Ruminations on Sepsis
“Except on a few occasions the patient appears to die from the body’s response to infection rather than from the infection itself.” William Osler, The Evolution of Modern Medicine, 1904.
Septic shock is difficult to diagnose and difficult to treat. One of the most daunting problems facing a clinician is a patient who is known to have good heart function, but who has refractory shock despite aggressive fluid resuscitation and the institution of vasopressor drugs, such as dopamine or norepinephrine. It is estimated that septic shock results in approximately 215,000 deaths per year in the United States, a number similar to the number of deaths from acute myocardial infarction.1
Physiology of steroids (in illness)
Addison first noted the essential role of the adrenal for survival in 1855.2 An increase in corticosteroid levels during illness is an important protective response.3 Free cortisol, rather than the protein bound hormone, is thought to be responsible for its physiological effects.4 Corticosteroid release is regulated by the hypothalamic-pituitary axis in a pulsatile manner.2 Vasopressin also plays an important physiological role in adrenal stimulation.5 In sepsis, inflammatory mediators such as cytokines, promote and maintain a high production of corticosteroids.2 The hypothalamic-pituitary-adrenal axis is itself massively activated in septic shock.4 There are many causes for adrenal insufficiency, and certainly this can occur with diseases, bleeding into the adrenal and infections, such as HIV.3 The concept of relative or functional adrenal insufficiency is somewhat controversial and difficult to define.2, 6
Rationale for using steroids in septic shock
Cortisol is commonly referred to as a “stress hormone” because it has several functions in the body’s response to stress. The role of cortisol in sepsis is associated with two factors: (1) hemodynamics and (2) inflammation. Cortisol maintains vascular tone by increasing the number of both alpha and beta adrenergic receptor as well as prevention of receptor desensitization, thereby enhancing vascular tone and cardiac contractility. In addition, cortisol improves vascular tone by blocking nitric oxide synthesis. Cortisol also suppresses the pro-inflammatory process that occurs in sepsis.
- Rashmi Rathor
Critically ill patients at some stage may develop adrenal insufficiency. Glucocorticoid insufficiency may be related to adrenal insufficiency or to a reduced delivery of glucocorticoid to target tissues and cells.7 Resistance of target tissues to steroids may also occur.2 In an animal model of sepsis, bacteremia was induced in male Sprague-Dawley rats by a intravenous injection of Escherichia coli.8 Dexamethasone alone afforded significant protection against Gram-negative bacteremic shock up to eight hours after the bacterial challenge.8 Dexamethasone plus antibiotics improved outcome more than either intervention alone.8 In a monkey model of endotoxemia, dexamethasone was shown to prevent lactic acidosis and hypoglycemia.9
Early experience with steroids in septic shock
In 1976, Schumer published a study in the Annals of Surgery that showed benefit from a short course of high dose corticosteroids in the treatment of septic shock.10 In a double-blind and randomized study, 172 patients in septic shock admitted over an 8-year period were treated with either high-dose steroid or saline. In the saline-treated patients, the mortality rate was 38.4% (33/86); in the steroid-treated patients, it was 10.4% (9/86).10 This was an impressive reduction in mortality.
Negative studies with steroids in septic shock
In a subsequent randomized controlled trial (RCT), however, which enrolled 59 patients with septic shock, high-dose steroids (30 mg/kg methylprednisolone or 6 mg/kg dexamethasone) were shown to improve hemodynamics, but not outcome.11 The same investigators were also unable to show that steroid administration decreased the likelihood of acute respiratory distress syndrome (ARDS).12 A study by Schein showed that plasma cortisol concentrations are increased in patients with septic shock, but that the degree of increase is variable.13 They found that neither patients who reversed their shock nor those who survived to hospital discharge had significantly different plasma cortisol concentrations from those who did not.13 High dose steroids were actually found to increase mortality through infectious complications.14 Among some intensivists, steroids acquired the reputation of accompanying the administration of “last rites.”15 The prudent warning was issued that current evidence provides no support for the use of corticosteroids in patients with sepsis or septic shock, and suggests that their use may be harmful.14 The warning was sounded that recent trials underscored the need for methodologically rigorous trials evaluating new immune-modulating therapies in well-defined critically ill patients with overwhelming infection.14
The pendulum swings, again
In 1992 the Society of Critical Care Medicine began to stratify patient with sepsis into different categories according to the severity of their disease (severe sepsis, septic shock). Around this same time, the role of the adrenal glands in the pathophysiology of severe sepsis and septic shock was becoming more clear. Accordingly, clinical trials designed after this time tended to investigate the affects of low-dose, long term steroid supplementation rather than high-dose steroids. The landmark study of this period was the Annane trial, “Effect of Treatment with Low Doses of Hydrocortisone and Fludrocortisone on Mortality in Patients with Septic Shock.” The study was a placebo-controlled, randomized, double-blind multi-center trial that enrolled patients between 1995 and 1999. Interestingly, the study was the first to separate patients into “responders” and “nonresponders” based on their response to the corticotrophin test (“nonresponders” were those found to be adrenally insufficient). Accordingly, the primary end point was 28 day survival in nonresponders and the study was powered to detect a 20% reduction in mortality between the control and study groups of nonresponders. Treatment consisted of placebo in the control group and 50 mg IV hydrocortisone Q6 hrs x 7 days and 50 micrograms of PO fludrocortisone daily x 7 days in both responders and nonresponders. Outcomes showed a statistically significant reduction in 28-day mortality, ICU mortality and hospital mortality among the nonresponders treated with steroids while the responders showed no statistically significant differences in outcome. Of note, this study failed to demonstrate a statistically significant difference in 1-year mortality of nonresponders and actually showed an increased mortality in responders receiving steroids (albeit, not a statistically significant difference) calling into question the wisdom of a broad-based treatment approach.
