Comparative evidence on harms in pediatric randomized clinical trials from less developed versus more developed countries is limited

Objectives: Evaluate comparative harm rates from medical interventions in pediatric randomized clinical trials (RCTs) from more developed (MDCs) and less developed countries (LDCs). Study Design and Setting: Meta-epidemiologic empirical evaluation of Cochrane Database of Systematic Reviews (June 2014) meta-analyses reporting clinically important harm-outcomes (severe adverse events [AEs], discontinuations due to AEs, any AE, and mortality) that included at least one pediatric RCT from MDCs and at least one from LDCs. We estimated relative odds ratios (RORs) for each harm, within each meta-analysis, between RCTs from MDCs and LDCs and calculated random-effects-summary-RORs (sRORs) for each harm across multiple meta-analyses. Results: Only 1% (26/2,363) of meta-analyses with clinically important harm-outcomes in the entire Cochrane Database of Systematic Reviews included pediatric RCTs both from MDCs and LDCs. We analyzed 26 meta-analyses with 244 data sets from pediatric RCTs, 116 from MDCs and 128 from LDCs (64 and 66 unique RCTs respectively). The summary ROR was 0.92 (95% confidence intervals: 0.78–1.08) for severe AEs; 1.13 (0.54–2.34) for discontinuations due to AEs; 1.10 (0.77–1.59) for any AE; and 0.99 (0.61–1.61) for mortality and for the all-harms-combined-end point 0.96 (0.83–1.10). Differences of ROR-point-estimates ≥2-fold between MDCs and LDCs were identified in 35% of meta-analyses. Conclusion: We found no major systematic differences in harm rates in pediatric trials between MDCs and LDCs, but data on harms in children were overall very limited.