Abstract & Commentary
This large, multi-institution, placebo-controlled, blinded study sought to identify the effect of replacement doses of adrenocortocosteroids on patients in septic shock with and without adrenal insufficiency. A total of 300 patients with clinical signs of severe infection and hypotension unresponsive to fluids and 5 mg/kg of dopamine (for at least 1 hour), requiring mechanical ventilation, and having oliguria or a PaO2/FiO2 ratio less than 280 mm Hg, were randomized to receive hydrocortisone 50 mg IV every 6 hours and 50 mg fludrocortisone by tube, or corresponding placebos, daily for 7 days. On entry, patients underwent a short cortrosyn stimulation test with a 250 mg bolus of tetracosactin and cortisol levels were determined before and at 30 and 60 minutes after drug administration. Patients who did not demonstrate a cortisol level increase of at least 9 mg/dL at 30 or 60 minutes were considered nonresponders (NR) indicating relative adrenal insufficiency. Study end points were 28-day, ICU, hospital, and 1-year mortality rates. Following complete, blinded data collection, 4 treatment groups were defined: Nonresponders treated with steroids (NR-S), nonresponders treated with placebo (NR-P), responders treated with steroids (R-S), and responders treated with placebo (R-P).
Of the 300 enrolled patients, 299 completed the study: in the one dropout, consent was withdrawn. The 4 groups of patients were well balanced in general characteristics and severity of illness. There were no significant differences in age, sex, co-morbidities, acute physiologic measures, time to treatment, dose or type of vasopressor, antibiotics or duration of antibiotic treatment, or level of mechanical ventilatory support. A total of 77% of the studied patients were classified as NR and only 23% (70 patients) were R. The use of steroids compared to placebo in all patients (NR-S plus R-S compared to NR-P plus R-P) resulted in a trend to lower mortality at all study points: 55% vs. 61% at 28 days; 60% vs. 68% at ICU discharge; 63% vs. 69% at hospital discharge; and 68% vs. 75% at 1 year. The difference was statistically significant when only the NR were considered: 53% vs. 63% at 28 days (P = .04); 58% vs. 70% at ICU discharge (P = .02); 61% vs. 72% at hospital discharge (P = .04); and 68% vs. 77% at 1 year (P = .07). The R, a much smaller group, showed no benefit at any point. There was no trend toward a worse outcome from steroid use in the R group. The improved mortality at 28 days from steroid treatment in all patients results in 1 life saved for every 8 patients treated.
In addition to the dramatic improvement in mortality, withdrawal from vasopressors was significantly shorter in the treatment group, 7 vs. 9 days. Adverse events possibly attributable to steroid use, including additional infections, GI bleed, and psychosis were not more frequent in the steroid-treated patients (Annane D, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA. 2002;288:862-871).
Comment by Charles G. Durbin, Jr., MD
After a period of enthusiastic use many years ago, steroid treatment of sepsis and septic shock has been challenged by many investigators and has been abandoned by most clinicians in the last decade. However, recent evidence suggests that there may be a subgroup of patients in septic shock whose adrenal response is insufficient. When these patients are given maintenance rather than "industrial" doses of steroids, their clinical status quickly improves and they can often be weaned from vasopressors. Effects on mortality have been equivocal. This large, well-executed study strongly supports the use of replacement/maintenance dose steroids in patients with severe sepsis. The effect on mortality of this simple replacement therapy is as great as that reported for the new drug, drotrecogin alpha (Xigris). The cost of steroid therapy is trivial compared to the cost of this new anticoagulant, anti-inflammatory agent.
On the basis of this study, it seems reasonable to provide a short course (7 days) of maintenance steroids to septic patients on vasopressors who require mechanical ventilation. While there were no adverse events (or benefits) in the responders, the number of patients in this group is small and it is possible that adverse events are more likely with steroids. It would be prudent to use an adrenal stimulation test to stop the steroids in those patients with a normal response to avoid any risk.
This is a remarkable study in many ways. All of the patients were followed for a year following their septic shock treatment and clinical and functional outcomes were collected—an impressive feat. Very few large clinical studies are successful in obtaining this degree of information and long-term follow-up. A wide variety of patients, both surgical and medical, were included allowing the findings to be applied to diverse patient groups.
There are many unanswered questions: What is the proper steroid drug and dose? Would a slightly higher dose further improve outcomes? When do the side effects and complications begin to increase? Is the mineralocorticoid agent (fludrocortisone) necessary? Will this therapy be useful earlier in the course of sepsis? What is the appropriate duration of therapy? Should the treatment be continued longer in patients still on vasopressors after 1 week?
Not reported in the paper was the effect of the steroid administration on blood glucose levels. Recent data indicate that tight control of serum glucose levels with insulin administration improves ICU outcomes:1,2 Does the use of steroids make this issue worse?
While this is only a single study, the impressively positive results, the lack of complications, and the insignificant cost of therapy suggest that vasopressor-dependent patients in septic shock be given the benefit of a week of steroid replacement therapy. n
1. van den Berghe G, et al. Intensive insulin therapy in the critically ill patients. N Engl J Med. 2001;345(19): 1359-1367.
2. van Den Berghe G. Beyond diabetes: Saving lives with insulin in the ICU. Int J Obes Relat Metab Disord. 2002;26(Suppl 3):S3-8.