of the first phase III randomised control trial are expected shortly. 


  1. Boyce A, Austen K. Allergies, Anaphylaxis, and Systemic Mastocytosis. In: Harrison’s Principles of Internal Medicine [Internet]. 19th ed. McGraw-Hill Education; 2015 [cited 2015 Oct 6]. Available from: http://goo.gl/f3Oaty
  2. Bork K. Recurrent Angioedema and the Threat of Asphyxiation. Dtsch Aerzteblatt Online. 2010;107:408–11.
  3. Moellman JJ, Bernstein JA, Lindsell C, Banerji A, Busse PJ, Camargo CA, et al. A Consensus Parameter for the Evaluation and Management of Angioedema in the Emergency Department. Goldstein JN, editor. Acad Emerg Med. 2014;21(4):469–84.
  4. Bernstein JA, Moellman J. Emerging concepts in the diagnosis and treatment of patients with undifferentiated angioedema. Int J Emerg Med. 2012;5(1):39.
  5. Simons FER, Ardusso LR, Bilò M, Cardona V, Ebisawa M, El-Gamal YM, et al. International consensus on (ICON) anaphylaxis. World Allergy Organ J. 2014;7(1):9.
  6. Cicardi M, Aberer W, Banerji A, Bas M, Bernstein JA, Bork K, et al. Classification, diagnosis, and approach to treatment for angioedema: consensus report from the Hereditary Angioedema International Working Group. Allergy. 2014;69(5):602–16.
  7. Zuraw BL, Bernstein JA, Lang DM, Craig T, Dreyfus D, Hsieh F, et al. A focused parameter update: hereditary angioedema, acquired C1 inhibitor deficiency, and angiotensin-converting enzyme inhibitor-associated angioedema. J Allergy Clin Immunol. 2013;131(6):1491–3.
  8. Lang DM, Aberer W, Bernstein JA, Chng HH, Grumach AS, Hide M, et al. International consensus on hereditary and acquired angioedema. Ann Allergy Asthma Immunol Off Publ Am Coll Allergy Asthma Immunol. 2012;109(6):395–402.
  9. Maurer M, Bader M, Bas M, Bossi F, Cicardi M, Cugno M, et al. New topics in bradykinin research. Allergy. 2011;66(11):1397–406.
  10. Han ED, MacFarlane RC, Mulligan AN, Scafidi J, Davis AE. Increased vascular permeability in C1 inhibitor-deficient mice mediated by the bradykinin type 2 receptor. J Clin Invest. 2002;109(8):1057–63.
  11. Zuraw BL. Clinical practice. Hereditary angioedema. N Engl J Med. 2008;359(10):1027–36.
  12. Zuraw BL, Christiansen SC. Pathophysiology of hereditary angioedema. Am J Rhinol Allergy. 2011;25(6):373–8.
  13. Prematta MJ, Kemp JG, Gibbs JG, Mende C, Rhoads C, Craig TJ. Frequency, timing, and type of prodromal symptoms associated with hereditary angioedema attacks. Allergy Asthma Proc Off J Reg State Allergy Soc. 2009;30(5):506–11.
  14. Bork K, Hardt J, Witzke G. Fatal laryngeal attacks and mortality in hereditary angioedema due to C1-INH deficiency. J Allergy Clin Immunol. 2012;130(3):692–7.
  15. Poquette C, Starner C, Hall S, Gleason P. Hereditary Angioedema Drug Utilization and Spend: A Medical and Pharmacy Integrated Analysis. J Manag Care Spec Pharm. 2015;21(4):S18–9.
  16. Ma J, Lee K-V, Stafford RS. Changes in antihypertensive prescribing during US outpatient visits for uncomplicated hypertension between 1993 and 2004. Hypertension. 2006;48(5):846–52.
  17. Garcia-Pavia P, Tomas JM, Alonso-Pulpón L. Late-onset angioedema due to an angiotensin-converting enzyme inhibitor. Can J Cardiol. 2007;23(4):315–6.
  18. Miller DR, Oliveria SA, Berlowitz DR, Fincke BG, Stang P, Lillienfeld DE. Angioedema incidence in US veterans initiating angiotensin-converting enzyme inhibitors. Hypertension. 2008 Jun;51(6):1624–30.
  19. Blais C, Rouleau JL, Brown NJ, Lepage Y, Spence D, Munoz C, et al. Serum metabolism of bradykinin and des-Arg9-bradykinin in patients with angiotensin-converting enzyme inhibitor-associated angioedema. Immunopharmacology. 1999 Sep;43(2-3):293–302.
  20. Bas M, Adams V, Suvorava T, Niehues T, Hoffmann TK, Kojda G. Nonallergic angioedema: role of bradykinin. Allergy. 2007;62(8):842–56.
  21. Bernstein JA. Update on angioedema: evaluation, diagnosis, and treatment. Allergy Asthma Proc Off J Reg State Allergy Soc. 2011;32(6):408–12.
  22. Ulmer JL, Garvey MJ. Fatal angioedema associated with lisinopril. Ann Pharmacother. 1992;26(10):1245–6.
  23. Banerji A, Clark S, Blanda M, LoVecchio F, Snyder B, Camargo CA. Multicenter study of patients with angiotensin-converting enzyme inhibitor-induced angioedema who present to the emergency department. Ann Allergy Asthma Immunol Off Publ Am Coll Allergy Asthma Immunol. 2008;100(4):327–32.
  24. Sondhi D, Lippmann M, Murali G. Airway compromise due to angiotensin-converting enzyme inhibitor-induced angioedema: clinical experience at a large community teaching hospital. Chest. 2004;126(2):400–4.
  25. Soo Hoo GW, Lin HK, Junaid I, Klaustermeyer WB. Angiotensin-converting Enzyme Inhibitor Angioedema Requiring Admission to an Intensive Care Unit. Am J Med. 2015;128(7):785–9.
  26. Charmillon A, Deibener J, Kaminsky P, Louis G. Angioedema induced by angiotensin converting enzyme inhibitors, potentiated by m-TOR inhibitors: successful treatment with icatibant. Intensive Care Med. 2014;40(6):893–4.
  27. Bas M, Greve J, Stelter K, Bier H, Stark T, Hoffmann TK, et al. Therapeutic Efficacy of Icatibant in Angioedema Induced by Angiotensin-Converting Enzyme Inhibitors: A Case Series. Ann Emerg Med. 2010 Sep;56(3):278–82.
  28. Illing EJ, Kelly S, Hobson JC, Charters S. Icatibant and ACE inhibitor angioedema. BMJ Case Rep. 2012;2012.
  29. Gallitelli M, Alzetta M. Icatibant: a novel approach to the treatment of angioedema related to the use of angiotensin-converting enzyme inhibitors. Am J Emerg Med. 2012;30(8):1664.e1–2.
  30. Bas M, Kojda G, Stelter K. [Angiotensin-converting enzyme inhibitor induced angioedema : new therapy options]. Anaesthesist. 2011;60(12):1141–5.
  31. Manders K, van Deuren M, Hoedemaekers C, Simon A. Bradykinin-receptor antagonist icatibant: possible treatment for ACE inhibitor-related angio-oedema. Neth J Med. 2012;70(8):386–7.
  32. Firazyr ® Production Information [Internet]. Shire Human Genetic Therapies; 2015.
    Available from: http://www.shireaustralia.com.au/documents/FirazyrPI.pdf
  33. Boccon-Gibod I, Bouillet L. Safety and efficacy of icatibant self-administration for acute hereditary angioedema: Icatibant self-administration to treat HAE. Clin Exp Immunol. 2012;168(3):303–7.
  34. Bork K, Frank J, Grundt B, Schlattmann P, Nussberger J, Kreuz W. Treatment of acute edema attacks in hereditary angioedema with a bradykinin receptor-2 antagonist (Icatibant). J Allergy Clin Immunol. 2007;119(6):1497–503.
  35. Cicardi M, Banerji A, Bracho F, Malbrán A, Rosenkranz B, Riedl M, et al. Icatibant, a New Bradykinin-Receptor Antagonist, in Hereditary Angioedema. N Engl J Med. 2010;363(6):532–41.
  36. Lumry WR, Li HH, Levy RJ, Potter PC, Farkas H, Moldovan D, et al. Randomized placebo-controlled trial of the bradykinin B2 receptor antagonist icatibant for the treatment of acute attacks of hereditary angioedema: the FAST-3 trial. Ann Allergy Asthma Immunol. 2011;107(6):529–37.e2.
  37. Cicardi M, Bork K, Caballero T, Craig T, Li HH, Longhurst H, et al. Evidence-based recommendations for the therapeutic management of angioedema owing to hereditary C1 inhibitor deficiency: consensus report of an International Working Group: HAE consensus report. Allergy. 2012;67(2):147–57.
  38. Floccard B, Hautin, Bouillet L, Coppere, Allaouchiche. An evidence-based review of the potential role of icatibant in the treatment of acute attacks in hereditary angioedema type I and II. Core Evid. 2012;105.
  39. A Multicenter, Open-Label, Non-Randomized Study to Assess the Pharmacokinetics, Tolerability, and Safety of a Single Subcutaneous Administration of Icatibant in Children and Adolescents With Hereditary Angioedema [Internet]. Available from: http://adisinsight.springer.com/trials/700242659
  40. Baş M, Greve J, Stelter K, Havel M, Strassen U, Rotter N, et al. A randomized trial of icatibant in ACE-inhibitor-induced angioedema. N Engl J Med. 2015;372(5):418–25.
  41. Bova M, Guilarte M, Sala-Cunill A, Borrelli P, Rizzelli GML, Zanichelli A. Treatment of ACEI-related angioedema with icatibant: a case series. Intern Emerg Med. 2015;10(3):345–50.
  42. CAMEO A study evaluating the safety and efficacy of icatibant as a treatment of Angiotensin-Converting Enzyme Inhibitor (ACE-I) induced angioedema in adults [Internet]. Shire; [cited 2015 Nov 3]. Available from: https://clinicaltrials.gov/show/NCT01919801
  43. Escalante FA, Del Baño F, Supervía A. MDMA-induced angioedema treated with icatibant. Clin Toxicol Phila Pa. 2015;1–2.
  44. Ishoo E, Shah UK, Grillone GA, Stram JR, Fuleihan NS. Predicting airway risk in angioedema: staging system based on presentation. Otolaryngol–Head Neck Surg Off J Am Acad Otolaryngol-Head Neck Surg. 1999;121(3):263–8.
  45. Bentsianov BL, Parhiscar A, Azer M, Har-El G. The role of fiberoptic nasopharyngoscopy in the management of the acute airway in angioneurotic edema. The Laryngoscope. 2000;110(12):2016–9.
  46. Pickering RJ, Good RA, Kelly JR, Gewurz H. Replacement therapy in hereditary angioedema. Successful treatment of two patients with fresh frozen plasma. Lancet Lond Engl. 1969;1(7590):326–30.
  47. Hill BJ, Thomas SH, McCabe C. Fresh frozen plasma for acute exacerbations of hereditary angioedema. Am J Emerg Med. 2004;22(7):633.
  48. Pekdemir M, Ersel M, Aksay E, Yanturali S, Akturk A, Kiyan S. Effective treatment of hereditary angioedema with fresh frozen plasma in an emergency department. J Emerg Med. 2007;33(2):137–9.
  49. Warrier MR, Copilevitz CA, Dykewicz MS, Slavin RG. Fresh frozen plasma in the treatment of resistant angiotensin-converting enzyme inhibitor angioedema. Ann Allergy Asthma Immunol Off Publ Am Coll Allergy Asthma Immunol. 2004;92(5):573–5.
  50. Bolton MR, Dooley-Hash SL. Angiotensin-converting enzyme inhibitor angioedema. J Emerg Med. 2012;43(4):e261–2.

Cite this article as follows:

Longley C. Icatibant and its Therapeutic Indications. Critical Care Horizons 2016;2:1-11.


Mitochondrial Function in Sepsis

Mitochondrial Function in Sepsis

Thiago D Corrêa1,2, Stephan M Jakoband Jukka Takala2
1Intensive Care Unit, Hospital Israelita Albert Einstein, São Paulo, Brazil
Department of Intensive Care Medicine, Inselspital, Bern University Hospital and University of Bern, Switzerland

Email: stephan.jakob@insel.ch

Full-Text PDF

AbstractFull-TextReference ListCitation

Sepsis syndrome represents a leading cause of intensive care unit (ICU) admission, morbidity and mortality. Several areas within the pathophysiology of sepsis remain controversial despite extensive pre-clinical and clinical research. It has been postulated that mitochondrial dysfunction may contribute to organ dysfunction and failure in sepsis. Nevertheless, many aspects of mitochondrial malfunction and the exact mechanisms as to how it may be linked to organ dysfunction remain unknown. Here, we briefly review some basic concepts of mitochondrial function in sepsis. The main available methods to assess mitochondrial function are presented and pre-clinical and clinical work summarized to show and explain some of the main controversies surrounding the role of mitochondrial function in sepsis. Finally, we propose future directions for new research in the field.

Keywords: Sepsis; Septic shock; Mitochondria; Multiple organ failure; Animal model; Resuscitation
Competing Interests: The authors declare that they have no competing interests


Despite being one of the oldest syndromes in critical care medicine, sepsis still represents a leading cause of morbidity and mortality in the intensive care unit (ICU),,.

. In sepsis, activation of immune and endothelial cells leads to a profound release of pro-inflammatory and anti-inflammatory mediators, increases adhesion molecule expression and upregulates complement and coagulation system activation. This leads to increased vascular permeability, fluid loss to the extravascular compartment, systemic vasodilatation, impaired myocardial function and derangements of microcirculatory blood flow, which impair tissue perfusion and oxygen delivery to the cells3.

The decreased availability of oxygen to the tissues secondary to various combinations of low arterial oxygen tension (hypoxic hypoxia), decreased hemoglobin levels (anaemic hypoxia), and/or hypoperfusion (stagnant hypoxia), inhibit cellular aerobic production of adenosine triphosphate (ATP). Nevertheless, findings of decreased ATP production in sepsis,,,,,

in the face of maintained or even increased tissue pO2 levels2,,,,, coupled with a variable pattern of systemic oxygen consumption, have given rise to the controversial theory of cytopathic hypoxia,. This theory suggests an acquired defect in cellular oxidative phosphorylation (i.e. mitochondrial dysfunction) prevents cells from different organs and tissues from using molecular oxygen for ATP production. Cytopathic hypoxia has been put forward as an explanation for sepsis-induced organ dysfunction and failure [16]. Several underlying mechanisms have been proposed to explain cytopathic hypoxia in sepsis. These include:

  • ultrastructural damage to the mitochondrial membranes, allowing a proton leak across the inner mitochondrial membrane,,;
  • derangements of the pyruvate dehydrogenase complex, which oxidatively decarboxylates pyruvate to acetyl-CoA;
  • inhibition of key mitochondrial enzymes of the tricar- boxylic acid cycle (Citric Acid cycle or Krebs cycle) and electron transport chain by nitric oxide (NO)6;
  • inhibition of mitochondrial enzyme complexes by peroxynitrite;
  • damage caused by reactive oxygen species (ROS); and
  • poly(ADP-ribose) polymerase (PARP-1) activation.

The aim of this review is to outline some basic concepts of mitochondrial function in the normal state and in sepsis, to present the main available methods to assess mitochondrial function, and to summarize pre-clinical and clinical work in an attempt to explain some of the main controversies surrounding the role of mitochondrial function in sepsis. Finally, we propose future directions for new research in the field.

Mitochondrial Structure and Function

A mitochondrion is surrounded by a double layer of membranes, which form an inter-membrane space (Figure 1). The outer mitochondrial membrane is composed of an equal proportion of protein and lipids and is permeable to most metabolites. In contrast, the inner mitochondrial membrane is composed mainly of proteins (80%) with a small proportion of lipids (20%), and is highly selectively permeable. The inner membrane is folded inwards to form cristae, projecting into the mitochondrial matrix, and concentrates the respiratory chain enzymes. The mitochondrial matrix contains the enzymes of the tricarboxylic acid cycle (TCA), either free or attached to the inner mitochondrial membrane, ß-oxidation enzymes, and the pyruvate dehydrogenase complex (Figure 1) [25]. The mitochondrion is the main structure responsible for energy production in animal cells. Energy production, i.e. ATP production, occurs in a three-step process which is intrinsically interconnected: glycolysis, TCA cycle and electron transport chain (oxidative phosphorylation)26. Glycolysis occurs in the cytosol of all cells and represents the major pathway for glucose, galactose and fructose metabolism.


Glycolysis is an oxygen-independent process. Under anaerobic conditions, the resulting end product, pyruvate, is reduced by lactate dehydrogenase into lactate, generating two molecules of ATP per molecule of glucose. More often, under aerobic conditions, pyruvate is transported into the mitochondria matrix, where it is oxidatively decarboxylated to Acetylcoenzyme A (Acetyl-CoA) by the pyruvate dehydrogenase complex26,27. This Acetyl-CoA then enters the TCA cycle (Figure 1).

Figure 1. Energy production pathways in animal cells.

Figure 1. Energy production pathways in animal cells.

Under anaerobic conditions (-O2), pyruvate is reduced by lactate dehydrogenase into lactate, generating two molecules of adenosine triphosphate (ATP) per molecule of glucose. More often, under aerobic conditions (+O2), pyruvate is transported into the mitochondria matrix, where it is oxidatively decarboxylated to Acetyl-CoA by pyruvate dehydrogenase complex. The tricarboxylic acid (TCA) cycle is a sequence of reactions that occur in the mitochondrial matrix and is always followed by oxidative phosphorylation. In the TCA cycle, Acetyl-CoA is oxidized and nicotinamide adenine dinucleotide (NAD) and flavin adenine dinucleotide (FAD) reduced, respectively, to NADH-H+ and FADH2. These coenzymes are subsequently reoxidized in the respiratory chain coupled with ATP production.


The TCA Cycle

The TCA cycle is a sequence of reactions that occur in the mitochondrial matrix, which is always followed by oxidative phosphorylation, the most important mechanism of ATP production (Figures 1 and 2)26, 27. In the TCA cycle, one molecule of Acetyl-CoA is oxidized and three molecules of nicotinamide adenine dinucleotide (NAD+) and one molecule of flavin adenine dinucleotide (FAD) are reduced, respectively, to three molecules of NADH-H+ and one molecule of FADH2. These coenzymes (or carriers) are subsequently utilized (reoxidized) in the respiratory chain in the key process of ATP production.

Figure 2. Summary of oxidative phosphorylation. The electrons flowing through the respiratory chain complexes generate adenosine triphosphate (ATP) in a process named oxidative phosphorylation. The respiratory chain comprises five protein complexes embedded on the inner mitochondrial membrane (complexes I to V). Electrons are transferred from NADH-H+ to complex I, also known as NADH-Q oxidoreductase, coupled with the transfer of 4H+ to intermembrane space. Electrons are also transferred from FADH2 to complex II (succinate-Q reductase). The electrons coming from both complexes are then transferred to coenzyme Q (ubiquinone), which is reduced to ubiquinol. Then, ubiquinol donates its electrons to complex III (Q-cytochrome c oxidoreductase), in which additional 4H+ are transferred to the intermembrane space. The reduced cytochrome c (Cyt c) shuttles electrons from complex III to complex IV (cytochrome c oxidase) simultaneously, with a reduction of molecular oxygen to water and the transfer of 2H+ to the intermembrane space. Finally, ATP synthase (complex V) drives the synthesis of ATP from adenosine diphosphate (ADP) and inorganic phosphate coupled with the release of energy from the passage of H+ back to the mitochondrial matrix. ATP synthase is composed of two subunits: F1, which projects into the matrix and contains the phosphorylation mechanism, and FO, which spans the inner mitochondrial membrane creating a proton channel. The flow of H+ through FO causes it to rotate, driving the production of ATP by the F1 subunit.

Figure 2. Summary of oxidative phosphorylation.

The electrons flowing through the respiratory chain complexes generate adenosine triphosphate (ATP) in a process named oxidative phosphorylation. The respiratory chain comprises five protein complexes embedded on the inner mitochondrial membrane (complexes I to V). Electrons are transferred from NADH-H+ to complex I, also known as NADH-Q oxidoreductase, coupled with the transfer of 4H+ to intermembrane space. Electrons are also transferred from FADH2 to complex II (succinate-Q reductase). The electrons coming from both complexes are then transferred to coenzyme Q (ubiquinone), which is reduced to ubiquinol. Then, ubiquinol donates its electrons to complex III (Q-cytochrome c oxidoreductase), in which additional 4H+ are transferred to the intermembrane space. The reduced cytochrome c (Cyt c) shuttles electrons from complex III to complex IV (cytochrome c oxidase) simultaneously, with a reduction of molecular oxygen to water and the transfer of 2H+ to the intermembrane space. Finally, ATP synthase (complex V) drives the synthesis of ATP from adenosine diphosphate (ADP) and inorganic phosphate coupled with the release of energy from the passage of H+ back to the mitochondrial matrix. ATP synthase is composed of two subunits: F1, which projects into the matrix and contains the phosphorylation mechanism, and FO, which spans the inner mitochondrial membrane creating a proton channel. The flow of H+ through FO causes it to rotate, driving the production of ATP by the F1 subunit.


The respiratory chain and oxidative phosphorylation

The flow of electrons through the complexes of the respiratory chain generates ATP by a process named oxidative phosphorylation (chemiosmotic theory) (Figure 2)26. Oxidative phosphorylation occurs on the inner mitochondrial membrane. Electrons from the NADH-H+ and FADH2 produced in the TCA cycle are transferred through the mitochondrial complexes (electron transport chain), re-oxidazing the carriers and generating ATP, a high-energy phosphate. Although oxygen is not necessary for the TCA cycle, it is crucial for oxidative phosphorylation26. The respiratory chain comprises five protein complexes embedded on the inner mitochondrial membrane (complexes I to V) (Figure 2). Electrons are transferred from NADH-Hto complex I, also known as NADH-Q oxidoreductase, coupled with the transfer of four H+ to the intermembrane space. Electrons are also transferred from FADH2 to complex II (succinate-Q reductase)26. The electrons coming from both complexes are then transferred to coenzyme Q (ubiquinone), which is reduced to ubiquinol (QH2, dihydroquinone). Then, ubiquinol donates its electrons to complex III (Q-cytochrome c oxidoreductase) in a process called Q cycle, in which an additional four H+ are transferred to the intermembrane space. The reduced cytochrome c shuttles electrons from complex III to complex IV (cytochrome c oxidase) simultaneously, with a reduction of molecular oxygen to water and the transfer of 2H+ to the intermembrane space. Therefore, oxygen is the final electron acceptor of the respiratory chain and water is the final product of oxygen reduction (Figure 2)26. While electrons are being carried through the complexes, the released energy in the process is used to pump H+ to the intermembrane space. This is a key process. This electro-chemical gradient of protons, concentrated in the intermembrane space, is then used by the complex V (ATP synthase or F1FO ATPase), located on the inner mitochondrial membrane, to produce ATP, coupled with the release of energy by the flux of H+ coming back to the mitochondrial matrix. In non-damaged mitochondria, protons return to the mitochondrial matrix from the intermembrane space through the complex V. The inner mitochondrial membrane must be physically intact to be able to create this proton gradient. Thus, the mitochondrion can control the re-entrance of protons into the mitochondrial matrix.

ATP synthase drives the synthesis of ATP from adenosine diphosphate (ADP) and inorganic phosphate, coupled with the release of energy from the passage of H+ back to the mitochondrial matrix. ATP synthase is composed of two subunits: F1, which projects into the matrix and contains the phosphorylation mechanism, and FO, which spans the inner mitochondrial membrane creating a proton channel. The flow of H+ through FO causes it to rotate, driving the production of ATP by the F1 subunit.

Impact of mitochondrial damage on function

In damaged mitochondria, in which an inner membrane is permeable to protons, ATP synthesis is not only reduced due to the reduced proton gradient, but it is also affected by the reverse action of ATP synthase. Under such circumstances, ATP synthase takes ATP from the mitochondrial matrix and works in a counterproductive way as an ATP hydrolase, reducing the ATP levels18.

How is Mitochondrial Function Assessed?

Over the years, the available methods to address mitochondrial function have evolved, becoming more user-friendly and robust. Mitochondrial function can be evaluated in-vivo or in-vitro, by spectrophotometric assays (enzymatic activity) or by polarography (Clark-type electrode or high-resolution respirometry). The technical aspects of each method are beyond the scope of this paper and have been described in detail elsewhere[30, 31].

Briefly, the spectrophotometric assay allows the assessment of the activity of respiratory chain complexes in human and animal tissues and cells. In a spectrophotometric assay, a tissue or cell homogenate is supplemented with different electron donors or acceptors, and the enzymatic activity of each complex is expressed as nanomole (nmol) cytochrome c reduced per minute per milligram protein (nmol cytochrome c·min-1·mg-1),.

The high-resolution respirometry (oxygraph) technique (Oxygraph-2k; Oroboros Instruments, Innsbruck, Austria) has been used in many studies (Figure 3)10,,,,,,,. Through this method, intact cells, permeabilized cells, or isolated mitochondria sampled from different human or animal tissues are processed, immersed into a medium, and placed inside a sealed chamber containing a known oxygen concentration. The high-resolution respirometry analysis is based on a continuous measurement by polarography of oxygen concentration inside the sealed chamber. As mitochondria consume oxygen inside the chamber, the oxygen concentration declines, and a plot of oxygen concentration by time is provided (Figure 3)31.

Figure 3. Representative diagram of measurement of respiration rates in isolated liver mitochondria from the left liver lobe of a healthy pig assessed by high-resolution respirometry. The high-resolution respirometry (Oxygraph-2k; Oroboros Instruments, Innsbruck, Austria) allows the evaluation of different states of respiratory control. By applying specific substrates, different complexes of the electron transport chain can be studied. Complex-I dependent respiration can be evaluated by adding glutamate and malate as substrate, which provides NADH-H+ to complex I. Complex-II-dependent respiration can be evaluated after inhibition of complex I by rotenone, by adding succinate as substrate, which provides FADH2 to complex II. Finally, ascorbate plus TMPD (N,N,N’,N’-Tetramethyl-p-phenylenediamine dihydrochloride) can be used to address complex-IV dependent respiration. The state 3 represents the maximal capacity of the respiratory chain itself when saturating concentrations of ADP and substrates are provided. State 4 represents the resting respiration, when ADP is depleted by its phosphorylation to ATP (ADP-limited resting state). The respiratory control ratio (RCR) can be obtained by dividing the rate of oxygen consumption at state 3 by the rate of oxygen consumption at state 4. The RCR represents a marker of oxidative phosphorylation efficiency (the coupling of phosphorylation to oxidation). The red line represents the oxygen consumption by liver mitochondria expressed as pmol oxygen per second per mg of mitochondrial protein and the blue line represents the oxygen concentration inside the sealed chamber expressed as pmol oxygen per ml.

