Comparison of the predictive performance of Chang and Baumgaertner reduction quality criteria for cutout complication in geriatric unstable intertrochanteric femur fractures

The aim of this study was to compare the predictive performance of Baumgaertner Reduction Quality Criteria (BRQC) and Chang Reduction Quality Criteria (CRQC) for lag screw cutout in geriatric unstable intertrochanteric femur fractures treated with proximal femoral nailing (PFN).

A retrospective study was conducted on 907 patients aged ≥ 65 years who underwent surgery for AO/OTA 31.A2 fractures between January 2018 and January 2026. Among these, 108 patients treated with the same implant (APFN) and meeting the inclusion criteria were included. To reduce the effect of known implant-position–related risk factors for cutout and thereby better isolate the predictive value of reduction quality classifications, patients with postoperative TAD values ≤ 30 mm and ≥ 20 mm and lag screws positioned center-center zone according to Cleveland were included in the study. Reduction quality was evaluated according to BRQC and CRQC using anteroposterior and lateral radiographs. Inter-system agreement was assessed using the McNemar–Bowker test and Cohen’s kappa. Cutout predictive performances were compared using ROC analysis and the DeLong test. Inter- and intra-observer reliability was evaluated using weighted Cohen’s kappa coefficients. A p-value < 0.05 was considered statistically significant.

The mean age of the patients was 79 ± 7.5 years, and 69.4% were female. 93 patients (86.1%) had bone union (non-cutout group), and 15 patients (13.9%) had cutout (cutout group). The mean bone union time was 12.54 ± 2.65 weeks. The mean time to cutout was 8.29 ± 5.52 weeks. No demographic differences were observed between groups. Both BRQC and CRQC significantly predicted cutout. However, cross-distribution and agreement analyses showed that reduction types associated with cutout were categorized into different risk groups, with CRQC demonstrating greater selectivity. ROC analysis showed AUC values of 0.79 (95% CI: 0.69–0.90) for CRQC and 0.68 (95% CI: 0.56–0.80) for BRQC, with a significant difference (ΔAUC = 0.10, p < .05). The reliability results demonstrated that CRQC exhibited consistently higher inter- and intra-observer agreement compared to BRQC.

CRQC more accurately and selectively identifies reduction types associated with cutout risk compared to BRQC. The use of CRQC can contribute to more effective intraoperative decision-making and postoperative risk stratification, especially in geriatric unstable intertrochanteric femur fractures.