Chronic renal failure in diabetic patients increases lipid risk factors for atherosclerosis

1. Introduction 

The incidence of diabetes is rapidly increasing and already affects a high number of subjects in all the world, with an incidence expected to increase to over two hundred millions by 2010 [1]. Diabetes is a serious and costly disease with micro and macrovascular complications representing the leading causes of morbidity and mortality associated with the condition. Among these complications, diabetic nephropathy affects 40% of type 1 and 10% of type 2 diabetic patients. Moreover, diabetic subjects without or under inadequate treatment may develop renal failure in about 23 years evolution [2].

Patients with diabetes mellitus are at high risk of cardiovascular disease and the risk is further increased when they are complicated with nephropathy [3]. This may be partially attributed to changes in plasma lipids and lipoproteins which may undergo quantitative and qualitative modifications [4], [5]. These alterations tend to increase the atherogenic potential of some lipoproteins and to decrease the antiatherogenic capacity of others. Moreover, lipid and lipoprotein profile may be affected in different ways by the presence of incipient nephropathy, glomerulopathy, mild or severe renal failure, and sustitutive treatment. Nevertheless, the consequences of these renal affections on lipids and lipoproteins have been scarcely explored in diabetes.

Different studies have evaluated the main features of dyslipidemia in type 1 and type 2 diabetes with or without nephropathy in comparison with healthy individuals [6], [7], [8]. Other studies have investigated lipoprotein profile in patients with non-diabetic nephropathy also compared with healthy subjects [9], [10].

The aim of our study was to evaluate whether chronic renal failure (CRF) changes lipoprotein profile in diabetic subjects with glomerulopathy in comparison with non-diabetic subjects with glomerulopathy and CRF. The specific aim was to study the changes in lipoprotein concentrations, apolipoprotein (apo) A-I plasma and urinary levels, and the activity of one HDL-associated enzyme named paraoxonase (PON).

3. Results 

We studied the effect of the association between diabetes and nephropathy on the lipoprotein profile. The three groups of subjects were not different with respect to age, sex distribution, body mass index, and waist/hip ratio (Table 1). Diabetic patients with glomerulopathy and without CRF showed shorter duration of diabetes than subjects with glomerulopathy and CRF.

General biochemical parameters are shown in Table 2. As it was expected, glucose plasma levels were significantly higher in diabetic patients in comparison to non-diabetic subjects, while albumin plasma concentration did not show statistically significant differences. When comparing both diabetic groups, no differences were detected in glucose, HbA1c, or fructosamine levels. Urea and creatinine plasma concentrations, as well as proteinuria, were significantly increased, while glomerular filtration rate was significantly lower in patients with CRF than subjects without CRF.

Table 2. General biochemical parameters from diabetic patients with glomerulopathy and without chronic renal failure (CRF), diabetic patients with glomerulopathy and CRF, and non-diabetic patients with glomerulopathy and CRF (mean±S.D.)

		Diabetic patients with glome rulopathy and without CRF	Diabetic patients with glomerulopathy and CRF	Non-diabetic patients with glomerulopathy and CRF
	Glucose (mg/dl)a	229±81b	209±105b	85±18
	HbA1c (%)	9.6±0.9	9.1±2.2	–
	Fructosamine (mg/dl)a	421±61	418±101	–
	Albumin (mg/dl)a	3917±522	3559±952	2535±1283
	Urea (mg/dl)a	37±8c,d	75±36	79±47
	Creatinine (mg/dl)a	0.8±0.2e,d	1.8±1.0	1.6±0.9
	Glomerular filtration rate (ml/min)	95±14f,b	43±18	46±16
	Proteinuria (g/24h)	1.1±0.6g,d	4.1±2.9	8.2±6.8

a Measured in serum samples. b p<0.0005 vs. non-diabetic patients with glomerulopathy and CRF. c p<0.005 vs. diabetic patients with glomerulopathy and CRF. d p<0.05 vs. non-diabetic patients with glomerulopathy and CRF. e p<0.001 vs. diabetic patients with glomerulopathy and CRF. f p<0.0001 vs. diabetic patients with glomerulopathy and CRF. g p<0.05 vs. diabetic patients with glomerulopathy and CRF.

As regards lipid, lipoprotein and apolipoprotein levels, results are shown in Table 3. Both diabetic and non-diabetic patients with glomerulopathy and CRF showed hypercholesterolemia due to a significant increase in LDL-C levels in comparison to diabetic individuals without CRF. This difference between groups was also evidenced in non-HDL-C concentration. Nevertheless, no statistically significant difference was observed between both groups with CRF. Triglyceride, HDL-C, HDL2-C, HDL3-C, HDL-phospholipid, IDL-C, apo A-I and apo B levels were similar in the three groups. HDL-triglycerides were significantly reduced in diabetic patients with glomerulopathy and without CRF than in the other two groups.