Figure 3. Representative diagram of measurement of respiration rates in isolated liver mitochondria from the left liver lobe of a healthy pig assessed by high-resolution respirometry.

The high-resolution respirometry (Oxygraph-2k; Oroboros Instruments, Innsbruck, Austria) allows the evaluation of different states of respiratory control. By applying specific substrates, different complexes of the electron transport chain can be studied. Complex-I dependent respiration can be evaluated by adding glutamate and malate as substrate, which provides NADH-H+ to complex I. Complex-II-dependent respiration can be evaluated after inhibition of complex I by rotenone, by adding succinate as substrate, which provides FADH2 to complex II. Finally, ascorbate plus TMPD (N,N,N’,N’-Tetramethyl-p-phenylenediamine dihydrochloride) can be used to address complex-IV dependent respiration. The state 3 represents the maximal capacity of the respiratory chain itself when saturating concentrations of ADP and substrates are provided. State 4 represents the resting respiration, when ADP is depleted by its phosphorylation to ATP (ADP-limited resting state). The respiratory control ratio (RCR) can be obtained by dividing the rate of oxygen consumption at state 3 by the rate of oxygen consumption at state 4. The RCR represents a marker of oxidative phosphorylation efficiency (the coupling of phosphorylation to oxidation). The red line represents the oxygen consumption by liver mitochondria expressed as pmol oxygen per second per mg of mitochondrial protein and the blue line represents the oxygen concentration inside the sealed chamber expressed as pmol oxygen per ml.


As part of this process, various respiratory states of mitochondria are referred to. A basic understanding of these becomes important when reviewing results of the studies on mitochondrial dysfunction in sepsis. The states include:

  • State 1: The first state in an oxygraph protocol, where mitochondria are present in a medium with oxygen and inorganic phosphate, but there is no ADP or respiratory substrate present as yet;
  • State 2: This refers to a substrate limited state of residual oxygen consumption (ADP has been added but respiratory substrates have not);
  • State 3: ADP stimulated respiration after addition of respiratory substrates – this represents the maximum capacity of the respiratory chain;
  • State 4: Following on from state 3, when the available ADP has been completely converted to ATP – this represents the resting level of respiration; and
  • State 5: Following on from state 4, when the available oxygen has been completely depleted i.e. anaerobic.

The ratio of state 3 to state 4 is called the respiratory control ration (RCR), and is an index of how efficiently coupled phsophorylation is to oxidation. The amount of tissue available, and whether fresh or frozen samples will be analysed, are critical aspects for the choice of the method used to address mitochondrial function30,31. While spectrophotometric analysis can be performed using either fresh or frozen tissue samples, polarographic analysis requires a prompt analysis of fresh samples30,31. The polarographic analysis has the advantage of allowing simultaneously evaluation of the electron transport chain and the TCA cycle, which more closely resembles the cell metabolism31.

Mitochondrial function in sepsis

Even though it has been over fifty years since the first studies about mitochondrial function were published,,, the role of mitochondrial dysfunction in sepsis remains controversial, and its contribution to the development of organ dysfunction is unknown29,,. Controversies exist given the observations that mitochondrial function might be depressed, improved, or unchanged in sepsis29. The putative mechanisms of mitochondrial dysfunction, the variable pattern of complexes involved, and the temporal sequence after the initial insult, i.e., early vs. late sepsis, are examples of the intriguing questions that remain to be answered29. Some authors have argued that the decreased mitochondrial function observed in experimental and/or clinical studies might represent an adaptive mechanism, occurring in response to varying degrees of tissue hypoperfusion and hypoxia. This state has been termed “mitochondrial hibernation”, which can be characterized by a cellular down-regulation of all non-essential functions, followed by a decreased global rate of oxygen and ATP consumption45. The hibernation phenomenon may explain the observations that organ dysfunction and failure in sepsis were seldom associated with histopathological damage.

Inhibitory effects of intravenously administered endotoxin on mitochondrial respiration and/or enzymatic activity in different animal species have been described29. Similar findings have also been reported when intact live bacteria are administrated through different routes, and in different species29. Moreover, in the few studies which addressed mitochondrial function in septic patients, isolated mitochondria from skeletal muscle and blood cells were assessed6,42,,,,,,,,,. Thus, it is important to emphasize that most of what is known about mitochondrial dysfunction in sepsis has been provided by experimental animal models29.

For instance, decreased complex I dependent respiration was reported in skeletal muscle and hepatocytes after twenty four hours of abdominal sepsis in rats8. Moreover, hepatic ATP content was shown to be lower in the most severe animals in comparison to the sham controls or mildly affected rats8. A decreased complex II-III activity was reported in the diaphragm of rats after twelve hours of faecal peritonitis, while a decreased activity of all mitochondrial complexes was demonstrated after forty-eight hours of sepsis. It has been demonstrated that isolated liver complex I and II dependent respiration, but not isolated skeletal muscle nor kidney mitochondrial respiration, were impaired when pigs were rendered septic by intravenous endotoxin (Escherichia coli lipopolysaccharide B0111:B4) infusion during a 24 hours period. On the other hand, Kozlov and colleagues demonstrated impaired kidney complex II dependent mitochondrial respiration after twelve hours of faecal peritonitis in pigs34. Impaired brain complex I dependent respiration after twenty-four hours of faecal peritonitis was described19, and decreased cerebellum, hippocampus, striatum and cortex mitochondrial complex I dependent respiration demonstrated after 24, 48 and 96 hours of faecal peritonitis in rats.

Evidence of mitochondrial dysfunction has also been reported in humans. Brealey and colleagues demonstrated that complex I activity was lower in patients with septic shock who died in the ICU, in comparison to non-septic patients6. In another study, a 60% reduction in complex I activity of intercostal muscle, but not of vastus lateralis muscle, of ten mechanically ventilated septic patients was reported47. Nevertheless, decreased ATP content was observed in vastus lateralis muscle while intercostal muscle ATP content was unchanged47.

Conversely, an increased skeletal muscle (vastus lateralis) complex I activity was demonstrated in seven human healthy volunteers two hours after intravenous endotoxin infusion49. Moreover, Sjövall and colleagues demonstrated an increased complex I and II state 3 dependent respiration in permeabilized platelets isolated from eighteen patients with severe sepsis/septic shock during the first seven days of the disease, both in comparison to days 1/2 and to non-septic (healthy) controls50. The same group also demonstrated increased complex I, II and IV state 3 dependent mitochondrial respiration in permeabilized peripheral blood immune cells obtained from patients with severe sepsis/septic shock, in comparison to healthy controls53.

Are the experimental models used to address mitochondrial function appropriate?

The results of studies across rodent and pig models of sepsis are summarised in Tables 1,,,,,,,,,,,,,,,,,,,,,,,, and 2,,. It is important to highlight that experimental models used to study mitochondrial function in sepsis differ in many respects from clinical sepsis in ICU patients,. Therefore, the reliability of experimental models has been questioned. Septic patients admitted to the ICU are often elderly, and exhibit multiple comorbidities, while experimental animals are usually young, of a single gender, with no comorbidities and from a similar genetic background87,88,90. Moreover, due to feasibility and costs, researchers very often set up short-term (shorter than 24 hours) models of sepsis, while the clinical course of human sepsis usually develops over several hours or days.

AuthorReferenceYearAnimalModelTime (h)RSTissueResults
Fry591981ratsi.v. LPS6NoLiverC-I and C-II state 3 & RCR increased
Liver RCR decreased.
SM RCR decreased
Garrison611982ratsCLP2, 4 or 6NoKidneyUnchanged
Geller621986ratsi.p. LPS18NoSMUnchanged
Dawson631988ratsi.v. LPS4Yes#Heart
Heart C-I RCR increased.
SM C-I RCR increased
Kopprasch641989ratsi.p. LPS6NoLiverC-I & C-II state 4 increased
Takayama651990ratsi.p. LPS24NoLiverC-I & C-II state 3 & RCR increased
Llesuy661994ratsCLP6, 12, 24YesLiver
Liver unchanged.
SM C-I and C-II State 3 and RCR decreased
Malaisse681997ratsi.p. LPS24NoLiverC-II state 3, state 4 & RCR increased
Kantrow691997ratsCLP16NoLiverC-I and C-II state 3 increased
Markley702002ratsi.p. LPS2NoLiverUnchanged
Fukumoto712003ratsi.p. LPS2 or 6No
Heart C-I RCR unchanged at 2 hrs but decreased at 6 hrs.
Kidney unchanged
Suliman722004ratsi.v. LPS6, 24 or 48NoHeartC-I State 3 decreased at 6 hrs, unchanged at 24 & 48 hrs
Heart RCR decreased.
Diaphragm RCR decreased
Kozlov742006ratsi.p. LPS16NoHeart
Heart C-I decreased.
Liver C-I & C-II increased
Larche752006miceCLPUp to 96NoHeartC-I state 3 & RCR decreased, C-IV unchanged
Mason762007ratsi.p. LPS6, 12 or 24NoHeartC-I state 3 unchanged at 6 or 12 but decreased at 24 hrs
Protti772007ratsFP48NoSMC-I decreased, C-II unchanged
Kozlov782007ratsi.p. LPS16NoLiverC-I and C-II state 3 increased
Duvigneau792008ratsi.v. LPS2, 4, 8, 12NoLiverC-I RCR increased at 2 hrs & 12 hrs, decreased 4-8 hrs
Hassoun802008ratsi.v. LPS4NoHeartDecreased C-1 state 3 and RCR. Increased state 4
Vanasco812008ratsi.p. LPS6NoLiver
Liver C-I State 3 decreased at 6 hrs and unchanged at 24 & 48 hrs.
Heart C-I State 3 decreased.
Diaphragm C-I State 3 decreased
Kozlov332009ratsi.v. LPS16NoLiverC-I and C-II RCR increased
Reynolds822009micei.p. LPS6, 24, 48, 72NoHeartC-I and C-II State 3 decreased at 24 hrs, increased at 72 hrs
Aguirre382012micei.p. LPS24NoSMC-II decreased
Vanasco832012ratsi.p. LPS6NoHeartC-I and C-II State 3 decreased
Table 1. The effect of sepsis on mitochondrial function in rodents addressed by polarography (Clark-type electrode or high-resolution respirometry).

RS = repeated samples, i.v. = intravenous, i.p. = intraperitoneal, SM = skeletal muscle, LPS = lipopolysaccharide, CLP = caecal ligation and puncture, FP = faecal peritonitis, # = postmortem period, C-I = Complex-I dependent respiration, C-II = Complex-II dependent respiration, C-IV = Complex-IV dependent respiration and RCR = respiratory control ratio (state 3 / state 4)


Hirai51984CLP0,2,4,7,12-14 daysNoFluidLiverC-I state 3 and RCR decreased by days 12-14
Porta572006i.v. LPS24hNoFluidsSM
SM unchanged.
Liver C-I state 4 increased & RCR decreased.
Kidney C-I, C-II and C-IV unchanged
Li $842007FP12hNoFluidsHeartC-I activity decreased
Regueira852008i.v. LPS10hNoFluids, pressorsLiver
Liver C-I and C-II RCR increased.
SM C-I state 3 increased
i.v, LPS
SM unchanged in FP.
Liver unchanged in FP. C-I state IV decreased in i.v. LPS
Liver C-I & C-II state 3, state 4 & RCR decreased.
Kidney C-I states 3 & 4 increased, RCR unchanged.
C-II states 3 & 4 increased, RCR decreased
Corrêa352012FP6, 12 or 24h sepsis no therapy + 48h of resuscitationYes*Fluids, pressors, inotropes, antibioticsSM
SM C-I RCR increased after 12 hrs of PI.
Brain C-II state 3 decreased after 72 hrs of PI.
Liver unchanged.
Heart unchanged.
Regueira102012FP0, 6 and end (max 24h)Yes*FluidsLiver
Liver C-I, C-II unchanged.
SM C-I RCR decreased.
Vuda862012FP27hYes*Fluids, pressorsLiver
Liver C-I, C-II and C-IV unchanged.
SM C-I, C-II and C-IV unchanged.
Corrêa362013FP12h sepsis no therapy + 48h of resuscitationYes*Fluids, pressors, inotropes, antibioticsSM
Both unchanged
Corrêa362013FP22h#YesNoSMC-I state increased
Table 2. The effect of sepsis on mitochondrial function in pigs addressed by polarography (Clark-type electrode or high-resolution respirometry).

RS = repeated samples, i.v. = intravenous, SM = skeletal muscle, LPS = lipopolysaccharide, CLP = cecal ligation and puncture, FP = fecal peritonitis, PI = peritonitis induction, # = postmortem period, C-I = Complex-I dependent respiration, C-II = Complex-II dependent respiration, C-IV = Complex-IV dependent respiration, RCR = respiratory control ratio (state 3 / state 4), * = for skeletal muscle analysis, $ = spectrophotometric analysis and # = median (range) survival time of 22 (16 to 28) hours.


Small rodents are the most common animals used in experimental sepsis29. Nevertheless, they have significant physiological and pharmacological differences in comparison to humans. Indeed, mice are more resistant to endotoxin infusion than humans, with a different pattern of inflammatory response to an infectious insult. Porcine models of sepsis, using medium-size pigs (30-40 kg), have been used to better reproduce the clinical aspects of sepsis and its treatment. Pigs share many aspects of human cardiovascular anatomy and physiology, and allow the reproduction of the clinical management of sepsis with full hemodynamic monitoring, fluid resuscitation, antibiotics and support with vasoactive drugs. In addition, repeated tissue samples for mitochondrial analysis are possible. Such interventions are usually not feasible in small-size experimental animals.

Another important aspect of sepsis models that may affect mitochondrial function is the severity of the disease. Three methods are commonly used to produce experimental sepsis: exogenous administration of a toxin (lipopolysaccharide; (LPS), methods that alter the endogenous protective barrier [caecal ligation and puncture (CLP) and colon ascendens stent peritonitis (CASP)], and exogenous administration of viable pathogens in the lungs, peritoneal cavity, subcutaneously or intravenously.

Endotoxin infusion models accounted for approximately 40% of studies of mitochondrial function in sepsis29. However, it is well known that the immune, inflammatory, and cardiovascular responses triggered by LPS infusion are completely different from those induced by a living pathogen92. This different pathophysiology may affect mitochondrial function94.

Models of faecal peritonitis were developed to overcome the pitfalls of the endotoxin models. Those models more closely resemble human sepsis, demonstrating several advantages in comparison to LPS models, including a polymicrobial infection caused by living pathogens and a well-defined focus of infection, which triggers an immune, inflammatory and cardiovascular response more comparable to human sepsis94. The severity of experimental sepsis, which may have important implications on mitochondrial function, can be adjusted by varying the amount of LPS infused, the size of the caecal punctures and/or the distance of cecum ligated, the diameter of the inserted stent into the ascending colon or the bacterial load infused into the abdominal cavity.

The Effect of Time on Mitochondrial Assessment

Sepsis is characterized by a biphasic inflammatory, immune, hormonal, and metabolic response triggered by an infection3. While early sepsis is characterized by a pronounced release of inflammatory mediators, increased release of stress hormones, metabolic activity and mitochondrial function, late sepsis is characterized by an anti-inflammatory immunosuppressive state, and impaired energy production secondary to mitochondrial inhibition and/or damage.

Thus, it is been postulated that a severe inflammation, endocrine and metabolic shutdown leads to a decreased energy production, i.e. cell “hibernation” or “stunning”, which may be a protective mechanism. This reduced cellular metabolism might boost the chances of cellular survival after an overwhelming insult46. Such different stages of the inflammatory, immune, hormonal, and metabolic response in septic animals and patients might explain why several groups have reported conflicting results (unchanged, impaired or improved) regarding mitochondrial function in sepsis, with marked organ-specific differences (Tables 1 and 2)29.

Impaired respiration in the heart but not in the kidney of endotoxemic rats71, in the hepatocytes but not in the heart of pigs rendered septic by intravenous infusion of Pseudomonas aeruginosa7, in the liver but neither in the kidney nor in skeletal muscle of endotoxemic pigs57, in small bowel mucosa but not in the muscular layer of challenged pigs with continuous infusion of endotoxin have all been reported.

We recently evaluated the impact of treatment delay on the development of sepsis-associated mitochondrial dysfunction in skeletal muscle, liver, heart and brain using a swine model of faecal peritonitis (peritoneal instillation of autologous faeces)35. After 6, 12 or 24 hours of untreated sepsis, all animals received 48 hours of protocolized resuscitation consisting of fluids, vasopressors and broad spectrum antibiotics. An increased skeletal muscle Complex I dependent respiration after 12 hours of untreated sepsis was the only sepsisassociated alteration observed before the beginning of resuscitation. At the end of study (i.e. 72 hours after peritonitis induction), a decreased maximal brain mitochondrial Complex II respiration was found in the animals resuscitated after 24 hours of untreated sepsis, while hepatic and myocardial mitochondrial respiration were not affected).

Are the Most Appropriate Organs Being Evaluated?

The contribution of each organ and system to the outcomes of critically ill patients is variable,. Although the liver, heart, kidneys, brain and bowel are easily assessed experimentally, most of the studies addressing the mitochondrial function in humans are limited to skeletal muscle, peripheral blood immune cells53, isolated platelets55,, monocytes and cultured human hepatocytes,, which may have limited clinical significance. Therefore, one can argue that the lack of association between mitochondrial dysfunction and organ dysfunction or failure in septic patients occurs because only organs with a non-critical impact on prognosis have been properly addressed in humans.

Additionally, the time necessary for different organs and systems to reach the maximum degree of dysfunction is variable. Therefore, serial sampling in such organs would be necessary to allow for early detection of mitochondrial dysfunction. Currently, such analysis is neither feasible nor ethically acceptable. Thus, most of our knowledge on the role of mitochondrial dysfunction in sepsis will continue to be provided by experimental studies. This fact highlights the importance of a considered approach to selection of the experimental model.

Future Directions

The development of easy to handle, non-invasive, devices to address mitochondrial function at the bedside may improve knowledge about the contribution of mitochondrial dysfunction to sepsis pathophysiology. Ideally, a better understanding of the mechanisms of mitochondrial damage and dysfunction in sepsis would be accompanied by new therapies aimed at decreasing the progression to organ dysfunction and failure, and ultimately decreasing sepsis mortality. Furthermore, there are currently no available therapies to treat mitochondrial dysfunction in sepsis. Nevertheless, it has been assumed that appropriate early management of sepsis might prevent the development of mitochondrial dysfunction in such a population of critically ill patients35. Finally, the development of standards for the performance and reporting of mitochondrial function analysis would help researchers and clinicians to more readily compare the results obtained by different investigators.


Despite decades of research, the pathophysiology of sepsis has not been completely elucidated. It seems that mitochondrial dysfunction may play a role in organ dysfunction and failure. Nevertheless, the clinical significance of mitochondrial dysfunction, and its association with organ failure, remain unclear. The development of user-friendly, non-invasive devices might allow us to address the role of mitochondrial dysfunction on vital organs in humans, and over appropriate time-scales. Lastly, it may also allow us to develop and study new therapies to improve the outcomes of septic patients.


