ORIGINAL ARTICLE Year : 2016  Volume : 27  Issue : 2  Page : 320325 Accuracy of GFR estimation formula in determination of glomerular filtration rate in kidney donors: Comparison with 24 h urine creatinine clearance Abdul Rauf Hafeez, Muhammad Khalid Idrees, Syed Fazal Akhtar Department of Nephrology, Sindh Institute of Urology and Transplantation, Karachi, Pakistan Correspondence Address: To determine the accuracy of estimated glomerular filtration rate (eGFR) using the modification of diet in renal disease (MDRD), CockcroftGault (CG), and chronic kidney disease epidemiology (CKDEPI) formulas in potential kidney donors compared with 24h urine creatinine clearance, we studied 207 potential live kidney donors in our center. There were 126 (60.9%) males and 81 (39.1%) females. Male:female ratio was 1.6:1. The age of the donors ranged from 1858 years, with mean age of 35.30 ± 9.23 years and most of the individuals were below 40 years of age. The body mass index (BMI) was calculated and venous blood samples were obtained for the measurement of serum creatinine and every study participant was instructed to collect 24h urine. GFR was calculated based on 24h urine creatinine clearance and the formulas. The accuracy of GFR estimation formula was taken as positive if the GFR calculated by the formulas and urine creatinine clearance fell between 90120 mL/min/1.73 m ^{2} . The accuracy of the MDRD formula was 48.8% and the CG formula was 41.5% whereas the accuracy of the CKDEPI formula was 78.2%. The accuracy of the eGFR using the MDRD formula was significantly higher in males than females (57.9% vs. 33.3% P = 0.001), while there was no statistically significant difference in the eGFR between them in case of the use of the CG and the CKDEPI formulas. BMI and obesity had no effect on the accuracy of eGFR by the use of the different formulas. The performance of GFR estimation formulas was sub optimal and these either underestimated and/or overestimated the GFR in healthy subjects. CKDEPI is closer to 24 h urinary creatinine clearance in the calculation of eGFR. However, none of the eGFR formulas can be used in renal transplant donors because of their low accuracy, and 24h urine creatinine clearance should be used for evaluation of the GFR in this population.
Introduction Chronic kidney disease (CKD) is a worldwide public health problem with rising incidence and prevalence as well as increasing cost of therapy. [1],[2] The exact incidence and prevalence of CKD in Pakistan is not known. A survey from Karachi found 25.3% CKD prevalence, while a recent community cohort study found 16.6% prevalence of kidney disease. [3],[4] Renal function is evaluated by the glomerular filtration rate (GFR) determined by several methods. Inulin, an inert, uncharged polymer of fructose, is the only ideal filtration marker that is, freely filtered from glomeruli. It is neither reabsorbed nor secreted by the tubules. Inulin clearance is the gold standard for measurement of GFR, but it is a time consuming and invasive method and the substance is expensive, difficult to dissolve, and maintained in a solution form with usually a short supply. [5] The most common method in clinical practice for the measurement of GFR is 24h urine collection for creatinine clearance. [5] This method has number of problems including difficulty in sample collection, inconvenience to patients, and waste of work time. It is also too cumbersome for screening of renal function when rapid decisions regarding the use of potentially nephrotoxic medications are required. [6] To overcome these problems, investigators have constructed GFR estimating equations that include plasma creatinine and additional variables such as age, sex, and race and body size. The most widely used formulas are those proposed by CockcroftGault (CG), chronic kidney disease epidemiology collaboration (CKDEPI) and the modification of diet in renal disease (MDRD) formula. [2] All formulas have primarily originated from Western Caucasian population and may not be directly applicable to other racial groups and need to be tested first before their application. [7],[8] In this study, we have assessed the accuracy of MDRD and other formulas to determine GFR in individuals evaluated for live related kidney donation and whose GFR is routinely measured by standard 24h urinary creatinine clearance, and the results can validate the use of GFR estimation formulas in a healthy Pakistani population. Materials and Methods This crosssectional study was conducted at Sindh Institute of Urology and Transplantation, Karachi, Pakistan from February 2012 to August 2012. Potential liverelated kidney donors of either gender, aged 1860 years, and with body mass index (BMI) <35 kg/m 2 were included in the study by nonprobability purposive sampling. Individuals who had a history of a primary renal or systemic disease known to affect kidneys function, those who were receiving potentially nephrotoxic drugs, pregnant women, and those who were unable to