Predictors of Glomerular Filtration Rate Decline in Type 2 Diabetic Patients: Two-Years Follow-up Study

Research Article

J Dis Markers. 2015;2(3): 1029.

Predictors of Glomerular Filtration Rate Decline in Type 2 Diabetic Patients: Two-Years Follow-up Study

Čabarkapa V1,2*, Stošić Z1,2, Đerić M1,2, Bećarević M¹, Radosavkić I² and Eremić Kojić N1,2

¹Medical Faculty, University of Novi Sad, Serbia

²Clinical Centre of Vojvodina, Center of Laboratory Medicine, Novi Sad, Serbia

*Corresponding author: Čabarkapa V, Center of Laboratory Medicine, Clinical Centre of Vojvodina, Serbia

Received: July 07, 2015; Accepted: August 07, 2015; Published: August 10, 2015

Abstract

Background: Diabetic nephropathy is one of the leading causes of chronic renal failure. The aim was to investigate the importance of specific biomarkers and clinical features in prediction of glomerular filtration rate (GFR) decline in type 2 diabetic patients during two-year follow-up.

Methods: Patients (n = 113) were divided into the following groups: I-41 with GFR reduction >20% (in relation to the reference value for given sex and age) and urinary albumin excretion (UAE) >30 mg/day; II- 34 with GFR reduction ≤20% and UAE >30 mg/day, and III-38 with GFR reduction ≤20% and UAE ≤30 mg/day. The control group included 30 healthy subjects. We determined albuminuria (sandwich-immunometric method); proteinuria (pirogalol red); C reactive protein, apolipoprotein A-I and B, lipoprotein (a), cystatin C (immunoturbidimetry); homocysteine (FPIA); fibrinogen (Clauss); oxidized LDL (ELISA); lipid parameters, creatinine, urea and uric acid (standard biochemical methods). GFR was estimated via creatinine clearance. We also evaluated the presence of chronic complications of diabetes.

Results: GFR reduction >10% compared to baseline values was more frequent in group I (53.13%), than in group II (36.67%) and III (23.14%). Regression analysis revealed that proteinuria >1 g/day increases the risk of progression of renal dysfunction fourfold, homocysteinemia >10 μmol/L by 3.42, systolic blood pressure (SBP) >130 mmHg by 5.08, hemoglobin <130 g/L by 3.05 and the presence of macrovascular complications (MC) by 6.25 times.

Conclusion: Homocysteinemia >10 μmol/L, presence of macrovascular complications, hemoglobin <130 g/L, and SBP >130 mmHg were independent predictors of progression of renal dysfunction in patients with DM type 2.

Keywords: Diabetic Nephropathy; Type 2 Diabetes Mellitus; Glomerular Filtration Rate; Biochemical Markers; Macroangiopathy

Introduction

Diabetic nephropathy (DN) implies a wide range of renal dysfunctions that involve the development of microalbuminuria (MiA), proteinuria and progressive reduction in renal functional reserve (RFR) [1]. Besides hypertensive nephropathy and glomerulonephritis, DN is a leading cause of chronic renal failure and end-stage renal disease [2,3,4].

Mortality in DN patients is 5-8 times higher than in the general population, and it is influenced considerably by the high cardiovascular mortality that increases with progression of renal dysfunction [5].

Diabetic nephropathy is a consequence of interaction between a range of hemodynamic and metabolic factors (systemic and intraglomerular hypertension, activation of vasoactive substances, activation of alternative metabolic pathways with the formation of polyols and advanced glycation end products with increased oxidative stress). Combined action of these factors leads to increased renal albumin permeability as well as to an accumulation of extracellular matrix, which result in proteinuria, glomerulosclerosis and tubulointerstitial fibrosis [6,7].

However, there is significant individual variation in the rate of reduction in RFR, i.e., reduction in glomerular filtration rate (GFR), in patients with DN [8,9,10]. It is worrisome that despite all the measures currently in use in the treatment of subjects suffering from diabetes, a significant number of patients still experience a progressive decline in GFR [11]. This finding indicates the necessity of determining risk factors for progression of renal dysfunction, especially in the early stages of nephropathy.

Therefore, in this two-year follow-up study we examined the importance of specific biomarkers (i.e. total homocysteine (tHcy), oxidized LDL (oxLDL), C-reactive protein (CRP), fibrinogen, lipid status parameters, apolipoprotein (apo) A-I, apoB, lipoprotein(a) (Lp(a)), proteinuria, albuminuria, cystatin C, hemoglobin and the clinical features (the presence of chronic vascular complications of diabetes) in prediction of GFR decline in patients with diabetes mellitus type 2 with varying degrees of RFR reduction.

Materials and Methods

Study population

This two-year follow-up study was carried out in the Clinical Centre of Vojvodina, Novi Sad, Serbia. The study had previously been approved by an institutional ethics committee and all included subjects had approved their participation. The study was performed according to the principles of the Declaration of Helsinki.

The study included 113 subjects Caucasians with diabetes mellitus type 2, with no urinary tract disease or non-diabetic renal disease. Patients were divided into the three groups according to GFR and urinary albumin excretion (UAE): group I 41 subjects (29 men, 12 women) with UAE over 30 mg/day and GFR reduction ≥20% compared with the reference values for given sex and age; group II 34 subjects (20 men and 14 women) with UAE over 30 mg/day and GFR reduction ≤20% compared with the reference values for given sex and age; group III 38 subjects (17 men and 21 women) with and GFR reduction ≤20% compared with the reference values for given sex and age and UAE ≤ 30mg/day.