  1. Funk DJ, Parrillo JE, Kumar A. Sepsis and septic shock: a history. Crit Care Clin. 2009;25(1):83–101, viii.
    Available from: http://dx.doi.org/10.1016/j.ccc.2008.12.003.
  2. Hotchkiss RS, Rust RS, Dence CS, Wasserman TH, Song SK, Hwang DR, et al. Evaluation of the role of cellular hypoxia in sepsis by the hypoxic marker [18F]fluoromisonidazole. Am J Physiol. 1991;261(4 Pt 2):R965–R972.
  3. Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med. 2013;369(9):840–851.
    Available from: http://dx.doi.org/10.1056/NEJMra1208623.
  4. Fink MP. Bench-to-bedside review: Cytopathic hypoxia. Crit Care. 2002;6(6):491–499.
  5. Hirai F, Aoyama H, Ohtoshi M, Kawashima S, Ozawa K, Tobe T. Significance of mitochondrial enhancement in hepatic energy metabolism in relation to alterations in hemodynamics in septic pigs with severe peritonitis. Eur Surg Res. 1984;16(3):148–155.
  6. Brealey D, Brand M, Hargreaves I, Heales S, Land J, Smolenski R, et al. Association between mitochondrial dysfunction and severity and outcome of septic shock. Lancet. 2002;360(9328):219–223.
    Available from: http://dx.doi.org/10.1016/S0140-6736(02)09459-X.
  7. Hart DW, Gore DC, Rinehart AJ, Asimakis GK, Chinkes DL. Sepsis-induced failure of hepatic energy metabolism. J Surg Res. 2003;115(1):139–147.
  8. Brealey D, Karyampudi S, Jacques TS, Novelli M, Stidwill R, Taylor V, et al. Mitochondrial dysfunction in a longterm rodent model of sepsis and organ failure. Am J Physiol Regul Integr Comp Physiol. 2004;286(3):R491–R497.
    Available from: http://dx.doi.org/10.1152/ajpregu.00432.2003.
  9. Huang LJ, Hsu C, Tsai TN, Wang SJ, Yang RC. Suppression of mitochondrial ATPase inhibitor protein (IF1) in the liver of late septic rats. Biochim Biophys Acta. 2007;1767(7):888–896.
    Available from: http://dx.doi.org/10.1016/j.bbabio.2007.03.009.
  10. Regueira T, Djafarzadeh S, Brandt S, Gorrasi J, Borotto E, Porta F, et al. Oxygen transport and mitochondrial function in porcine septic shock, cardiogenic shock, and hypoxaemia. Acta Anaesthesiol Scand. 2012;56(7):846–859.
    Available from: http://dx.doi.org/10.1111/j.1399-6576.2012.02706.x.
  11. Boekstegers P, Weidenhöfer S, Pilz G, Werdan K. Peripheral oxygen availability within skeletal muscle in sepsis and septic shock: comparison to limited infection and cardiogenic shock. Infection. 1991;19(5):317–323.
  12. VanderMeer TJ,Wang H, Fink MP. Endotoxemia causes ileal mucosal acidosis in the absence of mucosal hypoxia in a normodynamic porcine model of septic shock. Crit Care Med. 1995;23(7):1217–1226.
  13. Rosser DM, Stidwill RP, Jacobson D, Singer M. Oxygen tension in the bladder epithelium rises in both high and low cardiac output endotoxemic sepsis. J Appl Physiol. 1995;79(6):1878–1882.
  14. Sair M, Etherington PJ, Peter Winlove C, Evans TW. Tissue oxygenation and perfusion in patients with systemic sepsis. Crit Care Med. 2001;29(7):1343–1349.
  15. Friedman G, De Backer D, Shahla M, Vincent JL. Oxygen supply dependency can characterize septic shock. Intensive Care Med. 1998;24(2):118–123.
  16. Fink MP. Cytopathic hypoxia. Mitochondrial dysfunction as mechanism contributing to organ dysfunction in sepsis. Crit Care Clin. 2001;17(1):219–237.
  17. Fink MP. Cytopathic hypoxia. Is oxygen use impaired in sepsis as a result of an acquired intrinsic derangement in cellular respiration? Crit Care Clin. 2002;18(1):165–175.
  18. Crouser ED, Julian MW, Huff JE, Joshi MS, Bauer JA, Gadd ME, et al. Abnormal permeability of inner and outer mitochondrial membranes contributes independently to mitochondrial dysfunction in the liver during acute endotoxemia. Crit Care Med. 2004;32(2):478–488.
    Available from: http://dx.doi.org/10.1097/01.CCM.0000109449.99160.81.
  19. d’Avila JdCP, Santiago APSA, Amâncio RT, Galina A, Oliveira MF, Bozza FA. Sepsis induces brain mitochondrial dysfunction. Crit Care Med. 2008;36(6):1925–1932.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e3181760c4b.
  20. Takasu O, Gaut JP, Watanabe E, To K, Fagley RE, Sato B, et al. Mechanisms of cardiac and renal dysfunction in patients dying of sepsis. Am J Respir Crit Care Med. 2013;187(5):509–517.
    Available from: http://dx.doi.org/10.1164/rccm.201211-1983OC.
  21. Vary TC. Sepsis-induced alterations in pyruvate dehydrogenase complex activity in rat skeletal muscle: effects on plasma lactate. Shock. 1996;6(2):89–94.
  22. Radi R, Rodriguez M, Castro L, Telleri R. Inhibition of mitochondrial electron transport by peroxynitrite. Arch Biochem Biophys. 1994;308(1):89–95.
    Available from: http://dx.doi.org/10.1006/abbi.1994.1013.
  23. Galley HF. Oxidative stress and mitochondrial dysfunction in sepsis. Br J Anaesth. 2011;107(1):57–64.
    Available from: http://dx.doi.org/10.1093/bja/aer093.
  24. Goldfarb RD, Marton A, Szabó E, Virág L, Salzman AL, Glock D, et al. Protective effect of a novel, potent inhibitor of poly(adenosine 5’-diphosphate-ribose) synthetase in a porcine model of severe bacterial sepsis. Crit Care Med. 2002;30(5):974–980.
  25. Logan DC. The mitochondrial compartment. J Exp Bot. 2006;57(6):1225–1243.
    Available from: http://dx.doi.org/10.1093/jxb/erj151.
  26. Saraste M. Oxidative phosphorylation at the fin de siècle. Science. 1999;283(5407):1488–1493.
  27. Adeva-Andany M, López-Ojén M, Funcasta-Calderón R, Ameneiros-Rodríguez E, Donapetry-García C, Vila-Altesor M, et al. Comprehensive review on lactate metabolism in human health. Mitochondrion. 2014;17:76–100.
    Available from: http://dx.doi.org/10.1016/j.mito.2014.05.007.
  28. Brière JJ, Favier J, Gimenez-Roqueplo AP, Rustin P. Tricarboxylic acid cycle dysfunction as a cause of human diseases and tumor formation. Am J Physiol Cell Physiol. 2006;291(6):C1114–C1120.
    Available from: http://dx.doi.org/10.1152/ajpcell.00216.2006.
  29. Jeger V, Djafarzadeh S, Jakob SM, Takala J. Mitochondrial function in sepsis. Eur J Clin Invest. 2013;43(5):532–542.
    Available from: http://dx.doi.org/10.1111/eci.12069.
  30. Chretien D, Rustin P. Mitochondrial oxidative phosphorylation: pitfalls and tips in measuring and interpreting enzyme activities. J Inherit Metab Dis. 2003;26(2-3):189–198.
  31. Barrientos A, Fontanesi F, Díaz F. Evaluation of the mitochondrial respiratory chain and oxidative phosphorylation system using polarography and spectrophotometric enzyme assays. Curr Protoc Hum Genet. 2009;Chapter 19:Unit 19.3.
    Available from: http://dx.doi.org/10.1002/0471142905.hg1903s63.
  32. Brandt S, Regueira T, Bracht H, Porta F, Djafarzadeh S, Takala J, et al. Effect of fluid resuscitation on mortality and organ function in experimental sepsis models. Crit Care. 2009;13(6):R186.
    Available from: http://dx.doi.org/10.1186/cc8179.
  33. Kozlov AV, Duvigneau JC, Miller I, Nürnberger S, Gesslbauer B, Kungl A, et al. Endotoxin causes functional endoplasmic reticulum failure, possibly mediated by mitochondria. Biochim Biophys Acta. 2009;1792(6):521–530.
    Available from: http://dx.doi.org/10.1016/j.bbadis.2009.03.004.
  34. Kozlov AV, van Griensven M, Haindl S, Kehrer I, Duvigneau JC, Hartl RT, et al. Peritoneal inflammation in pigs is associated with early mitochondrial dysfunction in liver and kidney. Inflammation. 2010;33(5):295–305.
    Available from: http://dx.doi.org/10.1007/s10753-010-9185-4.
  35. Corrêa TD, Vuda M, Blaser AR, Takala J, Djafarzadeh S, Dünser MW, et al. Effect of treatment delay on disease severity and need for resuscitation in porcine fecal peritonitis. Crit Care Med. 2012;40(10):2841–2849.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e31825b916b.
  36. Corrêa TD, Vuda M, Takala J, Djafarzadeh S, Silva E, Jakob SM. Increasing mean arterial blood pressure in sepsis: effects on fluid balance, vasopressor load and renal function. Crit Care. 2013;17(1):R21.
    Available from: http://dx.doi.org/10.1186/cc12495.
  37. Corrêa TD, Jeger V, Pereira AJ, Takala J, Djafarzadeh S, Jakob SM. Angiotensin II in septic shock: effects on tissue perfusion, organ function, and mitochondrial respiration in a porcine model of fecal peritonitis. Crit Care Med. 2014;42(8):e550–e559.
    Available from: http://dx.doi.org/10.1097/CCM.0000000000000397.
  38. Aguirre E, López-Bernardo E, Cadenas S. Functional evidence for nitric oxide production by skeletal-muscle mitochondria from lipopolysaccharide-treated mice. Mitochondrion. 2012;12(1):126–131.
    Available from: http://dx.doi.org/10.1016/j.mito.2011.05.010.
  39. Chance B, Willaims GR. Respiratory enzymes in oxidative phosphorylation: III. The steady state. J Biol Chem. 1955;217:409–427.
  40. DePalma RG, Harano Y, Robinson AV, Holden WD. Structure and function of hepatic mitochondria in hemorrhage and endotoxemia. Surg Forum. 1970;21:3–6.
  41. Mela L, Bacalzo L Jr, White R 4th, Miller LD. Shock induced alterations of mitochondrial energy-linked functions. Surg Forum. 1970;21:6–8.
  42. Schumer W, Das Gupta TK, Moss GS, Nyhus LM. Effect of endotoxemia on liver cell mitochondria in man. Ann Surg. 1970;171(6):875–882.
  43. Duran-Bedolla J, Montes de Oca-Sandoval MA, Saldaña-Navor V, Villalobos-Silva JA, Rodriguez MC, Rivas-Arancibia S. Sepsis, mitochondrial failure and multiple organ dysfunction. Clin Invest Med. 2014;37(2):E58–E69.
  44. Singer M. The role of mitochondrial dysfunction in sepsis-induced multi-organ failure. Virulence. 2014;5(1):66–72.
    Available from: http://dx.doi.org/10.4161/viru.26907.
  45. Levy RJ. Mitochondrial dysfunction, bioenergetic impairment, and metabolic down-regulation in sepsis. Shock. 2007;28(1):24–28.
    Available from: http://dx.doi.org/10.1097/01.shk.0000235089.30550.2d.
  46. Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. N Engl J Med. 2003;348(2):138–150.
    Available from: http://dx.doi.org/10.1056/NEJMra021333.
  47. Fredriksson K, Hammarqvist F, Strigård K, Hultenby K, Ljungqvist O, Wernerman J, et al. Derangements in mitochondrial metabolism in intercostal and leg muscle of critically ill patients with sepsis-induced multiple organ failure. Am J Physiol Endocrinol Metab. 2006;291(5):E1044–E1050.
    Available from: http://dx.doi.org/10.1152/ajpendo.00218.2006.
  48. Belikova I, Lukaszewicz AC, Faivre V, Damoisel C, Singer M, Payen D. Oxygen consumption of human peripheral blood mononuclear cells in severe human sepsis. Crit Care Med. 2007;35(12):2702–2708.
  49. Fredriksson K, Fläring U, Guillet C, Wernerman J, Rooyackers O. Muscle mitochondrial activity increases rapidly after an endotoxin challenge in human volunteers. Acta Anaesthesiol Scand. 2009;53(3):299–304.
    Available from: http://dx.doi.org/10.1111/j.1399-6576.2008.01851.x.
  50. Sjövall F, Morota S, Hansson MJ, Friberg H, Gnaiger E, Elmér E. Temporal increase of platelet mitochondrial respiration is negatively associated with clinical outcome in patients with sepsis. Crit Care. 2010;14(6):R214.
    Available from: http://dx.doi.org/10.1186/cc9337.
  51. Japiassú AM, Santiago APSA, d’Avila JdCP, Garcia-Souza LF, Galina A, Castro Faria-Neto HC, et al. Bioenergetic failure of human peripheral blood monocytes in patients with septic shock is mediated by reduced F1Fo adenosine-5’-triphosphate synthase activity. Crit Care Med. 2011;39(5):1056–1063.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e31820eda5c.
  52. Garrabou G, Morén C, López S, Tobías E, Cardellach F, Miró O, et al. The effects of sepsis on mitochondria. J Infect Dis. 2012;205(3):392–400.
    Available from: http://dx.doi.org/10.1093/infdis/jir764.
  53. Sjövall F, Morota S, Persson J, Hansson MJ, Elmér E. Patients with sepsis exhibit increased mitochondrial respiratory capacity in peripheral blood immune cells. Crit Care. 2013;17(4):R152. Available from: http://dx.doi.org/10.1186/cc12831.
  54. Quoilin C, Mouithys-Mickalad A, Lécart S, Fontaine-Aupart MP, Hoebeke M. Evidence of oxidative stress and mitochondrial respiratory chain dysfunction in an in vitro model of sepsis-induced kidney injury. Biochim Biophys Acta. 2014;1837(10):1790–1800.
    Available from: http://dx.doi.org/10.1016/j.bbabio.2014.07.005.
  55. Protti A, Fortunato F, Artoni A, Lecchi A, Motta G, Mistraletti G, et al. Platelet mitochondrial dysfunction in critically ill patients: comparison between sepsis and cardiogenic shock. Crit Care. 2015;19:39.
    Available from: http://dx.doi.org/10.1186/s13054-015-0762-7.
  56. Peruchi BB, Petronilho F, Rojas HA, Constantino L, Mina F, Vuolo F, et al. Skeletal muscle electron transport chain dysfunction after sepsis in rats. J Surg Res. 2011;167(2):e333–e338. Available from: http://dx.doi.org/10.1016/j.jss.2010.11.893.
  57. Porta F, Takala J, Weikert C, Bracht H, Kolarova A, Lauterburg BH, et al. Effects of prolonged endotoxemia on liver, skeletal muscle and kidney mitochondrial function. Crit Care. 2006;10(4):R118.
    Available from: http://dx.doi.org/10.1186/cc5013.
  58. Comim CM, Rezin GT, Scaini G, Di-Pietro PB, Cardoso MR, Petronilho FC, et al. Mitochondrial respiratory chain and creatine kinase activities in rat brain after sepsis induced by cecal ligation and perforation. Mitochondrion. 2008;8(4):313–318.
    Available from: http://dx.doi.org/10.1016/j.mito.2008.07.002.
  59. Fry DE, Kaelin CR, Giammara BL, Rink RD. Alterations of oxygen metabolism in experimental bacteremia. Adv Shock Res. 1981;6:45–54.
  60. Tavakoli H, Mela L. Alterations of mitochondrial metabolism and protein concentrations in subacute septicemia. Infect Immun. 1982;38(2):536–541.
  61. Garrison RN, Ratcliffe DJ, Fry DE. The effects of peritonitis on murine renal mitochondria. Adv Shock Res. 1982;7:71–76.
  62. Geller ER, Jankauskas S, Kirkpatrick J. Mitochondrial death in sepsis: a failed concept. J Surg Res. 1986;40(5):514–517.
  63. Dawson KL, Geller ER, Kirkpatrick JR. Enhancement of mitochondrial function in sepsis. Arch Surg. 1988;123(2):241–244.
  64. Kopprasch S, Hörkner U, Orlik H, Kemmer C, Scheuch DW. Energy state, glycolytic intermediates and mitochondrial function in the liver during reversible and irreversible endotoxin shock. Biomed Biochim Acta. 1989;48(9):653–659.
  65. Takeyama N, Itoh Y, Kitazawa Y, Tanaka T. Altered hepatic mitochondrial fatty acid oxidation and ketogenesis in endotoxic rats. Am J Physiol. 1990;259(4 Pt 1):E498–E505.
  66. Llesuy S, Evelson P, González-Flecha B, Peralta J, Carreras MC, Poderoso JJ, et al. Oxidative stress in muscle and liver of rats with septic syndrome. Free Radic Biol Med. 1994;16(4):445–451.
  67. Taylor DE, Ghio AJ, Piantadosi CA. Reactive oxygen species produced by liver mitochondria of rats in sepsis. Arch Biochem Biophys. 1995;316(1):70–76.
    Available from: http://dx.doi.org/10.1006/abbi.1995.1011.
  68. Malaisse WJ, Nadi AB, Ladriere L, Zhang TM. Protective effects of succinic acid dimethyl ester infusion in experimental endotoxemia. Nutrition. 1997;13(4):330–341.
  69. Kantrow SP, Taylor DE, Carraway MS, Piantadosi CA. Oxidative metabolism in rat hepatocytes and mitochondria during sepsis. Arch Biochem Biophys. 1997;345(2):278–288.
    Available from: http://dx.doi.org/10.1006/abbi.1997.0264.
  70. Markley MA, Pierro A, Eaton S. Hepatocyte mitochondrial metabolism is inhibited in neonatal rat endotoxaemia: effects of glutamine. Clin Sci (Lond). 2002;102(3):337–344.
  71. Fukumoto K, Pierro A, Spitz L, Eaton S. Neonatal endotoxemia affects heart but not kidney bioenergetics. J Pediatr Surg. 2003;38(5):690–693.
    Available from: http://dx.doi.org/10.1016/jpsu.2003.50184.
  72. Suliman HB, Welty-Wolf KE, Carraway M, Tatro L, Piantadosi CA. Lipopolysaccharide induces oxidative cardiac mitochondrial damage and biogenesis. Cardiovasc Res. 2004;64(2):279–288.
    Available from: http://dx.doi.org/10.1016/j.cardiores.2004.07.005.
  73. Nin N, Cassina A, Boggia J, Alfonso E, Botti H, Peluffo G, et al. Septic diaphragmatic dysfunction is prevented by Mn(III)porphyrin therapy and inducible nitric oxide synthase inhibition. Intensive Care Med. 2004;30(12):2271–2278.
    Available from: http://dx.doi.org/10.1007/s00134-004-2427-x.
  74. Kozlov AV, Staniek K, Haindl S, Piskernik C, Ohlinger W, Gille L, et al. Different effects of endotoxic shock on the respiratory function of liver and heart mitochondria in rats. Am J Physiol Gastrointest Liver Physiol. 2006;290(3):G543–G549.
    Available from: http://dx.doi.org/10.1152/ajpgi.00331.2005.
  75. Larche J, Lancel S, Hassoun SM, Favory R, Decoster B,
    Marchetti P, et al. Inhibition of mitochondrial permeability transition prevents sepsis-induced myocardial dysfunction and mortality. J Am Coll Cardiol. 2006;48(2):377–385.
    Available from: http://dx.doi.org/10.1016/j.jacc.2006.02.069.
  76. Mason KE, Stofan DA. Endotoxin challenge reduces aconitase activity in myocardial tissue. Arch Biochem Biophys. 2008;469(2):151–156.
    Available from: http://dx.doi.org/10.1016/j.abb.2007.10.018.
  77. Protti A, Carré J, Frost MT, Taylor V, Stidwill R, Rudiger A, et al. Succinate recovers mitochondrial oxygen consumption in septic rat skeletal muscle. Crit Care Med. 2007;35(9):2150–2155.
  78. Kozlov AV, Gille L, Miller I, Piskernik C, Haindl S, Staniek K, et al. Opposite effects of endotoxin on mitochondrial and endoplasmic reticulum functions. Biochem Biophys Res Commun. 2007;352(1):91–96.
    Available from: http://dx.doi.org/10.1016/j.bbrc.2006.10.180.
  79. Duvigneau JC, Piskernik C, Haindl S, Kloesch B, Hartl RT, Hüttemann M, et al. A novel endotoxin-induced pathway: upregulation of heme oxygenase 1, accumulation of free iron, and free iron-mediated mitochondrial dysfunction. Lab Invest. 2008;88(1):70–77.
    Available from: http://dx.doi.org/10.1038/labinvest.3700691.
  80. Hassoun SM, Marechal X, Montaigne D, Bouazza Y, Decoster B, Lancel S, et al. Prevention of endotoxin-induced sarcoplasmic reticulum calcium leak improves mitochondrial and myocardial dysfunction. Crit Care Med. 2008;36(9):2590–2596.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e3181844276.
  81. Vanasco V, Cimolai MC, Evelson P, Alvarez S. The oxidative stress and the mitochondrial dysfunction caused by endotoxemia are prevented by alpha-lipoic acid. Free Radic Res. 2008;42(9):815–823.
    Available from: http://dx.doi.org/10.1080/10715760802438709.
  82. Reynolds CM, Suliman HB, Hollingsworth JW, Welty-Wolf KE, Carraway MS, Piantadosi CA. Nitric oxide synthase-2 induction optimizes cardiac mitochondrial biogenesis after endotoxemia. Free Radic Biol Med. 2009;46(5):564–572.
    Available from: http://dx.doi.org/10.1016/j.freeradbiomed.2008.11.007.
  83. Vanasco V, Magnani ND, Cimolai MC, Valdez LB, Evelson P, Boveris A, et al. Endotoxemia impairs heart mitochondrial function by decreasing electron transfer, ATP synthesis and ATP content without affecting membrane potential. J Bioenerg Biomembr. 2012;44(2):243–252.
    Available from: http://dx.doi.org/10.1007/s10863-012-9426-3.
  84. Li CM, Chen JH, Zhang P, He Q, Yuan J, Chen RJ, et al. Continuous veno-venous haemofiltration attenuates myocardial mitochondrial respiratory chain complexes activity in porcine septic shock. Anaesth Intensive Care. 2007;35(6):911–919.
  85. Regueira T, Bänziger B, Djafarzadeh S, Brandt S, Gorrasi J, Takala J, et al. Norepinephrine to increase blood pressure in endotoxaemic pigs is associated with improved hepatic mitochondrial respiration. Crit Care. 2008;12(4):R88.
    Available from: http://dx.doi.org/10.1186/cc6956.
  86. Vuda M, Brander L, Schröder R, Jakob SM, Takala J, Djafarzadeh S. Effects of catecholamines on hepatic and skeletal muscle mitochondrial respiration after prolonged exposure to faecal peritonitis in pigs. Innate Immun. 2012;18(2):217–230.
    Available from: http://dx.doi.org/10.1177/1753425911398279.
  87. Marshall JC, Deitch E, Moldawer LL, Opal S, Redl H, van der Poll T. Preclinical models of shock and sepsis: what can they tell us? Shock. 2005;24 Suppl 1:1–6.
  88. Dyson A, Singer M. Animal models of sepsis: why does preclinical efficacy fail to translate to the clinical setting? Crit Care Med. 2009;37(1 Suppl):S30–S37.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e3181922bd3.
  89. Marshall JC. From the bedside back to the bench: the role of preclinical studies in understanding clinical therapies. Crit Care Med. 2010;38(1):329–330.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e3181b9d4b4.
  90. Mouncey PR, Osborn TM, Power GS, Harrison DA, Sadique MZ, Grieve RD, et al. Trial of early, goal-directed resuscitation for septic shock. N Engl J Med. 2015;372(14):1301–1311.
    Available from: http://dx.doi.org/10.1056/NEJMoa1500896.
  91. Russell JA, Singer J, Bernard GR, Wheeler A, FulkersonW, Hudson L, et al. Changing pattern of organ dysfunction in early human sepsis is related to mortality. Crit Care Med. 2000;28(10):3405–3411.
  92. Copeland S,Warren HS, Lowry SF, Calvano SE, Remick D, Inflammation and the Host Response to Injury Investigators. Acute inflammatory response to endotoxin in mice and humans. Clin Diagn Lab Immunol. 2005;12(1):60–67.
  93. Goldfarb RD, Dellinger RP, Parrillo JE. Porcine models of severe sepsis: emphasis on porcine peritonitis. Shock. 2005;24 Suppl 1:75–81.
  94. Buras JA, Holzmann B, Sitkovsky M. Animal models of sepsis: setting the stage. Nat Rev Drug Discov. 2005;4(10):854–865.
    Available from: http://dx.doi.org/10.1038/nrd1854.
  95. Singleton KD,Wischmeyer PE. Distance of cecum ligated influences mortality, tumor necrosis factor-alpha and interleukin-6 expression following cecal ligation and puncture in the rat. Eur Surg Res. 2003;(6):486–491.
    Available from: http://dx.doi.org/10.1159/000073387.
  96. Cuesta JM, Singer M. The stress response and critical illness: a review. Crit Care Med. 2012;40(12):3283–3289.
    Available from: http://dx.doi.org/10.1097/CCM.0b013e31826567eb.
  97. Revelly JP, Liaudet L, Frascarolo P, Joseph JM, Martinet O, Markert M. Effects of norepinephrine on the distribution of intestinal blood flow and tissue adenosine triphosphate content in endotoxic shock. Crit Care Med. 2000;28(7):2500–2506.
  98. Le Gall JR, Klar J, Lemeshow S, Saulnier F, Alberti C, Artigas A, et al. The Logistic Organ Dysfunction system. A new way to assess organ dysfunction in the intensive care unit. ICU Scoring Group. JAMA. 1996;276(10):802–810.
  99. Moreno R, Vincent JL, Matos R, Mendonça A, Cantraine F, Thijs L, et al. The use of maximum SOFA score to quantify organ dysfunction/failure in intensive care. Results of a prospective, multicentre study. Working Group on Sepsis related Problems of the ESICM. Intensive Care Med. 1999;25(7):686–696.
  100. Fredriksson K, Rooyackers O. Mitochondrial function in sepsis: respiratory versus leg muscle. Crit Care Med. 2007;35(9 Suppl):S449–S453.
    Available from: http://dx.doi.org/10.1097/01.CCM.0000278048.00896.4B.
  101. Sjövall F, Morota S, Frostner EÅ, Hansson MJ, Elmér E. Cytokine and nitric oxide levels in patients with sepsis–temporal evolvement and relation to platelet mitochondrial respiratory function. PLoS One. 2014;9(7):e103756.
    Available from: http://dx.doi.org/10.1371/journal.pone.0103756.
  102. Merz TM, Hefti JP, Hefti U, Huber A, Jakob SM, Takala J, et al. Changes in mitochondrial enzymatic activities of monocytes during prolonged hypobaric hypoxia and influence of antioxidants: A randomized controlled study. Redox Rep. 2015.
    Available from: http://dx.doi.org/10.1179/1351000215Y.0000000007.
  103. Djafarzadeh S, Vuda M, Takala J, Ochs M, Jakob SM. Toll-like receptor-3-induced mitochondrial dysfunction in cultured human hepatocytes. Mitochondrion. 2011;11(1):83–88.
    Available from: http://dx.doi.org/10.1016/j.mito.2010.07.010.
  104. Djafarzadeh S, Vuda M, Takala J, Jakob SM. Effect of remifentanil on mitochondrial oxygen consumption of cultured human hepatocytes. PLoS One. 2012;7(9):e45195.
    Available from: http://dx.doi.org/10.1371/journal.pone.0045195.
  105. Zheng G, Lyu J, Huang J, Xiang D, Xie M, Zeng Q. Experimental treatments for mitochondrial dysfunction in sepsis: A narrative review. J Res Med Sci. 2015;20(2):185–195.

Cite this article as follows:

Author. Title. Critical Care Horizons 2015; 1: 31-41.


Optic Nerve Sheath Ultrasound for the Bedside Diagnosis of Intracranial Hypertension: Pitfalls and Potential

Optic Nerve Sheath Ultrasound for the Bedside Diagnosis of Intracranial Hypertension: Pitfalls and Potential

Claire Shevlin1
1Locum Consultant, Intensive Care Unit, Craigavon Area Hospital, 68 Lurgan Road, Portadown, United Kingdom

Email: claireshevlin@gmail.com

Full-Text PDF

AbstractFull-TextReference ListCitation

Raised intracranial pressure is a complication not just of traumatic brain injury and other acute cerebral insults, but also of a number of general medical conditions. Bedside diagnosis can be difficult; early clinical signs may be misinterpreted, and reliance on cross-sectional imaging studies may further delay diagnosis. Ultrasound is a readily available imaging modality in most critical care areas, and examination of the optic nerve sheath by bedside ultrasound allows detection of changes in diameter which may indicate intracranial hypertension. This paper reviews optic nerve sheath anatomy as a basis for its potential to provide a window on changes within the intracranial cerebrospinal fluid (CSF) space, the technique of sonographic measurement of optic nerve sheath diameter (ONSD), the evidence for correlation with intracranial pressure, and the comparison of ONSD with the ‘gold standard’ method of intracranial pressure assessment.

Keywords: Intracranial hypertension; Bedside ultrasound; Optic nerve sheath diameter; Ocular ultrasound
Competing Interests: The author declares that she has no competing interests


Bedside ultrasound scanning, particularly echocardiography and thoracoabdominal sonography, is widely used in the intensive care unit and emergency department to exclude life-threatening pathology. The use of ultrasound of the optic nerve and its sheath to diagnose time-critical raised intracranial pressure following traumatic brain injury is less widely practiced, but has been described in several small studies over the past two decades.

The diameter of the optic nerve sheath has been found to be a strong predictor of raised intracranial pressure, with a high sensitivity and specificity in multiple studies and in a systematic review. Raised intracranial pressure is a common emergency following brain injury, with prompt diagnosis having a significant impact on morbidity and mortalityUltrasound measurement of the optic nerve sheath diameter (ONSD) allows repeated noninvasive assessments of intracranial pressure and facilitates evaluation of the response to treatment. As invasive intracranial pressure monitoring is typically restricted to neurosurgical centres, this mode of investigation is particularly suited to patients suspected of raised intracranial pressure prior to transfer for definitive treatment, as well as patients who continue to be cared for in non-neurosurgical critical care units. In addition to its diagnostic goal, there is some, albeit limited, evidence to suggest that ONSD can also be used for prognostication,.

The majority of articles on ONSD measurement are in the setting of raised intracranial pressure secondary to traumatic brain injury. However, a few studies have used this measurement to diagnose or assess the severity of other pathologies, including meningitis, stroke, hepatic encephalopathy, epilepsy, and acute mountain sickness,,,,.

The potential value of this technique is reflected in the significant number of studies performed to date. Unfortunately, most have small patient numbers and hence low power, and could be criticised for potential observer bias. Efforts are ongoing to define the ONSD indicating ‘true’ raised intracranial pressure, the best sonographic approach to visualise the optic nerve sheath, the optimum axis to assess the optic nerve sheath, and the impact of operator experience on measurement variability.

Search Strategy

PubMed and Medline databases from 1950 through to April 2015 were searched, using the OvidSP interface. The reference terms used (along with synonyms and closely related words) were optic nerve sheath, optic nerve sheath diameter, optic ultrasound, optic nerve sonography, and optic nerve ultrasonography. These terms were searched for separately and in combination with the terms intracranial pressure and papilloedema. Further studies were identified by examining the reference lists of included articles. This yielded a total of over 100 articles. After review, forty-two articles were judged to be directly relevant to the topic discussed and are referenced.

The Anatomy of the Optic Nerve Sheath

The intraorbital section of the optic nerve extends from the globe, where it inserts medially, to the optic canal located in the lesser wing of the sphenoid bone. It is encased by a meningeal sheath consisting of dura mater, arachnoid mater and pia mater. Cerebrospinal fluid is contained in the trabeculated subarachnoid space and is continuously and slowly filtered. As a result the optic nerve sheath is in direct communication with the intracranial subarachnoid space. It is this relationship that forms the physiological basis for using the optic nerve sheath as a surrogate for intracranial pressure measurement. The anatomical relationships underpinning the use of ultrasound to measure ONSD can be readily appreciated on MRI (Figure 1).

Figure 1. MRI anatomy of the optic nerve and sheath Case courtesy of Dr Frank Gaillard, Radiopaedia.org. From the case Optic nerve and chiasm (MRI anatomy).

Figure 1.MRI anatomy of the optic nerve and sheath.

Case courtesy of Dr Frank Gaillard, Radiopedia.org. From the case Optic nerve and chiasm (MRI anatomy).


The optic nerve sheath is bound more loosely to the optic nerve closer to the globe. This loose binding creates a much larger, and potentially more distensible, subarachnoid space in this region, which can appear bulbous on ultrasound. While papilloedema may take time to develop, dilation of the optic nerve sheath occurs much earlier and may be a nearinstantaneous manifestation of raised intracranial pressure,.