- David McFarland
Experts suggested that the increased risk of infection with steroids was associated only with high dose steroids. It was argued that, based on the high proportion of patients who have relative adrenal insufficiency, the benefits of low doses of steroids (200-300 mg/day hydrocortisone) and the minimal risks, low dose steroids should be used to treat septic shock.16 The positive effects of steroids were thought to included reversal of shock, and trends toward decreased organ system dysfunction and decreased mortality.16 In effect a stalemate had been reached and experts were forced to concede that there was still equipoise in relation to the protracted steroid debate.17 In a seminal study published in JAMA, a 7-day treatment with low doses of hydrocortisone (50 mg every six hours) and fludrocortisone (50 mcg/day) reduced the risk of death in patients with septic shock and relative adrenal insufficiency without increasing adverse events.18 The experts agreed to compromise that low doses of corticosteroids are recommended in patients with septic shock requiring vasopressor support.19 In the retrospective CORTICUS study, patients with either baseline cortisol levels <15>
The Corticosteroid Therapy of Septic Shock (CORTICUS) trial was a multicenter, randomized, double-blind study that examined the use of hydrocortisone in patients with septic shock. 499 patients were randomly administered either a tapering dose of hydrocortisone (initial dose 50mg) or placebo q6hour for five days. The patients were also classified into two groups, those with or without adrenal reserve as defined by response to the high-dose ACTH stimulation test. Those with inadequate adrenal reserve had a maximum cortisol increase of less than/equal to 9 mcg/dL whereas those with adequate adrenal reserve had greater than 9 mcg/dL rise in serum cortisol levels. At 28 days, there was no significant difference in mortality between the two study groups (i.e., placebo vs. hydrocortisone) regardless of adrenal reserve. As such, it was determined that that hydrocortisone did not improve survival or reversal of shock in patients with septic shock. Hydrocortisone did, however, expedite the reversal of shock, which was defined as the “maintenance of a systolic blood pressure of at least 90 mm Hg without vasopressor support for at least 24 hours.” The group administered steroids also demonstrated that hydrocortisone was associated with more episodes of superinfection.
- Rashmi Rathor
Interestingly in the largest prospective clinical RCT, the CORTICUS trial, hydrocortisone did not improve survival or reversal of shock in patients with septic shock, either overall or in patients who did not have a response to corticotropin, although hydrocortisone hastened reversal of shock in patients in whom shock was reversed.21 This study did confirm the previous finding that adrenal insufficiency is associated with worse prognosis in septic shock, but checking adrenal response appears to be unhelpful.21 An important consideration is that measurement of total cortisol in critical illness may be misleading; during critical illness, glucocorticoid secretion markedly increases, but the increase is not discernible when only the serum total cortisol concentration is measured.4 Forty percent of critically ill patients with hypoproteinemia may have subnormal serum total cortisol concentrations, even though their adrenal function is normal.4 Measuring serum free cortisol concentrations, when this becomes available, in critically ill patients with hypoproteinemia may help prevent the unnecessary use of glucocorticoid therapy.4 The measurement of salivary cortisol may also prove useful.2 The CORTICUS study concluded with the sentiment, based on Annane’s earlier study18, that hydrocortisone may still have a role among patients who are treated early after the onset of septic shock who remain hypotensive despite the administration of high-dose vasopressors (vasopressor unresponsive).21 It is probable that we have misunderstood and possibly therefore abused adrenal stimulation tests to guide therapy in the past.2 Responses in critically ill subjects are higher than those of healthy volunteers.2 Therefore, the Cosyntropin-induced increment in serum total cortisol should probably not be used as a criterion for defining adrenal function in critically ill patients.2 The results of the CORTICUS study are such that it does not exclude the possibility of the same benefit found by Annane.18 In order to answer the question with more precision, an even larger study will be needed; to detect a 15% reduction in the risk of death, a 2600 patient study would probably be necessary.22 As a word of caution, The hydrocortisone dose 50 mg every six hours may mistakenly labeled as low-dose as it leads to excessive elevation in serum cortisol to values much greater than those in patients with normal adrenal function.2 Lower doses should perhaps be explored.
