Recovery Time of Diaphragm Paralysis After Congenital Heart Surgery: A Retrospective Single-Center Cohort Study

Objective(s): To estimate the time to recovery from diaphragm paralysis after congenital heart surgery using interval-censored survival analysis and to characterize cumulative recovery probability over 5 years. Methods We retrospectively reviewed 111 consecutive patients < 18 years who developed diaphragm paralysis after congenital heart surgery at a tertiary center in Japan (2009-2024); 110 were analyzed after excluding 1 patient without imaging follow-up. Recovery of diaphragmatic motion was the primary outcome. Because recovery was detectable only at discrete imaging assessments, interval-censored survival methods (Turnbull estimator) were used, followed by parametric accelerated failure time modeling and a Bayesian re-analysis incorporating informative priors from a comparable published cohort. Results Recovery was confirmed in 84 patients (76.4%); 26 (23.6%) were right-censored. The median recovery time was 44 days (95% confidence interval, 41-75 days). Cumulative recovery probability was 60.5% at 60 days, 74.5% at 90 days, and 84.9% at 1 year, plateauing near 85% after 2 years with a slight increase to 87.1% at 5 years. A log-logistic distribution provided the best parametric fit. The Bayesian model yielded a posterior median recovery time of 69 days (95% credible interval, 47-102 days), and sensitivity analyses yielded consistent estimates. Conclusions Diaphragm paralysis after congenital heart surgery resolves in approximately three of four patients, with most recovering within 3 months. However, approximately 15% may not recover within a clinically meaningful timeframe, supporting the need for long-term follow-up and multicenter studies to identify risk factors for non-recovery and establish generalizable recovery time estimates.