Department of Nephrology, Tokyo Metropolitan Komagome Hospital, 3-18-22 Honkomagome, Bunkyo-Ku, Tokyo 113-0021, Japan.
Clin Exp Nephrol. 2011 Apr;15(2):242-7. doi: 10.1007/s10157-010-0393-x. Epub 2011 Jan 20.
Human immunodeficiency virus (HIV)-infected individuals have concurrent risk factors for chronic kidney disease (CKD), but there are no prediction models to quantify the effect of multiple factors on the development of incident CKD.
A 1-year prognostic model was developed using prospective data between January 2008 and March 2009, from a derivation cohort of 623 patients with an estimated glomerular filtration rate (eGFR) exceeding 60 ml/min/1.73 m(2) at baseline. Incident CKD was defined as both an eGFR falling below 60 ml/min/1.73 m(2) and a decrease in eGFR of at least 25% during the period. Continuous variables were divided into categories using the mean value of the whole cohort. Multivariate logistic regression was used to determine baseline categories associated with incident CKD. The model was tested with prospective data between April 2009 and March 2010, from a validation cohort of 534 patients exceeding 60 ml/min/1.73 m(2) at baseline, consisting of part of the derivation cohort and newly enrolled patients. The discriminative ability was evaluated using the area under the receiver operating characteristic (AROC) curve.
Eighteen subjects developed incident CKD (2.9%). In a multivariate logistic regression analysis of the derivation cohort, 5 variables (age, CD4 cell count, diabetes, proteinuria, and eGFR at baseline) were independently associated with the incidence of CKD. The AROC curve was 0.841 (95% CI 0.799-0.894) in the validation cohort.
This prediction model may be a useful tool for identifying HIV-infected individuals with a high likelihood of new-onset CKD.
人类免疫缺陷病毒（HIV）感染者同时存在慢性肾脏病（CKD）的危险因素，但目前尚无预测模型来量化多种因素对新发 CKD 的影响。
使用 2008 年 1 月至 2009 年 3 月之间前瞻性数据，从基线估计肾小球滤过率（eGFR）超过 60 ml/min/1.73 m2 的 623 例患者中建立 1 年预后模型。新发 CKD 的定义为 eGFR 降至 60 ml/min/1.73 m2 以下和 eGFR 在此期间至少下降 25%。连续变量使用整个队列的平均值进行分类。使用多元逻辑回归确定与新发 CKD 相关的基线类别。该模型在 2009 年 4 月至 2010 年 3 月之间前瞻性数据进行验证，来自基线时 eGFR 超过 60 ml/min/1.73 m2 的 534 例患者的验证队列，其中包括部分衍生队列和新入组的患者。使用接收者操作特征（ROC）曲线下面积（AROC）评估区分能力。
18 例患者发生新发 CKD（2.9%）。在衍生队列的多元逻辑回归分析中，5 个变量（年龄、CD4 细胞计数、糖尿病、蛋白尿和基线时 eGFR）与 CKD 的发生独立相关。验证队列的 AROC 曲线为 0.841（95%CI 0.799-0.894）。
该预测模型可能是识别新发生 CKD 可能性较高的 HIV 感染者的有用工具。
