Predictors of incident albuminuria in the Framingham Offspring cohort

Abstract

Background: Predictors for incident albuminuria are not well known in population-based cohorts. The purpose of this study is to identify predictors of incident albuminuria in an unselected middle-aged population.

Study design: Observational cohort study.

Setting & participants: Framingham Offspring Study participants who attended both the sixth (baseline; 1995-1998) and eighth (2005-2008) examination cycles.

Predictors: Standard clinical predictors were used. Predictors of incident albuminuria were identified using stepwise logistic regression analysis with age and sex forced into the model.

Outcomes & measurements: Albuminuria was defined as urine albumin-creatinine ratio (UACR) ≥ 17 mg/g (men) or ≥ 25 mg/g (women). Individuals with albuminuria at baseline were excluded.

Results: 1,916 participants were available for analysis (mean age, 56 years; 54% women). Albuminuria developed in 10.0% of participants (n = 192) during 9.5 years. Age (OR, 2.09; P < 0.001), baseline diabetes (OR, 1.93; P = 0.01), smoking (OR, 2.09; P < 0.001), and baseline log UACR (OR per 1-SD increase in log UACR, 1.56; P < 0.001) were associated with incident albuminuria in a stepwise model. An inverse relationship with female sex (OR, 0.53; P < 0.001) and high-density lipoprotein (HDL) cholesterol level (OR, 0.80; P = 0.007) also was observed. Results were similar when participants with baseline chronic kidney disease (n = 102), defined as estimated glomerular filtration rate <60 mL/min/1.73 m(2), were excluded from the model. Age, male sex, low HDL cholesterol level, smoking, and log UACR continued to be associated with incident albuminuria when baseline diabetes (n = 107) was excluded. Age, male sex, and log UACR correlated with incident albuminuria after participants with baseline hypertension were excluded (n = 651).

Limitations: Causality may not be inferred because of the observational nature of the study. One-third of participants did not return for follow-up, potentially attenuating the observed risks of albuminuria.

Conclusions: The known cardiovascular risk factors of increasing age, male sex, diabetes, smoking, low HDL cholesterol level, and albuminuria within the reference range are correlates of incident albuminuria in the general population.