**Background:** The relationship between serum total cholesterol and bone mineral density (BMD) has been controversial, with prior studies reporting conflicting results—some showing inverse correlations, others positive or no associations. Most previous work focused on specific populations (e.g., postmenopausal women, type 2 diabetes patients) with small sample sizes. This study aimed to evaluate the association between total cholesterol and lumbar spine BMD in a large, nationally representative sample of older adults without cancer, using NHANES data.
**Methods:** Data were drawn from the National Health and Nutrition Examination Survey (NHANES) cycles 1999–2006. After exclusions (including cancer diagnosis), 3,290 participants aged 60–85 years were analyzed. Serum total cholesterol was measured at the Johns Hopkins University Lipoprotein Analytical Laboratory. Lumbar spine BMD was assessed by dual-energy X-ray absorptiometry (DEXA). Covariates included age, sex, race/ethnicity, physical activity, income-to-poverty ratio, blood urea nitrogen (BUN), total protein, serum uric acid, serum calcium, and body mass index (BMI). Weighted multiple linear regression models were used, with three levels of adjustment: crude model (unadjusted), Model I (adjusted for demographics), and Model II (fully adjusted). Smooth curve fitting and generalized additive models were employed to assess nonlinear relationships, with inflection points calculated using recursive algorithms and two-segmented linear regression.
**Key Results:** The study included 3,290 participants (50.6% male, 49.4% female; 52.2% aged 60–69 years, 47.8% aged ≥70 years). Baseline characteristics differed significantly across total cholesterol tertiles except for race/ethnicity. Participants in the highest tertile (≥225 mg/dL) were more likely to be female, aged 60–69, and sedentary, with lower income-to-poverty ratio, BUN, uric acid, and BMI, but higher total protein and serum calcium. Mean lumbar spine BMD was significantly lower in the highest tertile (0.982 g/cm²) compared to the low (1.044 g/cm²) and middle (1.044 g/cm²) tertiles (P < 0.001). In the crude model, total cholesterol was negatively associated with BMD (β = -0.007 per 10 mg/dL; 95% CI: -0.008 to -0.005; P < 0.0001). This association remained significant after full adjustment in Model II (β = -0.002 per 10 mg/dL; 95% CI: -0.004 to -0.001; P = 0.0034). The P for trend across tertiles was 0.020 in the fully adjusted model. Subgroup analyses showed the negative association was strongest in males (β = -0.003; P = 0.0058), participants aged ≥70 years (β = -0.003; P = 0.0115), Non-Hispanic Whites (β = -0.003; P = 0.0007), and those in the highest tertiles of BUN, serum calcium, and BMI. Smooth curve fitting revealed a generally linear negative relationship, but with inflection points in certain subgroups: participants ≥70 years had a turning point at 280 mg/dL (log-likelihood ratio test P = 0.002); those with moderate physical activity at 199 mg/dL (P = 0.005); those in the highest uric acid tertile at 277 mg/dL (P = 0.012); and those in the middle calcium tertile at 275 mg/dL (P = 0.023). These curves were U-shaped beyond the inflection points.
**Clinical Implications:** This large, multi-ethnic, population-based study provides evidence that higher serum total cholesterol is independently and negatively associated with lumbar spine BMD in older adults without cancer. The findings suggest that hypercholesterolemia may be a modifiable risk factor for reduced bone density, potentially informing osteoporosis screening and prevention strategies in elderly populations. The identification of inflection points (e.g., 280 mg/dL in those ≥70 years) may help identify thresholds for clinical intervention. However, due to the cross-sectional design, causality cannot be established, and residual confounding from unmeasured variables is possible. The authors recommend further large-scale prospective studies to clarify the role of cholesterol in bone metabolism.