IgA 腎症における腎生検後 5 年間の治療反応性と腎予後の関連性

Background: Immunoglobulin A nephropathy (IgAN) is the most common type of primary glomerulonephritis. Since most patients have a relatively benign renal prognosis, long-term follow-up is required. During such a long course of disease, relapse of IgAN is occasionally observed after upper respiratory tract infection or without any trigger. However, little is known about the impact of relapse on long-term renal outcomes. Methods: In this retrospective cohort study of biopsy-proven primary IgAN, we analyzed the association of 5-year therapeutic responsiveness (relapse) with the subsequent development of end-stage kidney disease (ESKD) using a 5-year landmark analysis (Cox model) and explored predictors of relapse from histological and clinical data at baseline. Results: Among 563 patients from the exploratory cohort, most relapses (13.7%) occurred within 5 years after treatment. Using 5-year landmark analysis, among 470 patients, 79 developed ESKD during a median follow-up period of 155 months. Even after adjustment for clinicopathological relevant confounders, hazard ratios (95% confidence intervals) in the relapse and non-responder groups compared with the remission group were 2.86 (1.41-5.79) and 2.74 (1.48-5.11), respectively. Among 250 patients who achieved remission within 5 years, proteinuria, eGFR, mesangial hypercellularity, endocapillary hypercellularity, segmental sclerosis, and crescent, but not interstitial fibrosis/tubular atrophy, were independent predictors of 5-year relapse in multivariable logistic regression analysis, CONCLUSIONS: Both relapsers and non-responders had similarly strong association with ESKD in patients with IgAN. We also confirmed the predictors of relapse 5 years after renal biopsy, which may guide the treatment strategies for patients with IgAN who occasionally relapse after remission.博士（医学）・乙第1530号・令和5年3月15日This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/s10157-022-02221-