Comparison Of Estimated Glomerular Filtration Rate Equations In Patients With Benign Arteriolar Nephrosclerosis

BackgroundChronic kidney disease (CKD) has been being a global public health problem. The high prevalence of CKD and poor prognosis bring high medical costs to society. In2007, National Institutes of Health embarked on a study on the National Health and Nutrition Survey from1999to2004. They found adults (≥20years) with CKD prevalence as high as13%, compared with the1988-1994, the rate of prevalence increased about30%. Similar results are not limited in developed counties. An epidemiological survey of Chinese in2012showed that the prevalence of CKD in China is about10.8%.The annual mortality of dialysis patients is about21%to23%. National Dialysis Transplant Registry of China in1999reported that the mortality rate of dialysis patients was47%and the main cause of death was cardiovascular events. The annually costs of dialysis may be hundreds of thousands which bring a heavy burden to society. The pathophysiology of CKD suggested that early diagnosis and treatment of CKD may be delay or prevent the progress of CKD to improve the prognosis. As we know, hypertension is one of the key causes of CKD. A chronic and non-communicable disease epidemiology survey in2002showed that the incidence of hypertension was up to17.7%in participates with age above15. However, the awareness rate of hypertension was only30%and only6%patients received effective treatment, which indicated most of patients did not known they underwent hypertension and got any treatment. The more poorly of blood pressure controlled long time, the higher risk of benign arteriolar nephrosclerosis will be occur in patients. The progress of benign arteriolar nephrosclerosis shows that the kidney injury in early is main renal tubular, then become glomerluar, which indicated the clinical intervention of benign arteriolar nephrosclerosis may be different due to kidney function. Now, glomerular filtration rate (GFR) is a common indicator of kidney function. Kidney Disease Outcomes Quality Initiative (K/DOQI) guidelines in2002suggested that the diagnosis criteria and stage of CKD mainly depended on the GFR. Therefore, evaluated accurately of GFR has an important role in prevention and prognostic evaluation of CKD and prevention of health policy formulation.As we known, GFR must be indirectly measured by the clearance rate of certain markers out, because it could not be measured directly.Although the clearance rate of inulin or radionuclides such as99mTc-DTPA,51Cr-EDTA is extremely accurately reflect the GFR, the test methods are complicated and expensive which is difficult to use in clinical practice. Therefore, different estimated glomerular filtration rate (eGFR) equations based on different markers has been developed and validated. Numerous clinical studies suggested that eGFR equations based on serum creatinine were better than that based on other markers. Serum creatinine test method is mature and simple, so it is widely used in clinical practice. However, serum creatinine could be secreted by renal tubular and then discharge. Hence, the24-hours urine creatinine clearance may be overestimate GFR about10～40%. The progress of benign arteriolar nephrosclerosis shows that the kidney injury in early is main renal tubular. Then creatinine secreted by tubular reduced, serum creatinine will correspondingly increase, which indicated eGFR equations based on serum creatinine would be underestimate the GFR. Although eGFR equations were validated in different population, it is not know whether these equations are suitable to be used in patients with benign arteriolar nephrosclerosis. Hence, in order to evaluate the eGFR of these patients more accurate and determine the applicability of each eGFR equation, we embark a study to validate currently eGFR equations in patients with benign arteriolar nephrosclerosis. Two-sample method plasma clearance of99mTc-DTPA was definded as the reference GFR.Part I Validation of estimated glomerular filtration rate equations in all patients with benign arteriolar nephrosclerosis.ObjectiveValidation of several estimated glomerular filtration rate equations in patients with benign arteriolar nephrosclerosis.Study design and methodsStudy populationWe enrolled179adult patients with benign arteriolar nephrosclerosis from January2010to June2012at the Guangdong General Hospital. All patients must underwent dual-plasma method for the determination of GFR, patients with hypertension history above8years and patients with hypertensive heart disease diagnosed by echocardiography and (or) with fundus lesions damaged by hypertension. Exclusion criteria were as follows:patients with other kidney disease, diabetes, ketoacidosis, heart dysfunction (defined as New York Heart Association function stage Ⅲ or Ⅳ), acute renal injury, edema, pleural effusion or ascites, muscle atrophy, limb absence, severely malnourished patients; taking some drugs which may interfere with serum creatinine excretion such as cimetidine, trimethoprim, or diuretics and as on.MethodsCollected data:(1) gender, age, height, weight;(2) biochemical variables:serum creatinine measured by automatic analyzer of Beckman DXC800(the alkaline picric power assay), blood urea nitrogen (urease method), serum albumin (bromocresol purple method);(3) auxiliary examination:GFR measured by dual plasma, ultrasound echocardiography, ophthalmoscopy, urine routine analysis.Estimated GFR methods:We used different equations to estimate GFR. The equations were as follows: Cockcroft-Gault (CG) equation, CG equation corrected by body surface area, the Modification of Diet in Renal Disease (MDRD) equation, the abbreviated MDRD equation, the modified MDRD equation, the modified abbreviated MDRD equation, Chronic