Thirty-five RCTs were included in the review (n=2,958). No individual Jadad scores were provided.
Systemic corticosteroids (10 RCTs, n=959, not all eligible for meta-analysis): Use of systemic corticosteroids (compared to placebo) reduced treatment failure by 46%, RR 0.54 (95% CI: 0.41, 0.71), length of hospital stay by 1.4 days, WMD -1.42 (95% CI: -2.18 to -0.65), and improved forced expiratory volume in one second (FEV1) by 0.13 litres after 3 days of therapy. However, systemic corticosteroids were associated with an increased risk of hyperglycaemia, RR 5.88 (95% CI: 2.40, 14.41).
Antibiotics (11 RCTs, n=1,020, not all eligible for meta-analysis): Use of antibiotics (compared to placebo) reduced in-hospital mortality by 78%, RR 0.22 (95% CI: 0.08, 0.62), and treatment failure by 46%, RR 0.54 (95% CI: 0.32, 0.92), although statistically significant heterogeneity (p=0.002) was seen. Stratification by patient type showed the treatment effect was significant for hospitalised patients, but not for outpatients.
NPPV (14 RCTs, n=979, not all eligible for meta-analysis): Use of NPPV (compared to standard therapy) reduced in-hospital mortality by 55%, RR 0.45 (95% CI: 0.30, 0.66), reduced length of hospital stay by 1.94 days, WMD -1.94 (95% CI: -3.87, -0.01), although there was significant heterogeneity (p=0.005), and reduced risk of intubation by 65%, RR 0.35 (95% CI: 0.26, 0.47), with the beneficial effect increasing as patient baseline pH decreased.