Remnant cholesterol and hypertension: how outcome definition and baseline blood pressure adjustment shape the pooled association — a systematic review and meta-analysis

Objective To systematically evaluate the association between remnant cholesterol (RC) and hypertension and examine how outcome definition, exposure scale, and adjustment for baseline blood pressure affect pooled estimates. Methods This review was registered in INPLASY (INPLASY202670098). PubMed, Embase, Web of Science Core Collection, and the Cochrane Library were searched through October 12, 2026. Reference lists of included studies and relevant systematic reviews and available supplementary materials were also examined. Observational studies evaluating RC and hypertension were included, with evidence organized separately for incident, prevalent, and maternal outcomes. Compatible HRs or ORs were pooled using inverse-variance random-effects models, with each potentially overlapping source population contributing one estimate per analysis. Uncertainty was assessed using modified Hartung–Knapp intervals, 95% prediction intervals when at least three independent sources were available, source-replacement analyses, and exploratory subgroup analyses. An exploratory two-stage linear dose–response analysis was also performed. Results Twenty-three study reports were included. For the highest versus lowest baseline RC categories, the pooled HR for incident hypertension from five independent source populations was 1.40 (95% CI: 1.05–1.85; I 2 = 97.5%), and the pooled OR for prevalent hypertension from three source populations was 1.60 (95% CI: 1.17–2.19; I 2 = 95.0%). The corresponding modified Hartung–Knapp intervals were 0.93–2.09 and 0.80–3.19. Pooled HRs were 1.92 and 1.17 in models without and with baseline blood pressure adjustment, respectively, although populations and other covariates also differed between groups. Four independent source populations contributed to the two-stage linear dose–response meta-analysis. The pooled HR for incident hypertension per 1 mmol/L higher RC was 1.32 (95% CI: 1.18–1.47; I 2 = 40.9%), with a modified Hartung–Knapp interval of 1.08–1.61 and a 95% prediction interval of 0.88–1.98. After replacing the CHARLS report, the pooled HR was 1.27 (95% CI: 1.14–1.40). A pooled nonlinear association was not evaluated. The pooled OR for preeclampsia was 1.83 (95% CI: 1.47–2.29), based on only two sources, with a modified interval of 0.44–7.65. Conclusions Higher RC was positively associated with incident and prevalent hypertension, and the exploratory dose–response analysis suggested a positive linear trend. High heterogeneity, few independent sources, differences in adjustment strategies, and approximations in dose–response reconstruction limit assessment of association magnitude and applicability across populations. Causality and clinical utility require further investigation. Systematic Review Registration https://doi.org/10.37766/inplasy2026.7.0098 , identifier INPLASY202670098.