Key Point

In sepsis and septic shock, long-term, low-dose therapy using corticosteroids provides a survival benefit.

PARIS—For severe sepsis, the overall hospital mortality rate is 30%, and for septic shock, it is 50% to 60%. A perennial debate on treating septic shock has centered on the potential benefits of corticosteroid therapy. A recent Cochrane meta-analysis of 16 randomized trials evaluated the effects of corticosteroids on patients with sepsis and septic shock. The results suggest that corticosteroids do not significantly decrease mortality—except when low doses are used for long periods.¹

Asked to explain the controversy, Djillali Annane, MD, PhD, a Professor of Medicine at the School of Medicine of Paris Ile de France Ouest, University of Versailles Saint Quentin en Yvelines, said that corticosteroids have been used in sepsis since the 1950s. “However, three randomized controlled trials in the 1980s, driven by worldwide experts, had negative results and cast serious doubt on the efficacy of corticosteroids.” He added that although the survival benefit seen with moderate doses of corticosteroids is becoming more obvious, many leaders in critical care are not yet ready to reconsider their position.

THE CURRENT STUDY

The meta-analysis included all studies reporting on intravenous treatment with any corticosteroid preparation. Treatment duration was considered long if it lasted five days or more at full dose and short if given for less than five days. Low-dose therapy was defined as 300 mg/day or less, and high-dose therapy was more than 300 mg/day. Control groups received standard treatment (antibiotics, fluid replacement, vasopressors, mechanical ventilation, or renal replacement therapy) as needed, given either alone or with a placebo.

The primary outcome was all-cause mortality at 28 days. Secondary outcomes included ICU and hospital mortality, as well as number of patients with reversal of shock at seven and 28 days.

All relevant studies up to 2003 were identified and graded for methodological quality. Data were organized so that a relative risk less than 1.0 favored corticosteroids—with the exception of shock reversal at days 7 and 28, for which a relative risk greater than 1.0 favored corticosteroids.

Sixteen trials were approved for the meta-analysis. The study population for evaluating all-cause 28-day mortality was 2,022 patients. Of the 1,033 patients receiving corticosteroids, 351 died (34%), compared with 329 of the 989 control patients (33%). The relative risk of dying at 28 days was 0.92.

In 425 ICU patients from four trials, 108 of 216 corticosteroid-treated patients died (50%), compared to 127 of 209 untreated patients (61%). Of 13 trials used to evaluate hospital mortality in 1,418 patients, 279 of 730 corticosteroid-treated patients died (38%), compared to 271 of 688 untreated patients (39%), yielding a relative risk of 0.89.

A subgroup analysis of 465 patients who received long courses of low-dose corticosteroids yielded a relative all-cause mortality risk at 28 days of 0.80, thus showing a reduction in mortality. Yet, a subgroup analysis of eight trials with a total of 1,115 patients showed no survival benefit with short courses of high-dose corticosteroids.

Again, a subgroup analysis of 425 patients looking at shock reversal after seven days showed an increased rate of reversal for those treated with long courses of low-dose corticosteroids (50% vs 31% without treatment). At 28 days, shock reversal occurred in 54% of patients treated with long courses of low-dose corticosteroids, compared to 43% of untreated patients.

Corticosteroids did not increase the rate of serious adverse events such as gastrointestinal bleeding, superinfections, or hyperglycemic episodes.

Overall, the analysis found no evidence that corticosteroids had a beneficial effect. However, for 28-day all-cause mortality and in-hospital mortality from severe sepsis or septic shock, long courses of low-dose corticosteroids reduced mortality in both the hospital and the ICU.

WHICH PATIENTS BENEFIT MOST?

Corticosteroids have various properties, including vascular, metabolic, and anti-inflammatory effects, noted Dr. Annane, who is also head of the ICU at Raymond Poincaré Hospital in Garches, France. “It has been recognized that both animals and patients with sepsis demonstrate dysfunction of cortisol production or uptake by tissue. This is why physicians have reconsidered using corticosteroids as hormone replacement therapy.” However, he continued, there is no consensus as to which patients will benefit most from this therapy.

The recommended dosage of 200 to 300 mg/day corresponds to the corticosteroid dose that results in serum cortisol levels similar to those achieved during maximal exercise stress in healthy adults, explained Dr. Annane.

The improvement seen with corticosteroids may be a result of reduced duration of shock, illustrated by the higher rates of shock reversal in steroid-treated patients, reduced inflammation, and fewer organ dysfunctions. Previous studies have shown that corticosteroids prevent the recruitment and proliferation of inflammatory cells.

“In my opinion,” said Dr. Annane, “there is no doubt that corticosteroids at low doses have a potent biological activity and that they may help in treating septic shock.” The true issue, he stressed, is who should be treated. “Corticosteroids are definitely not a magic bullet. Efforts should focus on identifying the patients who need to be treated.”

—Gale Jurasek

Reference
1. Annane D, Bellissant E, Bollaert PE, et al. Corticosteroids for severe sepsis and septic shock: a systematic review and meta-analysis. BMJ. 2004;329:480-488.

Return to table of contents