GFR was estimated through the calculation of creatinine clearance (CrCl), and the degree of reduction was assessed in relation to the reference value for a given sex and age.

The control group included 30 clinically healthy subjects (15 men and 15 women) matched for age, with serum glucose concentrations below 6.1 mmol/L, hemoglobin A1c (HbA1c) below 6.2%, CrCl within the normal range for given sex and age, UAE ≤30 mg/day and proteinuria ≤150 mg/day.

In the control group, blood samples were taken only once, while diabetic patients samples were taken three times: at the beginning of the study, after 12 months, and after 24 months. In all subjects blood was drawn after 12-hour fasting. We just followed the GFR during these 24 months, and other parameters have been doing at the beginning and the basis of the fall of GFR calculated their predictive value.

Analyses were performed immediately after sampling except for oxLDL (samples were kept frozen at -200C no longer than one month). In addition, 24-hour urine collection was performed in all subjects after the previously given instructions.

Smokers, chronic alcohol consumers and patients with acute infection, thyroid dysfunction, liver disease or malignancies were excluded from the study.

A relative reduction in GFR (rrGFR) was calculated as a ratio between the difference between GFR at baseline and GFR at the end (GFRb-GFRe) and GFRb. The value of 10% was an average of the reduction in GFR in all studied diabetic subjects. Subjects who after two-years follow-up had rrGFR of more than 10% were classified as progressors and those with rrGFR ≤10% as non-progressors.

Measurement of renal function parameters

Serum concentrations of creatinine, urea and uric acid were determined by standard biochemical methods (commercial kits- Beckman Coulter, Ireland on Olympus AU 400); cystatin C by the immunoturbidimetric method (commercial Dyazime kits, USA, on Olympus AU 400, reference range: 0.5-1.03 mg/L); albuminuria by the sandwich-immunometric method (commercial Nyco Card kits, Norvage, reference range ≤30 mg/day); proteinuria by pyrogalol red method (commercial Siemens kits, USA, on Advia 1800, reference range: ≤150 mg/day). Creatinine clearance (expressed as mL/ min/1.73 m2 of body surface area (BSA)) was calculated using the same 24-h urine used for proteinuria and albuminuria:

CrCl = (UCr x 24h urine volume (ml))/(SCr x 1440 (min/day))

UCr- urine creatinine (μmol/L), SCr– serum creatinine (μmol/L)

BSA was calculated by the formula [12]: BSA (m²) = 0.0235 x height (cm)0.42246 x weight (kg)0.51456.

Measurement of biochemical markers

Plasma tHcy concentration was determined by the fluorescence polarization immunoassay (FPIA) with commercial Abbott kits, USA, on AxSym analyzer (reference range: 5-12 μmol/L); lipid profile was measured by standard biochemical methods (commercial Siemens kits, USA, on Olympus AU 400); apo A-I and apoB (Beckman Coulter, Ireland), and Lp(a) serum concentrations (Sentinel, Italy) were determined by the immunoturbidimetric method on Olympus AU 400; oxLDL concentration was measured by enzyme-linked immunosorbent assay (ELISA) method (commercial Mercodia kits, Sweden); fibrinogen concentration were measured in citrate plasma by Claus method [13] on ACL system (Instrumentation Laboratory- Italy; reference range: 2.2-4.96 g/L); CRP levels were determined by immunoturbidimetric method (Beckman Coulter, Ireland, reference range 0-5 mg/L).

Assessment of glycemic control parameters

Serum concentration of glucose (hexokinase method, reference range 3.9-6.1 mmol/L) and glycated haemoglobin measured by immuno-inhibitory test as HbA1c (commercial Beckman-Coulter kits-Ireland, reference range <6.2%), were determined on Olympus AU 400.

Assessment of chronic complications of diabetes

Diabetic retinopathy was diagnosed after ophthalmological examination (standard fundus eye examination with presence of micro aneurysms, neovascularization, venous dilatation, cotton-wools spots or bleeding). Diabetic neuropathy was assessed by neurologic examination. Myocardial infarction (MI) was confirmed by a positive history of disease. Cerebrovascular disease was confirmed by cranial computerized tomography or magnetic resonance. Peripheral artery disease (PAD) was confirmed by vascular surgeon. Blood pressure values were determined with the patient sitting in an upright position after 10 minutes of rest.

Statistical analysis

Data was presented using descriptive statistical methods, such as mean values, standard deviation, median, percentage. Agreement with a normal distribution of data was tested by Kolmogorov Smirnov (n >50 by group) and Shapiro-Wilk’s test (n <50 by group).

Parametric (t-test, ANOVA) and non-parametric (Mann- Whitney, χ2 –test, Kruskal-Wallis test) statistical tests were used.

Univariate and multivariate logistic regression analyses were used to determine associations between a rrGFR of over 10% with baseline variables. A ROC curve and area under ROC curve were used for determining quality of the model obtained. P <0.05 was considered statistically significant.

Statistical analysis was performed using the Statistica 12 (Stat Soft Inc., Tulsa, OK, USA) software

Results

General characteristics and baseline laboratory parameters of all studied subjects are shown in Tables 1-3.