BACKGROUND: Pneumocystis jirovecii pneumonia in HIV-negative immunocompromised patients has a hospital mortality rate of 30-50%. Adjunctive corticosteroids improve outcomes of P jirovecii pneumonia in HIV-positive patients. The aim of this trial was to assess the effects of early adjunctive corticosteroid therapy for 21 days in HIV-negative patients with P jirovecii pneumonia responsible for acute hypoxaemia respiratory failure.
METHODS: This multicentre, double-blind, randomised controlled trial was conducted at 27 hospitals in France. We included patients with acute respiratory failure, aged 18 years or older with mild-to-severe hypoxaemia, microbiological documentation of P jirovecii pneumonia, and anti-Pneumocystis treatment duration of less than 7 days. Patients were randomly assigned (1:1) to the corticosteroid group (adjunctive corticosteroid therapy of methylprednisolone intravenously, 30 mg twice per day from days 1 to 5, 30 mg once per day from days 6 to 10, and 20 mg once per day until day 21) or placebo group (2 mL or 3 mL syringes of isotonic saline intravenously) using a web-based system. Permutation blocks of fixed size unknown to the local investigators were used. Stratification factors were centre, long-term corticosteroid treatment started more than 1 month before enrolling in the trial, underlying disease (malignancy vs other), and oxygen needs at randomisation (<6 vs =6 L per min). The primary outcome was all-cause 28-day mortality defined as the proportion of patients who died within 28 days, analysed in the intention-to-treat (ITT) population. This trial was registered on ClinicalTrials.gov, NCT02944045 (closed).
FINDINGS: From Feb 23, 2017, to Feb 23, 2024, 466 patients with acute respiratory failure were assessed for eligibility. Of those, 240 were excluded and 226 patients were randomly assigned (114 assigned to the placebo group and 112 assigned to the corticosteroid group). The ITT population included 111 patients in the placebo group and 107 in the corticosteroid group. Median age was 67 years (IQR 59-73). 126 (58%) patients were male and 92 (42%) were female. Nearly all patients (208 [95%]) were in the ICU or intermediate care at randomisation. The median time from P jirovecii pneumonia diagnosis to corticosteroid therapy initiation was 3 days (IQR 2-5). Patients received trial treatment for 13 days (range 7-20). All-cause 28-day mortality occurred in 36 (32·4%) patients in the placebo group versus 23 (21·5%) in the corticosteroid group (mean difference 10·9% [95% CI -0·9 to 22·5]; p=0·069). There were no significant differences in safety outcomes between groups, especially for all secondary infections (38 [34·2%; 95% CI 25·4 to 43·1] patients in the placebo group vs 25 [23·4%; 15·3 to 31·4] in the corticosteroid group) or insulin needs (25 [22·5%; 15·1 to 31·4] vs 33 [30·8%; 22·3 to 40·5]).
INTERPRETATION: In immunocompromised HIV-negative patients with P jirovecii pneumonia, adjunctive corticosteroid treatment did not significantly decrease 28-day mortality.
FUNDING: French Ministry of Health.
| Discipline Area | Score |
|---|---|
| Hospital Doctor/Hospitalists |  |
| Internal Medicine | |
| Intensivist/Critical Care |  |
| Respirology/Pulmonology | |
| Infectious Disease | |
These results hinge on the mortality assumptions being 40% with placebo and 20% with steroids. Would the results change if mortality assumption with placebo is 50%? Would the results change if patients received the trial treatment for 21 days as was intended? Is there a biological explanation for steroids benefiting HIV patients but not other immunocompromised patients? The absence of harm from steroids is reassuring. I think the jury is still out.
Although there was a trend toward lower 30-day mortality with steroids, it did not meet the criteria for statistical significance; however, the small sample size may have reduced power to detect a difference (90-day mortality was less). Given these factors and other limitations (long enrollment period, heterogeneous population for the underlying disease), it is unclear whether this study will be influential.
This is a sound and very relevant paper. but the conclusion is totally wrong. Early adjunctive corticosteroid therapy DOES IMPROVE the short-term outcome of 28-day mortality in HIV-negative immunocompromised patients with P jirovecii pneumonia. But the 11% mortality rate reduction they found was not statistically significant. This is a very different meaning. The assumption for sample size calculation was a 20% absolute risk reduction (from 40% to 20%), which obviously is overly optimistic, and the real reduction was from 32.4% to 21.5%. Look at the total of outcomes and adverse events in table 2. We need to be very clear about the difference between clinically and statistically significant results.
