Application of the International IgA Nephropathy Prediction Tool one or two years post-biopsy

International IgA Nephropathy Network

Research output: Contribution to journalArticleAcademicpeer-review

23 Citations (Scopus)

Abstract

The International IgA Nephropathy (IgAN) Prediction Tool is the preferred method in the 2021 KDIGO guidelines to predict, at the time of kidney biopsy, the risk of a 50% drop in estimated glomerular filtration rate or kidney failure. However, it is not known if the Prediction Tool can be accurately applied after a period of observation post-biopsy. Using an international multi-ethnic derivation cohort of 2,507 adults with IgAN, we updated the Prediction Tool for use one year after biopsy, and externally validated this in a cohort of 722 adults. The original Prediction Tool applied at one-year without modification had a coefficient of variation (R2) of 55% and 54% and four-year concordance (C statistic) of 0.82 but poor calibration with under-prediction of risk (integrated calibration index (ICI) 1.54 and 2.11, with and without race, respectively). Our updated Prediction Tool had a better model fit with higher R2 (61% and 60%), significant increase in four-year C-statistic (0.87 and 0.86) and better four-year calibration with lower ICI (0.75 and 0.35). On external validation, the updated Prediction Tool had similar R2 (60% and 58%) and four-year C-statistics (both 0.85) compared to the derivation analysis, with excellent four-year calibration (ICI 0.62 and 0.56). This updated Prediction Tool had similar prediction performance when used two years after biopsy. Thus, the original Prediction Tool should be used only at the time of biopsy whereas our updated Prediction Tool can be used for risk stratification one or two years post-biopsy.
Original languageEnglish
Pages (from-to)160-172
Number of pages13
JournalKidney International
Volume102
Issue number1
Early online date2022
DOIs
Publication statusPublished - 1 Jul 2022

Keywords

  • IgA nephropathy
  • disease progression
  • end-stage kidney disease
  • prediction tool
  • risk prediction

Cite this