Validation of the Toronto formula to predict progression in IgA nephropathy

Predicting outcome in IgA nephropathy (IgAN) is difficult. The Toronto formula uses average mean arterial blood pressure and proteinuria during the first 2 years of follow-up (MAP(0-2), UP(0-2)) to predict the subsequent slope of estimated creatinine clearance (eCrCl). We aimed to validate the Toronto formula in a Scottish cohort and test the hypothesis that adding the slope eCrCl over the first 2 years of follow-up (eCrCl(0-2)) would improve the predictive utility of a similar multivariate model.

Adultsfrom our centre with biopsy-proven IgAN (n = 169) and at least 2 years of follow-up (median 129.4 months) were included. Clinical data were used to calculate MAP(0-2),UP(0-2),slope eCrCl(0-2 )and predicted slope eCrCl (using the Toronto formula).

There was a significant correlation between predicted slope eCrCl using the Toronto formula and actual slope eCrCl (R(2 =) 0.21; p < 0.001). The formula predicted the actual rate of progression to within 4 ml/min/year in 75% of subjects, predicting patients with the most rapid deterioration with the greatest accuracy. The multivariate linear regression model created in our cohort using the same independent variables as the Toronto formula to predict the overall slope eCrCl had an R(2) of 0.22 (p < 0.001) and adding the slope CrCl(0-2) only increased this to 0.25.

The Toronto formula is valid in a European population and useful for identifying patients at high risk of future deterioration in renal function. Adding slope eCrCl(0-2) to a predictive model containing MAP(0-2), andUP(0-2 )does not appear to improve prediction of the overall slope of eCrCl.