Late referral defined by renal function: association with morbidity and mortality

Abstract

Background: Many patients with chronic renal failure are referred very late to nephrology units. Late referral (LR) is reported to be associated with increased morbidity and mortality.

Methods: We used glomerular filtration rate (GFR) at the first visit to a nephrologist to define early referral (ER) and LR in a retrospective analysis. Patients admitted with a GFR < 20 mL/min/1.73 m2 were classified as LR. The 75 patients with chronic renal failure beginning renal replacement therapy (RRT) at Innsbruck University Hospital between January 1999 and October 2000 were included. Patient characteristics were compared between the two groups. Survival analysis until the end of 2002 was carried out using Cox's proportional hazard model. To identify the influence of comorbidity on mortality a comorbidity score was applied.

Results: Thirty-three patients were classified as ER and 42 patients as LR. Diabetic nephropathy was more frequent in the LR group (18 vs. 6 patients, p = 0.005). ER patients were significantly younger (53 +/- 16 yrs) as compared to LR patients (62 +/- 14 yrs, p = 0.012). Comorbid conditions were more frequent in the LR group (comorbidity score 1.5 +/- 1.3 for LR and 0.7 +/- 1.1 for ER, p = 0.003). During follow-up, 27 patients died, 19 from the LR group and 8 from the ER group. In the univariate analysis, comorbidity score (p < 0.001) and age (p = 0.017) were significantly associated with mortality, whereas LR patients demonstrated higher mortality (p = 0.076). By multivariate analysis the comorbidity score (p < 0.001) only was associated with mortality within at least 2 yrs of RRT.

Conclusion: Over half of the patients with end-stage renal disease (ESRD) were referred too late, with a GFR < 20 mL/min/1.73 m2. Mortality during the 1st 2 yrs on RRT was mainly determined by comorbidity, acquired during the course of chronic renal failure. In comparison, the negative impact of LR seems to be minor and requires a larger sample size to be demonstrated.