Comparison of three different creatinine clearance calculation methods in patients with type 2 diabetes mellitus(tip 2 diyabetik bireylerde kreatinin klirensini hesaplamada kullanılan üç farklı yöntemin karşılaştırılması)

Abstract

Aim: To determine the most accurate and useful method for calculating creatinine clearance by comparing the results of different methods.

Methods: One hundred type 2 diabetic patients who have been followed by Okmeydani Training and Research Hospital internal medicine and/or diabetes outpatient clinics were included in this study. Individuals with hypertension, acute kidney disease and renal transplantation were excluded from the study.

Results: Glomerular filtration rate (GFR) calculated with Cockcroft-Gault formula was significantly affected by creatinine, weight,and age (p < 0,050). GFR measured with Modification of Diet in Renal Disease (MDRD) formula was significantly affected by creatinine and age (p < 0,050) in a univariate model; in a multivariate model, this was significantly independently affected by creatinine (p < 0,050). GFR measured with 24h urine was significantly affected by creatinine, weight,and age (p < 0,050) in a univariate model; in a multivariate model, this was significantly independently affected by weight (p < 0,050).

Conclusion: In this study, those three methods revealed similar results. All of three methods can be used for evaluating renal functions in Type II diabetic patients but creatinine clearance with 24 hours urine method requires two patient visits in a row and a more complex biochemistry laboratory; so in our opinion, this method may be used as an alternative to the other two methods.

Keywords: Creatinine clearance, Glomerular filtration rate, Type 2 Diabetes

Öz

Amaç: Diyabetik bireyler için kullanılabilecek en uygun kreatinin klirensi hesaplama metodunu belirlemek amaçlandı.

Yöntem: Çalışmaya Okmeydanı Eğitim Araştırma Hastanesi iç hastalıkları ve diyabet polikliniklerine başvurmuş 100 tip 2 diyabetik hasta dahil edildi. Hipertansiyon, akut böbrek yetersizliği tanısı almış veya böbrek nakil alıcısı olan diyabetik hastalar çalışma dışı bırakıldı.

Bulgular: Cockcroft-Gault değerini kestirmede tek ve çok değişkenli modellerde yaş, ağırlık, kreatininin anlamlı (p < 0,050) etkisi gözlenmiştir. MDRD değerini kestirmede tek değişkenli modelde yaş, kreatininin; çok değişkenli modelde ise yalnızca kreatininin anlamlı bağımsız (p < 0,050) etkisi gözlenmiştir.24 saatlik idrarda kreatinin klirensi değerini kestirmede tek değişkenli modelde yaş, ağırlık, kreatinin değerinin anlamlı (p < 0,050) etkisi gözlenmişken; çok değişkenli modelde ise yalnızca ağırlık değerinin anlamlı bağımsız (p < 0,050) etkisi gözlenmiştir.

Sonuç: Bu çalışmada üç yöntem de birbirleriyle uyumlu sonuç verdi. Tip II diyabetik hastalarda böbrek fonksiyonlarını değerlendirmek için üç yöntemin tamamı kullanılabilir, ancak 24 saatlik idrar yöntemiyle kreatinin klirensi, üst üste iki hasta ziyareti ve daha karmaşık bir biyokimya laboratuvarı gerektirir; bizim görüşümüze göre, bu yöntem diğer iki yönteme alternatif olarak kullanılabilir.

Anahtar kelimeler: Glomeruler fitrasyon hızı, Kreatinin klirensi, Tip 2 Diyabet

Introduction

Diabetes mellitus (DM) is a chronical and progressive disease. Approximately 150 million people are suffering from this disease and predicted the number for 2025 is 300 million [1,2].

Morbidity and mortality due to DM and its complications are increasing as the prevalence of type II DM increases [3]. Consequently, early diagnosis and effective treatment of type II DM is needed more and more every day. There are approximately 2.6 million type II DM patients in our country, and it is predicted that at least one-third of 1.8 million people still in impaired glucose tolerance stage will join to this group in the near future [4].