Table 3. Lipid, lipoprotein, and apolipoprotein profile from diabetic patients with glomerulopathy and without chronic renal failure (CRF), diabetic patients with glomerulopathy and CRF, and non-diabetic patients with glomerulopathy and CRF (mean±S.D.)

		Diabetic patients with glomerulopathy and without CRF	Diabetic patients with glomerulopathy and CRF	Non-diabetic patients with glomerulopathy and CRF
	TG (mg/dl)	147±99	159±88	173±97
	TC (mg/dl)	187±39a,b	241±72	287±116
	HDL-C (mg/dl)	48±17	53±11	63±23
	HDL2-C (mg/dl)	6±5	8±6	13±12
	HDL3-C (mg/dl)	42±13	49±6	50±17
	HDL-TG (mg/dl)	14±9c,d	19±6	20±6
	HDL-PL (mg/dl)	69±22	76±29	101±59
	TC/HDL-C	4.2±1.2	4.7±1.5	5.8±1.9
	Non-HDL-C (mg/dl)	140±32c,b	187±70	235±103
	LDL-C (mg/dl)	110±25c,b	160±76	199±105
	IDL-C (mg/dl)	13±1	15±9	9±6
	Apo A-I (mg/dl)	138±30	155±27	154±64
	Apo B (mg/dl)	126±41	150±47	154±46

TG, triglycerides; TC, total cholesterol; HDL, high density lipoprotein; PL, phospholipids; LDL, low density lipoprotein; IDL, intermediate density lipoprotein; Apo, apolipoprotein.

a p<0.005 vs. diabetic patients with glomerulopathy and CRF. b p<0.005 vs. non-diabetic patients with glomerulopathy and CRF. c p<0.05 vs. diabetic patients with glomerulopathy and CRF. d p<0.05 vs. non-diabetic patients with glomerulopathy and CRF.

Arylesterase and paraoxonase activities, two ways of evaluating an antioxidant enzyme bound to HDL fraction, did not show any difference between the studied groups (Fig. 1). Paraoxonase phenotypic distribution was also similar in the three groups of patients.

View Large Image Download to PowerPoint

Fig. 1. Arylesterase (ARE) (Panel A) and paraoxonase (PON) (Panel B) activities in diabetic patients with glomerulopathy and without chronic renal failure (CRF), diabetic patients with glomerulopathy and CRF, and non-diabetic patients with glomerulopathy and CRF (mean±S.D.).

Apo A-I was also evaluated in urine samples from all the participants. This apolipoprotein could not be detected in urine from diabetic patients without CRF. All diabetic subjects with glomerulopathy and CRF who presented proteinuria above 6.5g/24h (n=6) showed detectable urinary apo A-I (range=13.1–61.0mg/24h). Similarly, all non-diabetic patients with glomerulopathy and CRF who had proteinuria above 8.0g/24h (n=5) also evidenced detectable apo A-I in urine samples (range=25.6–557.3mg/24h). When the correlation coefficients were evaluated between urinary apo A-I and different parameters for the whole group (Table 4), positive and significant correlations were detected for urea and proteinuria, and a negative and significant correlation for albumin plasma levels.

Table 4. Correlation coefficients of urinary apo A-l with different parameters

		Urinary apo A-l
		r	p
	Albumina	−0.68	<0.05
	Ureaa	0.73	<0.05
	Creatininea	0.09	NS
	Glomerular filtration rate	0.24	NS
	Proteinuria	0.97	<0.0001

Apo, apolipoprotein; NS, not significant (p>0.05).

a Measured in serum samples.

4. Discussion 

Diabetes condition did not amplify the lipid risk factors associated with CRF. In this study, CRF patients with or without diabetes exhibited similar lipid and lipoprotein patterns, which were more atherogenic than the profile observed in diabetic patients with glomerulopathy but without CRF. This finding would highlight the fact that CRF represents an important atherogenic risk factor, beyond the presence of diabetes. Accordingly, Alebiosu et al. [19] found a high prevalence of major cardiovascular risk factors among Nigerian subjects with clinical diabetic nephropathy in comparison with type 2 diabetic subjects without nephropathy.

