Kidney Week

Abstract: SA-PO1115

The Japanese Histological Grade Classification and the Oxford Classification of IgA Nephropathy Show the Same Prediction Performance on Renal Functional Decline in 905 Japanese Patients

Session Information

• Pathology and Lab Medicine: Clinical
  October 27, 2018 | Location: Exhibit Hall, San Diego Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Pathology and Lab Medicine

• 1502 Pathology and Lab Medicine: Clinical

Authors

• Joh, Kensuke, Tohoku University Graduate School of Medicine, Sendai-city, Japan

Kensuke Joh,

• Hashiguchi, Akinori, Keio University School of Medicine, Tokyo, Japan

Akinori Hashiguchi,

• Shimizu, Akira, Nippion Medical School, Tokyo, Japan

Akira Shimizu,

• Hisano, Satoshi, University of Occupational and Environment Health, Kitakyushu, Japan

Satoshi Hisano,

• Katafuchi, Ritsuko, National Fukuoka-Higashi Medical Center, Koga, Fukuoka, Japan

Ritsuko Katafuchi,

• Kawamura, Tetsuya, Division of Kidney and Hypertension, Tokyo, Japan

Tetsuya Kawamura,

Background

The Japanese Histological Grade Classification (JHGC) (Gr1-Gr4) was established as a lumped system corresponding to <25%, 25-49%, 50-74% and ≥75% of total glomeruli exhibiting active crescents (C）, global sclerosis (G), segmental sclerosis (S), or fibrous crescents. In contrast, the Oxford classification is a split system constituting of dichotomized M, E, S, T1, T2, C1, and C2. The purpose was to compare the 2 classifications systems focusing on their ability to predict renal functional decline.

Methods

The 905 Japanese patients (pts) with IgA nephropathy (male：49%, median age: 38 yrs) were prospectively followed for a median of 48 months. The average amount of proteinuria (PU) was 1.07 g/day. Mean eGFR and median rate of decline in eGFR were 77.0 ml/min/1.73m² and -0.95 ml/min/1.73m²/y, respectively. The prediction performance was compared using the Harrell’s C statistic (HCS). The clinical and therapeutic variables were additionally collected including initial PU, initial square root eGFR (SReGFR), initial mean arterial pressure, steroid therapy (ST), RAS blockade (RASB) and tonsillectomy (TON).

Results

66 %, 54% and 42% of the patients were treated by ST, RASB, and TON, respectively. In stepwise Cox regression analysis for the full model, Gr4, Gr3, Gr2, PU, SReGFR, TON and ST for 1.5 time's increase of serum creatinine (sCr) were selected as independent predictors, whose HR were 28, 7, 4, 2, 1, 1 and 0, respectively. HCS was 0.87. In Oxford, T2, T1, M, RASB, C1, PU, TON, SReGFR, ST and E were selected, whose HR were 4, 3, 2, 2, 2, 2, 1, 1, and 0, respectively, HCS was 0.86, which was statistically not different from that of JHGC. For the limited model without clinical and therapeutic parameters, Gr4, Gr3, and Gr2 were selected, whose HR was 79, 14, and 5, respectively. HCS was 0.82. In Oxford, T2, T1, M, and E but not S, C1, and C2 were selected, whose HR were 20, 5, 3, and 1, respectively. HCS was 0.81.

Conclusion

Both the lumped system of JHGC and the split system of Oxford showed the same high prediction performance on renal functional decline in full and limited model. However, a flexibility in a lumped approach of JHGC as opposed to a split system of Oxford may be more robust when being applied to diverse cohorts.

Funding

• Government Support - Non-U.S.