The History of Imaging the Optic Nerve

The first report of ultrasound imaging of the eye was in 1956, but it was early cadaver studies which implicated the optic nerve sheath in the measurement of intracranial pressure. One such study noted the “bulbous portion of the optic nerve was seen to bulge or inflate somewhat as the intracranial pressure was created” with the infusion of crystalloid into the brain. The authors also noted this appeared to occur anteriorly, where the nerve sheath was at its thinnest and most expandable. These early studies did not measure the ONSD, relying on imprecise visual clues, and were hampered by the limited ultrasound modes available. These early modes made it difficult to locate a distinct point for measurement at a reproducible distance behind the globe.

As ultrasound modalities improved, the focus of most studies was the optimum distance behind the globe at which to best measure ONSD. A 1996 study using modern ultrasonographic techniques showed that ONSD increased by up to 60% at a distance of 3mm behind the globe in comparison to only 35% at 10 mm12. This has been confirmed in subsequent studies, indicating that a position 3 mm behind the globe is preferred for measurement. Measurements made at this point are more reproducible since ultrasound contrast is greater at this depth with a linear probe. Consistent with this, the optic nerve sheath is at its most distensible anteriorly, where it is potentially most reflective of raised intracranial pressure.

The Sonographic Appearance of the Optic Nerve Sheath

On ultrasound, the globe is visualised as a round, dark, fluid filled structure (see Figure 2). The anterior chamber is anechoic, as generally is the lens, while the iris appears bright and echogenic. The choroid and retina may be seen as a thin grey layer at the posterior aspect of the globe. The optic nerve is the ‘black stripe’ running away from the posterior aspect of the globe and optic disc, and should ideally be positioned in the centre of the ultrasound screen. The nerve sheath, as seen on ultrasound examination, has a high reflectivity compared to the homeogenous appearance of the nerve, and should be relatively easy to distinguish.

If the optic nerve sheath is markedly dilated, it may be possible to diagnose this from visual estimation alone. In general, however, the software calipers should be used to ensure accurate measurement and recording. In severely raised intracranial pressure, it may be possible to visualise a ‘crescent sign’, an echolucent circular artefact within the sheath separating the sheath from the nerve due to increased subarachnoid fluid.

There has been interest in using contrast enhanced ultrasound (CEUS) to help identify and recognise the anatomy surrounding the optic nerve, which is a small structure. The incorrect identification of artefacts as part of the sheath by an inexperienced sonographer is a criticism of the technique. A small proof of concept study, using a second generation contrast agent (Sonovue®, Bracco SpA), found good correlation between CEUS and MRI. This study suggests, by using nontoxic contrast, exact measurements can be more quickly and easily delineated, which may lessen the effect of operator inexperience.

Figure 2. Ultrasound anatomy on transverse imaging

Figure 2. Ultrasound anatomy on transverse imaging

Pros and Cons of Measurement of ONSD by Ultrasound

Sonographic ONSD assessment brings some clinical advantages but also some downsides that need to be considered when adopting the technique. Advantages include:

  • reproducibility of measurements
  • the non-invasive nature of the technique
  • ready availability of equipment
  • portability of equipment
  • rapid performance
  • relatively low costs
  • avoidance of ionising radiation
  • avoidance of patient transport for imaging

The primary clinical disadvantage, given the relative novelty of the technique, lies in the ongoing lack of a uniform cut-off value for the diagnosis of raised intracranial pressure (see below). Practical disadvantages are manageable and relate primarily to the need to acquire competence in the scanning technique to optimise accuracy, the potential risk of pressure injury to the globe if technique is poor, and the potential for injury resulting from thermal and non-thermal effects of ultrasound.

Ultrasound is generally acknowledged to be a safe technique. The largely hypothetical risks of ultrasound centre on the potential biological consequences of interaction between the scanned tissues and the ultrasound wave. These consequences may be thermal or non-thermal, and are measured by the safety indices Thermal Index (TI) and Mechanical Index (MI), which are displayed in real-time on the screen of most modern ultrasound machines. Ultrasound is presumed to be safe when the values of the TI and MI are less than 1.0. The TI is the ratio of the power used to the power required to produce a temperature rise of 1°C18.

Ultrasound energy from the probe passes into scanned tissues and is reflected from tissue interfaces; some energy is absorbed and converted to thermal energy, elevating the temperature of local tissues. Scanning time should be minimised to prevent possible thermal injury. It is advised that tissue temperature increase should be kept below 1.5 °C. The MI gives an approximate figure of the risk of the non-thermal effects. These include cavitation, which is the expansion and contraction of tissue gas bubbles during the cycle, and streaming, referring to the movement of complex fluids brought about by the ultrasound energy. The MI gives an approximate figure of the risk of the non-thermal effects. Optic nerve sheath ultrasound should not be used in the presence of evident or suspected rupture of the globe, or when there is significant periorbital injury. The technique is likely to be of limited incremental value in patients with chronically raised intracranial pressure or long-standing papilloedema.

The Technique of Optic Nerve Sheath Ultrasound

Although individual clinicians may vary in certain aspects of their examination technique, there are some general principles which will help optimise ocular ultrasound for assessment of the ONSD:

  • Select the high frequency linear array probe on the ultrasound machine as this provides the best compromise between footprint and resolution of superficial structures. 
  • Apply ultrasound gel liberally to the closed eyelid. If desired, a clear thin dressing (e.g. IV cannula dressing) can be used as a barrier between the closed eyelid and the gel medium although this is not strictly necessary.
  • Resting the probe hand on a bony structure such as the forehead or brow ridge stabilises the image and lowers the risk of inadvertent pressure on the globe. 
  • Place the ultrasound probe lightly over the gel in a transverse orientation initially. There should neither be any direct contact of the probe with the eyelid nor pressure exerted on the globe. The probe marker should be orientated laterally (Figure 3a).
  • With small, subtle movements scan from side to side (i.e. temporal to nasal), slowly angling the probe superiorly or inferiorly to bring the optic nerve into view. The nerve will appear as a ‘black stripe’ running posteriorly from the rear of the globe. The goal is to centre this on the monitor. If the lens or iris is not seen in your image, the imaging plane is likely off-axis and may result in an underestimation of ONSD.
  • The globe should also be scanned in the parasagittal plane, with the probe marker superiorly, towards the patient’s forehead (Figure 3b).
  • Both eyes should be scanned, in case of unilateral papilloedema. 
  • The time spent in active scanning should be minimised. Once the optimum view has been obtained, store the image either as a frame or a video loop and remove the probe from the eye. Measurements can then be performed without unnecessary exposure of the eye to ultrasound energy.
  • Use the caliper function on the ultrasound to enable precise measurement. First locate a point 3 mm posterior to the optic disk. At this point place the calipers at 90 degrees to the axis of the optic nerve to measure the diameter of optic nerve and optic nerve sheath (Figure 4).
  • Take the average of two or three measurements for each side.

A 1996 study by Helmke & Hansen suggested the optimal scanning orientation was longitudinal (axial), as this was associated with the least inter-observer variability12. However, aside from the variability findings, there was no significant difference in measurements between the two orientations.

Most patients will be scanned supine, or with a 20° to 30° head up tilt. A Nepalese study, which included 287 patients, examined the correlation between ONSD and acute mountain sickness. This study suggested ONSD does not change with patient positioning9. This was supported by results from a healthy adult study by Romagnuolo. In that study three investigators measured the ONSD in 10 separate volunteers and concluded the diameter measured by ultrasound does not change significantly with either standard Trendelenburg or reverse Trendelenburg, in comparison with a baseline supine position. The data on the impact of body position on ONSD should not be extrapolated beyond the clinical settings which have been studied, and more work remains to be done.

Figure 3. Ultrasound probe application for ONSD measurement. (a) transverse (b) parasagittal.

Figure 3. Ultrasound probe application for ONSD measurement. (a) transverse (b) parasagittal.


Figure 4. Caliper positions for ONSD measurement

Figure 4. Caliper positions for ONSD measurement

Differential Diagnosis

Although uncommon in critical care practice, there are alternative causes for a rise in ONSD and these should be kept in mind when diagnosing raised intracranial pressure in a patient with increased ONSD. The differential diagnosis of increased ONSD includes:

  • raised intracranial pressure
  • optic neuritis
  • arachnoid cyst of the optic nerve 
  • anterior orbital masses
  • cavernous sinus masses
  • trauma to the optic nerve
  • optic nerve sheath meningioma

Establishing the ONSD Cut-Off for Diagnosis of Raised Intracranial Pressure

Evidence from multiple studies supports correlation between ONSD and opening pressure of cerebrospinal fluid or intracranial pressure in both adults and children2,14,,,. 

Table 1 summarises the main clinical studies reporting an association between ONSD and intracranial pressure. Some studies show a high sensitivity and negative predictive value for the detection of further increases in intracranial pressure2,. A more recent study attempted to investigate the impact of brainstem death on ONSD, as compared to patients in a state of coma. In this study, ONSD was measured in 29 brainstem dead patients, 19 comatose patients (11 with indications of raised intracranial pressure), 20 patients with established neurological disease and 40 healthy subjects as controls. While the comatose and brainstem dead patients showed a markedly increased ONSD in comparison to the other two groups, there was limited evidence to suggest brainstem death could be reliably distinguished from coma on the basis of ONSD.

On review of earlier studies, a tentative association emerged between an ONSD of 5.0 mm and the presence of raised intracranial pressure. In a study of 59 emergency department patients, Tayal2 found an ONSD of 5.0 mm had a 100% sensitivity to detect patients with raised intracranial pressure. Another emergency department study by Qayyum, found a sensitivity and specificity of 100% and 75%, respectively, for a cutoff of 5.0 mm, with positive and negative predictive values of 94.5 and 100%23

However, there remains controversy about the exact diameter of the optic nerve sheath that best predicts elevated intracranial pressure. The bulk of studies recruited less than 100 patients and indicate a 5 mm cut-off. However, other studies have found evidence for different optimal cut-offs. Rajajee and colleagues used 5.2 mm as their upper limit of normal measurement, Soldatos used 5.7 mm, and Bäuerle used 5.8 mm20. The variation in cut-off across studies is significant, and ranges from 4.8 mm to 6.0 mm25,,,.

This makes it more difficult to establish a ‘set’ measurement above which raised intracranial pressure can be diagnosed based on ONSD. There are multiple explanations for the differences in the upper limit of normal. The studies from 1996-2015 focused on heterogeneous patient populations. The majority of studies were done on ICU and emergency department patients and compared a control group with a group that might reasonably be expected to have a high intracranial pressure. Not all studies attempted to stratify results by gender or ventilatory status, both factors which may affect results. A significant proportion were performed in medical inpatients requiring a lumbar puncture for neurological disease, or in outpatients undergoing treatment for intracranial hypertension. While it was standard in most studies (where stated) to use a 7.5 MHz linear ultrasound probe, this was not always the case. Lastly, the ‘confirmation’ of raised intracranial pressure was done with a variety of different methods – MRI, CT, invasive intracranial pressure monitors, and CSF opening pressure. This makes comparison between studies difficult. 

Of note, ethnic differences may need to be taken into account when measuring ONSD as a surrogate measure of ICP. A study in Chinese patients correlated an elevated opening pressure on lumbar puncture with a significantly lower ONSD than in Caucasian populations. 

ONSD measurement has not been widely examined in paediatric patients and no strong evidence to diagnose raised intracranial pressure exists. One study, in 64 paediatric patients, reported a very low specificity of ONSD for raised intracranial pressure.

Also, a small porcine study noted a lack of correlation between ONSD changes and central venous pressure. It is possible ONSD alone is a poor indicator of change in intracranial pressure when the rise in intracranial pressure is solely influenced by a rise in central venous pressure.

Comparison of ONSD by Ultrasound with Other Methods of Assessing Intracranial Pressure

Studies of sonographic ONSD have mainly correlated findings with clinical and radiological signs and symptoms of elevated intracranial pressure. Intraventricular measurement is the gold standard for measuring intracranial pressure. These measurement devices carry many risks, including haemorrhage and infection. These complications partly account for increasing interest in non invasive methods such as neuroimaging, transcranial Doppler sonography, ONSD ultrasonography and CT/MRI. In addition, invasive methods may not be available to individual patients due to contraindications, such as coagulopathy, thrombocytopaenia, or the lack of local facilities.

The correlation of ONSD with intraventricular pressure monitoring has not been widely examined, probably because of the invasive nature of the latter and the limited pool of patients available. Kimberly correlated ventriculostomy intracranial pressure measurements with ONSD21. They found a positive correlation with a Spearman rank correlation coefficient of ONSD and ICP of 0.59 (p < 0.0005). Similarly, the area under the ROC curve for ability of ONSD to distinguish an ICP greater than 20 cm H2O was 0.93. Their results indicated a reasonably high specificity of 93% and sensitivity of 88%. 

This compares well with the data from clinical and radiological studies. For example, Blaivas correlated ONSD with radiological signs of raised intracranial pressure, as defined by the presence of mass effect with midline shift greater or equal to 3 mm, a collapsed third ventricle, hydrocephalus or the effacement of sulci with evidence of significant oedema, and found a sensitivity of 100% and specificity of 95% for ONSD compared to CT, with positive and negative predictive values of 93% and 100% respectively. With regard to neuroimaging, Soldatos found good correlation of ONSD with CT findings in brain-injured adults (r = 0.68, P = 0.002)26. MRI is often spoken of as a reference test for ONSD, as it has higher spatial resolution, and the images offer a more representative calculation of the mean diameter, than CT. While Lagrèze and colleagues contend the accuracy of MRI exceeds sonographic methods for determining ONSD, Bauerle showed good scan-rescan reproducibility and good observer agreement in 15 healthy volunteers20.

Author (ref)Patients (n)Comparison groups (mean/range ONSD)Optimal cut-off (sensitivity/specificity)Population
Hansen and Helmke 1997 [11]39Control range 2.7-4 mm.
Patients with raised ICP 3.6-6.8 mm.
5.0 mmICU, children. 7.5 MHz probe
Blaivas et al 2003 [33]35Patients without raised ICP, mean 4.42 mm.
Patients with raised ICP, mean 6.27 mm.
5.0 mm. Sensitivity 100%, specificity 95%.ED, adults. 10 MHz probe
Geeraerts et al 2007 [27]31Patients without raised ICP, up to 5.1 mm.
Patients with raised ICP, up to 6.3 mm.
5.7 mm. Sensitivity 100%ICU, adults. 7.5 MHz probe
Tayal et al 2007 [2]59Patients with raised ICP, >5mm.5.0 mm. Sensitivity 100%, specificity 63%ED, adults. 7.5 MHz probe
Kimberly et al 2008 [21]15Patients without raised ICP, mean 4.4 mm.
Patients with raised ICP, mean 5.4 mm.
5.0 mm. Sensitivity 88%, specificity 93%ED/ICU adults. 10 MHz probe
Soldatos et al 2008 [26]76Control without raised ICP, mean 3.6 mm.
Patients with raised ICP, mean 6.1 mm.
5.7 mm. Sensitivity 74%, specificity 100%ICU, adults
Geeraerts et al 2008 [28]37Head injury patients requiring ICP monitor5.9 mm, AUC ROC 0.91. Sensitivity 90%, specificity 84%ICU, adults. 7.5 MHz probe
Goel et al 2008 [34]100Patients without raised ICP, mean 3.5 mm.
Patients with raised ICP, mean 5.8 mm.
5.0 mm. Sensitivity 98.6%, specificity 92.8%Trauma, adults. 7.5 MHz probe
Watanabe 2008 [35]12Pre-v-post-op chronic subdural collections.
Mean 6.1 mm pre-op, 4.8 mm post-op.
No cut-off givenNeurosurgery. MRI used
Moretti et al 2009 [36]535.2 mm. Sensitivity 93%, specificity 74%ICU, adults. 7.5 MHz probe
Moretti et al 2009 [22]63Controls, mean 4.9 mm.
Patients without raised ICP, mean 5 mm.
Patients with raised ICP, mean 6.2 mm.
5.2 mm. Sensitivity 94%, specificity 76%ICU, adults. 7.5 MHz probe
Bäuerle et al 2011 [20]10Controls, mean 5.4 mm.
Patients with raised ICP, mean 6.4 mm.
5.8 mm. Sensitivity 90%, specificty 84%Non-ICU neuro. Adults
Rajajee et al 2011 [25]65Patients without raised ICP, mean 4.0 mm.
Patients with raised ICP 5.3 mm.
4.8 mm. Sensitivity 96%, specificity 94%ICU, adults
Strumwasser et al 2011 [29]10Trauma patients requiring ICP monitoring6.0 mm. Sensitivity 26%, specificity 38%ICU, adults
Cammarata et al 2011 [37]21Controls, mean 5.51 mm.
Patients without raised ICP, mean 5.52 mm.
Patients with raised ICP, mean 7 mm.
No cut-off stated. r=0.74 for correlation of ONSD with ICPICU, adults
Amini et al 2013 [38]50Patients without raised ICP, mean 4.6 mm.
Patients with raised ICP, mean 6.7 mm.
5.5 mm. Sensitivity 100%, specificity 100%Neuro, adults
Qayyum et al 2013 [23]24Patients with suspected raised ICP.
Compared against CT findings.
5.0 mm. Sensitivity 100%, specificity 75%. Positive predictive value 95.4%ED, adults
Caffery et al 2014 [39]51ONSD against opening pressure on LP with
20 cmH2O chosen
5 mm pre-selected cut-off. Sensitivity 75%, specificity 44%.
AUC on ROC 0.69
ED, adults. Non-trauma
Shirodkar 2014 [6]101Control, male 4.8 mm, female 4.6 mm.
Patient with rasied ICP, mean 5.4 mm.
4.6 mm (females): sensitivity 84.6%, specificity 100%.
4.8 mm (males): sensitivity 75%, specificity 100%.
Medical, adults
Mehrpour 2015 [40]32Patients with raised ICP, mean 6.24 mm5.7 mm. Sensitivity 100%Neuro, adults. 7.5 MHz probe
Wang et al 2015 [30]279No high opening pressure 3.55 mm.
With high opening pressure 4.58 mm.
4.1 mm. Sensitivity 95%, specificity 92%.
AUC ROC curve 0.965
Neuro, Chinese adults. 9-3 probe
Topcuoglu et al 2015 [24]29Controls, mean 4.69 mm.
Neurological disease, mean 4.36 mm.
Coma raised ICP, mean 5.89 mm.
Coma no raised ICP, mean 5.16 mm.
Brainstem dead, mean 6.09 mm.
No cut-off (attempt to identify ONSD difference between brainstem dead and coma with raised ICP)ICU, adults
Table 1. Summary of clinical studies investigating ONSD


Transcranial doppler (TCD) has also been investigated, but sonography windows were often inadequate. While some studies have generally found it to be a poor predictor of intracranial pressure, others have found the method based on two depth TCD more reliable than ONSD measurement,,. This last study enrolled 92 patients and compared both non-invasive methods of measurement with CSF pressure obtained by lumbar puncture, using a cut-off for ICP of 14.7 mmHg and 5 mm for ONSD. The sensitivity and specificity of TCD were better than ONSD (sensitivity 37% for ONSD, 68% for TCD, specificity 58.5% for ONSD, 84.3% for TCD). The area under the ROC curve was 0.57 for ONSD and 0.87 for TCD. Skill acquisition, acoustic window availability, and ease of use may favour ultrasound assessment of ONSD by critical care clinicians over TCD, in spite of potential diagnostic advantages of TCD.

ONSD in Practice

From the available evidence, it is reasonable to conclude that optic nerve ultrasound has utility as part of the noninvasive assessment of intracranial pressure, and may offer potential means of reducing complications associated with intraventricular intracranial pressure monitors by better targeting these devices. Naturally, the technique will not be viable in some patients, such as those with severe ocular trauma, and it must be borne in mind that a number of aetiologies other than raised intracranial pressure can lead to dilatation of the optic nerve sheath.

Increased ONSD may also be of prognostic significance. Legrand found that ONSD, as measured on brain CT in 77 patients, was independently associated with ICU mortality, albeit at greater than 7 mm4. It may be argued the main value of this technique lies in early evaluation, during the initial assessment and resuscitation or transport phases, where cross-sectional imaging is unavailable. Similarly, the technique may be of value in bedside assessment in the ICU, although studies are needed to determine if this is a viable strategy to reduce the need for CT or MRI imaging.

For medical patients at risk of raised intracranial pressure, such as those with liver failure, this straightforward bedside assessment may be helpful in risk stratification and decision-making around the timing of, and need for, CT imaging. This assessment would also be beneficial in these patients prior to procedures such as central line insertion, as part of the decision-making process around target vein and patient position during access.


The studies to date have generally been small and poorly powered with the inevitable result that many questions remain unanswered. There is a clear opportunity for the critical care community to collaborate on larger scale studies to evaluate the potential impact of pathways using ultrasound assessment of ONSD. 

ONSD values greater than 5 mm, and certainly greater than 5.8 mm, have been shown to be highly specific and sensitive for the presence of raised intracranial pressure. Elevated ICP should be considered in the presence of an ONSD greater than 5mm, and if greater than 5.5 mm urgent consideration should be given to medical management pending further diagnostic workup. In adolescents and non-Caucasian populations, the thresholds are likely to vary and additional data is required before recommendations can be made. 

It is important that physicians using this technique are proficient in clinical ultrasound to minimise errors and adverse effects due to poor scanning technique. Fortunately ONSD measurement by ultrasound is readily taught. For governance, training, and quality monitoring purposes the indication for performance of ONSD assessment should be recorded along with the technique and findings. These should be reviewed in light of results of additional investigations such as CT or MRI.


