Albuminuria and racial disparities in the risk for ESRD

Abstract

The causes of the increased risk for ESRD among African Americans are not completely understood. Here, we examined whether higher levels of urinary albumin excretion among African Americans contributes to this disparity. We analyzed data from 27,911 participants in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study who had urinary albumin-to-creatinine ratio (ACR) and estimated GFR (eGFR) measured at baseline. We identified incident cases of ESRD through linkage with the United States Renal Data System. At baseline, African Americans were less likely to have an eGFR <60 ml/min per 1.73 m(2) but more likely to have an ACR ≥ 30 mg/g. The incidence rates of ESRD among African Americans and whites were 204 and 58.6 cases per 100,000 person-years, respectively. After adjustment for age and gender, African Americans had a fourfold greater risk for developing ESRD (HR 4.0; 95% CI 2.8 to 5.9) compared with whites. Additional adjustment for either eGFR or ACR reduced the risk associated with African-American race to 2.3-fold (95% CI 1.5 to 3.3) or 1.8-fold (95% CI 1.2 to 2.7), respectively. Adjustment for both ACR and eGFR reduced the race-associated risk to 1.6-fold (95% CI 1.1 to 2.4). Finally, in a model that further adjusted for both eGFR and ACR, hypertension, diabetes, family income, and educational status, African-American race associated with a nonsignificant 1.4-fold (95% CI 0.9 to 2.3) higher risk for ESRD. In conclusion, the increased prevalence of albuminuria may be an important contributor to the higher risk for ESRD experienced by African Americans.

Figure 1.

Age- and gender-related adjusted cumulative hazard for incident ESRD among African Americans and whites. The cumulative hazard during the follow-up is shown for 65-year-old men.

Figure 2.

Hazard ratio of incident ESRD for African Americans vs. whites with adjustments for various risk factors progressively added to the model. (♦), the hazard ratio; horizontal line, the 95% CI. Covariates in model: model 1: Race; model 2: model 1 and demographics (age, gender, and waist circumference); model 3: model 2 and log of ACR; model 4: model 2 and eGFR; model 5: model 2 and both eGFR and log of ACR; model 6: model 5 and diabetes, hypertension, and renin-angiotensin system blocking agents; model 7: model 6 and household income and educational attainment.