Development and internal validation of a diagnostic prediction model for osteoporosis in elderly people living with HIV receiving antiretroviral therapy: a cross-sectional study

BACKGROUND: Prolonged antiretroviral therapy (ART) in people living with HIV (PLWH) is associated with progressive bone mineral density (BMD) loss and an elevated risk of osteoporosis and fragility fractures. However, validated diagnostic tools for osteoporosis tailored to elderly PLWH remain scarce. This study aimed to develop and internally validate a nomogram-based diagnostic prediction model for osteoporosis in elderly PLWH (postmenopausal women and men aged ≥ 50 years) receiving ART. METHODS: This retrospective cross-sectional study enrolled elderly PLWH who had initiated ART at a single center between September 2010 and July 2022 and had received ART for ≥ 12 months. BMD was assessed by dual-energy X-ray absorptiometry (DXA) during a cross-sectional survey conducted from January to June 2022, and osteoporosis was defined according to the World Health Organization criteria (T-score ≤ - 2.5). Predictor selection was performed using the least absolute shrinkage and selection operator (LASSO) regression. Multivariable logistic regression was subsequently applied to construct a diagnostic prediction nomogram. Model performance was evaluated using the concordance index (C-index), calibration plots, and decision curve analysis (DCA). Internal validation was conducted using bootstrapping with 1,000 resamples. RESULTS: Among 256 eligible patients, 102 (39.8%) were diagnosed with osteoporosis. LASSO regression identified six predictors with nonzero coefficients, and one additional predictor (baseline CD4 count) was incorporated based on clinical evidence, yielding a seven-variable nomogram: Sex, Ratio of Age and BMI, Duration of ART, Baseline CD4 + count, tenofovir disoproxil fumarate (TDF) regimen, β-C-terminal telopeptide of type I collagen (β-CTX), and procollagen type I N-terminal propeptide (PINP). The model achieved a C-index of 0.709 (95% confidence interval [CI]: 0.645-0.773) with good calibration (mean absolute error: 0.014). Internal validation yielded a bias-corrected C-index of 0.695. DCA demonstrated positive net clinical benefit across threshold probabilities ranging from 0.0 to 0.7. CONCLUSION: We developed and internally validated a nomogram incorporating seven clinical and biochemical predictors including Ratio of Age and BMI, Duration of ART, TDF Regimen, β-CTX, PINP, Sex and Baseline CD4 count for the diagnosis of osteoporosis in elderly PLWH receiving ART. The model demonstrated acceptable discrimination, satisfactory calibration and favorable clinical utility. External validation in independent, multi-center cohorts is warranted before clinical implementation.