Steroids in Cardiac Surgery
Steroid administration in this setting is theoretically appealing because they can be given before a predictable potent inflammatory-inducing event, cardiopulmonary bypass. Recent evidence suggests that perioperative steroids may decrease atrial fibrillation without increasing the risk of infection.23 Other studies have not found this. In a study evaluating low dose dexamethasone (8 mg in divided doses), it was found to be beneficial in reducing emetic symptoms and improving appetite after cardiac surgery. However, this dose of the corticosteroids did not seem to prevent atrial fibrillation or to have analgesic-sparing properties.24 Steroids do appear to decrease the need for vasopressors after CPB, but whether this is important is debatable. Large RCTs with steroids in cardiac surgery would be interesting.
Physiology of vasopressin
Vasopressin is a peptide hormone released from the posterior pituitary gland that has multiple physiological effects. It induces vasoconstriction by activating V1 receptors on vascular smooth muscle, a mechanism distinct from that of adrenergic vasoconstriction.1 The most potent stimuli to vasopressin release are increased plasma osmolality, hypotension and hypovolemia.5 It takes a couple of hours to synthesize new vasopressin; it may be depleted with sustained stimuli for release.5 Vasopressin is also responsible for volume and osmolity regulation through its action on V2 receptors.5 Interestingly, vasopressin may cause vasodilatation in some vascular beds through oxytocin receptor stimulation, which results in nitric oxide realease.5
Rationale for using vasopressin in septic shock
Shock is associated with an initial spike in plasma vasopressin levels followed by a sustained fall.5 The reasons for the deficiency are not well understood. The rationale for the use of vasopressin is its relative deficiency in patients with septic shock and the hypothesis that exogenously administered vasopressin can restore vascular tone and blood pressure, thereby decreasing the need for the use of catecholamines, such as dopamine, epinephrine and norepinephrine.5, 25
Early experience with vasopressin in septic shock
Vasopressin for the treatment of septic shock is a new tool in the critical care arsenal with significant interest developing only within the last 10 years. One of the first studies to examine the effects of low-dose vasopressin in patients with septic shock was published in Circulation in 1997. Landry and colleagues enrolled 19 patients meeting standardized criteria for either septic shock or cardiogenic shock in two separate ICUs. All patients received vasopressin. In the septic shock cohort, starting plasma levels of AVP averaged 3.1 pg/mL while the cardiogenic shock cohort’s starting levels averaged 22.7 pg/mL. This suggested a relatively low state of AVP existed in the septic cohort. This study also observed an increase in SBP from a mean of 92 to 146 in the septic shock cohort suggesting a powerful vasoconstrictive effect.