  1. Dubourg J, Javouhey E, Geeraerts T, Messerer M, Kassai B. Ultrasonography of optic nerve sheath diameter for detection of raised intracranial pressure: a systematic review and meta-analysis. Intensive Care Med. 2011;37(7):1059–1068.
    Available from: http://dx.doi.org/10.1007/s00134-011-2224-2.
  2. Tayal VS, Neulander M, Norton HJ, Foster T, Saunders T, Blaivas M. Emergency department sonographic measurement of optic nerve sheath diameter to detect findings of increased intracranial pressure in adult head injury patients. Ann Emerg Med. 2007;49(4):508–514.
    Available from: http://dx.doi.org/10.1016/j.annemergmed.2006.06.040.
  3. Hwan Kim Y, Ho Lee J, Kun Hong C,Won Cho K, Hoon Yeo J, Ju Kang M, et al. Feasibility of optic nerve sheath diameter measured on initial brain computed tomography as an early neurologic outcome predictor after cardiac arrest. Acad Emerg Med. 2014;21(10):1121–1128.
    Available from: http://dx.doi.org/10.1111/acem.12477.
  4. Legrand A, Jeanjean P, Delanghe F, Peltier J, Lecat B, Dupont H. Estimation of optic nerve sheath diameter on an initial brain computed tomography scan can contribute prognostic information in traumatic brain injury patients. Crit Care. 2013;17(2):R61.
    Available from: http://dx.doi.org/10.1186/cc12589.
  5. Nabeta HW, Bahr NC, Rhein J, Fossland N, Kiragga AN, Meya DB, et al. Accuracy of noninvasive intraocular pressure or optic nerve sheath diameter measurements for predicting elevated intracranial pressure in cryptococcal meningitis. Open Forum Infect Dis. 2014;1(3):ofu093.
    Available from: http://dx.doi.org/10.1093/ofid/ofu093.
  6. Shirodkar CG, Rao SM, Mutkule DP, Harde YR, Venkategowda PM, Mahesh MU. Optic nerve sheath diameter as a marker for evaluation and prognostication of intracranial pressure in Indian patients: An observational study. Indian J Crit Care Med. 2014;18(11):728–734.
    Available from: http://dx.doi.org/10.4103/0972-5229.144015.
  7. Kim YK, Seo H, Yu J, Hwang GS. Noninvasive estimation of raised intracranial pressure using ocular ultrasonography in liver transplant recipients with acute liver failure -A report of two cases-. Korean J Anesthesiol. 2013;64(5):451–455.
    Available from: http://dx.doi.org/10.4097/kjae.2013.64.5.451.
  8. Manno E, Motevallian M, Mfochive A, Navarra M. Ultrasonography Of The Optic Nerve Sheath Suggested Elevated Intracranial Pressure In Epilepsy: Case Report. Internet J Anesthesiol. 2013;26(1).
    Available from: https://ispub.com/IJA/26/1/11214.
  9. Fagenholz PJ, Gutman JA, Murray AF, Noble VE, Camargo CA Jr, Harris NS. Optic nerve sheath diameter correlates with the presence and severity of acute mountain sickness: evidence for increased intracranial pressure. J Appl Physiol (1985). 2009;106(4):1207–1211.
    Available from: http://dx.doi.org/10.1152/japplphysiol.01188.2007.
  10. Hansen HC, Helmke K. The subarachnoid space surrounding the optic nerves. An ultrasound study of the optic nerve sheath. Surg Radiol Anat. 1996;18(4):323–328.
  11. Hansen HC, Helmke K. Validation of the optic nerve sheath response to changing cerebrospinal fluid pressure: ultrasound findings during intrathecal infusion tests. J Neurosurg. 1997;87(1):34–40.
    Available from: http://dx.doi.org/10.3171/jns.1997.87.1.0034.
  12. Helmke K, Hansen HC. Fundamentals of transorbital sonographic evaluation of optic nerve sheath expansion under intracranial hypertension. I. Experimental study. Pediatr Radiol. 1996;26(10):701–705.
  13. Liu D, Kahn M. Measurement and relationship of subarachnoid pressure of the optic nerve to intracranial pressures in fresh cadavers. Am J Ophthalmol. 1993;116(5):548–556.
  14. Newman WD, Hollman AS, Dutton GN, Carachi R. Measurement of optic nerve sheath diameter by ultrasound: a means of detecting acute raised intracranial pressure in hydrocephalus. Br J Ophthalmol. 2002;86(10):1109–1113.
  15. Marchese RF, Mistry RD, Scarfone RJ, Chen AE. Identification of optic disc elevation and the crescent sign using point-of-care ocular ultrasound in children. Pediatr Emerg Care. 2015;31(4):304–307.
    Available from: http://dx.doi.org/10.1097/PEC.0000000000000408.
  16. Bergauer A, Prosen G, Flis V, Seruga T, Brvar M, Kobilica N. Contrast enhanced ultrasound imaging of the optic nerve sheath diameter – what are we really measuring? Crit Ultrasound J. 2012;4((Suppl 1)):A2.
    Available from: http://dx.doi.org/10.1186/2036-7902-4-S1-A2.
  17. Joy J, Cooke I, Love M. Is ultrasound safe? The Obstetrician & Gynaecologist. 2006;8(4):222–227.
    Available from: http://dx.doi.org/10.1576/toag.
  18. Bigelow TA, Church CC, Sandstrom K, Abbott JG, Ziskin MC, Edmonds PD, et al. The thermal index: its strengths, weaknesses, and proposed improvements. J Ultrasound Med. 2011;30(5):714–734.
  19. Romagnuolo L, Tayal V, Tomaszewski C, Saunders T, Norton HJ. Optic nerve sheath diameter does not change with patient position. Am J Emerg Med. 2005;23(5):686–688.
    Available from: http://dx.doi.org/10.1016/j.ajem.2004.11.004.
  20. Bäuerle J, Nedelmann M. Sonographic assessment of the optic nerve sheath in idiopathic intracranial hypertension. J Neurol. 2011;258(11):2014–2019.
    Available from: http://dx.doi.org/10.1007/s00415-011-6059-0.
  21. Kimberly HH, Shah S, Marill K, Noble V. Correlation of optic nerve sheath diameter with direct measurement of intracranial pressure. Acad Emerg Med. 2008;15(2):201–204.
    Available from: http://dx.doi.org/10.1111/j.1553-2712.2007.00031.x.
  22. Moretti R, Pizzi B, Cassini F, Vivaldi N. Reliability of optic nerve ultrasound for the evaluation of patients with spontaneous intracranial hemorrhage. Neurocrit Care. 2009;11(3):406–410. Available from: http://dx.doi.org/10.1007/s12028-009-9250-8.
  23. Qayyum H, Ramlakhan S. Can ocular ultrasound predict intracranial hypertension? A pilot diagnostic accuracy evaluation in a UK emergency department. Eur J Emerg Med. 2013;20(2):91–97.
    Available from: http://dx.doi.org/10.1097/MEJ.0b013e32835105c8.
  24. Topcuoglu MA, Arsava EM, Bas DF, Kozak HH. Transorbital Ultrasonographic Measurement of Optic Nerve Sheath Diameter in Brain Death. J Neuroimaging. 2015;p. Epub ahead of print. Available from: http://dx.doi.org/10.1111/jon.12233.
  25. Rajajee V, Vanaman M, Fletcher JJ, Jacobs TL. Optic nerve ultrasound for the detection of raised intracranial pressure. Neurocrit Care. 2011;15(3):506–515.
    Available from: http://dx.doi.org/10.1007/s12028-011-9606-8.
  26. Soldatos T, Karakitsos D, Chatzimichail K, Papathanasiou M, Gouliamos A, Karabinis A. Optic nerve sonography in the diagnostic evaluation of adult brain injury. Crit Care. 2008;12(3):R67. Available from: http://dx.doi.org/10.1186/cc6897.
  27. Geeraerts T, Launey Y, Martin L, Pottecher J, Vigué B, Duranteau J, et al. Ultrasonography of the optic nerve sheath may be useful for detecting raised intracranial pressure after severe brain injury. Intensive Care Med. 2007;33(10):1704–1711.
    Available from: http://dx.doi.org/10.1007/s00134-007-0797-6.
  28. Geeraerts T, Merceron S, Benhamou D, Vigué B, Duranteau J. Non-invasive assessment of intracranial pressure using ocular sonography in neurocritical care patients. Intensive Care Med. 2008;34(11):2062–2067.
    Available from: http://dx.doi.org/10.1007/s00134-008-1149-x.
  29. Strumwasser A, Kwan RO, Yeung L, Miraflor E, Ereso A, Castro-Moure F, et al. Sonographic optic nerve sheath diameter as an estimate of intracranial pressure in adult trauma. J Surg Res. 2011;170(2):265–271.
    Available from: http://dx.doi.org/10.1016/j.jss.2011.03.009.
  30. Wang L, Feng L, Yao Y, Wang Y, Chen Y, Feng J, et al. Optimal optic nerve sheath diameter threshold for the identification of elevated opening pressure on lumbar puncture in a Chinese population. PLoS One. 2015;10(2):e0117939.
    Available from: http://dx.doi.org/10.1371/journal.pone.0117939.
  31. Le A, Hoehn ME, Smith ME, Spentzas T, Schlappy D, Pershad J. Bedside sonographic measurement of optic nerve sheath diameter as a predictor of increased intracranial pressure in children. Ann Emerg Med. 2009;53(6):785–791.
    Available from: http://dx.doi.org/10.1016/j.annemergmed.2008.11.025.
  32. Hamilton DR, Sargsyan AE, Melton SL, Garcia KM, Oddo B, Kwon DS, et al. Sonography for determining the optic nerve sheath diameter with increasing intracranial pressure in a porcine model. J Ultrasound Med. 2011;30(5):651–659.
  33. Blaivas M, Theodoro D, Sierzenski PR. Elevated intracranial pressure detected by bedside emergency ultrasonography of the optic nerve sheath. Acad Emerg Med. 2003;10(4):376–381. Available from: http://dx.doi.org/10.1111/j.1553-2712.2003.tb01352.x.
  34. Goel RS, Goyal NK, Dharap SB, Kumar M, Gore MA. Utility of optic nerve ultrasonography in head injury. Injury. 2008;39(5):519–524.
    Available from: http://dx.doi.org/10.1016/j.injury.2007.09.029.
  35. Watanabe A, Kinouchi H, Horikoshi T, Uchida M, Ishigame K. Effect of intracranial pressure on the diameter of the optic nerve sheath. J Neurosurg. 2008;109(2):255–258.
    Available from: http://dx.doi.org/10.3171/JNS/2008/109/8/0255.
  36. Moretti R, Pizzi B. Optic nerve ultrasound for detection of intracranial hypertension in intracranial hemorrhage patients: confirmation of previous findings in a different patient population. J Neurosurg Anesthesiol. 2009;21(1):16–20.
    Available from: http://dx.doi.org/10.1097/ANA.0b013e318185996a.
  37. Cammarata G, Ristagno G, Cammarata A, Mannanici G, Denaro C, Gullo A. Ocular ultrasound to detect intracranial hypertension in trauma patients. J Trauma. 2011;71(3):779–781. Available from: http://dx.doi.org/10.1097/TA.0b013e3182220673.
  38. Amini A, Kariman H, Arhami Dolatabadi A, Hatamabadi HR, Derakhshanfar H, Mansouri B, et al. Use of the sonographic diameter of optic nerve sheath to estimate intracranial pressure. Am J Emerg Med. 2013;31(1):236–239.
    Available from: http://dx.doi.org/10.1016/j.ajem.2012.06.025.
  39. Caffery TS, Perret JN, Musso MW, Jones GN. Optic nerve sheath diameter and lumbar puncture opening pressure in nontrauma patients suspected of elevated intracranial pressure. Am J Emerg Med. 2014;32(12):1513–1515.
    Available from: http://dx.doi.org/10.1016/j.ajem.2014.09.014.
  40. Mehrpour M, Oliaee Torshizi F, Esmaeeli S, Taghipour S, Abdollahi S. Optic nerve sonography in the diagnostic evaluation of pseudopapilledema and raised intracranial pressure: a cross-sectional study. Neurol Res Int. 2015;2015:146059.
    Available from: http://dx.doi.org/10.1155/2015/146059.
  41. LagrèzeWA, Lazzaro A,Weigel M, Hansen HC, Hennig J, Bley TA. Morphometry of the retrobulbar human optic nerve: comparison between conventional sonography and ultrafast magnetic resonance sequences. Invest Ophthalmol Vis Sci. 2007;48(5):1913–1917.
    Available from: http://dx.doi.org/10.1167/iovs.06-1075.
  42. Bolesch S, vonWegner F, Senft C, LorenzMW. Transcranial ultrasound to detect elevated intracranial pressure: comparison of septum pellucidum undulations and optic nerve sheath diameter. Ultrasound Med Biol. 2015;41(5):1233–1240.
    Available from: http://dx.doi.org/10.1016/j.ultrasmedbio.2014.12.023.
  43. Edouard AR, Vanhille E, Le Moigno S, Benhamou D, Mazoit JX. Non-invasive assessment of cerebral perfusion pressure in brain injured patients with moderate intracranial hypertension. Br J Anaesth. 2005;94(2):216–221.
    Available from: http://dx.doi.org/10.1093/bja/aei034.
  44. Ragauskas A, Bartusis L, Piper I, Zakelis R, Matijosaitis V, Petrikonis K, et al. Improved diagnostic value of a TCD-based non-invasive ICP measurement method compared with the sonographic ONSD method for detecting elevated intracranial pressure. Neurol Res. 2014;36(7):607–614.
    Available from: http://dx.doi.org/10.1179/1743132813Y.0000000308.

Cite this article as follows:

Shevlin C. Optic Nerve Sheath Ultrasound for the Bedside Diagnosis of Intracranial Hypertension: Pitfalls and Potential. Critical Care Horizons 2015; 1: 22-30.


Antimicrobial Therapeutics in Critical Care

Antimicrobial Therapeutics in Critical Care

Katie Fong1 and Ronan McMullan1,2
1Department of Medical Microbiology, Royal Victoria Hospital, 274 Grosvenor Road, Belfast BT12 6BA, United Kingdom
2Centre for Infection and Immunity, Queen’s University Belfast, 97 Lisburn Road, Belfast BT9 7BL, United Kingdom

Email: Katie.Fong@belfasttrust.hscni.net

Full-Text PDF

AbstractFull-TextReference ListCitation

The purpose of this manuscript is to summarise current issues in antimicrobial therapeutics, specifically relating to antibacterial drugs, in the critical care setting. Antimicrobial taxonomy, mode of action, pharmacokinetics and pharmacodynamics are important considerations in choosing the most appropriate agent for treatment of infections. Adverse effects of antibiotics can range from mild gastrointestinal disturbance to anaphylactic shock. Prompt escalation and appropriate de-escalation of antimicrobials are central to effective treatment of sepsis associated with important infections in critical care including ventilator-associated pneumonia (VAP) and catheter-related blood stream infection (CRBSI). Improved diagnostic techniques for earlier detection of pathogens may aid such stewardship of antibiotics. The emergence, and continued rise, of antibiotic-resistant bacteria in recent years has highlighted the importance of antimicrobial stewardship.

Keywords: Sepsis; Healthcare-associated infection; Antimicrobial resistance; Antibiotic stewardship

Antimicrobial Pharmacology

The aim of antimicrobial therapy is to achieve selective toxicity by inhibition of the microorganism without damage to the host. This is achieved by exploiting the differences between metabolism or structure of the microorganism and the corresponding features of human cells. Antimicrobials can be divided into different categories according to their mode of action. These include cell wall inhibition, nucleic acid synthesis inhibition, protein synthesis inhibition and cell membrane damage. Major drug classes, sites of action, antimicrobial spectra, and adverse effects are summarised in Figure 1 and Table 1.


Fong Figure 1

Figure 1. Antibacterial Sites of Action


DrugClass & Mode of ActionSpectrumAdverse Effects
BenzylpenicillinBeta-lactam (cell wall)Gram-positive, mainly streptococcalAll beta-lactams: Hypersensitivity (rashes, fever, eosinophilia, angio-oedema, serum sickness, anaphylaxis), vomiting, diarrhoea, convulsions, nephrotoxicity, cholestatic jaundice, Stevens-Johnson syndrome
FlucloxacillinBeta-lactam (cell wall)Gram-positive, mainly staphylococcal
AmoxicillinBeta-lactam (cell wall)Gram-positive, Non
beta-lactamase gram negative
Co-amoxiclavBeta-lactam (cell wall)Gram-positive, Gram-negative, Anaerobes
Piperacillin/tazobactamBeta-lactam (cell wall)Gram-positive, Gram-negative, Anaerobes
CephalosporinsBeta-lactam (cell wall)Gram-positive, Gram-negative, Anaerobes
CarbapenemsBeta-lactam (cell wall)Gram-positive, Gram-negative +
Pseudomonas spp, Anaerobes
VancomycinGlycopeptide (cell wall)Gram-positive + MRSARenal toxicity, red man syndrome, oto-toxicity, thrombocytopaenia, rash
GentamicinAminoglycoside (protein synthesis)Gram-negative + Pseudomonas
species, some Gram-positive
Renal toxicity, ototoxicity
ChloramphenicolAmphenicol (protein synthesis)Gram-positive, Gram-negative,
Marrow suppression, aplastic anaemia, hypersensitivity
LinezolidOxazolidinones (protein synthesis)Gram-positive + MRSA,
Haemopoietic disorders, optic neuropathy
TigecyclineGlycylcycline (protein synthesis)Gram-positive + MRSA,
Gram-negative, Anaerobes
Nausea and vomiting, hepatotoxicity
DaptomycinLipopeptide (cell membrane)Gram-positive + MRSAMyalgia, nausea and vomiting
CiprofloxacinFluoroquinolone (nucleic acid synthesis)Gram-negative including
Pseudomonas species
Diarrhoea, nausea and vomiting, tendon damage, seizures
Co-trimoxazoleAnti-folates (nucleic acid synthesis)Gram-positive, Gram-negative
+ Stenotrophomonas species,
Pneumocystis jiroveci
Haemopoietic disorders, Stevens-Johnson syndrome, nausea
MetronidazoleNitroimidazoles (nucleic acid synthesis)AnaerobesNausea, disulfiram-like reaction with alcohol, peripheral neuropathy
Table 1. Synopsis of drug classes, sites of action, antimicrobial spectra, and adverse effects.

Agents acting on the cell wall


All beta-lactam agents feature a beta-lactam ring and act on the bacterial cell wall by binding to proteins involved in cell wall construction. Benzylpenicillin, or penicillin G, has excellent activity against streptococci. However, it is easily degraded by gastric acid and has a short half-life; therefore, it must be given intravenously and at frequent intervals1.

Flucloxacillin is a narrow spectrum antimicrobial used to treat staphylococcal infections; it also has short half-life and is poorly absorbed following oral administration. It is not active against methicillin-resistant Staphylococcus aureus (MRSA), because of alteration to its cell target (penicillin-binding protein).

Amoxicillin is a further development of the penicillin family and has activity against non-beta-lactamase-producing Gram-positive and Gram-negative organisms. Several bacteria produce a beta-lactamase enzyme which hydrolyses the beta-lactam ring causing it to be ineffective. Clavulanic acid, added to amoxicillin in co-amoxiclav, inhibits this enzyme and hence breakdown of the beta-lactam. Piperacillin/tazobactam acts in a similar manner, with beta-lactamase inhibition provided by tazobactam, and has a broader spectrum of coverage, including activity against Pseudomonas species1(Figure 2).

Figure 2. Summary of Antibacterial Spectra

Figure 2. Summary of Antibacterial Spectra

The first generation cephalosporins (e.g. cefazolin) primarily have activity against Gram-positive organisms with a limited Gram-negative spectrum. The second (e.g. cefuroxime) and third generation (e.g. ceftriaxone) cephalosporins have improved Gram-negative cover and cannot be hydrolysed by all beta-lactamases, with a notable exception being the extended-spectrum beta-lactamases (ESBL). The fourth (e.g. cefepime) and fifth (e.g. ceftaroline) generations have activity against an extended Gram-negative spectrum, and MRSA, respectively2.

Carbapenems are structurally different from penicillins and cephalosporins. They are more stable as they are not affected by most beta lactamases. This class of beta-lactams are extremely broad spectrum with activity against Gram-negative bacilli which produce ESBL enzymes as well as Pseudomonas aeroginosa. However, carbapenem resistance is now emerging with a recent rise in carbapenemase-producing enterobacteriaceae.


Glycopeptides, like beta-lactams, act on the cell wall; however, they inhibit the last stages of cell wall assembly by preventing cross linking reactions between constituent peptidoglycans. Their antimicrobial spectrum is exclusively Gram-positive organisms. Intravenous teicoplanin and vancomycin are widely used for treatment of staphylococcal infections, including MRSA2.

Protein synthesis inhibitors

The protein synthesis inhibitors are constituted by a large number of antibiotics; perhaps the most commonly prescribed examples are macrolides and tetracyclines. A relatively new member of this group, tigecycline, has a very broad spectrum of activity which includes glycopeptide resistant enterococci (GRE). Its main use is for skin/soft tissue and intra-abdominal infections1. It is important to note that, because of its pharmacokinetics, tigecycline is not indicated in the treatment of septic shock, primary bacteraemia, and urinary tract infections.

Aminoglycosides also act by inhibiting protein synthesisand are bactericidal drugs that are especially active against Gram-negative bacilli. They exhibit synergy when used with beta-lactams to target both Gram-positive and -negative organisms; hence, they are useful in treatment of severe infections with bacteraemia.

Chloramphenicol acts against a broad range of organisms, however, it is indicated for treatment of only a limited number of infections because of toxicity concerns; these include meningitis in patients for whom beta-lactams are contraindicated e.g. severe allergy.

Linezolid and tedizolid are oxazolidinones with very broad Gram-positive activity. These antimicrobials also target protein synthesis and can be used to treat skin/soft tissue infections, especially when involving MRSA. Linezolid is effective in treating lower respiratory chest infections, particularly MRSA pneumonia.

Agents acting on the cell membrane

Daptomycin (a lipopeptide) is a bactericidal antibiotic acting on the cell membrane and, like the oxazolidinones, has a very broad Gram-positive spectrum of activity including MRSA and GRE. It is effective in treating bacteraemia due to Staphylococcus aureus and Enterococcus species.

Nucleic acid inhibitors

Fluoroquinolones act by inhibiting bacterial nucleic acid synthesis, providing broad Gram-negative cover, with some members having good Gram-positive activity. Ciprofloxacin is the most widely prescribed member of this group in critical care and has excellent oral bioavailability. It is the only oral antimicrobial with action against Pseudomonas species and also has activity against some ESBL-producing organisms. Levofloxacin has better Gram-positive cover because of its anti-streptococcal activity1.

Cotrimoxazole contains sulphamethoxazole and trimethoprim (in a 5:1 ratio), both of which inhibit DNA synthesis by acting on the folate pathway. It is used to treat a variety of bacterial, fungal and protozoal infections and is particularly useful in treatment of Pneumocystis jirovecii pneumonia and Stenotrophomonas maltophilia infections (such as catheter-related bacteraemia)1.

Metronidazole has cytotoxic effects on anaerobes, though its mode of action is not fully understood. It is well absorbed when administered orally and is a first-line treatment for non-severe Clostridium difficile infections.

Adverse Effects

Several antimicrobials have non-specific adverse effects (e.g. headache and gastro-intestinal disturbances) as well as specific side effects relating to certain antibiotics. Antimicrobial allergy may occur in the form of immediate or non-immediate (delayed) hypersensitivity reactions. Immediate reactions are Ig-E mediated whereas delayed reactions are non-IgE, or T-cell, mediated.


The penicillin family of drugs are generally well tolerated, but they have been associated with a wide range of hypersensitivity reactions. True penicillin allergy occurs in 7-23% of patients who give a history of penicillin allergy. IgE-mediated hypersensitivity reactions are the most feared adverse event and are a well-recognised effect attributed to beta-lactam agents, especially the penicillin family. These reactions manifest with urticaria, pruritus, hypotension, bronchospasm, and laryngeal oedema.

Non-immediate reactions may be immune-mediated or direct-toxicity-mediated, usually occurring more than one hour after drug administration. The main non-immediate reaction is maculopapular exanthem, particularly during treatment with a penicillin. High intravenous doses can cause direct central nervous system toxicity, with myoclonic jerks, seizures or coma; this is a particular problem in patients with renal impairment.

Platelet-mediated bleeding caused by ticarcillin and piperacillin is often duration-related and, while not particularly common, can be substantial. Incidence of delayed adverse reactions to beta-lactams increases sharply when parenteral treatment is extended beyond 2 weeks. Beta-Lactam-induced neutropenia occurs on average after 3 weeks of treatment, and may be due to either immunologic or toxic effects of treatment. Patients with a late adverse reaction to penicillin can safely be re-treated with penicillin subsequently, with close surveillance13.

Cephalosporins are considered safe for the patient who experiences a penicillin-induced maculopapular rash – but not an urticarial skin eruption, or other immediate reaction, indicating IgE-mediated allergy. Cross-reactivity between penicillins and second/third generation cephalosporins is low and may be even lower than the cross reactivity between penicillins and some unrelated antibiotics: anaphylaxis with cephalosporins is uncommon8. In life threatening infections such as bacterial meningitis, septicaemia, and severe respiratory tract infections, consideration should be given to using second and third generation cephalosporins, even in patients with a history of penicillin allergy, when the risk of different options is being weighed-up8.

Cephalosporin, clindamycin and fluoroquinolone use has been associated with diarrhoea and pseudo-membranous colitis,,. Uncommon, but important, adverse effects of fluoroquinolones include seizures, elevation of liver enzymes, and tendinopathy.

Meropenem demonstrates cross-reactivity with penicillin and is contraindicated for the patient with a history of immediate or accelerated hypersensitivity reactions to penicillin. A retrospective analysis demonstrated the incidence of patients with a reported or documented penicillin allergy experiencing an allergic-type reaction to a carbapenem was 11%, which is over five times greater than the risk in patients who were reportedly not allergic to penicillin.

Aztreonam, a monocyclic beta-lactam, does not appear to cross-react with penicillin, and has been safely administered to penicillin-allergic patients.


The most dramatic adverse effect of intravenous vancomycin is the red man syndrome: a non-immunologically mediated reaction consisting of pruritus and erythema involving the upper body, with or without hypotension. It appears to be dependent on dose, frequency of administration, and rate of infusion, and is thought to be mediated by histamine release.


Concerns regarding administration of aminoglycosides include nephrotoxicity, ototoxicity (both the auditory and vestibular components) and neuromuscular blockade (particularly in patients with myasthenia gravis). Factors contributing to these adverse effects include duration of therapy, age, liver disease, shock and the co-administration of drugs that have the potential to cause ototoxicity or nephrotoxicity.


Linezolid is a weak inhibitor of monoamine oxidase (MAO) and may potentially cause ‘serotonin syndrome’ in patients taking selective serotonin reuptake inhibitors (SSRIs), tricyclic antidepressants, MAO inhibitors, cocaine and other recreational drugs such as 3,4-methylenedioxymethamphetamine (MDMA). Linezolid has been notably associated with reversible myelosuppression which manifests in three ways: thrombocytopaenia, anaemia, and neutropenia. Patients should be monitored appropriately, especially those with pre-existing bone marrow suppression or diminished bone marrow reserves. For those being treated for longer than 14 days, the clinical benefit of linezolid in serious Gram-positive infections should be weighed against potential, but reversible, haematological effects. Linezolid-associated toxic optic neuropathy is recognised complication of prolonged treatment. These adverse reactions have led to the recommendation to limit treatment courses to a maximum of 28 days.

Dosing Strategies and Monitoring

Many antibiotic-related adverse events are precipitated by an extension of the drug’s normal pharmacology and are predictable complications of high serum levels; hence, many adverse reactions may be avoided by appropriate dosage adjustment. The major determinants of bacterial killing include the maximum serum concentration reached (Cmax), and the time the antibiotic concentration remains over an efficacy threshold such as the minimum inhibitory concentration (MIC), T>MIC. The area under the serum concentration-time curve (AUC) after a dose of antibiotic is a measure of total drug exposure and is reflected by both Cmax and T>MIC. Therefore, the AUC accounts for both how high, and for how long, the antibiotic levels remain above the MIC of the target organism during any one dosing interval2(Figure 3).

Figure 3. Pharmacodynamic parameters relevant to antimicrobial therapy

Figure 3. Pharmacodynamic parameters relevant to antimicrobial therapy

For time-dependent antibiotics, such as beta-lactams and glycopeptides, optimal bacterial kill is achieved by maximizing the duration of time drug concentration is greater than the MIC of the target organism for that particular antibiotic (T >MIC). For this reason, frequent dosing with penicillins is more effective than administering higher unit doses less frequently.

With concentration-dependent antibiotics, such as aminoglycosides, a high initial concentration is required to ensure maximum bacterial kill. The efficacy of these agents is related to the achievement of a high Cmax/MIC ratio, hence, gentamicin is usually best given in a large dose, once daily. This high initial concentration may also aid tissue penetration. For most antibiotics, however, the ratio of the AUC (over 24-h) to MIC of the target organism is most important in determining the extent of antibacterial effect. Therefore, for many antibiotics, adjusting either the unit dose or dosing interval increases the antibacterial effect.

The target vancomycin therapeutic trough concentration for treatment of infections is usually between 10-15 mg/litre26. A higher vancomycin concentration at 15-20 mg/litre is recommended for severe infections such as bacteraemia, endocarditis, osteomyelitis, meningitis and hospital-acquired pneumonia caused by Staphylococcus aureus. However, a systematic review suggested a strong relationship exists between higher vancomycin trough concentrations and nephrotoxicity. Patients with vancomycin troughs in excess of 15 mg/litre were found to have a greater risk of nephrotoxicity than those with concentrations of <15 mg/litre. Hence, there is a trade-off between efficacy and risk of toxicity, guided by the severity of infection.

Knowledge of antibiotic pharmacodynamic properties and the potential altered antibiotic pharmacokinetics in critically ill patients can allow clinicians to develop tailored dosing regimens. Measured creatinine clearance can be used to drive many dose adjustments for renally cleared drugs. Antibiotics can be broadly categorized according to their solubility characteristics which can help describe possible altered pharmacokinetics that can be caused by the pathophysiological changes common to critical illness. Hydrophilic antibiotics (e.g. beta-lactams, glycopeptides) are mostly affected with the pathophysiological changes observed in critically ill patients with increased volumes of distribution and altered drug clearance (related to changes in creatinine clearance). Lipophilic antibiotics (e.g. fluoroquinolone, tigecycline) have lesser volume of distribution alterations, but may develop altered drug clearances. Using antibiotic pharmacodynamic bacterial kill characteristics, dosing regimens can be altered to optimize treatment31.

Treatment of Patients with Sepsis

Mortality figures for severe sepsis and septic shock have commonly been quoted as ranging from 20% to 50%. A retrospective analysis of 17,990 patients with septic shock noted that the time to initiation of appropriate antimicrobial therapy was the strongest predictor of mortality. Intravenous antibiotic therapy should be initiated as a matter of urgency, ideally after obtaining appropriate cultures. The choice of antibiotics can be complex and should reflect the patient’s history (e.g. recent antibiotics received, comorbidities), the clinical context (such as community or hospital acquired infection), Gram stain result, previously cultured organisms, and local resistance patterns32. When the potential pathogen or infection source is not immediately obvious, broad-spectrum antimicrobial coverage directed against both Gram-positive and Gram-negative bacteria are favoured.

