A tricontinental view of IgA nephropathy

Background. The purpose of this retrospective study was to analyse patients from four centres in three continents to determine if differences in long-term outcome of IgA nephropathy (IgAN) are explained by clinical and laboratory features at presentation. Methods. The study included 711 adults with biopsy-proven IgAN from Glasgow, UK (n = 112), Helsinki, Finland (n=204), Sydney, Australia (n=121) and Toronto, Canada (n = 274). Data collected from time of presentation to a nephrologist were age, gender, 24-h urine protein excretion (UP0),mean arterial pressure (MAP(0)) and creatinine clearance (CrCl0). Outcomes were slope of creatinine clearance (CrCl) and renal survival. Results. At presentation there was significant variability in baseline clinical features with patients from Helsinki having the lowest median UP0, lowest MAPO and highest CrCl0, all suggesting milder disease. There was significant variability in renal survival between centres with 10-year actuarial survival of 95.7 87.0, 63.9 and 61.6% in Helsinki, Sydney, Glasgow and Toronto, respectively (P < 0.0001; log rank). Cox proportional hazards model revealed lower age(0) and lower CrCl0 were significant independent predictors of reduced renal survival. In addition, patients from Helsinki and Sydney but not Glasgow had significantly longer renal survival than patients from Toronto. Median slope of CrCl varied by region from -1.24ml/min/year in Helsinki, to -3.99ml/min/ year in Toronto (Kruskal-Wallis H test P < 0.0001). By multivariate analysis older age(0), higher CrCl0 and lower UP0 were independently associated with slower progression. Subjects from Helsinki had a significantly slower deterioration independent of the other clinical parameters at presentation. When the 269 patients presenting with CrCl0 <75ml/min were analysed separately there was no independent centre effect. Conclusions. The findings are consistent with the hypothesis that geographical variability in long-term outcome of IgAN is explained by lead-time bias and inclusion of milder cases in centres with apparent good outcome, but do not exclude the possibility that some of the variability is due to other factors such as genetics, diet or treatment.