In 1999 Malay and colleagues performed the first RCT to assess the efficacy of vasopressin in septic shock. In all, 10 patients were enrolled in a single institution. They were diagnosed by standardized criteria for septic shock and randomized to receive either vasopressin (o.04 units / min) or placebo in combination with other vasopressors. The primary end point was a rise in systolic arterial blood pressure. They found that patients randomized to the vasopressin group had a mean increased systolic arterial pressure of 27 mmHg (98 +/- 5 to 125 +/- 8 mm Hg, p <>
- David McFarland
Patel and colleagues published a study in 2002 examining vasopressin as an adjunct to spare the use of other vasopressors. In this study 24 patients were admitted to an ICU and randomized to either a 4-hour vasopressin or norepi gtt in a double-blinded fashion. Open label vasopressors were titrated to achieve an acceptable BP and physiologic endpoints (urine output, creatine clearance, gastric mucosal CO2 tension and ST segment analysis) were used to assess end-organ perfusion. At the end of the 4 hour infusions the group randomized to norepi decreased their infusion rate from 20 mcg/min to 17 mcg/min where the AVP group decreased their norepi rate from 25 to 5.3 mcg/min (p <>
Low dose vasopressin (less than 0.1 unit/minute) has been shown to increase blood pressure in patients with septic shock.26 In a small study of septic shock where, in one group vasopressin was added to high dose norepinephrine, the results were compelling. Vasopressin was associated with a massive reduction in norepinephrine requirements, and, intriguingly, in increased urine output and creatinine clearance.26
Negative studies with vasopressin in septic shock
The VASST trial was a multi-center RCT that was conducted to determine whether vasopressin use was associated with decreased mortality in patients with septic shock, especially those who had most severe shock.25 Almost 800 patients with septic shock were randomized to receive vasopressin of escalating doses of norepinephrine. There was no statistically significant difference in mortality between the groups (35.4% vs. 39.3% respectively).25 Interestingly, the 95% confidence interval for absolute risk reduction includes a 10% reduction in mortality attributable to vasopressin.25 Curiously, in post hoc analysis, there was an unexpected finding that patients with less severe sepsis tended to do better with vasopressin. This hypothesis merits further study.25 Importantly, the VASST trial did not evaluate vasopressin in situations where shock is refractory to norepinephrine; it is possible that vasopressin specifically has a role in such circumstances.1 My interpretation of the VASST study is that it does not put the use of vasopressin to rest in sepsis. If anything, it pricks my curiosity further and eases concerns about the safety of vasopressin.
Vasopressin in cardiac surgery
There is limited evidence surrounding the use of vasopressin in cardiac surgery. Anecdotally, we use vasopressin, both during cardiopulmonary bypass and after, when patients have refractory shock. We have no idea what impact this intervention has on outcome. One small (unconvincing) study suggests that the addition of vasopressin is not associated with increased predicted mortality associated with cardiac surgery.27 Specifically, and with scant evidence, vasopressin has been advocated for refractory shock in this setting, especially when there is coexisting pulmonary hypertension.28 In early animal studies, vasopressin was shown to cause pulmonary vasodilation. However, in a well-conducted dog study, vasopressin was shown to cause pulmonary vascular constriction and to exert an important negative inotropic effect on the right ventricle.29 This was not the case when phenylephrine was used to augment systemic BP.29 It may be that at high doses vasopressin causes constriction, but at low doses causes nitric oxide mediated pulmonary vasodilatation.5 One interesting study showed that vasopressin started before CPB in a dose that did not affect BP was associated with decreased vasopressor requirements post-CPB and a shorter ICU stay.30 The results from studies in septic patients may not translate to the cardiac surgery setting. Well-conducted studies in these patients would be valuable.
There are strong anti-vasopressor sentiments, with norepinephrine having been given the nickname “leave-‘em-dead.” There is a widely held view that vasopressor use causally increases the likelihood of organ dysfunction, such as gut and kidney.31 It was thought that if shock resolved and vasopressor dose decreased, outcome would improve. The VASST and CORTICUS trials suggest that shock and vasopressor requirements, as surrogate measures, may not sufficiently reflect severity of the underlying disease process and the association with mortality is not straightforward. It is important to insert the caveat that both of these trials were not sufficiently powered to detect a lesser reduction in mortality that remains clinically important.21, 25 We are, after all, talking about mortality! In concluding about steroids and vasopressin in sepsis, it is tempting to ask, as one article did32, whether steroids and vasopressin are even more effective than activated protein C for the treatment of sepsis. The answer to this question may be that treatments for sepsis are like red wines; they have good years and bad years.
1. Parrillo JE. Septic shock--vasopressin, norepinephrine, and urgency. The New England journal of medicine 2008;358(9):954-6.
2. Arafah BM. Hypothalamic pituitary adrenal function during critical illness: limitations of current assessment methods. The Journal of clinical endocrinology and metabolism 2006;91(10):3725-45.
3. Cooper MS, Stewart PM. Corticosteroid insufficiency in acutely ill patients. The New England journal of medicine 2003;348(8):727-34.
4. Hamrahian AH, Oseni TS, Arafah BM. Measurements of serum free cortisol in critically ill patients. The New England journal of medicine 2004;350(16):1629-38.
5. Holmes CL, Patel BM, Russell JA, Walley KR. Physiology of vasopressin relevant to management of septic shock. Chest 2001;120(3):989-1002.
6. Burchard K. A review of the adrenal cortex and severe inflammation: quest of the "eucorticoid" state. The Journal of trauma 2001;51(4):800-14.
7. Gonzalez H, Nardi O, Annane D. Relative adrenal failure in the ICU: an identifiable problem requiring treatment. Critical care clinics 2006;22(1):105-18, vii.
