Intrapartum determinants of early postpartum pelvic floor electrophysiological dysfunction and clinical symptoms: a retrospective cohort study linking delivery records and pelvic floor assessment data

Abstract Background Vaginal delivery is the principal modifiable determinant of postpartum pelvic floor dysfunction, yet how specific intrapartum exposures relate to objective pelvic floor electrophysiology and to early clinical symptoms remains incompletely quantified. We aimed to quantify the association between perineal injury severity and early postpartum pelvic floor surface electromyography (sEMG), and to compare the determinants of electrophysiological impairment with those of clinical symptom diagnoses. Methods Retrospective cohort study at a single tertiary maternity center in China. Integrated delivery–pelvic floor assessment records (2023–2026) were analyzed for women with vaginal delivery and a valid Glazer-protocol sEMG assessment, grouped by perineal status: intact/first-degree tear (n = 134), second-degree tear (n = 567), or episiotomy (n = 428). Outcomes were five sEMG phases, manual muscle testing, and clinician-recorded diagnoses (pelvic organ prolapse, stress urinary incontinence [SUI], dyspareunia) available for 1,095 women. Multivariable linear, logistic, and ordinal models, restricted cubic splines, and propensity-score matching were used, adjusting for the delivery-to-assessment interval (median 49 days). Results Among 1,129 women, contractile sEMG values declined across perineal categories in crude analyses (flicks maximum 30.1 vs 30.5 vs 27.4 µV, P = 0.003), but the gradient was fully attenuated after adjustment (episiotomy vs intact: β=+0.03, 95% CI − 3.11 to 3.17, P = 0.99). Forceps delivery (β=−5.36, 95% CI − 10.01 to − 0.71, P = 0.024) and second-stage duration (β=−1.08 µV per 30 minutes, P = 0.001) were the independent correlates of impaired contractility—confirmed in propensity-score-matched (β=−5.86, P = 0.030) and ordinal models. Second-stage durations ≥ 90 minutes were associated with reduced tonic contraction (β=−2.07, P = 0.010). In contrast, symptom diagnoses showed a reversed crude gradient (SUI 65.4% vs 56.6% vs 47.2%, P < 0.001) that likewise attenuated after adjustment (episiotomy OR 0.64, P = 0.056); multiparity, not intrapartum exposures, independently predicted SUI (OR 1.50, P = 0.018), prolapse (OR 1.57, P = 0.023), and any symptom (OR 1.72, P = 0.005). Conclusions Perineal injury severity is a marker of difficult labor rather than an independent driver of early postpartum pelvic floor impairment. Objective electrophysiological dysfunction tracks the intrapartum process (forceps, prolonged second stage), whereas early symptom burden tracks cumulative obstetric history (parity)—two complementary windows on childbirth-related pelvic floor injury that may warrant distinct referral strategies.