Abstract / Summary
Abstract Background. Early postoperative pain may help identify patients whose functional recovery after unilateral biportal endoscopic lumbar interbody fusion (ULIF) is falling behind, but the most informative pain measure is uncertain. We compared four early pain measures and then examined the association, candidate threshold, and incremental information of the best-performing measure for failure to achieve the 1-year Oswestry Disability Index (ODI) minimum clinically important difference (MCID). Methods. We retrospectively screened 282 surgery-system records of possible ULIF procedures performed from January 2023 through December 2024. After procedure confirmation and hierarchical eligibility review, 234 patients were clinically eligible; 189 had complete required pain and ODI data and formed the 3-month landmark cohort. Back- and leg-pain numerical rating scale (NRS) scores at 2 weeks and 3 months were compared using receiver operating characteristic analysis. The measure with the highest apparent area under the curve (AUC) was then examined with Firth logistic regression and restricted cubic splines. Fixed post-selection models underwent 1,000-sample conditional bootstrap validation, while a separate 10,000-sample bootstrap repeated marker and threshold selection. Results. Of the 189 patients, 77 (40.7%) did not achieve the ODI MCID. Three-month leg pain had the highest apparent AUC (0.891; 95% CI, 0.850–0.933) and was selected in 9,999 of 10,000 marker-selection bootstraps. The probability of MCID non-achievement rose with higher leg-pain scores and followed an apparently nonlinear pattern. The selection-aware corrected AUC for the continuous score was 0.891, and the corrected balanced accuracy of the NRS ≥ 3 rule was 0.791. In the fixed post-selection models, concurrent 3-month ODI had a corrected AUC of 0.973. Adding the leg-pain spline increased corrected AUC by 0.0125 but moved the calibration slope farther from 1. Conclusions. Three-month leg pain separated ODI-MCID achievers from non-achievers better than the other three early pain measures, although it added little beyond concurrent 3-month ODI. An NRS score of 3 or higher may be useful as a signal for closer reassessment, but not as a stand-alone prediction or treatment rule. External transportability remains unknown.