How accurate are published recurrence rates after rectal prolapse surgery? A meta-analysis of individual patient data

Background The literature shows wide variations of recurrence rates (RRs) after abdominal surgery for rectal prolapse. The aim of this meta-analysis was to evaluate the accuracy of published RRs. Methods An electronic search was performed with no restrictions. Inclusion criterion was abdominal surgery in at least 10 adults with follow-up evaluation of any length of time. Two reviewers screened 1669 references. A total of 190 investigators were asked to provide individual patient data that should be the same used at the time their reports were written. The RR was estimated by actuarial analysis. Investigators were asked for comments on results. Results Individual patient data from 6 reports published with 273 patients (186 women, 87 men) with a median age of 54 years (range, 18–88 y) were available. Abdominal surgery included mobilization with pexy (88%), with additional resection (7%), or mobilization only (5%). There were 16 recurrences at a median follow-up period of 3.94 years (range, .05–15.11 y). The effect of age (hazard ratio [HR], 2.010; P = .3443), sex (HR, 2.070; P = .4260), surgical technique (HR, .743; P = .7669), and publication (HR, 1.014; P = .8747) on RR was not significant. Two publications reported a RR of 0. In another report, the published and estimated RRs of 15% did not differ. Published RRs differed from estimated RRs in 3 reports (2.5% vs 4%; 7% vs 54%; and 9.6% vs 36%). The pooled odds ratios of 6 reports revealed a borderline significant difference between the published and estimated RRs (P = .066). Conclusions Published RRs differed by as much as 47% from the RRs estimated by actuarial analysis depending on event definition and how the data were censored.