Single-visit adiposity-kidney phenotypes and mortality among adults without known baseline cardiovascular disease: a survey-weighted analysis of NHANES 1999–2018

Background Albuminuria and a reduced estimated glomerular filtration rate (eGFR) reflect related but distinct forms of kidney injury. Epidemiologic studies nevertheless combine the two into a single chronic kidney disease indicator. We examined whether this simplification obscures mortality differences across eight mutually exclusive obesity–albuminuria–eGFR phenotypes within a cardiovascular-kidney-metabolic (CKM) framework. Methods We analyzed NHANES 1999–2018 examinations linked to the 2019 mortality files. The survey-weighted cohort comprised adults aged ≥ 20 years with complete phenotype, mortality linkage, and baseline CVD information. Eight phenotypes combined obesity (body-mass index ≥ 30 kg/m²), albuminuria (urine albumin-to-creatinine ratio ≥ 30 mg/g), and reduced kidney function (2021 CKD-EPI race-free eGFR < 60 mL/min/1.73 m²). Single-visit abnormalities did not by themselves establish clinical CKD. Survey-weighted Cox models assessed all-cause mortality as the primary outcome and mortality from heart disease or cerebrovascular disease (hereafter, heart/cerebrovascular mortality) as the key secondary outcome. Results The cohort included 42,496 adults and 4,792 all-cause deaths, including 1,293 deaths from heart disease or cerebrovascular disease. Fully adjusted models included 37,533 adults, 4,180 all-cause deaths, and 1,125 deaths from these causes. Phenotypes containing albuminuria generally had higher mortality, and the point estimates were largest when both kidney abnormalities were present. Compared with the reference phenotype, all-cause HRs were 1.28 (95% CI 1.12–1.46) for reduced-eGFR-only and 1.86 (1.63–2.13) for albuminuria-only. The HR was 2.28 (1.89–2.76) for both abnormalities without obesity and 2.79 (2.15–3.61) for all three features. Corresponding heart/cerebrovascular HRs were 1.69, 2.13, 2.21, and 3.57. Albuminuria-only versus reduced-eGFR-only HRs were 1.46 (1.23–1.73) without obesity and 1.71 (1.32–2.21) with obesity; their ratio was 1.17 (0.87–1.58). A five-year internal split showed no clinically important gain in discrimination. Conclusions Among adults without known baseline CVD participating in NHANES, separating albuminuria from reduced eGFR across eight obesity–kidney phenotypes revealed meaningful heterogeneity in mortality associations that a composite kidney indicator can obscure. Albuminuria-only was more strongly associated with all-cause mortality than reduced-eGFR-only in both obesity strata, whereas coexisting abnormalities identified the phenotypes with the highest mortality associations. These findings support retaining UACR and eGFR as separate descriptors in population-level CKM research.