Geographic variation in CKD prevalence and ESRD incidence in the United States: results from the reasons for geographic and racial differences in stroke (REGARDS) study

Abstract

Background: It is not known whether geographic differences in the prevalence of chronic kidney disease exist and are associated with end-stage renal disease (ESRD) incidence rates across the United States.

Study design: Cross-sectional and ecologic.

Setting & participants: White (n = 16,410) and black (n = 11,109) participants from across the continental United States in the population-based Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study.

Predictor: Geographic region, defined by the 18 networks of the US ESRD Network Program.

Outcomes & measurements: Albuminuria, defined as albumin-creatinine ratio ≥30 mg/g, and decreased estimated glomerular filtration rate (eGFR), defined as <60 mL/min/1.73 m(2), were measured in the REGARDS Study. ESRD incidence rates were obtained from the US Renal Data System.

Results: For whites, the network-specific prevalence of albuminuria ranged from 8.4% (95% CI, 3.3%-13.5%) in Network 15 to 14.8% (95% CI, 8.0%-21.6%) in Network 3, and decreased eGFR ranged from 4.3% (95% CI, 2.0%-6.6%) in Network 4 to 16.7% (95% CI, 12.7%-20.7%) in Network 7. For blacks, the prevalence of albuminuria ranged from 12.1% (95% CI, 8.7%-15.5%) in Network 5 to 26.5% (95% CI, 16.7%-36.3%) in Network 4, and decreased eGFR ranged from 6.7% (95% CI, 5.0%-8.4%) in Network 17/18 to 13.4% (95% CI, 7.8%-19.1%) in Network 12. Spearman correlation coefficients for the prevalence of albuminuria and decreased eGFR with network-specific ESRD incidence rates were 0.49 and 0.24, respectively, for whites and 0.29 and 0.25, respectively, for blacks.

Limitations: There were few cases of albuminuria and decreased eGFR in some geographic regions.

Conclusions: In the United States, substantial geographic variations in the prevalence of albuminuria and decreased eGFR exist, but were correlated only modestly with ESRD incidence, suggesting the chronic kidney disease burden may not explain the geographic variation in ESRD incidence.

Figure 1

End-stage renal disease incidence rates per million population in 2005 by end-stage renal disease network among whites (left) and blacks (right) ≥45 years. The table is limited to the continental US as the REGARDS study did not recruit participants from Alaska, Hawaii, or US protectorates.

Figure 2

Prevalence of albuminuria^† (left), reduced estimated glomerular filtration rate (eGFR; center)^‡, and both albuminuria/reduced eGFR (right) by end-stage renal disease network of residence among white (top panel) and blacks (bottom panel) REGARDS participants. N/A: Estimates not shown due to insufficient sample size (n<50) for blacks in these regions. Prevalence is weighted to represent the US population. ^† Albuminuria defined as albumin-to-creatinine ratio (ACR) ≥30 mg/g ^‡ Reduced eGFR defined as levels <60 mL/min/1.73m²

Figure 3

Correlation of end-stage renal disease network-specific prevalence of albuminuria^† (left), reduced estimated glomerular filtration rate^‡ (center), and both albuminuria/reduced eGFR (right) among REGARDS participants with USRDS-derived end-stage renal disease incidence rates for whites (top panel) and blacks (bottom panel). r= Spearman correlation coefficient ^†Albuminuria defined as an albumin to creatinine ratio ≥30 mg/g ^‡Reduced eGFR defined as levels <60 ml/min/1.73m² Networks 15 and 16 excluded for blacks due to small sample size (n<50) for blacks in these regions