Combination therapy can be used to broaden the spectrum of antimicrobial coverage when used empirically to increase the chance of adequately targeting the causative organism. Secondly, some combinations may possess an enhanced potential for synergism, when compared with the additive effect of each of the antibiotics assessed separately. Synergy between betalactam antibiotics and aminoglycoside antibiotics has been shown in vitro for Gram-negative bacteria and specifically for Pseudomonas aeruginosa. A meta-analysis suggested combination antimicrobial therapy improves survival and clinical response of high-risk, life-threatening infections, particularly those associated with septic shock, but may be detrimental to low-risk patients.

In observational studies of almost 10,000 critically-ill patients with community-acquired pneumonia, macrolide use was associated with a significant relative reduction in mortality (18%) compared with non-macrolide therapies. These results suggest macrolides be considered first-line combination treatment in critically ill patients with community-acquired pneumonia and support current guidelines. Furthermore, it may be inferred this benefit is not limited to patients with atypical pathogens and a speculative mechanism for this effect is immunological.

Optimal management of patients with sepsis includes early goal-directed therapy, lung-protective ventilation, and adequate antimicrobials. Later in the course of sepsis, appropriate management may necessitate organ support and prevention of nosocomial infection. Studies focused on novel targets, mechanisms of action, and combination therapy may improve current treatment.

The Surviving Sepsis Campaign guideline clearly endorses de-escalation to the most appropriate single therapy as soon as the susceptibility profile is available32. A prospective observational study published recently concluded de-escalation therapy for severe sepsis and septic shock is a safe strategy associated with lower mortality; efforts to increase the practice of early de-escalation are, therefore, justified.

Treatment of Community-Acquired Pneumonia

Approximately 10% of hospitalized patients with CAP require ICU admission. The most common aetiologies of community-acquired pneumonia are Streptococcus pneumoniae, Staphylococcus aureus, Legionella species, Gram-negative bacilli and Haemophilus influenzae.

The Infectious Diseases Society of America (IDSA) recommends using a beta-lactam plus a fluoroquinolone for critically ill patients. Aztreonam and a respiratory fluoroquinolone is the recommended alternative regimen for penicillin allergic patients. Once the aetiology of CAP has been identified on the basis of reliable microbiological methods, antimicrobial therapy should be directed at that pathogen.

For all critically ill patients, coverage for S. pneumoniae and Legionella species should be ensured by using a potent anti-pneumococcal beta-lactam and either a macrolide or a fluoroquinolone43,. Macrolide use has been associated with decreased mortality in patients with severe sepsis due to pneumonia even with macrolide-resistant pathogens. A recent trial demonstrated non-inferiority of a beta-lactam alone compared with a beta-lactam and macrolide combination in moderately severe community-acquired pneumonia.

Reduction in patient exposure to antibiotics may limit the increasing rates of antimicrobial drug resistance, decrease cost, and improve patient adherence and tolerability. It has been suggested adults with mild to moderate community-acquired pneumonia can be safely and effectively treated with an antibiotic regimen of 7 days or less47. IDSA recommends patients with CAP should be treated for a minimum of 5 days, should be afebrile for 48–72 h, and should be clinically stable before discontinuation of therapy43.

Treatment of hospital and ventilator-associated pneumonias (HAP/VAP)

Ventilator-associated pneumonia (VAP) contributes up to half of all cases of hospital-acquired pneumonia, is estimated to occur in up to 27% all mechanically ventilated patients and is the most common nosocomial infection in this group,,. Risk of VAP is greatest during the first 5 days of ventilation, with the mean time period between intubation and development of VAP being 3.3 days. Over the years, the attributable risk of death from VAP has decreased and was recently estimated at 9-13%. Despite surveillance data from  recent years suggesting a much lower incidence of confirmed VAP, suspected VAP remains challenging and also drives use of empirical antimicrobials in intensive care units.

The type of organism that causes VAP usually depends on the duration of mechanical ventilation. In general, early VAP is caused by pathogens sensitive to antibiotics, whereas late onset VAP is caused by multi-drug resistant and more difficult to treat bacteria. These include MRSA, Acinetobacter baumanii, Pseudomonas aeroginosa and ESBL-producing organisms. However, this is by no means a hard and fast rule and is merely a guide to initial antibiotic therapy until further clinical information is available.

Not all organisms isolated from respiratory specimens should be regarded as pathogens that necessarily require therapy; they should be interpreted and treated in the light of the full clinical picture. The choice of empirical antibiotic therapy of patients with HAP in an individual unit should, ideally, be based on the nature and susceptibility patterns of the pathogens prevalent on that unit and should also take account of such variables as duration of hospital stay (i.e. early- or late-onset infection), recent administration of antimicrobial therapy and co-morbidities. Similarly, definitive therapy should be determined by culture and susceptibility test results. For patients with early-onset infections who have not previously received antimicrobial, and in the absence of other risk factors, the use of coamoxiclav would be appropriate. For patients who have recently received antibiotics and/or who have other risk factors, a third-generation cephalosporin, a fluoroquinolone or piperacillin/tazobactam may be appropriate.

Nebulised antimicrobial therapy may have an adjunctive role in treatment of VAP, although its incremental efficacy is uncertain. Direct pulmonary administration might be effective by achieving high local concentrations by nebulisation. Nebulised colistimethate sodium (CMS), used as adjunctive therapy for Gram-negative VAP, seems to be safe although in one study no beneficial effect on clinical outcomes was seen with adjunctive nebulised CMS for treatment of Gram-negative VAP.

Aminoglycosides have a suitable antimicrobial spectrum for the treatment of serious Gram-negative infections. However, their use in pneumonia is limited by the risks of toxicity and by poor penetration into infected lung tissues when administered intravenously. A recent study concluded that nebulised amikacin, for the treatment of Gram-negative pneumonia in mechanically ventilated patients, warrants further clinical evaluation. It is worth noting traditional breakpoints used to determine whether pathogens are susceptible or resistant are not readily interpretable in the nebulised setting.

Treatment of complicated intra-abdominal infections

The empiric use of antimicrobial regimens with broad spectrum activity against gram-negative organisms such as meropenem or piperacillin/tazobactam is recommended for patients with high-severity community-acquired intra-abdominal infection. Ciprofloxacin or aztreonam plus metronidazole is an alternative in the context of beta-lactam allergy, but addition of an agent effective against gram-positive cocci is recommended57. In these high-risk patients, antimicrobial regimens should be adjusted according to culture and susceptibility.

Antifungal therapy for patients with severe community acquired or healthcare-associated infection is recommended if Candida is grown from intra-abdominal cultures57. For the critically ill patient, initial therapy with an echinocandin (caspofungin, micafungin, or anidulafungin) instead of a triazole has recommended when invasive infection has been proven,.

Empiric therapy directed against glycopeptide-resistant Enterococcus faecium (GRE) is usually not recommended except when the patient is at very high risk for an infection due to this organism, such as a liver transplant recipient with an intraabdominal infection originating in the hepatobiliary tree or a patient known to be colonized with GRE57.

An appropriate source control procedure to drain infected foci, control ongoing peritoneal contamination by diversion or resection, and restoration of anatomic and physiological function is recommended for nearly all patients with intraabdominal infection57.

Antimicrobial therapy for established infection should be limited to 4–7 days, unless it is difficult to achieve adequate source control,. Longer durations of therapy have not been associated with improved outcome57. Appropriate diagnostic investigation should be undertaken In patients who have persistent or recurrent clinical evidence of intra-abdominal infection after 4–7 days of therapy. Antimicrobial therapy effective against the organisms initially identified should be continued. Extra-abdominal sources of infection and non-infectious inflammatory conditions should also be investigated if the patient is not experiencing a satisfactory clinical response to a microbiologically adequate initial empiric antimicrobial regimen.

Treatment of catheter-related bloodstream infection (CRBSI)

Catheter-related bloodstream infection (CRBSI) accounts for 10-20% of hospital-acquired infections in the UK and is associated with both increased ICU stay and mortality. Coagulase-negative staphylococci, Staphylococcus aureus, aerobic Gram-negative bacilli, and Candida albicans most commonly cause CRBSI62. After appropriate cultures are taken, empirical IV antimicrobial therapy should be initiated on the basis of clinical suspicion, the severity of the patient’s acute illness, underlying disease, and the potential pathogens involved. The central venous catheter (CVC) should be removed in most cases of non-tunnelled CVC-related bacteraemia62.

Coagulase-negative staphylococci, such as Staphylococcus epidermidis, are the most common cause of catheter-related infections. Five to seven days antimicrobial duration is recommended for uncomplicated CRBSI if the catheter is removed. If the catheter is retained, 10–14 days duration, in combination with intraluminal antibiotic therapy, is recommended62. Glycopeptide antibiotics are the usual choice for these infections1.

Patients with S. aureus CRBSI should have the infected catheter removed, and receive 2–6 weeks of antimicrobial therapy. Patients can be considered for the shorter duration of antimicrobial therapy (a minimum of 14 days) when62:

  • the patient is not diabetic or immunosuppressed
  • the infected catheter is removed
  • the patient has no prosthetic intravascular device
  • there is no evidence of endocarditis
  • fever and bacteraemia resolve within 72 hours after initiation of appropriate antimicrobial therapy
  • there is no evidence of metastatic infection

Resistant Organisms

Antimicrobial resistance is a problem that is becoming increasingly challenging and is a predictable consequence of antibiotic use. Methicillin-resistant Staphylococcus aureus (MRSA) is resistant
to almost all beta-lactam antimicrobials (with the exception of fifth generation cephalosporins). Therefore options are 
limited and glycopeptides are the mainstay of treatment for serious infections1. A trial in 2012 demonstrated improved clinical and microbiological outcomes of linezolid for MRSA nosocomial pneumonia compared with vancomycin. However a meta-analysis demonstrated linezolid and vancomycin have similar efficacy and safety profiles and concluded neither drug is superior for the treatment of hospital-acquired pneumonia. A randomised trial in 2006 showed that high dose daptomycin is non-inferior to standard therapy (beta lactam or glycopeptide) for treatment of Staphylococcus aureus bacteraemia. Similarly, daptomycin has not displaced betalactams and glycopeptides as first-line therapy; the relatively wide non-inferiority margin reported in this trial in conjunction with the principle of antibiotic conservation may have contributed to this.

In the UK, Glycopeptide resistant enterococci (GRE) were first detected in 1986. The most frequent site of colonisation is the bowel, hence, GRE may cause infections in patients with severe, and often complicated, intra-abdominal pathology in critical care. Treatment options are limited and include daptomycin or linezolid for patients who have bacteraemia, with tigecycline an option for patients with intra-abdominal infections who are not bacteraemic.

Extended spectrum beta-lactamases (ESBL) are enzymes carried by Gram-negative bacteria that mediate resistance to extended-spectrum cephalosporins. Carbapenems are the mainstay of therapy for these infections in critical care. Alternatives are needed because of the emergence of carbapenemase-producing enterobacteriaceae (CPE). Such alternatives may comprise aminoglycosides, colistin, and in some circumstances, tigecycline.

There has been a marked increase in the incidence of CPE in the UK in recent years. Over the last decade, CPE have spread throughout the world and are now endemic in healthcare facilities in many countries. Infections can be extremely difficult to treat, as these isolates are highly resistant. CPE are typically resistant to carbapenems, as well as many other classes of antimicrobials, leaving few choices for treatment of infections with combination therapy being usual. Mortality from infections due to Klebsiella pneumoniae carbapenemase (KPC)-producing Klebsiella pneumoniae, an exemplar CPE, has been reported to be at least 50%. The optimal antimicrobial therapy for infections with KPC-producing organisms has not been well established. In recent studies patients who received monotherapy with either colistin or tigecycline had poorer outcomes compared to those who had combination therapy,. The most commonly used combination therapy was either colistin or tigecycline with a carbapenem in isolates that are not fully resistant. Other combinations which have been used are colistin with tigecycline or colistin with an aminoglycoside71. Mortality was reported to be up to 50% for patients receiving tigecycline-gentamicin combination and up to 64% for tigecycline-colistin; in patients treated with tigecycline monotherapy, mortality was up to 80%. In vitro studies provide experimental evidence for the use of ceftazidime combined with avibactam (a novel synthetic beta-lactamase inhibitor) to treat infections due to certain types of CPE e.g. Klebsiella pneumoniae with OXA-48 enzyme.

Multiple drug-resistant Acinetobacter baumanii (MDRAB) may cause serious infections in critically ill patients for which colistin often remains the only therapeutic option. Addition of rifampicin to colistin is synergistic in vitro and has been widely reported to confer a clinical outcome benefit in cohort studies. However, a randomised trial concluded that 30-day mortality is not reduced by addition of rifampicin to colistin in serious MDRAB infections. Nonetheless, the increased rate of A. baumannii eradication with combination treatment could still bring clinical benefit by limiting transmission. The optimal dosing of colistin has not been established. A loading dose of 9 MU and a 9 MU twice-daily maintenance dose have been suggested to be beneficial without causing significant toxicity.

Antimicrobial Stewardship

Current evidence demonstrates that the widespread use of broad-spectrum antibiotics is associated with antimicrobial resistance (e.g. ESBL-producers, MRSA, CPE) and the induction of Clostridium difficile infection (CDI). Antimicrobial stewardship aims to achieve optimal clinical outcomes related to antimicrobial use, minimize toxicity, reduce the costs of health care, and limit the selection for antimicrobial resistant strains.

Early detection of the pathogen is essential for optimal antimicrobial treatment. A diagnostic test that could reduce time to adjustment of antibiotic treatment, in comparison with conventional blood cultures, would be favourable in critically ill patients with sepsis. SeptiFast®is a CE-marked multi-pathogen real-time PCR system capable of detecting DNA sequences of bacteria and fungi present in blood samples within a few hours; such a test has potential to impact stewardship processes. However, a recent systematic review and meta-analysis of diagnostic accuracy studies of SeptiFast®, in the context of suspected sepsis, concluded that firm recommendations about the clinical utility of SeptiFast®within this setting could not be made.

Early identification of pathogens from blood cultures using matrix-assisted laser desorption ionization time-offlight (MALDI-TOF) mass spectrometry may optimize the choice of empirical antibiotic therapy in the setting of bloodstream infections. A prospective observational study in 2010 demonstrated MALDI-TOF had an impact on clinical management in 35.1% of all Gram-negative bacteraemia cases, showing a greater utility than Gram’s stain reporting. Thus, MALDI-TOF could become a useful second step beside Gram stain in guiding the empirical treatment of patients with bloodstream infection.

Measurement of procalcitonin levels to guide antibiotic decisions in patients with respiratory tract infections and sepsis appears to reduce antibiotic exposure without worsening clinical outcomes. Procalcitonin (PCT)-based algorithms have been used to guide antibiotic therapy in several clinical settings. In a trial of a PCT-based algorithm to reduce antibiotic exposure in secondary peritonitis, antibiotics were discontinued if PCT was <1.0 ng/mL or decreased by 80%, with resolution of clinical signs. The PCT-based algorithm safely reduced antibiotic exposure in this study. A more recent study demonstrated that, in critically ill adults with undifferentiated infections, a PCT algorithm including 0.1 ng/ml cut-off did not achieve the expected 25% reduction in duration of antibiotic treatment. However, it is worth noting that a target of 25% reduction in antibiotic exposure is ambitious, especially in the critical care setting. 

Rapid biomarker-based exclusion of infection may improve antibiotic stewardship in VAP; however, successful validation of the usefulness of potential markers in this setting is rare. A prospective, multicentre study validated the ability of specific host inflammatory mediators to exclude pneumonia in patients with suspected VAP. The study concluded low bronchoalveolar lavage fluid levels of Interleukin-1-beta, in combination with interleukin-8, confidently excludes VAP and could form a rapid biomarker-based rule-out test, with potential to improve antibiotic stewardship.

The optimal duration of antimicrobial treatment for VAP has not been firmly established. A randomised trial published in 2003 compared 8 and 15 days of antibiotic treatment in microbiologically-proven VAP and reported comparable clinical effectiveness. Therefore, both excluding VAP when suspected and shortening duration of treatment in confirmed infection may be useful strategies to support stewardship. With current pressures to find alternatives to carbapenems, a study in 2013 reported that extended-infusion of cefepime provides increased clinical and economic benefits in the treatment of invasive Pseudomonas aeruginosa infections. This type of research is needed to allow confident diversification of antimicrobials in common use, to reduce selection pressure in key drug classes.


There are well-rehearsed points regarding antimicrobial use that every clinician should consider. As a general rule, there should be a clear indication for starting antibiotic treatment. Withholding antimicrobials should always be considered when they are not clearly necessary, though this is particularly difficult in severely ill patients. Improved diagnostic techniques may aid clinical decisions on the initiation, and early cessation, of antimicrobials. Furthermore, appropriate initial selection of antimicrobials is crucial for adequate treatment of time-critical infections, as is ensuring the appropriate dosage, taking into account the pharmacokinetics and pharmacodynamics of different drug classes. Lastly, prompt initiation of broad spectrum antimicrobials should be complemented by early de-escalation strategies to control the antimicrobial resistance burden driven by antibiotic exposure.