Diabetic nephropathy (DN) is a serious health problem causing end-stage renal failure. In the United States of America, DN causes 40 % of newly developed end-stage renal failure. DN defined as positive urine albumin stick test or excretion of albumin more than 300 mg in a diabetic patient who is not suffering from other renal diseases. DN, as appears a late finding of diabetes has some physiological, pathological and clinical symptoms. That made some researchers consider DN into stages [5].

Creatinine clearance measurement is the most common method for evaluating renal functions. Creatinine clearance may be measured with 24-hour urine collection and also with Cockcroft-Gault formula and MDRD.

In this study, we aimed to to determine the most accurate and useful method for calculating creatinine clearance by comparing the results of different methods. It was aimed to improve feasibility by determining the most suitable method to be possible.

Materials and Methods

This retrospective study approved by Okmeydani Training and Research Hospital (OEAH) Clinical Researchs Ethical Board Presidency with a number of 178 at 09.09.2014.  Files of type 2 DM patients who applied to one of internal medicine outpatient clinics between 2012 and 2014 were retrospectively screened. From a total of 184 patients; patients with hypertension (n=74), acute renal failure (n=6) or renal transplantion (n=4) were excluded from the study. The remaining 100 patients (56 female, 44 male) included to the case group. Median age of the patients was 56 years with a range from 20 to 82 years.

Patients’ characteristics (age, gender and weight (kilograms)) and laboratory findings (serum creatinine level (mg/dl), fasting blood glucose (mg/dl), postprandial blood glucose (mg/dl), HbA1c (%) and 24-hour urine creatinine clearance (GFR24) (mg/24 hours)) were evaluated. Roche-Hitachi Cobas 8000 (Serial number: 1349-09, 2014,Japan) was used to evaluate laboratory findings. The prediction of creatinine clearance (in ml/min) by the Cockcroft-Gault formula (GFRC&G) was calculated as (140 − age) × body weight/plasma creatinine × 72 (× 0.85 if female) [6]. The abbreviated MDRD (GFRMDRD) estimate of the kidney function was calculated as 175 × plasma creatinine−1.154 × age−0.203 (× 0.742 if female) [7]. Grading of the patients with regard to renal failure were performed according to the KDIGO 2017 guideline using GFR values (G1-G5) (Table 1) [8].

IBM SPSS for Windows 21.0 (Armonk, New York, USA) statistics package program was used for analyzes. Mean, median, minimum, maximum, frequency values and standard deviation were used for defining statistics of data. Distribution of the variables was controlled with Kolmogorov Simirnov test. Unpaired t-test and Mann-Whitney U test were used for quantitative data analysis. Chi-square test was used for qualitative data analysis. Spearman correlation test was used for correlation analysis. Univariate and multivariate regression tests were performed. Level of significance determined as p≤0.050 for all analyzes.

Results

A total of 100 patients were staged by GFR. Sixty-nine patients (69%) had GFR greater than 90 mL/min. staged as G1, 22 patients (22%) had GFR between 60-89 mL/min staged as G2  and 9 patients (9%) had GFR between 30-59 mL/min staged as G3. None of the patients staged as G4 and G5.

Creatinine clearance of the patients was calculated by Cockcroft-Gault formula (GFRC&G), Modification of Diet in Renal Disease (GFRMDRD) and 24h urine collection method (GFR24).

Mean ± Standart Deviation values of these three methods were; GFR24:96.4 ± 28.8 mL/min, GFRC&G:104.5± 29.8 mL/min, GFRMDRD:86.2± 24.7 mL/min .

Table 2 shows statistical values of patients’ in terms of gender, weight, fasting blood glucose, postprandial blood glucose, HbA1c, GFR24, GFRC&G and GFRMDRD (Table 2).

