A predictive nomogram based on sub-anesthetic esketamine and clinical variables for estimating the risk of postoperative depression following laparoendoscopic single-site hysterectomy

Objective: To construct a nomogram combining sub-anesthetic esketamine administration with perioperative clinical characteristics for predicting postoperative depression in patients undergoing laparoendoscopic single-site (LESS) hysterectomy.

Methods: Clinical records from 326 patients who underwent LESS hysterectomy were retrospectively reviewed. Participants were randomly assigned to training ( n =228) and internal validation ( n = 98) cohorts at a 7:3 ratio. Using the 17-item Hamilton Depression Rating Scale (HAMD-17) on postoperative day 7, patients in the training cohort were classified into depression (HAMD-17 ≥ 7) and non-depression (HAMD-17 < 7) groups. A nomogram was established, and its discrimination, calibration, and clinical utility were evaluated using receiver operating characteristic (ROC) analysis, calibration curves, and decision curve analysis (DCA).

Results: Postoperative depression occurred in 23.62% (77/326) of patients: 22.81% (52/228) and 25.51% (25/98) in the training and internal validation cohorts, respectively. Multivariable logistic regression identified the 12-hour postoperative visual analogue scale (VAS) score (OR = 3.165, 95% CI: 1.223–8.192), increasing age (OR = 1.126, 95% CI: 1.053–1.205), and prolonged operative duration (OR = 1.046, 95% CI: 1.014–1.078) as independent risk factors (all P < 0.05), whereas higher preoperative serum albumin (OR = 0.785, 95% CI: 0.684–0.900), single-dose esketamine (OR = 0.290, 95% CI: 0.118–0.714), and continuous-dose esketamine (OR = 0.306, 95% CI: 0.116–0.803) were protective (all P < 0.05). The nomogram achieved areas under the ROC curve of 0.884 (95% CI: 0.829–0.942) and 0.772 (95% CI: 0.637–0.906) in the training and validation cohorts. Calibration was good in the training cohort (slope = 1.000, intercept = 0.000) but was poor in the internal validation cohort (slope = 0.467, 95% CI: 0.215–0.719; intercept = −0.925, 95% CI: −1.386 to −0.464; Brier score = 0.167), indicating systematic overestimation of predicted risk and overfitting. DCA indicated favorable clinical utility. SHapley Additive exPlanations ranked predictors as operative time, age, preoperative albumin, esketamine administration, and 12-hour VAS score.

Conclusion: The nomogram demonstrated acceptable discrimination but poor calibration in internal validation, systematically overestimating postoperative depression risk; it should be regarded as an exploratory tool for postoperative risk stratification and requires external validation and recalibration before any clinical application. Chinese Clinical Trial Registry Number ChiCTR2500103453