Rate of change of end-stage renal disease treatment incidence 1978-1987--has there been selection?

End-stage renal disease (ESRD) treatment rates in the United States have increased steadily since 1973. Decreasing selection against elderly patients with a poor prognostic primary cause of ESRD (i.e., diabetic nephropathy) may partly account for this increase in rates. To test this hypothesis, we calculated log ESRD treatment incidence (ESRDI) rates by four major primary causes of ESRD (diabetic nephropathy (DN), hypertensive nephropathy (HN), glomerulonephritis (GN), and cystic kidney disease (PC); two age groups (old (O), greater than 65 and young (Y), 15 to 44 yr of age) for black and white, male and female, new ESRD patients from 1978 to 1987. As predicted, summary log ESRDI slopes (produced by analysis of covariance) occurred in the following decreasing order, ODN (0.19), OGN = OHN = YDN (0.134). YHN = YPC = YGN (in white patients) = slope not significantly different from 0. Log ESRDI slopes for young black males and females with GN increased significantly between 1978 and 1987, possibly as a result of an increased incidence of GN. In conclusion, decreasing selection may be a factor in the continuing increase in the U.S. ESRD population.