steroids in sepsis
This is more in-depth explanation of the steroid debate - for a more concise explanation + case study, see the Severe CAP podcast we did on the steroid debate - click here
One thing you should know working in critical care medicine that the word steroid should make you shudder. While there are some clear cut indications, steroids have been involved in a lot of gray-area topics that spark debate between providers. But anyone new to the ICU - or naive to the history of the topic - may be asking, why?
The idea is that normal patients who are critically ill are under large amounts of stress - and not just emotional. Physiological stress from hypotension, severe infection and critical illness causes the HPA (hypothalamic-pituitary-adrenal) axis to be activated; this causes diurnal variation to be lost and serum cortisol to increase via secretion from the adrenal glands. When a patient responds the WRONG way in response to physiological stress and cortisol production is sub-optimal this is deemed “relative adrenal insufficiency”. Specifically in the ICU this is also referred to as “critical-illness-related corticosteroid insufficiency”.
This idea of sepsis being an inflammatory issue - and steroids getting rid of inflammation - caused both serum cortisol response + exogenous administration of cortisol to be a focus of study.
Based on our physiologic knowledge of the stress response, the theoretical benefits of steroids could be:
Decease in transcription of pro-inflammatory cytokines
Prevention of neutrophil aggregation
Inhibit inducible form of nitric oxide
Prevent compliment activation
Increase transcription of anti-inflammatory cytokines, and more
It makes sense that steroids could help in sepsis - but the research has taken us on a roller coaster.
History of Research
Although all the trials below can be daunting - there are hundreds on steroids alone - getting a general idea about how controversial steroids are and how the medical community’s opinion of them has waxed and waned can give you background into how we got where we are now.
The First trials in 1960s through early 1980s - steroids fall into favor; they reverse shock and short term survival
The earliest clinical trials on steroids started in the 1960s.
1963 (Bennett) - first prospective randomized double blind placebo-controlled studies of steroid admin; 100mg of oral hydro-cortisone vs placebo for patients w/severe sepsis and septic shock
Found no significant survival difference between treatment/control groups
So basically, no benefit. Things with steroids stayed relatively quiet until Schumer shook up the steroid game in the late 1970s. This is when the idea of 30 cc/kg IV methylprednisolone came into favor (a dose of a specific drug seen in a lot of future trials).
1976 (Schumer) - combined a prospective trial with a retrospective analysis contrasting two steroid regimens with placebo in adult surgical patients w/septic shock
Findings: Mortality rates of 11.6% and 9.3% were seen for patients treated with methylprednisolone (one or two doses of 30 mg/kg IV) and dexamethasone (one or two doses of 3 mg/kg IV), respectively, vs 38.4% for patients receiving placebo
This paper caused high dose corticosteroids to be the standard practice in septic shock throughout the late 1970s/early 1980s; although it has a lot of limitations based on our current understanding of sepsis + the evolution of study designs.
The Late 1980s - steroids fall out of favor, they are not beneficial and may even do harm (incr complications/mortality)
Curiosity began to peak, and in the 1980s, more studies were done; none showed any survival benefit.
1984 (Sprung) - 59 patients in single center either received methylprednisolone 30 cc/kg IV, dexamethasone 6 mg/kg IV or placebo
Findings: more rapid shock reversal but no survival benefit
1987 (Veterans Administration Trial) - RCT w/ ~200 patients randomized to methylprednisolone 30 cc/kg followed by 5 cc/kg, or placebo, within 3 hours of dx
Findings: no difference in 14 day mortality or complications
1987 (Bone) - RCT w/ 381 patients who received methylprednisolone 30 cc/kg or placebo
Findings: higher mortality in steroid group
1988 (Luce) - 87 patients received either 30 cc/kg methylprednisolone or placebo (mannitol) for a total for 4 doses following dx of septic shock
Findings: no improvement in survival or prevalence of ARDS
Because of these four studies, steroid use in septic shock began to dwindle again in the late 1980s/1990s. High dose (supraphysiological) were recognized to have no benefit; and based on the Bone study, potentially be harmful
Moving onto 2000s - The Annane study gives steroids a good rep, but then CORTICUS shuts it back down
After steroids began to fall out of favor, the landmark Annane study came back to reignite the debate about potential benefits. It was the most definitive trial of stress dose steroids in septic shock done up to that point; it was a small study but VERY well designed which greatly influenced intensivists and made steroids the new standard of care once again. These positive results became the basis for national and international guidelines.
