Hypocholesterolemia is a consequence, not a cause, of kidney cancer: A bidirectional Mendelian randomization study with NHANES 2001–2018
Menée à l'aide d'une méthode de randomisation mendélienne et de données portant sur 45 366 témoins et 102 patients atteints d'un cancer du rein, cette étude analyse l'association entre les profils lipidiques et le risque de développer la maladie
Background: Kidney cancer patients often exhibit reduced circulating cholesterol, yet whether dyslipidemia is a cause or consequence of the disease remains unclear due to inherent limitations of observational studies. We integrated population-based data from NHANES with bidirectional Mendelian randomization (MR) to investigate the causal direction between lipid profiles and kidney cancer.
Methods: Observational data were derived from National Health and Nutrition Examination Survey (NHANES) 2001–2018, comprising 102 self-reported kidney cancer cases and 45,366 cancer-free controls. All analyses accounted for the complex sampling design using survey weights, stratification, and primary sampling units. Bidirectional two-sample MR was conducted using genetic instruments for four lipid traits (low-density lipoprotein [LDL], high-density lipoprotein [HDL], triglycerides, total cholesterol [TC]) from IEU OpenGWAS and clear cell renal cell carcinoma (ccRCC) data from the largest available genome-wide association study (Purdue et al., n = 752,817), with no sample overlap. The primary analysis used inverse-variance weighted (IVW) estimation, complemented by sensitivity analyses including weighted median, MR-Egger, MR-PRESSO, and MR-RAPS.
Results: Weighted linear regression (adjusted for age, sex, BMI, statin use, smoking, hypertension, diabetes) showed kidney cancer associated with lower LDL (β = –21.08 mg/dL, Bonferroni-corrected P = 0.0016); TC, HDL, and TG showed no significant associations. Forward MR found no causal effect of any lipid trait on ccRCC risk. Reverse MR demonstrated genetic predisposition to ccRCC associated with lower LDL (β = –0.020, P = 0.011), HDL (β = –0.016, P = 0.022), and TC (β = –0.009, P = 0.037), with robust results across sensitivity analyses.
Conclusions: These findings suggest genetic liability to ccRCC is associated with lower circulating cholesterol, not dyslipidemia predisposing to kidney cancer, providing insight into the directionality question and implications for etiologic understanding and clinical surveillance without overstating causality.
Cancer Epidemiology , article en libre accès, 2026