Evaluate The Four Different Pathological Classification Of Primary IgA Nephropathy

Background and objective: IgA nephropathy (IgAN) is the most common primarykidney disease, and at present it is also the most important cause leading to end-stage renaldisease (ESRD) in our country. It has been confirmed that early diagnosis, accurateassessment and effective treatment are the important measures to slow the progression ofIgAN. The characteristic and diagnostic lesion of IgAN is the deposition of glomerular IgA.However, pathological lesions of IgAN are so complicated that more information related todisease such as prognosis and treatment is expected. A lot of researches developed manypathological classification system, which have not been comparatively analyzed. Therefore,this study aims to analyze the four pathological classification of primary IgAN, comparingthe prediction ability of prognosis and the curative effect and reliability and determiningthe correlation among them.Methods:(1) We examined the patient records of newly biopsy-confirmed cases ofIgA nephropathy between January2002and December2007in the Department ofNephrology, Chinese PLA General Hospital. Inclusion criteria for the study were primaryIgAN, age≥14years, with an estimated glomerular filtration rate (eGFR)≥15ml/min/1.73m2at biopsy. According to the different pathologic classification to reevaluatetheir pathological lesions, we analyzed the correlation of the different pathologicalclassification with decline in eGFR≥50%and progress to ESRD to evaluate the predictionability for the prognosis.(2) We examined a cohort who had participated in a prospective,randomized, double-blind, double simulation clinical trial-new treatments in patients withprimary IgAN. Inclusion criteria for the study were biopsy-confirmed primary IgAN in theDepartment of Nephrology, Chinese PLA General Hospital. We analyzed the correlation ofthe different pathological classification with the curative effect to evaluate the predictionability for the curative effect.(3) Using pathological data of patients from the previous twostudies, we analyzed repeatability of different pathological classification to assess their reliability and compared the correlation of different pathological classification to determinetheir correspondence.Results:(1) A total of412IgAN patients with no less than3years followed-up wereincluded. Average follow-up time was80.62±23.63months,44(10.68%) casesprogressed to ESRD,70(16.99%) patients had decline in eGFR≥50%. Multivariable Coxproportional hazards regression analysis suggested four pathological classification wereindependently associated with an increased risk of decline in eGFR≥50%and progress toESRD. Comparing Harrell’s concordance index and continuous net reclassification indexof different models, we found that prediction ability of four pathological classification forprognosis were comparable and were higher than models with clinical factors only, exceptHaas class.(2) A total of162IgAN patients were included. Multivariable logisticregression analysis suggested Lee’s grade, the scoring system of Katafuchi and Oxfordclassification could predict the curative effect. Comparing the area under thereceiver-operating curve of models of four pathological classification, we found that theability of four pathologic classification to predict the curative effect were limited and werenot higher than models with clinical factors only. The Oxford classification could provideevidence for choosing treatment by interaction effect between lesions and drugs.(3) Usingpathological data of574patients from the previous two studies, we found reliability of fourpathological classification were comparable by cacluating the consistency coefficient andLee’s grade, Haas class, the scoring system of Katafuchi and some of lesions of Oxfordclassification were positively related to each other.Conclusion: The four different pathological classification of primary IgAN hadcomparable prediction ability of decline in eGFR≥50%and progress to ESRD, whichwere higher than models with clinical factors at biopsy only, while their prediction abilityof the curative effect were limited. The Oxford classification could provide evidence forchoosing treatment. Their reliability were comparable and positively related to each other.