8. Pitcairn M, Schuler J, Erve PR, Holtzman S, Schumer W. Glucocorticoid and antibiotic effect on experimental gram-negative bacteremic shock. Arch Surg 1975;110(8):1012-5.
9. Schuler JJ, Erve PR, Schumer W. Glucocorticoid effect on hepatic carbohydrate metabolism in the endotoxin-shocked monkey. Annals of surgery 1976;183(4):345-54.
10. Schumer W. Steroids in the treatment of clinical septic shock. Annals of surgery 1976;184(3):333-41.
11. Sprung CL, Caralis PV, Marcial EH, et al. The effects of high-dose corticosteroids in patients with septic shock. A prospective, controlled study. The New England journal of medicine 1984;311(18):1137-43.
12. Schein RM, Bergman R, Marcial EH, et al. Complement activation and corticosteroid therapy in the development of the adult respiratory distress syndrome. Chest 1987;91(6):850-4.
13. Schein RM, Sprung CL, Marcial E, Napolitano L, Chernow B. Plasma cortisol levels in patients with septic shock. Critical care medicine 1990;18(3):259-63.
14. Cronin L, Cook DJ, Carlet J, et al. Corticosteroid treatment for sepsis: a critical appraisal and meta-analysis of the literature. Critical care medicine 1995;23(8):1430-9.
15. Matot I, Sprung CL. Corticosteroids in septic shock: resurrection of the last rites? Critical care medicine 1998;26(4):627-30.
16. Goodman S, Sprung CL. The International Sepsis Forum's controversies in sepsis: corticosteroids should be used to treat septic shock. Critical care (London, England) 2002;6(5):381-3.
17. Annane D, Briegel J, Keh D, Moreno R, Singer M, Sprung CL. Clinical equipoise remains for issues of adrenocorticotropic hormone administration, cortisol testing, and therapeutic use of hydrocortisone. Critical care medicine 2003;31(8):2250-1; author reply 2-3.
18. Annane D, Sebille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone and f ludrocortisone on mortality in patients with septic shock. Jama 2002;288(7):862-71.
19. Keh D, Sprung CL. Use of corticosteroid therapy in patients with sepsis and septic shock: an evidence-based review. Critical care medicine 2004;32(11 Suppl):S527-33.
20. Lipiner-Friedman D, Sprung CL, Laterre PF, et al. Adrenal function in sepsis: the retrospective Corticus cohort study. Critical care medicine 2007;35(4):1012-8.
21. Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients with septic shock. The New England journal of medicine 2008;358(2):111-24.
22. Finfer S. Corticosteroids in septic shock. The New England journal of medicine 2008;358(2):188-90.
23. Halonen J, Halonen P, Jarvinen O, et al. Corticosteroids for the prevention of atrial fibrillation after cardiac surgery: a randomized controlled trial. Jama 2007;297(14):1562-7.
24. Halvorsen P, Raeder J, White PF, et al. The effect of dexamethasone on side effects after coronary revascularization procedures. Anesthesia and analgesia 2003;96(6):1578-83, table of contents.
25. Russell JA, Walley KR, Singer J, et al. Vasopressin versus norepinephrine infusion in patients with septic shock. The New England journal of medicine 2008;358(9):877-87.
26. Patel BM, Chittock DR, Russell JA, Walley KR. Beneficial effects of short-term vasopressin infusion during severe septic shock. Anesthesiology 2002;96(3):576-82.
27. Suojaranta-Ylinen RT, Vento AE, Patila T, Kukkonen SI. Vasopressin, when added to norepinephrine, was not associated with increased predicted mortality after cardiac surgery. Scand J Surg 2007;96(4):314-8.
28. Tayama E, Ueda T, Shojima T, et al. Arginine vasopressin is an ideal drug after cardiac surgery for the management of low systemic vascular resistant hypotension concomitant with pulmonary hypertension. Interactive cardiovascular and thoracic surgery 2007;6(6):715-9.
29. Leather HA, Segers P, Berends N, Vandermeersch E, Wouters PF. Effects of vasopressin on right ventricular function in an experimental model of acute pulmonary hypertension. Critical care medicine 2002;30(11):2548-52.
30. Morales DL, Garrido MJ, Madigan JD, et al. A double-blind randomized trial: prophylactic vasopressin reduces hypotension after cardiopulmonary bypass. The Annals of thoracic surgery 2003;75(3):926-30.
31. Bomzon L, Rosendorff C. Renovascular resistance and noradrenaline. The American journal of physiology 1975;229(6):1649-53.
32. Bradley C. Steroids in sepsis--more effective than activated protein C? Intensive Crit Care Nurs 2001;17(4):242-4.