Predictors for incident albuminuria are not well-known in population based cohorts. The purpose of this study was to identify predictors of incident albuminuria in an unselected, middle-aged population.
Observational cohort study
Setting and participants
Framingham Offspring Study participants who attended both the sixth (baseline; 1995–1998) and eighth (2005–2008) examination cycles.
Standard clinical predictors were used. Predictors of incident albuminuria were identified by stepwise logistic regression analysis with age and sex forced into the model.
Outcomes and 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.
1916 participants were available for analysis (mean age 56 years, 54% women). Albuminuria developed in 10.0% of participants (n=192) over 9.5 years. Age (odds ratio [OR], 2.09; p-value<0.001), baseline diabetes (OR, 1.93; p-value= 0.01), smoking (OR, 2.09; p-value <0.001) and baseline log UACR (OR per standard deviation increase in log UACR, 1.56; p-value <0.001) were associated with incident albuminuria in a stepwise model. An inverse relationship with female sex (OR, 0.53; p <0.001) and HDL cholesterol (OR, 0.80; p-value=0.007) was also observed. Results were similar when participants with baseline chronic kidney disease (n=102), defined as eGFR < 60 mL/min/1.73 m2, were excluded from the model. Age, male sex, low HDL-cholesterol, 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).
Causality may not be inferred due to the observational nature of the study. One-third of participants did not return for follow-up, potentially attenuating the observed risks for albuminuria.
The known cardiovascular risk factors of increasing age, male sex, diabetes, smoking, low HDL cholesterol, and albuminuria within the “normal range” are correlates of incident albuminuria in the general population.