provide accurate timed urine collection were excluded from the study. After approval from the Ethical Review Committee of the Institution, individuals who fulfilled the study criteria were explained the purpose and procedure of the study and they signed informed consents for the study. Detailed history was obtained and physical examination was done to rule out the possibility of any renal or related systemic diseases. Height was measured in centimeters and weight was recorded in kilograms (Kg) on standard clinical height and weight balance to calculate the BMI. All the information was recorded on the predesigned proforma. Venous blood samples were obtained by trained phlebotomist in vacuum tube and serum creatinine was measured using an automated chemistry analyzer. GFR was calculated using the MDRD, CG, and CKDEPI formulas. Every study participant was instructed to collect 24h urine in a container (provided by institution) without any additive/preservative. The collected urine was sent to institutional lab and the GFR (based on 24 h urine creatinine clearance) was reported by experienced pathologist. The study participants who had undercollections or overcollections were asked to collect the urine again. The accuracy of the formulas was considered as positive if the GFR calculated by formulas (MDRD, CG, and CKDEPI) and 24h urine creatinine clearance fell between 90120 mL/min/1.73 m 2 . Statistical Analysis Data were entered and analyzed on Statistical Package for the Social Sciences (SPSS) Version 16.0 developed by IBM Corporation (originally released by SPSS Inc. Chicago in 2007). The means and standard deviations of the continuous variables (age, weight, height, BMI, serum creatinine, GFR by MDRD, CG, and CKDEPI formulas, and 24h urinary creatinine clearance) were calculated. Frequencies and percentages were computed for gender, obesity, and accuracy. Stratification was performed with regards to age, gender and obesity to evaluate the effects of these outcomes. Chisquare test was applied to assess the difference in these strata. P <0.05 was considered significant. Results We studied 207 potential liverelated kidney donors, 126 (60.9%) males and 81 (39.1%) females. Most of the individuals were below 40 years of age. The mean age of the donors was 35.3 ± 9.23 years (95% CI: 34.036.6). Among the study participants, 9.7% were obese, having BMI ≥30 kg/m 2 . Demographic and descriptive characteristics of study participants are presented in [Table 1].{Table 1} The mean serum creatinine level was 0.89 ± 0.14 mg/dL, ranging from 0.59 to 1.50 mg/dL. GFR calculated by the MDRD formula was 89.03 ± 11.26 mL/min/1.73 m 2 with a range of 59114 mL/min/1.73 m 2 , while GFR measured by 24h urinary creatinine clearance was 104.7 ± 9.19 mL/min/1.73 m 2 with a range of 90 120 mL/min/1.73 m 2 . On average, there were differences of 15.7, 8.59, and 3.88 mL/min/1.73 m 2 in estimated glomerular filtration rate (eGFR) calculated by the MDRD, CG, and CKDEPI formulas, respectively, when compared with the 24h urine creatinine clearance. Accuracy of the MDRD formula was found to be 48.8%. Similarly with respect to age groups accuracy of MDRD was 69.1%, 38.8%, 28.8%, and 42.9% for the individuals below 30 years of age, 3140 years, 4150 years and above 50 years, respectively. Accuracy of MDRD formula was significantly high in individuals below 30 years of age. The accuracy of the CKDEPI formula remained above 50% in all the age groups, while the accuracy of CG formula was 14.3% in the age group above 50 years. Accuracy of CKDEPI was the highest (78.2%) while that of CG was the lowest (41.5%) and the accuracy of CKDEPI was statistically significant as compared with both MDRD and CG formulae ( P ≤0.05). Accuracy of MDRD formula was significantly higher in males than in females (57.9% vs. 33.3% P = 0.001), while there was no statistically significant difference (P >0.05) between the two genders with regards to accuracy of CG and CKDEPI formula. Furthermore, obesity had no effect on the accuracy of the GFR estimation formulas. Accuracy of measured and calculated GFR (eGFR) is presented in [Table 2]. The GFR was overestimated in 14.5% and 1.45% of the individuals using the CG and the CKDEPI formulas, respectively, while the MDRD formula did not overestimate GFR. The group of donors having GFR underestimated by CG formula was older as compared to those who had normal or overestimated GFR (36.9 ± 9.58 vs. 34.00 ± 9.04 years) and this difference was statistically significant (P = 0.03).