Significant (p < 0,050) negative correlation was observed between creatinine levels and GFRC&G, GFRMDRD, GFR24. Significant (p < 0,050) positive correlation was observed betweenGFRC&G, GFRMDRD and GFR24. Significant (p < 0,050) positive correlation revealed between fasting blood glucose, postprandial blood glucose and HbA1c. (Table3)

In both univariate and multivariate models age, weight, and creatinine had significant (p < 0,050) association on determining GFRC&G value (Table 4). Although in a univariate model age and creatinine had significant (p < 0,050) association on determining GFRMDRD value; in a multivariate model only creatinine had independently significant (p < 0,050) association (Table 5). Although in a univariate model age, weight and creatinine had significant (p < 0,050) association on determining GFR24 value; in a multivariate model only weight had independently significant (p < 0,050) association on determining GFR24 value (Table 6).

Discussion

The incidence of DN is increasing in proportion to DM incidence and increased lifetime in diabetics. Our study showed that 73% of patients had GFR under 120 mL/min. However, in our study, there was no significant correlation between fasting blood glucose, postprandial blood glucose and HbA1c and GFR values ​​measured by three different methods.

This study compared the most popular three methods for calculating creatinine clearance. One of those methods, Cockcroft & Gault formulation uses serum creatinine, age, weight, and gender to calculate creatinine clearance by the unit ml/min [6].The second one is MDRD formulation uses race, age, serum creatinine and gender [7]. The last method is to evaluate the creatinine level in urine patient collected for 24 hours without interruption.

A study compared Cockcroft & Gault formulation, and MDRD formulation suggested that Cockcroft & Gault formulation calculated the lowest creatinine clearance in patients above age 70; while MDRD formulation is the most valuable method to estimate mortality rate in patients above age 85 [9]. In this study, the median age was under 70. GFRMDRD was slightly lower than GFRC&G without statistical signification. Yet another study published in 2007 suggested that Cockcroft&Gault formulation achieved more accurate results than other methods [10]. Another study published in 2010 suggested that Cockcroft&Gault formulation is superior to the MDRD formulation in patients with normal creatinine clearance and diabetics with normal or close to normal GFR. Otherwise, MDRD formulation had more accurate results [11].

Our study revealed serum creatinine levels are increasing with age. Yet increased age resulted with lower GFRC&G, GFRMDRD, and GFR24h. Those results pointed out that age may be a prognostic factor for diabetic nephropathy. A study published at 2002 including 98.688 patients age between 20 and 94 years showed progressively increasing serum creatinine levels in male patients from age 60 and female patients from age 40 [12]; results of our study are consistent with this study.

In our study, independent factors that significantly affected GFRC&G increase are age, weight and serum creatinine. This result was expected as they all are variables in the Cockroft-Gault formulation. This result is consistent with the findings of two other studies. [13, 14]; and being in association with weight, is seemed to be the weakness of Cockroft-Gault formulation. Because of this deficit, another study recommended of estimating a CrCl range with the lower boundary defined by using ideal body weight in the Cockcroft-Gault equation and the upper boundary by using total body weight [15].

Independent factors significantly affected GFRMDRD increase are age and serum creatinine. This is consistent with the previous studies [13, 16]. This result was expected as they are also variables in the MDRD formulation. It is not surprising that there is no effect of weight on the GFRMDRD since the MDRD formula does not use weight.

Independent factors significantly affected GFR24h increase are age, weight, and serum creatinine. As creatinine is released from the muscles and muscles are the big part of our weight; weight should be considered normal to affect the GFR24h.

An increase in GFR24h had a positive correlation with GFRMDRD and GFRC&G. This result indicated these three methods are consistent among themselves.

Major limitations of this study are being retrospective and the small sample size: Because of the retrospective design of the study some important clinical features could not be recorded. The small sample size may have limited our ability to detect statistically significant results.

In conclusion, there was no statistically significant difference between Cockcroft-Gault formulation, MDRD formulation and creatinine clearance with 24 hours urine method; they are all equally useful in clinical practice. So all of three methods can be used for evaluating renal functions in Type II diabetic patients but creatinine clearance with 24 hours urine method requires two patient visits in a row and a more complex biochemistry laboratory, and it might give incorrect results because of lack of communication between physician-patient-laboratory triangles especially in an outpatient clinic. In our opinion, this method may remain in the background as a result of the process.