  1. Levinson WE. Review of Medical Microbiology and Immunology. 13th ed. New York: McGraw-Hill; 2014.
  2. Greenwood D. Medical Microbiology: A Guide to Microbial Infections: Pathogenesis, Immunity, Laboratory Diagnosis and Control. 18th ed. New York: Churchill Livingstone/Elsevier; 2012.
  3. Cai Y, Wang R, Liang B, Bai N, Liu Y. Systematic review and meta-analysis of the effectiveness and safety of tigecycline for treatment of infectious disease. Antimicrob Agents Chemother. 2011;55(3):1162–1172.
  4. Chavanet P. The ZEPHyR study: a randomized comparison of linezolid and vancomycin for MRSA pneumonia. Med Mal Infect. 2013;43(11):451–455.
  5. Updated guidance on the management and treatment of Clostridium difficile infection. Public Health England; 2013.
  6. Johansson SGO, Bieber T, Dahl R, Friedmann PS, Lanier BQ, Lockey RF, et al. Revised nomenclature for allergy for global use: Report of the Nomenclature Review Committee of the World Allergy Organization, October 2003. J Allergy Clin Immunol. 2004;113(5):832–836.
  7. Lin RY. A perspective on penicillin allergy. Arch Intern Med. 1992;152(5):930–937.
  8. Pegler S, Healy B. In patients allergic to penicillin, consider second and third generation cephalosporins for life threatening infections. BMJ. 2007;335(7627):991.
  9. Anderson JA. Allergic reactions to drugs and biological agents. JAMA. 1992;268(20):2844–2857.
  10. Shapiro S, Siskind V, Slone D, Lewis GP, Jick H. Drug rash with ampicillin and other penicillins. Lancet. 1969;2(7628):969–972.
  11. Snavely SR, Hodges GR. The neurotoxicity of antibacterial agents. Ann Intern Med. 1984;101(1):92–104.
  12. Fass RJ, Copelan EA, Brandt JT, Moeschberger ML, Ashton JJ. Platelet-mediated bleeding caused by broad-spectrum penicillins. J Infect Dis. 1987;155(6):1242–1248.
  13. Olaison L, Belin L, Hogevik H, Alestig K. Incidence of betalactam-induced delayed hypersensitivity and neutropenia during treatment of infective endocarditis. Arch Intern Med. 1999;159(6):607–615.
  14. Campagna JD, Bond MC, Schabelman E, Hayes BD. The use of cephalosporins in penicillin-allergic patients: a literature review. J Emerg Med. 2012;42(5):612–620.
  15. Hensgens MPM, Goorhuis A, Dekkers OM, Kuijper EJ. Time interval of increased risk for Clostridium difficile infection after exposure to antibiotics. J Antimicrob Chemother. 2012;67(3):742–748.
  16. Thomas C, Stevenson M, Riley TV. Antibiotics and hospital-acquired Clostridium difficile-associated diarrhoea: a systematic review. J Antimicrob Chemother. 2003;51(6):1339–1350.
  17. Pépin J, Saheb N, Coulombe MA, Alary ME, Corriveau MP, Authier S, et al. Emergence of fluoroquinolones as the predominant risk factor for Clostridium difficile-associated diarrhea: a cohort study during an epidemic in Quebec. Clin Infect Dis. 2005;41(9):1254–1260.
  18. Stahlmann R, Lode H. Safety overview: Toxicity, adverse effects and drug interactions. In: The Quinolones. Academic Press; 1998. p. 369–414.
  19. Prescott WA, DePestel DD, Ellis JJ, Regal RE. Incidence of carbapenem-associated allergic-type reactions among patients with versus patients without a reported penicillin allergy. Clin Infect Dis. 2004;38(8):1102–1107.
  20. Sodhi M, Axtell SS, Callahan J, Shekar R. Is it safe to use carbapenems in patients with a history of allergy to penicillin? J Antimicrob Chemother. 2004;54(6):1155–1157.
  21. Wallace MR, Mascola JR, Oldfield EC. Red man syndrome: incidence, etiology, and prophylaxis. J Infect Dis. 1991;164(6):1180–1185.
  22. Appel GB. Aminoglycoside nephrotoxicity. Am J Med. 1990;88(3):16S–20S; discussion 38S–42S.
  23. Vinh DC, Rubinstein E. Linezolid: a review of safety and tolerability. J Infect. 2009;59 Suppl 1:S59–74.
  24. Kuter DJ, Tillotson GS. Hematologic effects of antimicrobials: focus on the oxazolidinone linezolid. Pharmacotherapy. 2001;21(8):1010–1013.
  25. Lee E, Burger S, Shah J, Melton C, Mullen M, Warren F, et al. Linezolid-associated toxic optic neuropathy: a report of 2 cases. Clin Infect Dis. 2003;37(10):1389–1391.
  26. Joint Formulary Committee. British National Formulary. 69th ed. BMJ Group and Pharmaceutical Press; 2015.
  27. Panidis D, Markantonis SL, Boutzouka E, Karatzas S, Baltopoulos G. Penetration of gentamicin into the alveolar lining fluid of critically ill patients with ventilator-associated pneumonia. Chest. 2005;128(2):545–552.
  28. Shah S, Barton G, Fischer A. Pharmacokinetic considerations and dosing strategies of antibiotics in the critically ill patient. Journal of the Intensive Care Society. 2015;16(2):147–153.
    Available from: http://inc.sagepub.com/content/16/2/147.abstract.
  29. Rybak M, Lomaestro B, Rotschafer JC, Moellering R, Craig W, Billeter M, et al. Therapeutic monitoring of vancomycin in adult patients: a consensus review of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists. Am J Health Syst Pharm. 2009;66(1):82–98.
  30. van Hal SJ, Paterson DL, Lodise TP. Systematic review and meta-analysis of vancomycin-induced nephrotoxicity associated with dosing schedules that maintain troughs between 15 and 20 milligrams per liter. Antimicrob Agents Chemother. 2013;57(2):734–744.
  31. Roberts JA, Lipman J. Pharmacokinetic issues for antibiotics in the critically ill patient. Crit Care Med. 2009;37(3):840–851; quiz 859.
  32. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med. 2013;41(2):580–637.
  33. Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, Townsend S, Dellinger RP, et al. Empiric Antibiotic Treatment Reduces Mortality in Severe Sepsis and Septic Shock From the First Hour: Results From a Guideline-Based Performance Improvement Program. Crit Care Med. 2014-08;42(8):1749–1755.
  34. Schmidt G, Mandel J. Evaluation and management of severe sepsis and septic shock in adults. In: UpToDate. UpToDate; 2015.
  35. Klastersky J, Zinner SH. Synergistic Combinations of Antibiotics in Gram-Negative Bacillary Infections. Rev Infect Dis. 1982;4(2):294–301.
  36. Kumar A, Safdar N, Kethireddy S, Chateau D. A survival benefit of combination antibiotic therapy for serious infections associated with sepsis and septic shock is contingent only on the risk of death: A meta-analytic/meta-regression study:. Crit Care Med. 2010;38(8):1651–1664.
  37. Sligl WI, Asadi L, Eurich DT, Tjosvold L, Marrie TJ, Majumdar SR. Macrolides and Mortality in Critically Ill Patients With Community-Acquired Pneumonia: A Systematic Review and Meta-Analysis. Crit Care Med. 2014;42(2):420–432.
  38. Dellinger RP, Carlet JM, Masur H, Gerlach H, Calandra T, Cohen J, et al. Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock. Crit Care Med. 2004;32(3):858–873.
  39. Russell JA. Management of Sepsis. N Engl J Med. 2006;355(16):1699–1713.
  40. Garnacho-Montero J, Gutiérrez-Pizarraya A, Escoresca-Ortega A, Corcia-Palomo Y, Fernández-Delgado E, Herrera-Melero I, et al. De-escalation of empirical therapy is associated with lower mortality in patients with severe sepsis and septic shock. Intensive Care Med. 2014;40(1):32–40.
  41. Rosón B, Carratalà J, Dorca J, Casanova A, Manresa F, Gudiol F. Etiology, Reasons for Hospitalization, Risk Classes, and Outcomes of Community-Acquired Pneumonia in Patients Hospitalized on the Basis of Conventional Admission Criteria. Clin Infect Dis. 2001;33(2):158–165.
  42. File TM. Community-acquired pneumonia. Lancet. 2003;362(9400):1991–2001.
  43. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, et al. Infectious Diseases Society of America/American Thoracic Society Consensus Guidelines on the Management of Community-Acquired Pneumonia in Adults. Clin Infect Dis. 2007;44:S27–S72.
  44. Ruiz M, Ewig S, Torres A, Arancibia F, Marco F, Mensa J, et al. Severe Community-acquired Pneumonia: Risk Factors and Follow-up Epidemiology. Am J Respir Crit Care Med. 1999;160(3):923–929.
  45. Restrepo MI, Mortensen EM,Waterer GW,Wunderink RG, Coalson JJ, Anzueto A. Impact of macrolide therapy on mortality for patients with severe sepsis due to pneumonia. Eur Respir J. 2009;33(1):153–159.
  46. Garin N, Genné D, Carballo S, Chuard C, Eich G, Hugli O, et al. b-Lactam Monotherapy vs b-Lactam-Macrolide Combination Treatment in Moderately Severe Community-Acquired Pneumonia: A Randomized Noninferiority Trial. JAMA Intern Med. 2014;174(12):1894.
  47. Li JZ, Winston LG, Moore DH, Bent S. Efficacy of Short-Course Antibiotic Regimens for Community-Acquired Pneumonia: A Meta-analysis. Am J Med. 2007;120(9):783–790.
  48. American Thoracic Society and Infectious Diseases Society of America. Guidelines for the Management of Adults with Hospital-acquired, Ventilator-associated, and Healthcare-associated Pneumonia. Am J Respir Crit Care Med. 2005;171(4):388–416.
  49. Afshari A, Pagani L, Harbarth S. Year in review 2011: Critical Care – infection. Crit Care. 2012;16(6):242.
  50. Chastre J, Fagon JY. Ventilator-associated Pneumonia. Am J Respir Crit Care Med. 2002;165(7):867–903.
  51. Rello J. Epidemiology and Outcomes of Ventilator-Associated Pneumonia in a Large US Database. Chest. 2002;122(6):2115–2121.
  52. Melsen WG, Rovers MM, Groenwold RH, Bergmans DC, Camus C, Bauer TT, et al. Attributable mortality of ventilator-associated pneumonia: a meta-analysis of individual patient data from randomised prevention studies. Lancet Infect Dis. 2013;13(8):665–671.
  53. Kalanuria A, Zai W, Mirski M. Ventilator-associated pneumonia in the ICU. Crit Care. 2014;18(2):208.
  54. Masterton RG, Galloway A, French G, Street M, Armstrong J, Brown E, et al. Guidelines for the management of hospital-acquired pneumonia in the UK: Report of the Working Party on Hospital-Acquired Pneumonia of the British Society for Antimicrobial Chemotherapy. J Antimicrob Chemother. 2008;62(1):5–34.
  55. Rattanaumpawan P, Lorsutthitham J, Ungprasert P, Angkasekwinai N, Thamlikitkul V. Randomized controlled trial of nebulized colistimethate sodium as adjunctive therapy of ventilator-associated pneumonia caused by Gram-negative bacteria. J Antimicrob Chemother. 2010;65(12):2645–2649.
  56. Niederman MS, Chastre J, Corkery K, Fink JB, Luyt CE, García MS. BAY41-6551 achieves bactericidal tracheal aspirate amikacin concentrations in mechanically ventilated patients with Gram-negative pneumonia. Intensive Care Med. 2012;38(2):263–271.
  57. Solomkin JS, Mazuski JE, Bradley JS, Rodvold KA, Goldstein EJ, Baron EJ, et al. Diagnosis and Management of Complicated Intra-abdominal Infection in Adults and Children: Guidelines by the Surgical Infection Society and the Infectious Diseases Society of America. Clin Infect Dis. 2010;50(2):133–164.
  58. Pfaller MA, Boyken L, Hollis RJ, Kroeger J, Messer SA, Tendolkar S, et al. In Vitro Susceptibility of Invasive Isolates of Candida spp. to Anidulafungin, Caspofungin, and Micafungin: Six Years of Global Surveillance. J Clin Microbiol. 2008;46(1):150–156.
  59. Pappas PG, Rotstein CMF, Betts RF, Nucci M, Talwar D, De Waele JJ, et al. Micafungin versus caspofungin for treatment of candidemia and other forms of invasive candidiasis. Clin Infect Dis. 2007;45(7):883–893.
  60. Solomkin JS, Mazuski JE, Baron EJ, Sawyer RG, Nathens AB, DiPiro JT, et al. Guidelines for the Selection of Anti-infective Agents for Complicated Intra-abdominal Infections. Clin Infect Dis. 2003;37(8):997–1005.
  61. Mazuski JE, Sawyer RG, Nathens AB, DiPiro JT, Schein M, Kudsk KA, et al. The Surgical Infection Society guidelines on antimicrobial therapy for intra-abdominal infections: an executive summary. Surg Infect (Larchmt). 2002;3(3):161–173.
  62. O’Grady NP, Alexander M, Burns LA, Dellinger EP, Garland J, Heard SO, et al. Guidelines for the Prevention of Intravascular Catheter-related Infections. Clin Infect Dis. 2011;52(9):e162–e193.
  63. Wunderink RG, Niederman MS, Kollef MH, Shorr AF, Kunkel MJ, Baruch A, et al. Linezolid in Methicillin- Resistant Staphylococcus aureus Nosocomial Pneumonia: A Randomized, Controlled Study. Clin Infect Dis. 2012;54(5):621–629.
  64. Kalil AC, Klompas M, Haynatzki G, Rupp ME. Treatment of hospital-acquired pneumonia with linezolid or vancomycin: a systematic review and meta-analysis. BMJ Open. 2013;3(10):e003912–e003912.
  65. Fowler VG, Boucher HW, Corey GR, Abrutyn E, Karchmer AW, Rupp ME, et al. Daptomycin versus Standard Therapy for Bacteremia and Endocarditis Caused by Staphylococcus aureus. N Engl J Med. 2006;355(7):653–665.
  66. Enterococcus species and glycopeptide-resistant enterococci (GRE). Public Health England; 2008.
  67. Vardakas KZ, Tansarli GS, Rafailidis PI, Falagas ME. Carbapenems versus alternative antibiotics for the treatment of bacteraemia due to Enterobacteriaceae producing extended-spectrum beta-lactamases: a systematic review and meta-analysis. J Antimicrob Chemother. 2012;67(12):2793–2803.
  68. Rodriguez-Bano J, Navarro MD, Retamar P, Picon E, Pascual A, the Extended-Spectrum Beta-Lactamases-Red Espanola de Investigacion en Patologia Infecciosa/Grupo de Estudio de Infeccion Hospitalaria Group. b-Lactam/b-Lactam Inhibitor Combinations for the Treatment of Bacteremia Due to Extended-Spectrum b-Lactamase-Producing Escherichia coli: A Post Hoc Analysis of Prospective Cohorts. Clin Infect Dis. 2012;54(2):167–174.
  69. Laboratory Detection and Reporting of Bacteria with Carbapenem-Hydrolysing β-lactamases (Carbapenemases). UK Standard for Microbiology Investigations. Public Health England; 2014.
  70. Neuner EA, Yeh JY, Hall GS, Sekeres J, Endimiani A, Bonomo RA, et al. Treatment and outcomes in carbapenem-resistant Klebsiella pneumoniae bloodstream infections. Diagn Microbiol Infect Dis. 2011;69(4):357–362.
  71. Lee GC, Burgess DS. Treatment of Klebsiella pneumoniae carbapenemase (KPC) infections: a review of published case series and case reports. Ann Clin Microbiol Antimicrob. 2012;11:32.
  72. Akova M, Daikos GL, Tzouvelekis L, Carmeli Y. Interventional strategies and current clinical experience with carbapenemase-producing Gram-negative bacteria. Clin Microbiol Infect. 2012;18(5):439–448.
  73. Qureshi ZA, Paterson DL, Potoski BA, Kilayko MC, Sandovsky G, Sordillo E, et al. Treatment Outcome of Bacteremia Due to KPC-Producing Klebsiella pneumoniae: Superiority of Combination Antimicrobial Regimens. Antimicrob Agents Chemother. 2012;56(4):2108–2113.
  74. Falagas ME, Lourida P, Poulikakos P, Rafailidis PI, Tansarli GS. Antibiotic Treatment of Infections Due to Carbapenem-Resistant Enterobacteriaceae: Systematic Evaluation of the Available Evidence. Antimicrob Agents Chemother. 2014;58(2):654–663.
  75. Akta¸s Z, Kayacan C, Oncul O. In vitro activity of avibactam (NXL104) in combination with beta-lactams against Gram-negative bacteria, including OXA-48 beta-lactamase-producing Klebsiella pneumoniae. Int J Antimicrob Agents. 2012;39(1):86–89.
  76. Durante-Mangoni E, Signoriello G, Andini R, Mattei A, De Cristoforo M, Murino P, et al. Colistin and Rifampicin Compared With Colistin Alone for the Treatment of Serious Infections Due to Extensively Drug-Resistant Acinetobacter baumannii: A Multicenter, Randomized Clinical Trial. Clin Infect Dis. 2013;57(3):349–358.
  77. Dalfino L, Puntillo F, Mosca A, Monno R, Spada ML, Coppolecchia S, et al. High-Dose, Extended-Interval Colistin Administration in Critically Ill Patients: Is This the Right Dosing Strategy? A Preliminary Study. Clin Infect Dis. 2012;54(12):1720–1726.
  78. Antimicrobial stewardship: Start smart – then focus. Department of Health, UK; 2011.
  79. Dark P, Blackwood B, Gates S, McAuley D, Perkins GD, McMullan R, et al. Accuracy of LightCycler®SeptiFast for the detection and identification of pathogens in the blood of patients with suspected sepsis: a systematic review and meta-analysis. Intensive Care Med. 2015;41(1):21–33.
  80. Clerc O, Prod’hom G, Vogne C, Bizzini A, Calandra T, Greub G. Impact of Matrix-Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry on the Clinical Management of Patients With Gram-negative Bacteremia: A Prospective Observational Study. Clin Infect Dis. 2013;56(8):1101–1107.
  81. Schuetz P, Chiappa V, Briel M, Greenwald JL. Procalcitonin algorithms for antibiotic therapy decisions: a systematic review of randomized controlled trials and recommendations for clinical algorithms. Arch Intern Med. 2011;171(15):1322–1331.
  82. Schuetz P, Briel M, Mueller B. Clinical Outcomes Associated With Procalcitonin Algorithms to Guide Antibiotic Therapy in Respiratory Tract Infections. JAMA. 2013;309(7):717.
  83. Shehabi Y, Sterba M, Garrett PM, Rachakonda KS, Stephens D, Harrigan P, et al. Procalcitonin Algorithm in Critically Ill Adults with Undifferentiated Infection or Suspected Sepsis. A Randomized Controlled Trial. Am J Respir Crit Care Med. 2014;190(10):1102–1110.
  84. Hellyer TP, Morris AC, McAuley DF, Walsh TS, Anderson NH, Singh S, et al. Diagnostic accuracy of pulmonary host inflammatory mediators in the exclusion of ventilator-acquired pneumonia. Thorax. 2015;70(1):41–47.
  85. Chastre J, Wolff M, Fagon JY, Chevret S, Thomas F, Wermert D, et al. Comparison of 8 vs 15 Days of Antibiotic Therapy for Ventilator-Associated Pneumonia in Adults: A Randomized Trial. JAMA. 2003;290(19):2588.
  86. Bauer KA, West JE, O’Brien JM, Goff DA. Extended-Infusion Cefepime Reduces Mortality in Patients with Pseudomonas aeruginosa Infections. Antimicrob Agents Chemother. 2013;57(7):2907–2912.

Cite this article as follows:

Fong K, McMullan R. Antimicrobial Therapeutics in Critical Care. Critical Care Horizons 2015;1:11-21.


Airway Management of the Critically Ill Patient: Modifications of Traditional Rapid Sequence Induction and Intubation

Airway Management of the Critically Ill Patient: Modifications of Traditional Rapid Sequence Induction and Intubation

Tim Leeuwenburg
Kangaroo Island Medical Centre and MedSTAR Retrieval, South Australia
Email: drtim@wrongsideoftheroad.com.au

Full text PDF

AbstractFull-TextReference ListCitation

Discussion about airway management is common amongst clinicians involved in critical care, regardless of background. The technique of rapid sequence induction and intubation (RSI) was described in 1970 and there are many accepted variations in modern day practice. Sadly this can lead to difficulty, particularly in the event of an airway misadventure, as clinicians may be subject to post hoc critique from expert opinion in other disciplines, and often held to a ‘standard’ of RSI that no longer exists. Experts may differ in opinions, and expertise in one arena may not translate to another. This paper outlines the variations in RSI practice and the rationale for deviation. Such discussion is necessary, as expert opinion referring to a ‘standard’ RSI may be inappropriate for the critically ill patient, exposing practitioners to medico-legal risk. Acknowledgement of variations in RSI practice allows the development of institutional procedures, with potential for future consensus recommendations guided by both published studies and expert opinion.

Keywords: Airway management; Rapid sequence induction; Endotracheal intubation; Emergency intubation; Difficult airway


Rapid sequence induction and intubation (RSI) has been considered the gold standard in emergency airway management. Core elements of the classical RSI include rapid induction of anaesthesia followed by administration of a paralysing agent, techniques to minimise aspiration risk and a goal of first pass placement of a cuffed endotracheal tube in the trachea.

There is evidence for variation in how individuals, institutions and nations practice RSI. The technique of RSI is centred around reduction of risk; that of regurgitation/aspiration, and that associated with the procedure itself, including failure to rapidly secure the airway, hypoxia, airway trauma, and hypotension from induction agents. Analysis of airway complications reveals a higher incidence of difficulty in intensive care unit (ICU) and emergency department (ED) intubations than in the operating theatre (OT) (incidence of death or brain damage 38-fold higher in the ED and 58-fold higher in the ICU compared with OT),,.

RSI is a technique utilised by clinicians in anaesthesia, emergency medicine, and intensive care, both in hospital and in the prehospital environment. Variations in RSI are inevitable given the heterogeneous mix of patient pre-morbid physiology, RSI operators, teams, environment and available options. Indeed it is appropriate that RSI is modified to the circumstances, particularly in the critically ill patient . Unfortunately, the existence of such appropriate heterogeneity in practice can lead to criticism, whether between clinical experts, between health institutions, between medical specialties or in the medico-legal arena,,.

It is not the purpose of this paper to outline a uniform standard for RSI; rather to explore issues pertaining to expertise, to discuss recognised variations in components of the RSI technique and to advocate for pragmatic modifications for RSI in the critically ill patient.

Individual organisations may wish to use this as a guide to formulate institutional standard operating procedures for RSI of the critically ill, as well as for training programs for those involved in emergency RSI, thus helping to mitigate recognised complications of airway management.

The dilemma of defining an expert

Anaesthetists are traditionally regarded as the experts in airway management, reflecting the duration of their training and pivotal role in airway management. Nevertheless, airway management is a core skill of staff in other disciplines, particularly those who are actively involved in resuscitation and emergency management, or in circumstances where a specialist anaesthetist is not immediately available. The appropriate degree of diligence and expertise is expected from all providers caring for the critically ill, whether anaesthetist, intensivist, emergency physician, rural proceduralist, retrieval nurse practitioner or prehospital paramedic. Inevitably there will be differences of opinion between experts, with previously ‘indisputable truths’ in difficult airway management having been challenged in recent times. This lack of scientific consensus is problematic. In the case of an airway catastrophe, expert witnesses are usually drawn from those who may refer to classical RSI, reflecting their own traditional teaching, not the current practice of modified RSI in the critical care arena. There are documented medicolegal critiques leading to ensure, including expert opinion that it was negligent to fail to pass a nasogastric tube pre-RSI, inappropriate use of a bolus dose of induction agent, and negligent to omit cricoid force during RSI,. Post hoc expert criticism can be catastrophic for individuals. It is more appropriate to refer to the expertise of highly-trained peers regularly practicing in a similar environment. This requires acknowledgement of variations in practice in emergency airway management, whether in operating theatre, emergency department, intensive care, prehospital or in situations limited by available resources. Moreover there is heterogeneity in expert opinion and Cook et al have previously described the difficulties of contrary expert opinion in airway management, with implications for incident review, medico-legal claims and closed claim analysis.

Rapid Sequence Induction and Intubation: a standardised process or not?

Basic airway management (maintenance of oxygenation and ventilation) by use of adjuncts such as suction, oro- and nasopharyngeal airways, bag-mask ventilation and even placement of a laryngeal mask in the truly obtunded is well within the expected competency of all clinicians working in acute care. However, RSI is expected of advanced airway practitioners, with indications including:

  • failure to maintain airway patency by other means
  • failure of airway protection
  • failure of ventilation or oxygenation
  • for anticipated clinical course
  • to facilitate transportation
  • for humanitarian reasons

The original 15-step technique of RSI was described in 1970, yet this form of RSI is not uniformly applied in modern practice and nor should we expect it to be. Advances in equipment, induction drugs and paralysing agents have allowed refinement of the technique over time, with RSI modifications made as appropriate to the clinical circumstances of individual patients, to the skill mix of airway teams and to the environment in which airways are managed. Therein lies the difficulty. Despite the universal acceptance of RSI as the ‘gold standard’ in securing the airway in a critically ill patient, the actual components of RSI are known to differ markedly between individuals, institutions and countries, as well as between practitioners in different arenas (prehospital, ED, ICU or OT),. Documented modifications to RSI technique include patient position, preoxygenation strategies, pre-RSI decompression of gastric contents with a nasogastric tube,  choice and method of administration of induction agent, application of cricoid pressure, choice of paralysing agent, use of manual ventilation and options for failed RSI (not least whether wakening is an option). However the key elements of RSI remain, namely:

  • pre-oxygenation or denitrogenation to prolong time to critical desaturation
  • prevention of hypoxia and hypotension during the induction and intubation sequence
  • passage of a cuffed endotracheal tube with confirmation of placement

In short, a refinement of the classical RSI technique as defined by Stept and Safar14 is called for, with a need for a consensus position allowing for variation in the practice of RSI between experts, as governed by the requirements of the patient, team and clinical circumstances. Accepted practice variation should be understood in the context of both the need to minimise aspiration.

RSI of the Critically Unwell Patient

Airway Team and Dynamics

Regardless of the individual expertise of the intubator, team factors will impact on performance of the RSI process. Team members should be adequately trained prior to involvement in airway management, preferably involving simulation training under increasing degrees of cognitive load to allow a degree of ‘stress inoculation’ and to reinforce the importance of human factors in performance.

Use of a standardised approach to RSI may be appropriate within an institution or service6,. A challenge-response RSI checklist is recommended, but any such checklist should be short, clear and contain a check only of essential items. Checklists should be designed to be read aloud to verify ’essential items completed’ rather than being presented as a ’how to cookbook’. Completion should take no more than 60 seconds and the process can be completed during preoxygenation. For austere environments, or RSI where assistants may be unfamiliar, this checklist can be combined with a shadow board kit dump (Figure 1) to ensure all of the required equipment is readily available

Figure 1. Shadow Board Kit Dump with Challenge-Response Checklist

Figure 1. Shadow Board Kit Dump with Challenge-Response Checklist

Roles should be clearly assigned prior to performing RSI and will usually include: intubator, airway assistant, provision of manual in-line immobilisation (if required) and someone responsible for giving medications. A ‘reader’ may be assigned for reading of both a pre-RSI challenge-response checklist and for crisis management checklists in case of difficulty.

Airway teams should regularly engage in simulation training, using their own equipment and personnel, simulating both common and uncommon scenarios. This may include common critical care presentations, but will also incorporate changes in team members, equipment failure and other measures to encourage understanding of human factors in team performance.

Airway Planning

Pre-RSI briefing should include planning for anticipated difficulties. Difficult airway plans usually include direct laryngoscopy for endotracheal intubation as the primary plan, with backup which may include alternative devices such as videolaryngoscopes or an intubating laryngeal mask to maintain oxygenation and facilitate subsequent intubation. Rescue ventilation via bag-mask or supra-glottic devices may be required as a bridge, but if they fail the team should be prepared to perform an emergency surgical airway.

Guidelines exist for management of the difficult airway and airway plans should be tailored to the availability of such equipment within an institution or location, as well to the anticipated clinical course. Standard difficult airway plans which incorporate options to ‘awaken the patient and abandon the procedure’ may be wholly inappropriate due to the immediate need to secure an airway in the critically ill. An example of an institutional airway plan is shown in Figure 2.

Figure 2

Figure 2. Example Airway Plan

In certain circumstances, such as anticipated failed intubation and rapid desaturation, it may be necessary to consider a ’double set up’ approach, with one initial attempt at laryngoscopy and intubation before progressing to an emergency surgical airway. In cases of anticipated difficulty the cricothyroid membrane may be identified prior to RSI (clinically or with ultrasound) and marked using an indelible marker. Identification of the cricothyroid membrane has obvious advantages, both to overcome the recognised cognitive hurdle to establishing an emergency surgical airway and to aid identification of the cricoid cartilage if cricoid pressure is to be applied by an inexperienced assistant.

Patient Positioning, Optimisation and Monitoring

Stept and Safar described RSI with the patient in a recumbent position, with legs raised (an attempt to counteract hypotension) and the trunk raised 30 degrees (to counteract regurgitation)14. However Sellick described the procedure of cricoid pressure in a steep head-down position with head and neck extended, ostensibly to tether the oesophagus to vertebral bodies in order to minimise aspiration.

Most clinicians perform RSI in the supine position. In the bariatric patient, ‘ramping’ of the upper body to around 45 degrees may be required to improve functional residual capacity, via displacing the weight of the anterior chest wall off the thoracic cavity and the weight of the intra-abdominal contents off the diaphragm. This ramped position is often referred to as the ear-to-sternum position as it results in the external auditory meatus being at the same horizontal level as the sternum. Head up positioning may be preferable for the non-hypotensive head-injured patient, to improve venous outflow from the brain, thus helping to reduce intracranial pressure. It may also improve respiratory dynamics for pre-oxygenation.

The head up position has been suggested as an alternative, or in addition to, the application of cricoid pressure in reducing passive regurgitation. However, if the patient vomits it theoretically increases the chances of aspiration from the effects of gravity, rather than particulate matter draining from the mouth.

Pregnant patients should be positioned in left lateral tilt and/or the uterus manually displaced to avoid aortocaval compression. Regardless of whether positioned supine, head up to limit regurgitation, head down to limit aspiration or in a left lateral position if pregnant, working suction should always be available.

For those in whom ear-to-sternum positioning is contraindicated (suspected spinal injury or musculoskeletal abnormality limiting spinal mobility), head position should minimise unnecessary flexion-extension or lateral rotation. Spinal precautions should be observed for the trauma patient, which may include manual inline stabilisation or use of an occipital pad to optimise laryngoscopy and minimise movement of the cervical spine. Airway team members, monitors and equipment should be appropriately positioned to maximise visual cues and not hinder 360 degree access to the patient. Alarm limits should be pre-determined as appropriate for patient age and anticipated difficulties. Standard monitoring (oximetry, waveform end-tidal carbon dioxide, blood pressure and ECG) should be applied and abnormal physiology optimised pre-RSI wherever possible (where time permits this may include commencement of vasopressor infusions). A timer should be used both for pre-oxygenation and to facilitate rapid progression through agreed airway plans. Figure 3 illustrates an example set up.

Figure 3. Example Equipment and Team Set-up for Intubation

Figure 3. Example Equipment and Team Set-up for Intubation

Avoidance of hypoxia and hypotension is essential in the critically ill patient. Pre-oxygenation strategies are discussed below. With regards to hypotension, many clinicians will consider the impact of induction and paralysis on both cardiac contractility and venous tone, along with effects of positive pressure ventilation, as mandating an intravenous fluid bolus to improve preload prior to induction, unless contraindicated.

Measurements of blood pressure should be made regularly, with either non-invasive blood pressure set to cycle at 1 minute intervals or placement of an arterial line if sufficient time allows.


The purpose of pre-oxygenation is to denitrogenate the lungs and create a reservoir of oxygen to allow a margin of safety before critical desaturation during attempts to secure the airway. An excellent summary of methods to maximise pre-oxygenation and prevent desaturation during emergency airway management is described by Weingart and Levitan29. Key steps include mandatory use of pre-oxygenation to extend safe apnoea time during RSI, along with appropriate positioning, and may involve the use of positive endexpiratory pressure (PEEP). The delivery of oxygen via nasal cannulae during intubation (apnoeic diffusion oxygenation) is increasingly being adopted for airway management of the critically ill patient. The period of pre-oxygenation should adequately denitrogenate the lungs. An empiric approach applying high-flow oxygen for three minutes or eight vital capacity breaths is common anaesthetic practice. However, critically ill patients may require a longer period to denitrogenate and are often unable to perform eight vital capacity breaths. If available, measurement of expired end-tidal oxygen should be used as a guide to adequate pre-oxygenation, aiming for a value of at least 90% (FeO2 of 0.9). Pre-oxygenation technique may be governed by available equipment, personnel and patient requirements. Valid techniques include:

  • use of a Mapleson B or C anaesthetic circuit. These lack the separate inspiratory or expiratory ports of traditional bag-valve-mask (BVM) devices, with exhaled gas flushed out of the circuit by high fresh gas flow via the pressure-release valve, ensuring maximal oxygen delivery.
  • use of standard reservoir face masks on maximal oxygen flow and supplemented with nasal cannulae on maximal flow. This may be a preferred in the prehospital environment, where limitations of personnel preclude alternatives.
  • use of standard bag-valve-mask devices commonly used in ED, ICU or by emergency medical services. Caution is needed as such devices may entrain room air during spontaneous ventilation. Addition of a PEEP valve to the expiratory port of BVM assembly obviates this.
  • use of existing non-invasive ventilation modes. For many critically ill patients, RSI may represent the end result of a failure of non-invasive ventilation (NIV). NIV masks may be left in situ and used to pre-oxygenate. CPAP/NIV may be very useful in pre-oxygenation of the morbidly obese patient.

On occasions, the combative patient (e.g. intoxicated, head injured, hypoxic) will thwart best attempts at both positioning and pre-oxygenation. Pre-treatment with small titrated aliquots of a sedative agent can be effective (so-called ‘delayed sequence intubation’), with ketamine the preferred agent to facilitate assessment, monitoring, positioning and preoxygenation.

Choice and Timing of Induction Agent

Commonly used induction agents include thiopentone (as originally described by Stept and Safar), etomidate (not available in all countries), propofol, benzodiazepines such as midazolam (relatively slow onset compared to other agents) and ketamine. Ketamine is gaining favour within emergency and critical care circles due to relative cardiovascular stability. It should be noted that all induction agents (including midazolam and ketamine) have potential for cardiovascular depression and hypotension if too high a dose is used. In addition, combinations of agents may be synergistic with amplification of effect. Previous concerns of deleterious effects of ketamine on intracranial pressure in head injury have been challenged and as such, use of ketamine has much to commend it for RSI in the critically ill patient.

