<- Back to digest

International Orthopaedics - 2026-10-17 - Journal Article

Development of a clinical risk score for common peroneal nerve palsy after total knee arthroplasty using a Japanese nationwide database.

Kurihara S, Furusato S, Hatayama K, Fushimi K, Shimizu S

retrospective cohortLOE IIIn = 328,314N/A (perioperative complication; index hospitalization only)

Topics

arthroplasty
PMID: 42891828DOI: 10.1007/s00264-026-07057-xView on PubMed ->

Key Takeaway

A 4-variable risk score (valgus deformity 3pts, contracture 2pts, BMI<18.5 1pt, spinal disorder 1pt) stratifies CPNP risk after TKA from 0.09% to 0.68%, but achieves only AUC 0.55 in 328,314 patients.

Summary Depth

Choose how much analysis to show on this article page.

Summary

This study used a Japanese nationwide administrative database (2014–2024) to develop a preoperative risk score for CPNP following primary TKA, with CPNP identified by ICD-10 codes occurring in 340 of 328,314 patients (0.10%). LASSO regression selected four predictors—valgus deformity, knee contracture, low BMI, and spinal disorder—yielding a 0–7 point score. Observed CPNP risk increased from 0.09% at score 0 to 0.68% at score ≥3, with good calibration (slope 1.03) but poor discrimination (AUC 0.55).

Key Limitation

ICD-10-based CPNP identification in an administrative database precludes differentiation of transient neurapraxia from permanent palsy and excludes critical intraoperative variables (correction angle, tourniquet duration, ligament release extent) that likely drive nerve injury.

Original Abstract

PURPOSE

This study aimed to develop a clinical risk score to stratify the risk of post-operative common peroneal nerve palsy (CPNP) following total knee arthroplasty (TKA), based on pre-operative factors.

METHODS

We conducted a retrospective cohort study of adult patients undergoing primary TKA using a Japanese nationwide database (from 2014 to 2024). The outcome was post-operative CPNP identified by International Classification of Diseases, Tenth Revision codes. Candidate predictors included 24 clinical and hospital-level variables. Least Absolute Shrinkage and Selection Operator (LASSO) regularization was applied for predictor selection. Multivariable logistic regression was used as the primary model, with penalized and Firth logistic models as sensitivity analyses. Model performance and internal validation were assessed using the area under the curve (AUC), calibration, and bootstrap resampling.

RESULTS

A total of 328,314 patients were eligible and CPNP occurred in 340 cases (0.10%). LASSO identified four predictors, and a 0-7 point risk score was developed: knee valgus deformity (3 points), knee contracture (2 points), low BMI < 18.5 (1 point), and spinal disorder (1 point). The observed risk rose from 0.09% (score 0) to 0.56% (score 3), and to 0.68% for scores ≥ 3, while categories of score ≥ 4 were sparsely populated. The risk score showed limited discrimination (AUC = 0.55) but favourable calibration (intercept = - 0.001, slope = 1.03).

CONCLUSION

We developed a clinical risk score for CPNP after TKA based on knee valgus deformity, knee contracture, low BMI, and spinal disorder. This score may support pre-operative risk assessment and informed consent in high-risk patients.