{Table 2} Discussion All of our study participants were healthy adults who had come from all over the country to donate kidneys for their close relatives who had been affected by renal failure. Males predominated as potential live related donors and most were below 40 years of age. [9] In our study, the accuracy of the MDRD and CG formulas against 24h urine creatinine clearance was 48.3% and 41.5%, respectively. Both MDRD and CG formula underestimated the GFR, while CG formula overestimated it in 14.5% cases. Kaitwatcharachai [10] compared the performance of GFR estimation formulas (MDRD and CG) with 24h urine creatinine clearance in healthy Thai adults and found that both the MDRD and CG formulas underestimated the GFR in about 60% of healthy Thai adults. The results of our study closely resemble the results of the Thai study. Lin et al [11] evaluated 117 kidney donors and found out that the MDRD formula consistently underestimated GFR but was more accurate than the CG formula. Chaurasia et al [12] evaluated the performance of GFR estimation equations among 51 healthy donors and found that both the CG and MDRD formulas underestimated GFR. Shastry et al [13] also found estimation of eGFR in 431 healthy South Indian males by using the CG and MDRD. It is well known that the CG formula overestimate GFR, while the MDRD formula underestimate GFR. [11] However, 44% of our study subjects had underestimation of GFR by the CG formula; it has also been observed in some studies from India, and this may be because underweight subjects predominated the studies. [14] Jafar et al [15] reported low estimated GFR both by the CG and MDRD equations among migrant population in Karachi. In contrast to the findings of our study, Zubairi et al [16] who compared the performance of the MDRD and CG formula against 24 h urinary creatinine clearance in Pakistani healthy individuals and those with CKD and found that the MDRD and the CG formula overestimated GFR; the overestimation of GFR by the MDRD formula could be due to inclusion of patients with CKD. A recent crosssectional study from Karachi [17] compared GFR calculated by MDRD and CKDEPI (modified) formula with inulin clearance and concluded that CKDEPI creatinine equation was more accurate and precise than the MDRD equation in estimating GFR in Pakistani population and could be used for reporting estimated GFR. Our study also showed that CKDEPI formula had better accuracy than that of the MDRD. We did not use inulin clearance because of financial constraints and its limited availability. In our study, the highest accuracy of the MDRD formula (69.1%) was found in the youngest age group, i.e., ≤30 years old. With increasing age, the accuracy of MDRD formula declined because of greater degree of underestimation of GFR in older individuals. Michels et al [18] assessed the influence of age on the performance of MDRD formula and found out that the accuracy of the MDRD formula was the highest in youngest and oldest subgroups (82.6% for the youngest and 81.8% for the oldest group). This resembles our study findings. In contrast to our study findings, Cirillo et al [19] did not find any significant difference in the GFR estimation by the MDRD formula among different age groups (1888 years). The accuracy of CG formula was 39.7% and 43% in donors <30 years and 3150 years of age, respectively, and it dropped to 14% in those above 50 years of age. The accuracy of GFR decreased from 92.5% to 57% with the use of the CKDEPI formula. It is well known that GFR decline with increasing age. [20] The degree of decline is highly variable and probably reflects on the accuracy of the GFR estimation. In our study, the performance of the MDRD formula was found to be better in males than females. Michels et al [18] also had similar observations. Chaurasia et al [12] from Nepal had found higher estimated and measured GFR among female donors. But that study included comparatively younger females than males (mean age 40.03 ± 12.578 vs. 42.33 ± 11.39). Higher GFR among males than females is probably due to greater body surface area among men besides a possible effect of sex hormones. [21] We found in our study that obesity had no effect on the performance of the formulas in the calculations of eGFR. A study from Bangladesh [22] also concurs with our results and found that the MDRD formula did not overestimate GFR among obese subjects. Besides the MDRD and the CKDEPI formulas, cystatin C is emerging as useful marker for assessing the kidney function among healthy adults and potential kidney donors. [23] However, it is too early to recommend its routine use in kidney donors. GFR estimation equations have limitations and can be used with confidence only in populations in which they have been validated and cannot be used with confidence in those with normal renal function. [24] However, the CKDEPI formula performed much better than the CG and MDRD formulas among healthy Pakistani populations for reporting of estimated GFR. We conclude that the performance of GFR estimation formula was suboptimal in healthy Pakistani population and these underestimated and/or overestimated the GFR. These may misclassify the healthy individuals into chronic kidney disease population. The CKDEPI formula perform better than the MDRD and the CG formulas in healthy Pakistani individuals and none of these can be recommended for evaluation of renal function in potential kidney donors. Conflict of interest: None declared. References