2002 (Annane) - a french study in 19 ICUs; formed our basis for our use of exogenous glucocorticoids in shock for several years. Patients received placebo or hydrocortisone (50 mg IV qh) + fludrocortisone (50 mcg enterally) for 7 days
Findings: Hydrocortisone administration was associated with decreased 28-day mortality (53% vs. 63 %), ICU mortality (58% vs. 70%), hospital mortality (61% vs. 72%) and shorter duration of vasopressor use in those patients with inadequate adrenal reserve (the ‘non responders’)
Limitations - 24% of patients received etomidate (inhibits CS synthesis), timing of antibiotics were delayed compared to current standards
Although the Annane study had such positive results there will still questions about the timing and duration of steroid doses and the utility of hydrocortisone/fludrocortisone. One of the largest studies to date to address the role of corticosteroids in septic shock was done - the CORTICUS study.
It was aimed at confirming the survival benefit found in the Annane study, but the results ended up being quite different. The CORTICUS study outcomes sharply contrasted those of the Annane Trial. It helped re-establish the idea that corticosteroids should not be routinely used in septic shock. It also helped fuel the refuel controversy of the sepsis/steroids debate.
2008 CORTICUS - evaluated the efficacy and safety of low-dose hydrocortisone therapy in a broad population of patients with septic shock, including patients who responded to a corticotropin test, in whom a benefit was unproven. The use of low-dose hydrocortisone had no significant effect on 28-day mortality, regardless of the patients' adrenal responsiveness to corticotropin. Click here to see our more detailed review
2010s - Now; still a lot of debate
2010 (COIITSS) - multi-center (11 ICUs in France) RCT involving 509 adults with septic shock and SOFA>8 looking at insulin therapy in patients treated with hydrocortisone as well as the benefit of fludrocortisone. Compared with conventional insulin therapy, intensive insulin therapy did not improve in-hospital mortality among patients who were treated with hydrocortisone for septic shock. The addition of oral fludrocortisone did not result in a statistically significant improvement in in-hospital mortality
2012 (Sherwin) - sepsis sub-committee of american academy of EM clinical practice committee; identified 7 relevant trials and concludes low dose corticosteroids may reverse shock but not improve mortality
2014 (Wang) - 8 RCTs, low dose hydrocortisone; no change in mortality but decreased shock at 7/28 days
2015 (Cochrane Review) - 33 RCTs accounting for ~4000 patients with sepsis; 3 trials included children, and remaining 30 included adults. Low quality evidence indicates steroids reduce mortality among patients with sepsis; mod quality evidence supports that a long course of low dose steroids reduced 28 day mortality without major complications.
2016 (VANISH) -421 patients who had 2/4 SIRS + known infection (sepsis) who required vasopressors despite IVF; multicenter and double blinded RCT. Studied two drugs; drug 1 (noradrenaline or vasopressin) and drug 2 (hydrocortisone or placebo). Found vasopressin reduces levo requirement and RRT requirement but does not reduce RRT free days or mortality rate; no clinical interaction with corticosteroids
2016 (Hydro-cortisone, Vitamin C and Thiamine for Treatment of Severe Sepsis and Septic Shock) - a single center unblinded retrospective observational study of 47 patients studying vitamin C + Hydrocortisone + thiamine in addition to standard patient; improvement of hospital mortality (8.5%) compared to control group (40.4%), p<0.001
2016 (HYPRESS) - a randomized double blind multicenter trial in 34 sites in Germany; hydrocortisone compared to placebo preventing development of septic shock. Was found that hydrocortisone did not prevent the development of shock in patients wit with severe sepsis
2017 (Gibbison) - 22 studies, no evidence that one corticosteroid drug or treatment regimen is more likely to be effective in reducing mortality or reducing incidence of GI bleeding/superinfection in septic shock. Hydrocortisone as a bolus/infusion was more likely than placebo + methylprednisolone to result in shock reversal
2017 (Early initiation of low-dose hydrocortisone treatment for septic shock in adults: a randomized clinical trial) - 118 patients with septic shock received hydrocortisone OR 0.9% saline initiated simultaneously with vasopressors; early initiation of low dose hydrocortisone did not decrease risk of mortality and the length of stay in the ICU/hospital in adults with septic shock
2018 ADRENAL Trial -