Caravaca F, Arrobas M, Luna E, Naranjo M, Pizarro J L, Sánchez-Casado E
Servicio de Nefrología Hospital Infanta Cristina 06080 Badajoz.
Nefrologia. 2002;22(5):432-7.
From the results of the Modification of Diet in Renal Disease (MDRD) study, a prediction equation for a more accurate estimate of glomerular filtration rate (GFR), was developed. The present study ais to compare the GFR estimated by MDRD formula and that calculated by the average of creatinine and urea clearances in unselected patients with advanced renal failure. The study group consisted of 320 (163 males) with advanced renal failure not yet on dialysis. Their mean age was 63 +/- 14 years. Diabetic nephropathy was the most common etiology of renal failure (25%). Significant comorbidity was observed in 115 patients. Serum creatinine (Cr), urea and albumin were determined in all patients. Creatinine (Ccr) and urea clearance (Cu) were calculated on a 24-hour urine collection. The GFR was estimated by summing Ccr and Cu, and dividing by two (Ccr-Cu). THe clearances were corrected for a body surface area of 1.73 m2. The MDRD formula for the estimation of GFR included the following parameters: serum Cr, BUN, age, gender and serum albumin. Linear regression analysis and Bland-Altmann plot were utilized to establish the degree of correlation and agreement between both estimations of GFR. The percent differences between the two estimations of GFR was especially analyzed in those subgroups of patients which were not included in the MDRD study (patients older than 70 years, diabetics and those with comorbid conditions). The mean GFR estimated by Ccr-Cu and by MDRD formula were 10.04 +/- 3.10 ml/min and 10.55 +/- 3.60 ml/min, respectively (p < 0.0001). The two parameters correlated significantly (R = 0.76, p < 0.0001). GFR by the MDRD formula tended to overestimate the highest values of Ccr-Cu. The mean percent difference between both methods was 6.5 +/- 23.6. MDRD predictive equation overestimated significantly Ccr-Cu in patients older than 70 years (mean overestimation of 15%), males (10%), diabetics (10%), and mainly in patients with comorbidity (17%). In conclusion, the GFR estimated by MDRD formula is very similar to Ccr-Cu in young uremic patients without comorbidity. However, major discrepancies between these two methods could be observed in older patients, and mainly in those with comorbidity.
根据肾脏病饮食改良（MDRD）研究结果，开发了一个用于更准确估算肾小球滤过率（GFR）的预测方程。本研究旨在比较MDRD公式估算的GFR与未选择的晚期肾衰竭患者肌酐清除率和尿素清除率平均值计算所得的GFR。研究组由320例（163例男性）尚未接受透析的晚期肾衰竭患者组成。他们的平均年龄为63±14岁。糖尿病肾病是肾衰竭最常见的病因（25%）。115例患者存在显著的合并症。对所有患者测定血清肌酐（Cr）、尿素和白蛋白。通过收集24小时尿液计算肌酐清除率（Ccr）和尿素清除率（Cu）。GFR通过将Ccr和Cu相加并除以2（Ccr-Cu）来估算。清除率校正至体表面积为1.73 m²。估算GFR的MDRD公式包括以下参数：血清Cr、血尿素氮（BUN）、年龄、性别和血清白蛋白。采用线性回归分析和Bland-Altmann图来确定两种GFR估算值之间的相关程度和一致性。尤其对MDRD研究未纳入的患者亚组（年龄大于70岁的患者、糖尿病患者和合并症患者）分析了两种GFR估算值之间的百分比差异。通过Ccr-Cu和MDRD公式估算的平均GFR分别为10.04±3.10 ml/min和10.55±3.60 ml/min（p<0.0001）。这两个参数显著相关（R=0.76，p<0.0001）。MDRD公式估算的GFR往往高估了Ccr-Cu的最高值。两种方法之间的平均百分比差异为6.5±23.6。MDRD预测方程在年龄大于70岁的患者（平均高估15%）、男性（10%）、糖尿病患者（10%）以及主要是合并症患者（17%）中显著高估了Ccr-Cu。总之，MDRD公式估算的GFR与无合并症的年轻尿毒症患者的Ccr-Cu非常相似。然而，在老年患者中，尤其是合并症患者中，这两种方法之间存在较大差异。