Diabetic patients with glomerulopathy and without CRF showed LDL-C and non-HDL-C levels slightly higher than the recommended values for diabetic subjects [20]. On the other hand, both groups of patients with CRF presented significantly higher total cholesterol, LDL-C and non-HDL-C than diabetic patients without CRF. No statistically significant differences were detected in IDL-C levels among the three groups of patients. In contrast, Kimoto et al. [21] found increased IDL-C concentration in patients with end-stage renal disease which was more pronounced in the presence of diabetes. This adverse effect was accounted for, at least partly, by hypertriglyceridemia associated with chronic hyperglycemia. It must be noted that in our study, triglyceride levels were similar in the three groups. Therefore, the main underlying disturbance in both diabetic and non-diabetic patients with nephropathy from the present study is the accumulation of atherogenic apo B-containing lipoproteins [22].

When we compared both groups with CRF, no differences were detected in the different parameters evaluated. Gonzalez et al. [23] have also shown that lipid and lipoprotein levels were similar in diabetic and non-diabetic patients, both of them with end-stage hemodialysed renal disease. Nevertheless, in those groups of subjects, hemodialysis seemed to improve the lipoprotein profile. Similarly, Joven et al. [24] studied patients with nephrotic syndrome grouped according to the absence or presence of renal failure and/or diabetes mellitus and they found that diabetes mellitus did not affect the pattern of hyperlipoproteinemia of nephrotic syndrome while the characteristic lipoprotein and apoprotein pattern of uremia was present irrespective of nephrosis in uremic, non-diabetic patients. On the other hand, Attman et al. [5] found that diabetic patients with nephropathy presented more pronounced dyslipidemia than non-diabetic subjects with nephropathy. However, those patients had proteinuria levels considerably lower than the subjects evaluated in our study.

Even if plasma levels of HDL-C and HDL-phospholipids showed no statistically significant differences among the studied subjects, a tendency towards higher values may be observed in both groups with CRF. It has been shown that the decrease in plasma colloid osmotic pressure, consequence of the reduction in total protein plasma concentration and hypoalbuminemia, was responsible for the increase in apo A-I and HDL synthesis by the liver and that this synthesis was transcriptionally regulated [25]. This process could be counteracting apo A-I loss via urine in patients with massive proteinuria. Furthermore, Jüngst et al. [26] suggested that the amount of HDL excreted in 24h urine from patients with nephrotic syndrome was too small to significantly decrease HDL plasma concentration. Moreover, HDL-triglycerides showed a significant increase in both groups with CRF in comparison with diabetic subjects without CRF. In previous studies, we have demonstrated that HDL enrichment in triglycerides reduces its capacity to promote reverse cholesterol transport, the antiatherogenic pathway by which excess cholesterol is transported from peripheral tissues to the liver for excretion [27]. HDL is also able to protect low density lipoprotein (LDL) against in vitro oxidation [28] and the consequent monocyte-endothelial cell interaction [29]. HDL antioxidant potential can be assigned to different factors such as its chemical composition, the content of liposoluble antioxidants and the presence of associated enzymes like paraoxonase (PON) 1 (EC 3.1.1.2) and platelet-activating factor acethylhydrolase (PAF-AH; EC 3.1.1.47) [30]. In a previous study, we have shown that in type 2 diabetic patients, PON1 activity was similar than in control subjects [31]. In the present work, PON1 activity, evaluated with two different substrates, did not differ among the studied subjects, which would evidence that CRF development did not worsen HDL antioxidant capacity. In contrast, Ikeda et al. [32] showed that PON1 activity was different between type 2 diabetic patients with and without overt proteinuria.

Given that measurement of urinary apo A-I has been shown to reflect the excretion of HDL particles in urine [33], we evaluated this parameter in all the subjects. Only patients with CRF and massive proteinuria eliminated detectable amounts of apo A-I via urine. Accordingly, urinary apo A-I was positively associated with plasma urea and proteinuria, and negatively related to plasma albumin. HDL apolipoproteins may be excreted as small HDL particles or possibly filtered as individual apolipoproteins that recombine to form HDL particles [33]. Horowitz et al. [34] demonstrated that HDL enrichment in triglycerides and further action of lipases increases apo A-I fractional catabolic rate. Accordingly, patients with CRF and detectable urinary apo A-I also had higher HDL-triglycerides. Thus, in these individuals a higher proportion of apo A-I would be present in a more disociable form which may be easily excreted by the kidney.

In conclusion, the presence of CRF in diabetic patients was associated with a more atherogenic lipid and lipoprotein profile consistent of higher total cholesterol, LDL-C, non-HDL-C, and HDL-triglyceride levels, the latter being related with an increase in urinary apo A-I excretion. These results highlight the relevance of preventing CRF development in diabetic patients which clearly increases the risk of atherosclerotic cardiovascular disease.