Effect of postoperative hormonal therapy on pain following endometriosis surgery: a retrospective cohort study

Endometriosis is heterogeneous, and evidence for individualised postoperative management is limited. We assessed whether postoperative hormonal therapy exposure is associated with one-year changes in pain domains after first-time endometriosis surgery and explored factors related to treatment allocation.

Retrospective cohort study (Erlangen, 09/2019-06/2024) of premenopausal women (18-45 years) undergoing first minimally invasive surgery for endometriosis-associated pain with one-year follow-up. Patients were classified into four hormonal therapy (HT) patterns based on baseline and follow-up status: HT none (n = 73), HT initiation (n = 57), HT discontinuation (n = 21), and HT continuation (n = 38). For each pain domain, adjusted analysis of covariance (ANCOVA) models estimated one-year follow-up Numeric Rating Scale (NRS) differences, and therapy-group effects were tested using robust Wald omnibus tests with Holm multiplicity adjustment across domains, and post-hoc contrasts were Holm-adjusted within domains and interpreted only for domains with significant omnibus tests. HT-group membership was analysed using multinomial logistic regression.

After multiplicity adjustment across domains, therapy-group differences were observed only for dysmenorrhoea/pelvic pain (p < 0.001). Within this domain, only HT initiation was associated with lower one-year follow-up scores versus HT none (adjusted difference -2.39 NRS points, 95% confidence interval (CI) -3.58 to -1.21; p < 0.001), whereas other contrasts were not significant after within-domain multiplicity correction. No domain-level evidence of between-group differences was found for dyspareunia, dysuria, or dyschezia. Younger age was associated with HT initiation (odds ratio (OR) 0.92, 95% CI 0.86-0.98; p = 0.012) and continuation (OR 0.82, 95% CI 0.74-0.91; p < 0.001).

HT initiation was associated with lower one-year dysmenorrhoea/pelvic pain scores, while other domains showed no consistent differences. Younger patients were more likely to initiate or continue hormonal therapy, suggesting non-random allocation. Given the observational design and potential residual confounding, findings are hypothesis-generating and support symptom-profile-guided postoperative management; prospective studies with standardised regimens and domain-specific outcomes are needed.