Associations of non-traditional lipid-inflammatory parameters with all-cause mortality in middle-aged and older adults: Insights from two nationally representative cohorts

Background Non-traditional lipid-inflammatory parameters are emerging biomarkers integrating lipid metabolism and inflammation. However, their association with all-cause mortality in middle-aged and older adults has not been systematically evaluated. Methods This study enrolled 9,943 participants from the National Health and Nutrition Examination Survey (NHANES) and 9,409 participants from the China Health and Retirement Longitudinal Study (CHARLS). Survey weights were applied to account for the complex sampling design. Kaplan–Meier analyses were performed to estimate cumulative survival probabilities. Cox proportional hazards models were used to examine the associations. Restricted cubic splines were fitted to explore potential dose–response relationships. Time-dependent ROC analyses were conducted to evaluate predictive performance at follow-up time points. DeLong tests were used to assess the statistical significance of differences in the areas under the curve. Calibration curves were constructed to assess the model calibration. Subgroup and sensitivity analyses were conducted to explore heterogeneity and assess the robustness of the results. Results Over a mean follow-up of 80.3 months in NHANES and 100.2 months in CHARLS, 2,377 and 1,246 deaths were recorded, respectively. Cumulative survival probabilities decreased across quartiles for all parameters except LCI-CRP (Log-rank P =0.960) in NHANES. After adjusting for confounders, all parameters were positively associated with all-cause mortality in continuous analyses. In quartile analyses, the highest quartile carried a higher risk than the lowest, except for AIP-CRP (HR=1.16, 95% CI:0.88–1.52, P =0.295) and LCI-CRP (HR=1.28, 95% CI:0.99–1.64, P =0.056) in CHARLS. AIP-CRP exhibited a nonlinear dose–response relationship in both cohorts. The predictive performance of parameters improved with longer follow-up time in NHANES, whereas it fluctuated in CHARLS. DeLong tests showed that models incorporating parameters had better discriminative ability than the base model. Calibration curves indicated good calibration of the parameter-incorporating models. Hypertension significantly modified the associations of certain parameters in CHARLS. The results remained robust across all sensitivity analyses. Conclusions Non-traditional lipid-inflammatory parameters demonstrate significant independent associations with an elevated risk of all-cause mortality among middle-aged and older adults. Based on routine lipid and inflammatory tests, these parameters are simple, low-cost tools for early risk stratification.