Serum 25(OH)D concentration and risk of all-cause and cause-specific mortality in patients with chronic kidney disease: systematic review and dose-response meta-analysis of observational cohort studies

To summarize the evidence on the associations between 25-hydroxyvitamin D (25(OH)D) and risk of mortality in patients with chronic kidney disease (CKD). A comprehensive search of PubMed, Scopus, Web of Science, and Google Scholar for relevant papers published up to November 2025 that assessed the association between serum vitamin D levels and the risk of all-cause, cardiovascular disease (CVD), non-CVD, and cancer mortality was conducted. The lowest-versus-highest analysis and the linear and non-linear dose-response analyses were performed using a random-effects model. Overall, 36 publications (35 studies) with a total sample size of 129,135 participants, aged between 18 and 90 years, were included in the current meta-analysis. During the follow-up periods ranging between 3 months and 18 years, 117,403 cases of all-cause mortality, 2,568 cases of CVD mortality, 886 cases of non-CVD mortality, and 289 cases of cancer mortality were identified. The summary relative risk (RR) and 95% confidence intervals (CIs) comparing lowest versus highest levels of 25(OH)D was 1.61 (95% CI: 1.41–1.84, I2 = 89%, n = 28 studies, very low certainty) for all-cause mortality, 1.68 (95% CI: 1.41-2.00, I2 = 20%, n = 9, very low certainty) for CVD mortality, 1.33 (95% CI: 0.94–1.89, I2 = 26%, n = 3, very low certainty) for non-CVD mortality, and 1.51 (95% CI: 1.09–2.09, I2 = 26%, n = 1, very low certainty) for cancer mortality. There was evidence of non-linearity in the analysis of all-cause and CVD mortality, with a greater reduction in risk from serum levels of 12.5 nmol/L up to 60 nmol/L compared to higher levels, but with slight further reductions in risk with serum levels up to 110 nmol/L. This meta-analysis provides further evidence that lower levels of 25(OH)D are associated with a higher risk of all-cause, CVD, and cancer mortality in CKD patients. However, all outcomes were graded as very low certainty, and the observed associations may be influenced by confounding, small-study effects, and wide prediction intervals that include the null. Observed low-risk range around ~ 60 nmol/L in the non-linear dose-response analyses, should be interpreted cautiously and requires validation in randomized controlled trials. The protocol for this review was registered in PROSPERO (2025 CRD42025105350).