Lumbar spine-hip T-score discordance in atypical femoral fracture: An exploratory systematic review and meta-analysis.

Objective

Atypical femoral fracture is rare. We evaluated whether lumbar spine-hip T-score discordance, a dual-energy X-ray absorptiometry-derived densitometric phenotype, is enriched among atypical femoral fracture cases compared with controls.

Methods

PubMed, Embase, Cochrane Library, Web of Science, KoreaMed, and RISS were searched from inception to December 31, 2025. Comparative studies of adults with American Society for Bone and Mineral Research-defined atypical femoral fracture that reported lumbar spine and femoral neck T-scores with discordance data were included. Study-level grouped 2 × 2 data were pooled using random-effects models. T-score timing was extracted; it was not uniformly reported. Certainty was assessed using GRADE.

Results

Three Korean retrospective case-control studies including 160 atypical femoral fracture cases and 513 controls were eligible. Discordance definitions were non-equivalent: World Health Organization category mismatch, Delta T-score ≥1.0, and Delta T-score ≥1.5. Discordance prevalence was 103/160 (64.4%) in cases and 170/513 (33.1%) in controls. The pooled random-effects odds ratio was 4.16 (95% CI 2.30-7.52; I2 56.9%). The low-lumbar-spine pattern showed greater case enrichment (odds ratio 7.76; 95% CI 1.70-35.51; I2 86.8%). GRADE certainty was Very Low.

Conclusions

Across three Korean studies with non-equivalent discordance definitions and heterogeneous comparators, lumbar spine-hip T-score discordance was more frequent among atypical femoral fracture cases. This exploratory, hypothesis-generating association should not be interpreted as a diagnostic threshold or individual-level prediction. Because atypical femoral fracture is rare and case-control sampling does not represent population prevalence, odds ratios cannot yield valid positive or negative predictive values.