Mise Koki, Hoshino Junichi, Ueno Toshiharu, Hazue Ryo, Sumida Keiichi, Hiramatsu Rikako, Hasegawa Eiko, Yamanouchi Masayuki, Hayami Noriko, Suwabe Tatsuya, Sawa Naoki, Fujii Takeshi, Hara Shigeko, Ohashi Kenichi, Takaichi Kenmei, Ubara Yoshifumi
Nephrology Center, Toranomon Hospital, Tokyo, Japan.
Department of Pathology, Toranomon Hospital, Tokyo, Japan.
Diabetes Metab Res Rev. 2015 Sep;31(6):572-81. doi: 10.1002/dmrr.2633. Epub 2015 Mar 5.
The effect of clinical and pathological parameters on the estimated glomerular filtration rate (eGFR) decline has not been investigated in patients with type 2 diabetes and overt proteinuric biopsy-proven diabetic nephropathy.
Among 198 patients with type 2 diabetes who underwent renal biopsy and were confirmed to have pure diabetic nephropathy according to the recent classification, 128 patients with overt proteinuria were enrolled. Receiver operating characteristic, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) analyses were performed using models adjusted for various clinical and pathological covariates to determine the best predictors of rapid eGFR decline [defined as >14.9%/year (median eGFR decline)].
A model that incorporated proteinuria showed the largest area under the curve (AUC) among clinical models, which suggested that proteinuria was the best clinical predictor. Although a model incorporating interstitial fibrosis and tubular atrophy (IFTA) score did not display a significantly larger AUC than the model with proteinuria (0.843 vs 0.812, respectively, p = 0.47), a model with both IFTA score and proteinuria had a significantly larger AUC than the model with proteinuria alone (0.875 vs 0.812, respectively, p = 0.014). Similarly, the addition of IFTA score resulted in a significantly greater net reclassification improvement and integrated discrimination improvement than the model with proteinuria alone [NRI: 0.78 (95% CI: 0.43-1.13; p < 0.001), IDI: 0.13 (95% CI: 0.07-0.19; p < 0.001)].
Our results suggest that not only proteinuria but also tubulointerstitial lesions should be assessed to predict rapid eGFR decline in patients with type 2 diabetes who have overt proteinuria and biopsy-proven diabetic nephropathy.
在2型糖尿病且经活检证实有明显蛋白尿的糖尿病肾病患者中，尚未研究临床和病理参数对估计肾小球滤过率（eGFR）下降的影响。
在198例接受肾活检且根据最新分类确诊为单纯糖尿病肾病的2型糖尿病患者中，纳入128例有明显蛋白尿的患者。使用针对各种临床和病理协变量进行调整的模型进行受试者操作特征、净重新分类改善（NRI）和综合辨别改善（IDI）分析，以确定eGFR快速下降[定义为>14.9%/年（eGFR下降中位数）]的最佳预测因素。
在临床模型中，纳入蛋白尿的模型显示出最大的曲线下面积（AUC），这表明蛋白尿是最佳的临床预测因素。虽然纳入间质纤维化和肾小管萎缩（IFTA）评分的模型的AUC并未显著大于含蛋白尿的模型（分别为0.843和0.812，p = 0.47），但同时包含IFTA评分和蛋白尿的模型的AUC显著大于仅含蛋白尿的模型（分别为0.875和0.812，p = 0.014）。同样，与仅含蛋白尿的模型相比，加入IFTA评分导致净重新分类改善和综合辨别改善显著更大[NRI：0.78（95%CI：0.43 - 1.13；p < 0.001），IDI：0.13（95%CI：0.07 - 0.19；p < 0.001）]。
我们的结果表明，对于有明显蛋白尿且经活检证实为糖尿病肾病的2型糖尿病患者，若要预测eGFR快速下降，不仅应评估蛋白尿，还应评估肾小管间质病变。
