Nonlinear Association of Estimated Pulse Wave Velocity With Bone Health Outcomes: Evidence From the NHANES Database

The prevalence and incidence of impaired bone health have continued to rise globally, imposing a substantial public health burden. Cardiovascular-related factors, including vascular aging, may be involved in the regulation of bone health. However, the association between the estimated pulse wave velocity (ePWV) and bone health outcomes remains unclear. This study aimed to investigate the relationship between ePWV and bone health using data from the National Health and Nutrition Examination Survey (NHANES).

Data were derived from five NHANES survey cycles spanning 2005-2010, 2013-2014, and 2017-2018, comprising 13,068 eligible participants. Osteoporosis and femoral neck bone mineral density (BMD) were selected as bone health outcomes. ePWV was calculated using a validated formula based on age and mean arterial pressure. Descriptive statistics were used to summarize baseline characteristics. Group differences were assessed using nonparametric rank-sum tests and Rao-Scott chi-square tests. Multivariable weighted logistic regression models were applied to evaluate the association between ePWV and osteoporosis, while multivariable weighted linear regression models were used to examine the relationship between ePWV and femoral neck BMD. Restricted cubic spline (RCS) modeling was applied to assess potential nonlinear relationships, and subgroup analyses were conducted to assess effect modification across different population characteristics. Weighted receiver operating characteristic (ROC) analysis was performed to evaluate the discriminative ability of ePWV for osteoporosis, and percentile-based trimming sensitivity analysis was conducted to assess the robustness of the findings.

Elevated ePWV levels were significantly linked to a higher likelihood of osteoporosis and reduced femoral neck BMD. In fully adjusted models (Model 3), each one-unit increase in ePWV was associated with a 22% higher odds of prevalent osteoporosis (odds ratio (OR) = 1.22, 95% confidence interval CI: 1.14, 1.32, p < 0.001) and a decrease of 0.030 g/cm² in femoral neck BMD (regression coefficients (β) = -0.030, 95% CI: -0.035, -0.024, p < 0.001). RCS and threshold effect analyses revealed significant nonlinear associations with identifiable inflection points between ePWV and both bone health outcomes. Subgroup analyses demonstrated generally consistent associations across most population subgroups, although effect heterogeneity was observed according to sex, age, body mass index (BMI), diabetes status, and alcohol consumption. Weighted ROC analysis showed moderate discriminatory performance of ePWV for osteoporosis, with an AUC of 0.774 and an optimal cutoff value of 8.865 m/s. Percentile-based trimming sensitivity analysis yielded results consistent with the primary findings.

ePWV is nonlinearly associated with osteoporosis and femoral neck BMD, and these associations appear to be robust across most population subgroups.