Comparative safety of sugammadex versus neostigmine for reversal of neuromuscular blockade in children: a systematic review of randomized and non-randomized studies with an integrated pharmacovigilance strand
Abstract
Sugammadex reverses aminosteroidal neuromuscular blockade faster than neostigmine, and its pediatric efficacy is established, but its comparative safety for rare, serious harms is far less studied. This review synthesized the comparative safety of sugammadex versus neostigmine in children (birth to under 18 years), integrating randomized trials, non-randomized comparative studies, and a descriptive pharmacovigilance and case-report strand. Free-access databases, trial registries, citation chasing, and the FDA Adverse Event Reporting System were searched until 18 July 2026, restricted to freely accessible sources by pre-registered decision. Screening, eligibility assessment, and risk-of-bias appraisal each used two independent passes with third-reviewer adjudication, using Cochrane RoB 2, ROBINS-I, and the Murad 2018 tool. Given rare, heterogeneously ascertained harm, a Synthesis Without Meta-analysis was performed with per-study absolute-risk tabulation, an effect-direction plot, and GRADE certainty rating; no meta-analysis was undertaken. Of 719 records identified, 44 eligible reports (43 studies) were included: 20 comparative and 23 uncontrolled descriptive. No randomized trial was at low risk of bias for its harm outcomes. No anaphylaxis occurred in any of five estimable comparative studies (0/196 versus 0/168); severe bradycardia was rare and numerically lower with sugammadex in the two best-reporting trials; respiratory adverse events were equivalent where blinded and pre-registered (25/86 versus 26/86; risk ratio 0.96, 95% confidence interval 0.61â"1.52); recurarization, measured objectively in one non-randomized study, favored sugammadex (0% versus 29.7%). All four primary outcomes were very low certainty. Comparative safety for serious pediatric harms is not established; this is an evidence gap, not demonstrated equivalence.