Trends in initiation of dialysis in an urban dialysis clinic in the United States: a long way from dialysis outcomes quality initiative guidelines.

The guidelines issued for peritoneal dialysis by the National Kidney Foundation Dialysis Outcomes Quality Initiative (NKF-DOQI) suggest initiation of dialysis at a glomerular filtration rate (GFR) of 10.5 mL/min/1.73 m2. We undertook this study to determine trends at a center in Philadelphia. Using the MDRD (Modification of Diet in Renal Disease) formula, we estimated the GFR (mL/min/1.73 m2) at the time of first dialysis for patients starting on dialysis at the center between 1994 and 1996. Data on 72 of the 86 new patients were available. Of the 72, 69 patients (96%) were black and 29 (40%) were diabetic. The estimated GFR at the time of dialysis initiation was 6.3 +/- 0.3 mL/min/1.73 m2. Only 3 patients (4%) had a GFR > 10.5; 25 patients (35%) had GFR < 5.0. Patients with a lower GFR had more severe acidosis (HCO3 15.8 +/- 0.5 mmol/L vs 19.5 +/- 0.9 mmol/L, p = 0.0023) and greater impairment in divalent ion metabolism (PO4: 7.6 +/- 0.2 mg/dL vs 5.6 +/- 0.3 mg/dL, p < 0.0001). Diabetic patients were initiated on dialysis at a higher level of GFR than non diabetic patients (7.2 +/- 0.3 mL/min/1.73 m2 vs 5.7 +/- 0.4 mL/min/1.73 m2, p = 0.0087). Even though diabetic patients had higher GFR and lower serum creatinine (8.6 +/- 0.5 mg/dL vs 11.8 +/- 0.9 mg/dL, p = 0.0021) than non diabetic patients, blood urea nitrogen (BUN) was similar in the two groups (89 +/- 3.7 mg/dL vs 96.8 +/- 5.8 mg/dL, p = 0.73). This difference may indicate that nondiabetic patients had a greater decline in protein intake. Finally, a trend towards initiation of dialysis at higher levels of renal function was seen with time (GFRs 5.5 +/- 0.4 mL/min/1.73 m2, 6.4 +/- 0.4 mL/min/1.73 m2, and 6.9 +/- 0.6 mL/min/1.73 m2 for 1994, 1995, and 1996 respectively; p = 0.058, 1994 vs 1996). The trend was stronger for diabetic patients (GFR 6.3 +/- 0.4 mL/min/1.73 m2 and 8.1 +/- 0.7 mL/min/1.73 m2 for 1994 and 1996 respectively; p = 0.06) than for non diabetic patients (GFR 5.1 +/- 0.5 mL/min/1.73 m2 and 6.2 +/- 0.7 mL/min/1.73 m2 for 1994 and 1996 respectively; p = 0.22). We conclude that initiation of dialysis is delayed in urban centers, particularly in non diabetic patients, and that the current practice is far below the DOQI recommendations.