Whilst the original description of RSI involved a bolus of thiopentone based on patient weight, such weight-based calculations may not be appropriate in critical illness due to adverse haemodynamic effects. Doses should be adjusted according to pre-RSI physiology, requiring dose-reductions to as little as 10% of standard induction doses for the critically ill patient with haemodynamic compromise. Both bolus dosing and titration of induction agent to loss of consciousness have been described. Bolus dosing from a predrawn syringe has the advantage of rapidity; however, there is potential for either under- or overdosing, the former perhaps contributing to increased reports of awareness during RSI in trauma and obstetric patients, the latter risking haemodynamic compromise. Currently, there are no data to compare the potential aspiration risks of a longer induction time via dose titration versus the risks of either awareness or haemodynamic instability with a predetermined bolus technique.

Clinicians will determine the optimal choice of induction agent for the situation, often guided by personal expertise, institutional guidelines, available agents and appropriate patient selection. Regardless of induction agent used, delay between administration, loss of consciousness and administration of paralysing agent may prolong the period of aspiration risk and increased the risk of desaturation.

Adjunct Opioid Agents

Adjunct agents are not described in the traditional teaching of RSI, yet many practitioners incorporate rapid acting opioids to attenuate the reflex sympathetic responses to laryngoscopy and intubation. This may be especially useful in critically ill patients with head injuries. Arguments against use of opioids include historical concerns due to slow onset and longer duration with older opioids, as well as concerns of decreased respiratory drive if intubation fails. This is less of a concern in the critically ill patient, as options to awaken the patient are generally not appropriate.

Lyon et al describe a modification of RSI technique using adjunctive fentanyl, along with ketamine induction and rocuronium paralysis within their prehospital service. They note both superior intubating conditions and a more favourable haemodynamic response to intubation. Development of protocols for modified RSI within an institution, and their subsequent publication, is to be encouraged.

It should be noted that the use of opioids such as alfentanil and fentanyl may produce synergistic effects in combination with induction agents, and cautious dosing should be used in haemodynamically unstable patients to minimise hypotension.

Cricoid Force

Cricoid force has become an area of contention in airway management. Sellick’s original description was of a ‘firm’ amount of pressure applied to the cricoid cartilage of a cadaver whilst in a steep head-down position to occlude the oesophagus and prevent regurgitation of fluid into the oropharynx27. The procedure was repeated during induction of 26 patients deemed at high-risk of aspiration. None experienced regurgitation with application of cricoid force; 3 experienced immediate reflux upon release of cricoid force after tracheal intubation. Cricoid force was incorporated into Stept and Safar’s description of RSI and has since been considered an essential component. Refinements describe a force of 10N applied at the commencement of induction, increased to 30N with loss of consciousness. Application of cricoid force remains a recommendation during RSI from the authors of the NAP4 audit in the United Kingdom2.

However, application of cricoid force is not considered routine practice in some countries or organisations. There are concerns that cricoid force does not effectively occlude the oesophagus and thus prevent aspiration, is variably applied by assistants (often incorrect timing, incorrect position or force) and that cricoid force can impede view at laryngoscopy thus delaying first pass success,,,.

Some have proposed that cricoid force is a low-risk procedure that works in a proportion of patients but is confounded by poor technique and relative infrequency of regurgitation. Thus, they propose application of cricoid force and early removal if this impedes laryngoscopy, if there is active vomiting, or if there is impediment of rescue ventilation via laryngeal mask airway or BVM,. It can be argued that in certain arenas, particularly prehospital or with limited/untrained personnel (rural, small ED or ICU) application of cricoid pressure is more likely to hinder laryngoscopy and that the policy of ‘apply, then release’ adds additional cognitive load to an already high-stakes tightly-coupled procedure. On this basis, some airway experts may opt to omit cricoid force in such circumstances, based on limited evidence of efficacy and risk-benefit balance in regard to optimising first-pass intubation success6,. Meanwhile trials are under-way to test the hypothesis that use of cricoid force during RSI in ED does not prevent aspiration and investigate the effect of such force on difficult or failed intubation.

A decision not to apply cricoid force may be reasonable in airway management of the critically ill patient. It is recommended that any decision to use or omit cricoid force be supported by an institutional policy. Practitioners with clinical expertise in resuscitation are responsible for shaping such policy, mindful that this may differ from published national or international guidelines. Hence, despite a lack of absolute evidence of benefit, cricoid force may continue to be applied; reflecting medico-legal concerns as individual clinicians have been criticised for failing to apply cricoid force in post-event medico-legal dissection of airway catastrophes12. It is essential that any expert opinion on cricoid force, as indeed any other matter in RSI, acknowledges the existing variation in practice. At this point in time, the literature does not support evidence either for or against the application of cricoid force.


Use of succinylcholine (a depolarising neuromuscular blocker) as the preferred agent to facilitate vocal cord relaxation and endotracheal tube passage has been the accepted norm for RSI, with traditional teaching being that the short duration of action will allow return of spontaneous ventilation in the case of a failed RSI. Whilst awakening may be an option for some patients in the operating theatre, it is rarely the case for the unfasted, haemodynamically-compromised patient for whom RSI represents a commitment to securing the airway.

Rocuronium at a dose of 1.6 mg/kg gives the same onset of muscle relaxation as succinylcholine and is suggested as the preferred choice of non-depolarising neuromuscular blockers for RSI in the critically ill. A commitment to full paralysis and rapid progression to a surgical airway in the case of failed intubation and ventilation in the critically ill patient is congruent with pre-agreed airway plans between team members, appropriate for the patient (whose pathology requires a cuffed tube in the trachea by whatever means) and avoids the possibility of attempting a surgical airway in a combative, coughing patient.

Manual Ventilation between Induction and Intubation

Manual ventilation has traditionally been avoided in classical RSI, due to concerns of gastric insufflation and aspiration. However, gentle ventilation has been advocated in both obstetric and paediatric RSI due to concerns of rapid desaturation in these populations. Anecdotal evidence from experienced resuscitationists includes use of gentle manual ventilation whilst awaiting onset of paralysis as a ‘do least harm’ approach. A decision on whether to gently ventilate will be guided by aspiration risk – the patient with ileus, with gastroparesis or with upper gastrointestinal bleeding is clearly at higher risk than the fasted patient. For the critically ill patient, risks of hypoxia and hypercapnia may require gentle manual ventilation. Critically ill patients commonly have an existing metabolic acidosis with respiratory compensation, and periods of apnoea can result in significant reductions in pH which amplify haemodynamic risk. Manual ventilation attempts should be initiated with pressures less than 15 cmH2O to minimise gastric insufflation. There is potential for the use of adjuncts such as automatic, low pressure, constant flow ventilation devices to minimise ventilation pressures during RSI.

Maximising First-Pass Success

It is important to appreciate that repeated attempts at laryngoscopy may increase rates of aspiration. Thus, maximising the potential for first pass success is essential in RSI of the critically ill patient.

Direct laryngoscopy using an appropriate blade and light source (modern day LED optics offer excellent illumination and contrast) remains the cornerstone of intubation. Careful and sequential visualisation of landmarks and avoidance of repeated attempts causing airway trauma are key skills.

Adjuncts such as a bougie or malleable stylet are commonly used in cases of difficult intubation. For intubation of the critically ill patient, such adjuncts should be used routinely. Understanding appropriate use is vital as infrequent users may not appreciate the nuances of these devices, which are designed to facilitate navigation to the laryngeal inlet in difficult cases. Stylets, if used, should be shaped ‘straight-to-cuff’ i.e. the stylet should remain straight as far as the proximal part of the endotracheal tube cuff where it should be angled to no more than 35 degrees (angles beyond 35 degrees increase difficulty). Traditional teaching has been to avoid preloading endotracheal tubes onto bougies, as the weight of the tube may impair control of the bougie tip; however, hang-up of the bougie on the endotracheal tube connector may impede smooth rail-roading of the endotracheal tube, causing delay in tube passage and risking a loss of situational awareness in the operator. A refinement is to pre-load an endotracheal tube onto a bougie and hold them in such a grip that control of the bougie is maintained during navigation to the laryngeal inlet (Figure 4).

Figure 4. Endotracheal Tube Pre-loaded on Bougie

Figure 4. Endotracheal Tube Pre-loaded on Bougie

It is not uncommon for the leading edge of bevel-shaped endotracheal tubes to hang up on the right arytenoid cartilage; gentle slight withdrawal and a counter-clockwise rotation of the endotracheal tube/bougie complex allows the free-edge to enter the glottic opening and advance.

Some advocate the use of video-laryngoscopy over direct laryngoscopy, particularly for a known or anticipated difficult airway. Currently there is a plethora of available devices available. Cited advantages include improved view of the glottic opening for difficult airways, allowing other members of the team to visualise tube passage, and potential for recording of intubation procedures for audit and training. Videolaryngoscopy may afford better visualisation of the glottic opening in a difficult airway; caution is recommended as a better view with some devices does not translate to easier tube passage unless the operator is experienced in use of the particular device. Additional caveats include cost, poor performance in the presence of blood/secretions and many require a different technique to traditional direct laryngoscopy. The optimal video-laryngoscope would be low cost, have the same technique as standard direct laryngoscopy, have similar blade geometry and tube passage, perform well in the presence of both a soiled airway and in the presence of bright sunlight. At present no such device exists. If a videolaryngoscope is used, operators must be fully aware of nuances of the device and be trained to use in elective settings prior to an emergency.

Failed RSI

A difficult airway plan should be discussed and a checklist should be completed prior to RSI such that a shared mental model of actions to be undertaken exists between team members, both for routine and in case of difficulty13,. Many such difficult airway plans exist,. Another cognitive aid showing promise is The Vortex Approach (Figures 5 and 6). The Vortex approach is designed to optimise rescue techniques whether through endotracheal intubation, placement of a supra-glottic airway or rescue bag-mask ventilation,. Time-limited drills should be agreed prior to RSI and then completed sequentially. In a ‘cannot intubate, cannot oxygenate’ situation the operator is prompted towards establishment of a surgical airway.

Standards exist for equipment to manage the difficult airway and such equipment should be available wherever airways are managed. Regular practice of RSI competency and airway planning using simulation is a hallmark of a well-functioning airway team. Such rehearsal may facilitate swift transition through airway plans and crisis algorithms, with early use of appropriate equipment and decisions24. In particular, rescue surgical airway techniques must be regularly practiced as they are infrequently used in anger and as such remain a common area of unease.




Figure 5. The Vortex Cognitive Aid

Figure 5. The Vortex Cognitive Aid


Figure 6. Vortex Optimisation Strategies. OPA oropharyngeal airway; NPA nasopharyngeal airway; FM facemask; VL video-laryngoscope; BZD benzodiazepine; NMBD neuromuscular blocking drugs; GZ green zone

Figure 6. Vortex Optimisation Strategies. OPA
oropharyngeal airway; NPA nasopharyngeal airway; FM facemask; VL video-laryngoscope; BZD benzodiazepine; NMBD neuromuscular blocking drugs; GZ green zone


Post-Intubation Care

Once the trachea has been intubated and endotracheal tube cuff inflated, placement should be confirmed with waveform endtidal CO2 (colorimetric devices, although inferior, will suffice if waveform end-tidal CO2 is unavailable). Potential exists for haemodynamic instability post-RSI; whilst efforts may have been made to mitigate against this (e.g. preloading fluid and dose reduction in the haemodynamically compromised, use of adjunct opioid to blunt response to intubation in the head injured), post-RSI monitoring of heart rate and blood pressure is vital. The combination of worsening acidosis from peri-intubation apnoea, the presence of hypovolaemia, the impact of induction agents on cardiac contractility and vasomotor tone, and the effects of over-zealous post-intubation ventilator settings on right ventricular preload (reduction) and afterload (increase) is a potent trigger for haemodynamic collapse. Postintubation ventilation and sedation plans should be previously agreed during airway planning, and should be enacted. RSI of the critically ill patient may pose a challenge for the most experienced operators and care must be taken to avoid clinical inertia and to continue resuscitation and vigilant monitoring for complications34.


Rapid sequence induction and intubation has evolved since the original description by Stept and Safar in 1970, with many practitioners using a modified RSI. Variations in technique exist between individuals, specialties, institutions and countries. Whilst some components of RSI are unchanged, refinements may be made as appropriate to the needs of individual patients, composition of airway team and the clinical environment. No doubt some of the current controversies in RSI will be resolved in time; meanwhile, the evidence-base for practice remains predominantly based on tradition and expert opinion (Level V evidence).

Although standardisation in procedures is to be applauded for the purposes of training, quality control and audit, the existence of variation between expert practitioners should not be a cause for inappropriate concern nor litigation. Sadly, post hoc analysis of adverse outcomes in emergency airway management may fail to acknowledge the accepted variations in RSI practice, with expert opinion on the same case differing widely due to individual preference, discordance in expertise between arenas and low quality evidence in the literature.

This paper discusses the variation in RSI practice and highlights specific measures for consideration in the critically ill. Acknowledgement and thorough understanding of available options in airway management of the critically ill patient should form a central component when training clinicians. In the absence of an agreed international standard for RSI and with documented variation in practice, this paper may form the basis for development of agreed procedures at the level of institutions or organisations, as well as guide future medico-legal opinion. Opportunity exist for development of consensus recommendations for airway management in the critically ill, based on both published literature and Delphi methodology.


The author would like to thank the following for reading and commenting on the manuscript prior to submission:
Nicholas Chrimes, Minh le Cong & Casey Parker (Australia);
Daniel Kornhall (Norway);
Natasha Burley, Kirsty Challen, Marietjie Slabbert & Alistair Steel (UK);
Salim Rezaie & Anand Swaminathan (USA).


  1. Dagal A, Joffe AM, Treggiari MM, Sharar SR, Tansley J, Moppett IK, et al. Rapid sequence induction practices in the United States and the United Kingdom: a comparative survey study. Internet Journal of Anesthesiology. 2012;30(2).
    Available from: https://ispub.com/IJA/30/2/13832.
  2. Cook TM, Woodall N, Frerk C, Fourth National Audit Project. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1: anaesthesia. Br J Anaesth. 2011 May;106(5):617–631.
  3. Cook TM, Woodall N, Harper J, Benger J, Fourth National Audit Project. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 2: intensive care and emergency departments. Br J Anaesth. 2011 May;106(5):632–642.
  4. Cook TM, MacDougall-Davis SR. Complications and failure of airway management. Br J Anaesth. 2012 Dec;109 Suppl 1:i68–i85.
  5. Kovacs G, Law JA, Ross J, Tallon J, MacQuarrie K, Petrie D, et al. Acute airway management in the emergency department by non-anesthesiologists. Can J Anaesth. 2004 Feb;51(2):174–180.
  6. Lockey DJ, Crewdson K, Lossius HM. Pre-hospital anaesthesia: the same but different. Br J Anaesth. 2014 Aug;113(2):211–219.
  7. Huitink JM. Developing expert opinion in airway management. Anaesthesia. 2011 Dec;66(12):1174; author reply 1174–1175.
  8. Stevens A. Reliability and cogency of expert witness evidence in modern civil litigation. Anaesthesia. 2011 Sep;66(9):764–768.
  9. White S. Problems with expert opinion. Anaesthesia. 2011 Dec;66(12):1172–1173; author reply 1173–1174.
  10. Greenland KB, Irwin MG. Airway management–’spinning silk from cocoons’ ( – Chinese idiom). Anaesthesia. 2014 Apr;69(4):296–300.
  11. Fitness to Practise Panel of the Medical Practitioner’s Tribunal Service 17 – 26 March 2014. Medical Practitoner’s Tribunal Service; 2014.
    Available from: http://webcache.gmc-uk.org/minutesfiles/Minutes%20PUBLISHABLE%206146949%20Mar%202014.docx.
  12. Hospital bosses apologise for surgery error that led to woman’s death. Boston Standard. 2013 Jul;
    Available from: http://goo.gl/80RrXN.
  13. Cook TM, Morgan PJ, Hersch PE. Equal and opposite expert opinion. Airway obstruction caused by a retrosternal thyroid mass: management and prospective international expert opinion. Anaesthesia. 2011 Sep;66(9):828–836.
  14. Stept WJ, Safar P. Rapid induction-intubation for prevention of gastric-content aspiration. Anesth Analg. 1970 Aug;49(4):633–636.
  15. Koerber JP, Roberts GEW, Whitaker R, Thorpe CM. Variation in rapid sequence induction techniques: current practice in Wales. Anaesthesia. 2009 Jan;64(1):54–59.
  16. Morris J, Cook TM. Rapid sequence induction: a national survey of practice. Anaesthesia. 2001 Nov;56(11):1090–1097.
  17. Cook T, Behringer EC, Benger J. Airway management outside the operating room: hazardous and incompletely studied. Curr Opin Anaesthesiol. 2012 Aug;25(4):461–469.
  18. El-Orbany M, Connolly LA. Rapid sequence induction and intubation: current controversy. Anesth Analg. 2010 May;110(5):1318–1325.
  19. Kim J, Neilipovitz D, Cardinal P, Chiu M, Clinch J. A pilot study using high-fidelity simulation to formally evaluate performance in the resuscitation of critically ill patients: The University of Ottawa Critical Care Medicine, High-Fidelity Simulation, and Crisis Resource Management I Study. Crit Care Med. 2006 Aug;34(8):2167–2174.
  20. Standard operating procedure : Rapid sequence intubation. UK HEMS; 2007.
    Available from: http://www.uk-hems.co.uk/18%20UKHEMS%20CP%20Rapid%20Sequence%20Induction%2007.pdf.
  21. Booth A, Steel A, Klein J. Anaesthesia and pre-hospital emergency medicine. Anaesthesia. 2013 Jan;68 Suppl 1:40–48.
  22. Astin J, Cook TM, King EC, Bellchambers E, Bradley T. Timely safe airway management in critically ill patients. Br J Anaesth. 2013 Feb;110(2):315–316.
  23. Mackenzie R, French J, Lewis S, Steel A. A prehospital emergency anaesthesia pre-procedure checklist. Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine. 2009;17(Suppl 3):O26.
    Available from: http://www.sjtrem.com/content/17/S3/O26.
  24. Sherren PB, Tricklebank S, Glover G. Development of a standard operating procedure and checklist for rapid sequence induction in the critically ill. Scand J Trauma Resusc Emerg Med. 2014;22:41.
  25. Davies L, Benger JR. Audit of advanced airway management in UK Emergency Departments following the Fourth National Audit Project of the Royal College of Anaesthetists and Difficult Airway Society. Emerg Med J. 2013 May;30(5):427.
  26. Mallin M, Curtis K, Dawson M, Ockerse P, Ahern M. Accuracy of ultrasound-guided marking of the cricothyroid membrane before simulated failed intubation. Am J Emerg Med. 2014 Jan;32(1):61–63.
  27. Sellick BA. Cricoid pressure to control regurgitation of stomach contents during induction of anaesthesia. Lancet. 1961 Aug;2(7199):404–406.
  28. Fessler RD, Diaz FG. The management of cerebral perfusion pressure and intracranial pressure after severe head injury. Ann Emerg Med. 1993 Jun;22(6):998–1003.
  29. Weingart SD, Levitan RM. Preoxygenation and prevention of desaturation during emergency airway management. Ann Emerg Med. 2012 Mar;59(3):165–175.e1.
  30. Robitaille A. Airway management in the patient with potential cervical spine instability: continuing professional development. Can J Anaesth. 2011 Dec;58(12):1125–1139.
  31. Moran C, Karalapillai D, Darvall J, Nanuan A. Is it time for apnoeic oxygenation during endotracheal intubation in critically ill patients? Crit Care Resusc. 2014 Sep;16(3):233–235.
  32. Chrimes N. Demonstration of Factors Influencing Performance of Oxygen Delivery Devices; 2013.
    Available from: http://monashanaesthesia.org/fio2/.
  33. Kwei P, Matzelle S,Wallman D, Ong M,Weightman W. Inadequate preoxygenation during spontaneous ventilation with single patient use self-inflating resuscitation bags. Anaesth Intensive Care. 2006 Oct;34(5):685–686.
  34. Habig K, Reid C, Hanrahan B. Prehospital RSI Manual v2.01. Greater Sydney Area HEMS; 2012.
    Available from: http://nswhems.files.wordpress.com/2012/12/rsimanual2-1-oct-2012.pdf.
  35. Sehdev RS, Symmons DAD, Kindl K. Ketamine for rapid sequence induction in patients with head injury in the emergency department. Emerg Med Australas. 2006 Feb;18(1):37–44.
  36. Filanovsky Y, Miller P, Kao J. Myth: Ketamine should not be used as an induction agent for intubation in patients with head injury. CJEM. 2010 Mar;12(2):154–157.
  37. Reich DL, Hossain S, Krol M, Baez B, Patel P, Bernstein A, et al. Predictors of hypotension after induction of general anesthesia. Anesth Analg. 2005 Sep;101(3):622–628.
  38. Lyon RM, Perkins ZB, Chatterjee D, Lockey DJ, Russell MQ, Kent, Surrey & Sussex Air Ambulance Trust. Significant modification of traditional rapid sequence induction improves safety and effectiveness of prehospital trauma anaesthesia. Crit Care. 2015;19(1):134.
  39. Stewart JC, Bhananker S, Ramaiah R. Rapid-sequence intubation and cricoid pressure. Int J Crit Illn Inj Sci. 2014 Jan;4(1):42–49.
  40. Fenton PM, Reynolds F. Life-saving or ineffective? An observational study of the use of cricoid pressure and maternal outcome in an African setting. Int J Obstet Anesth. 2009 Apr;18(2):106–110.
  41. Harris T, Ellis D, Lockey D, Foster L. Cricoid pressure – friend or foe? Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine. 2009;17(Suppl 1):O5.
    Available from: http://www.sjtrem.com/content/17/S1/O5.
  42. Maltby JR, Beriault MT. Science, pseudoscience and Sellick. Can J Anaesth. 2002 May;49(5):443–447.
  43. Priebe HJ. Use of cricoid pressure during rapid sequence induction: Facts and fiction. Trends in Anaesthesia and Critical Care. 2012 Jun;2(3):123–127.
    Available from: http://linkinghub.elsevier.com/retrieve/pii/S2210844012000196.
  44. Harris T, Ellis DY, Foster L, Lockey D. Cricoid pressure and laryngeal manipulation in 402 pre-hospital emergency anaesthetics: essential safety measure or a hindrance to rapid safe intubation? Resuscitation. 2010 Jul;81(7):810–816.
  45. Sinha AC. Cricoid pressure: An enigma wrapped in a mystery or a hand wrapped around a throat? If I can’t disprove a lie, does it become the truth? J Anaesthesiol Clin Pharmacol. 2014 Jan;30(1):1–2.
  46. Ellis DY, Harris T, Zideman D. Cricoid pressure in emergency department rapid sequence tracheal intubations: a risk-benefit analysis. Ann Emerg Med. 2007 Dec;50(6):653–665.
  47. Trethewy CE, Burrows JM, Clausen D, Doherty SR. Effectiveness of cricoid pressure in preventing gastric aspiration during rapid sequence intubation in the emergency department: study protocol for a randomised controlled trial. Trials. 2012;13:17.
  48. Curley GF. Rapid sequence induction with rocuronium – a challenge to the gold standard. Crit Care. 2011;15(5):190.
  49. Neuhaus D, Schmitz A, Gerber A, Weiss M. Controlled rapid sequence induction and intubation – an analysis of 1001 children. Paediatr Anaesth. 2013 Aug;23(8):734–740.
  50. Bouvet L, Albert ML, Augris C, Boselli E, Ecochard R, Rabilloud M, et al. Real-time detection of gastric insufflation related to facemask pressure-controlled ventilation using ultrasonography of the antrum and epigastric auscultation in nonparalyzed patients: a prospective, randomized, double-blind study. Anesthesiology. 2014 Feb;120(2):326–334.
  51. Hu X, Ramadeen A, Laurent G, So PPS, Baig E, Hare GMT, et al. The effects of an automatic, low pressure and constant flow ventilation device versus manual ventilation during cardiovascular resuscitation in a porcine model of cardiac arrest. Resuscitation. 2013 Aug;84(8):1150–1155.
  52. Mort TC. Emergency tracheal intubation: complications associated with repeated laryngoscopic attempts. Anesth Analg. 2004 Aug;99(2):607–613.
  53. Levitan RM. Airway Cam Textbook Guide to Intubation and Practical Emergency Airway Management. Airway Cam Technologies; 2004.
  54. Levitan RM, Pisaturo JT, Kinkle WC, Butler K, Everett WW. Stylet bend angles and tracheal tube passage using a straight-to-cuff shape. Acad Emerg Med. 2006 Dec;13(12):1255–1258.
  55. Rothfield KP, Russo SG. Videolaryngoscopy: should it replace direct laryngoscopy? a pro-con debate. J Clin Anesth. 2012 Nov;24(7):593–597.
  56. De Jong A, Clavieras N, Conseil M, Coisel Y, Moury PH, Pouzeratte Y, et al. Implementation of a combo videolaryngoscope for intubation in critically ill patients: a before-after comparative study. Intensive Care Med. 2013 Dec;39(12):2144–2152.
  57. Kovacs G. Airway management: “the times they are achangin”. CJEM. 2013;15(0):1–4.
  58. Tobin JM, Grabinsky A, McCunn M, Pittet JF, Smith CE, Murray MJ, et al. A checklist for trauma and emergency anesthesia. Anesth Analg. 2013 Nov;117(5):1178–1184.
  59. Apfelbaum JL, Hagberg CA, Caplan RA, Blitt CD, Connis RT, Nickinovich DG, et al. Practice guidelines for management of the difficult airway: an updated report by the American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Anesthesiology. 2013 Feb;118(2):251–270.
  60. Henderson JJ, Popat MT, Latto IP, Pearce AC, Difficult Airway Society. Difficult Airway Society guidelines for management of the unanticipated difficult intubation. Anaesthesia. 2004 Jul;59(7):675–694.
  61. Chrimes N. The Vortex Approach: Management of the Unanticipated Difficult Airway; 2013. Available from: http://www.vortexapproach.com/Vortex_Approach/Vortex.html.
  62. Sillén A. Cognitive tool for dealing with unexpected difficult airway. Br J Anaesth. 2014 Apr;112(4):773–774.
  63. Baker PA, Flanagan BT, Greenland KB, Morris R, Owen H, Riley RH, et al. Equipment to manage a difficult airway during anaesthesia. Anaesth Intensive Care. 2011 Jan;39(1):16–34.
  64. Rowe G, Wright G. The Delphi technique as a forecasting tool: issues and analysis. International Journal of Forecasting. 1999;15:353–375.

Cite this article as follows:

Leeuwenburg T. Airway Management of the Critically Ill Patient: Modifications of Traditional Rapid Sequence Induction and Intubation. Critical Care Horizons 2015; 1: 1-10.