Analgesic-associated nephropathy in the West of Scotland: a 12-year observational study

Background. Analgesic-associated nephropathy (AAN) is an important and preventable cause of chronic renal failure (CRF). Although its incidence is falling in some countries, others are witnessing an increase in the number of new cases. Methods. The aim of this study was to evaluate the natural history of AAN, determine the correlates of the rate of decline in renal function and examine factors conferring an increased risk of death or dialysis in such patients. A prospective observational cohort study of all patients with AAN attending a single-centre was conducted. Results. Seventy-eight patients (25 male), with at least 24 months of follow-up for analysis, were diagnosed as having AAN over the 10-year period 1989-1999. During follow up, the mean (+/- SD) rate of change in estimated creatinine clearance (ECC) was -1.2ml/min/year (+/- 5.28). By multiple linear regression three variables were found to independently predict the rate of deterioration in ECC; continuing analgesic use (P < 0.001), degree of proteinuria at presentation (P = 0.002) and male sex (P = 0.03). A Cox's model revealed a 6-fold increase in the hazard of reaching the combined end-point of death or dialysis in those patients with AAN who continue to use analgesics. This was independent of the other two significant risk factors of pre-existing vascular disease (HR 3.93, 1.36-11.29) and ECC at presentation (HR 0.95, 0.91-0.98 per ml/min). Conclusions. In patients with CRF due to AAN ongoing analgesic use, male gender and increasing proteinuria predict a more rapid decline in renal function. Patients who continue analgesics, those with pre-existing vascular disease and those with more advanced renal impairment at presentation, are at a significantly increased risk of reaching the combined end-point of death or end-stage renal failure requiring dialysis. The design of this study, however, leaves it open to the criticism that selection bias may account for some of its effects, and as with all work on AAN the possible confounding issue of reverse causality is difficult to dismiss.