Development and internal validation of a clinical score to predict persistent anterior knee pain after anterior cruciate ligament reconstruction.

Abstract

Persistent anterior knee pain remains one of the most frequent causes of postoperative dissatisfaction after anterior cruciate ligament reconstruction (ACLR). Despite extensive literature on potential etiologic factors, no validated clinical tool currently exists to predict which patients are at greatest risk. A retrospective cohort of 450 consecutive primary ACLR procedures was analyzed. Patients with revision surgery, multiligament injury, or previous osteotomy were excluded. The primary outcome was persistent anterior knee pain at 12 months, defined as a visual analogue scale (VAS) score >3/10 or Kujala score <80. Candidate predictors included demographic factors, graft type, posterior tibial slope, notchplasty, meniscal injury, injury chronicity, and early quadriceps activation. Multivariable logistic regression with backward elimination was performed. Model discrimination and calibration were assessed using the area under the receiver operating characteristic curve (AUC-ROC) and Hosmer-Lemeshow (HL) test, with internal validation by 1000 bootstrap replications. Five independent predictors were identified: use of bone-patellar-tendon-bone graft (odds ratio (OR) 2.8, P < 0.01), female sex (OR 1.9, P = 0.03), posterior tibial slope >12° (OR 2.2, P = 0.01), absence of notchplasty (OR 1.7, P = 0.04), and delayed quadriceps activation >2 weeks (OR 2.4, P < 0.01). The model showed excellent discrimination (AUC = 0.82, 95% confidence interval 0.78-0.87) and good calibration (HL P = 0.46). The simplified 0- to 10-point Anterior Knee Pain Score (AKPS-ACL) identified patients at high risk with a cutoff ≥6 (sensitivity 80%, specificity 75%, negative predictive value 91%). The AKPS-ACL represents an internally validated clinical score for predicting persistent anterior knee pain after ACLR, integrating anatomical, technical, and rehabilitation factors into a practical risk-stratification tool.