Kidney Disease Epidemiology (CKD-EPI) equation. The two-sample method99mTc-DTPA plasma clearance was definded as reference GFR (rGFR). The equations for calculating body surface area:body surface area=0.0061X height (cm)+0.0128×body weight (kg)-0.15299.Statistical analysisMedcalc software and SPSS version13.0were used for data analysis. Data are expressed as mean±standard deviation. The deviation was definded as the difference between eGFR and rGFR measured with two-sample plasma clearance method; the absolute deviation was definded as the absolute difference between eGFR and rGFR measured with two-sample plasma clearance method. We used Pearson correlation and linear regression to describe the relationship between eGFR and rGFR.We drew a graph using Bland-Altman between deviation and the average of eGFR and rGFR for regression. The area between the horizontal0and the regression line represented the difference between each eGFR equation and rGFR. That is to say, the larger area means the more bias between each eGFR equation and rGFR. The95percentiles of the distribution of the regression line represented the accuracy of each eGFR equation. The accuracies of GFR estimated by using each equation were expressed as percentage of points deviating less than15%,30%,50%from rGFR and the chi-square test test the difference accuracy of the each eGFR.ResultsA total of179patients with benign arteriolar nephrosclerosis were enrolled in the study, including121males and58women and the average age was (66.39±15.06) years, the mean serum creatinine was (292.67±293.77) umol/L, the average urea nitrogen was (12.22±9.30) mmol/L, mean serum albumin was (33.06±4.89) g/L. The average rGFR of patients was (40.47±25.11) ml±min-1±1.73m-2The stage of CKD in patients were5.03%,15.08%,37.43%,28.49%and13.97%, respectively.CG equation, CG equation modified by body surface area, MDRD equation, the abbreviated MDRD equation and CKD-EPI equation underestimated GFR. The difference between eGFR based on CG equation modified by body surface area and GFR was largest among them (3.54ml·min-1·1.73m-2). However, the difference between eGFR based on the abbreviated MDRD equation and rGFR was less than that based on others. Modified MDRD equation and the modified abbreviated MDRD equation overestimated GFR. The difference between eGFR based on them and GFR were6.43ml·min-1·1.73m-2,8.01ml·min-1·1.73m-2, respectively. The difference performance between eGFR based on the abbreviated MDRD equation and rGFR was not statistically significant (P=0.191), others were statistically significant (P<0.05).The differences of correlation coefficient and determination coefficient among eGFR equations were not significant. The deviation of abbreviated MDRD equation was significantly less than that of other equations (P<0.05). The absolute deviation of abbreviated MDRD equation was slightly above than that of CG equation modified by body surface area and CKD-EPI equation. There were significant differences of absolute deviation between abbreviated MDRD equation and modified MDRD equation and the abbreviated modified MDRD equation, however, that of others were not significant (P<0.05). There were significant differences of15percentiles of the accuracy between abbreviated MDRD equation and CG equation, modified MDRD equation and the abbreviated modified MDRD equation (P<0.05). The difference of30percentiles of the accuracy between abbreviated MDRD equation and the modified abbreviated MDRD equation was statistically significant (P<0.05). However, there were no significant differences of50percentiles of the accuracy between abbreviated MDRD equation and other equation (P>0.05). The precision and degree of deviation of abbreviated MDRD equation,using by Bland-Altman analysis, were better than that of CG equation, modified MDRD equation and the abbreviated modified MDRD equation. However, the precision and degree of deviation between abbreviated MDRD equation and others were relatively close.ConclusionOur study suggested that eGFR based on the abbreviated MDRD equation should be more suitable to be used to in patients with benign arteriolar nephrosclerosis.Part Ⅱ Validation of estimated glomerular filtration rate equations in patients with benign arteriolar nephrosclerosis and different chronic kidney disease stages. ObjectiveValidation of several estimated glomerular filtration rate equations in patients underwent benign arteriolar nephrosclerosis with CKD stages.Study design and methodsStudy populationWe enrolled179adult patients with benign arteriolar nephrosclerosis from January2010to June2012at the Guangdong General Hospital. All patients must underwent dual-plasma method for the determination of GFR, patients with hypertension history above8years and patients with hypertensive heart disease diagnosed by echocardiography and (or) with fundus lesions damaged by hypertension. Exclusion criteria were as follows:patients with other kidney disease, diabetes, ketoacidosis, heart dysfunction (defined as New York Heart Association function stage Ⅲ or Ⅳ), acute renal injury, edema, pleural effusion or ascites, muscle atrophy, limb absence, severely malnourished patients; taking some drugs which may interfere with serum creatinine excretion such as cimetidine, trimethoprim, or diuretics and as on.MethodsCollected data:(1) gender, age, height, weight;(2) biochemical variables:serum creatinine measured by automatic analyzer of Beckman DXC800(the alkaline picric power assay), blood urea nitrogen (urease method), serum albumin (bromocresol purple method);(3) auxiliary examination:GFR measured by dual plasma, ultrasound echocardiography, ophthalmoscopy, urine routine analysis. estimated GFR methods:We used different equations to estimate GFR. The equations were as follows: Cockcroft-Gault (CG) equation, CG equation corrected by body surface area, the Modification of Diet in Renal Disease (MDRD) equation, the abbreviated MDRD equation, the modified MDRD equation, the modified abbreviated MDRD equation, Chronic Kidney Disease Epidemiology (CKD-EPI) equation. The two-sample method99mTc-DTPA plasma clearance was definded as reference GFR (rGFR). The equations for calculating body surface area:body surface area=0.0061X height (cm)+0.0128X body weight (kg)-0.15299.Statistical analysisMedcalc software and SPSS version13.0were used for data analysis. Data are expressed as mean±standard deviation. The deviation was definded as the difference between eGFR and rGFR measured with two-sample plasma clearance method; the absolute deviation was definded as the absolute difference between eGFR and rGFR measured with two-sample plasma clearance method. We used Pearson correlation and linear regression to describe the relationship between eGFR and rGFR.We drew a graph using Bland-Altman between deviation and the average of eGFR and rGFR for regression. The area between the horizontal0and the regression line represented the difference between each eGFR equation and rGFR. That is to say, the larger area means the more bias between each eGFR equation and rGFR. The95percentiles of the distribution of the regression line represented the accuracy of each eGFR equation. The accuracies of GFR estimated by using each equation were expressed as percentage of points deviating less than15%,30%,50%from rGFR and the chi-square test test the difference accuracy of the each eGFR. In order to more accurately evaluate the applicability of the eGFR equations, we grouped the patients according the rGFR into CKD stage1-2group and CKD stage3-5group, then we used similar Statistical analysis to validate eGFR equations.ResultsA total of36patients, among172patients, underwent CKD stage1or2were included in study, including27males and9females and the average age was (56.00±17.96) years, the mean serum creatinine was (93.33±25.25) umol/L, the average urea nitrogen was (4.71±1.70) mmol/L, mean serum albumin was (35.98±4.00) g/L. The average rGFR of patients was (79.44±18.02) ml·min-1·1.73m-2. However, A total of143patients, among172patients, underwent CKD stage3-5were included in study, including94males and49females and the average age was (69.00±13.06) years, the mean serum creatinine was (342.86±308.89) umol/L, the average urea nitrogen was (14.11±9.47) mmol/L, mean serum albumin was (32.33±4.82) g/L. The average rGFR of patients was (30.66±15.12) ml·min-1·1.73m-2.In CKD stage1-2and CKD stage3-5groups, CG equation modified by body surface area, MDRD equation, the modified MDRD equation and CKD-EPI equation were underestimated GFR. The difference between eGFR based on CG equation modified by body surface area and GFR was largest among them. In CKD stage1-2and CKD stage3-5groups, modified MDRD equation and the modified abbreviated MDRD equation were overestimated GFR. The difference between eGFR based on the modified abbreviated MDRD equation and GFR was largest among them. Compared with CG equation underestimated GFR in CKD stage1-2groups, CG equation overestimated GFR in CKD stage3-5groups.Compared the deviation of abbreviated MDRD equation with other equations in CKD stage1-2group, there were significantly differences between abbreviated MDRD equation and modified MDRD equation and the modified abbreviated MDRD equation (P<0.05). The absolute deviation of abbreviated MDRD equation was less than that of other equations. Compared the absolute deviation of abbreviated MDRD equation with other equations, there were significantly differences between CG equation, CG equation modified by body surface area, modified MDRD equation, the modified abbreviated MDRD equation and CKD-EPI equation (P<0.05). There was significant difference of15percentiles of the accuracy between abbreviated MDRD equation and the modified abbreviated MDRD equation (P<0.05). The differences of30percentiles of the accuracy between abbreviated MDRD equation and modified MDRD equation and the modified abbreviated MDRD equation were statistically significant (P<0.05). However, there were no significant differences of50percentiles of the accuracy between abbreviated MDRD equation and other equations (.P>0.05). The precision between abbreviated MDRD equation and MDRD equation was relatively close. However, the precision of abbreviated MDRD equation was better than that of CG equation, CG equation modified by body surface area, modified MDRD equation, the abbreviated Chinese MDRD equation and CKD-EPI equation.There was significantly difference of deviation between abbreviated MDRD equation and other equations (P<0.05). Compared the absolute deviation of abbreviated MDRD equation with other equations in CKD stage3-5group, there were significantly differences between abbreviated MDRD equation and modified MDRD equation and the modified abbreviated MDRD equation (P<0.05). There was significant difference of15percentiles of the accuracy between MDRD equation and the modified abbreviated MDRD equation (P<0.05). The difference of30percentiles of the accuracy between MDRD equation and the modified abbreviated MDRD equation was also statistically significant (P<0.05). However, compared the difference of50percentiles of the accuracy of MDRD equation with that of other equations, there was significant difference between MDRD equation and CKD-EPI equation (P<0.05).ConclusionIn order to evaluate the performances of eGFR equations in CKD stages, the eGFR equations were validated in patients according to renal function. We found that eGFR based on the abbreviated MDRD equation would be more suitable to be used to in patients with CKD stage1or2, however, that based on MDRD equation in patients with CKD stage3-5.