Systemic corticosteroids reduce need for ICU or mechanical ventilation in patients with community-acquired pneumonia

Référence

Saleem N, Kulkarni A, Snow TA, et al. Effect of corticosteroids on mortality and clinical cure in community-acquired pneumonia: a systematic review, meta-analysis, and meta-regression of randomized control trials. Chest 2023;163(3):484-497.

Sommaire

According to current UK guidelines and current Infectious Diseases Society of America guidelines, systemic corticosteroids are not recommended for any patients with community-acquired pneumonia (CAP). However, another July 2023 POEM presented the results of the largest study to date, finding a significant mortality reduction with 8 to 14 days of intravenous hydrocortisone for patients with severe CAP. This study is a methodologically sound meta-analysis of randomized trials comparing a systemic corticosteroid with a placebo or usual care and reporting all-cause mortality. The authors identified 16 randomized trials with a total of 3842 hospitalized patients. A variety of corticosteroids were used: most commonly, intravenous hydrocortisone (6 studies). Steroids were given for 7 or more days in 10 trials and for 5 or fewer days in 6 trials. No mortality benefit was seen (9.5% vs 10.8%; 95% CI 0.67 - 1.07; I2 = 14%). The authors performed meta-regression to evaluate the impact of study design and population factors on mortality. They found an association between greater baseline severity of illness and lower mortality, but not with dose or duration of treatment. However, with only 16 trials for meta-regression, the power to detect associations is limited. In 6 studies with 2619 patients that reported the need for intensive care unit (ICU) admission, the outcome was lower in the corticosteroid groups (3.1% vs 4.7%; P = .04; I2 = 0%; number needed to treat [NNT] = 62). In 8 smaller studies with 1457 patients, the need for mechanical ventilation was also reduced (4.2% vs 7.1%; P = .001; NNT = 34). The authors found no difference in the likelihood of treatment failure, gastrointestinal bleeding, or secondary infections, but found an increase in hospital readmission (5 studies; 2853 patients; 21.5% vs 17.7%; P = -.008; number needed to treat to harm = 26) and hyperglycemia.