Non-albuminuric renal disease among subjects with advanced stages of chronic kidney failure related to type 2 diabetes mellitus

Abstract

Urinary albumin excretion has been consistently found to be normal in a significant number of subjects with early stages of diabetic kidney disease. This study was aimed to estimate the prevalence and characteristics of non-albuminuric chronic kidney disease associated with type 2 diabetes mellitus among subjects who reach advanced stages of renal failure. Study population was composed of incident patients with advanced chronic kidney disease (glomerular filtration rate <30 mL/min) related to type 2 diabetes in a tertiary hospital from Gran Canaria (Spain) during a period of 2 years. Subjects were classified as normoalbuminuric (urinary albumin-to-creatine ratio [UACR] <30 mg/g), microalbuminuric (UACR ≥30 and <300 mg/g), or proteinuric (UACR ≥300 mg/g). Of 78 eligible patients, 21.8% had normoalbuminuria, 20.5% had microalbuminuria, and 57.7% had proteinuria. Individuals with normoalbuminuria were mostly women and had a lower prevalence of smoking and polyneuropathy than subjects with microalbuminuria or proteinuria. They also presented greater measures of body mass index and waist circumference, higher values of total and LDL cholesterol, and lower values of HbA_1c and serum creatinine than subjects with microalbuminuria or proteinuria. Multivariate analysis demonstrated that female sex (positively) and HbA_1c and polyneuropathy (negatively) were independently associated with absence of albuminuria. In conclusion, around 20% of subjects with diabetes-related advanced chronic kidney disease, characteristically women, have normal urinary albumin excretion. HbA_1c and polyneuropathy are inversely related to this non-albuminuric form of nephropathy.

• Diabetic kidney disease
• diabetic nephropathy
• microalbuminuria
• proteinuria
• type 2 diabetes mellitus

Introduction

Proteinuria has been generally considered a hallmark of diabetic nephropathy. According to the model proposed by Mogensen et al. in the early eighties,1 increases in urinary excretion of albumin precede and accompany the decline of renal function of patients with diabetes-related kidney disease. However, this classic concept has been challenged by several investigations published during the last decade, reporting that a considerable number of patients with chronic kidney disease and diabetes, especially of type 2, present normal rates of urinary albumin.2–10 In the majority of these studies, renal impairment was defined as a glomerular filtration rate (GFR) less than 60 mL/min/1.73 m², with most patients having GFR between 30 and 60 mL/min/1.73 m². Because of urinary protein excretion is strongly related to the risk for renal disease progression,11 it remains possible that the prevalence of normoalbuminuric nephropathy diminishes in patients with more advanced kidney failure, but information on this kind of patients is scarce.

The present investigation is aimed to estimate the prevalence and characteristics of the non-albuminuric form of diabetic kidney disease in a population of subjects with type 2 diabetes and advanced stages of renal failure from the Gran Canaria Island, an area with a high incidence and prevalence of diabetes-related end-stage renal disease.12

Material and methods

Study subjects

Since February 2011 new (incident) patients attending the Advanced Chronic Kidney Disease Office of the Hospital Universitario Insular are invited to participate in the CERCA-Diabetes Study, a prospective observational survey directed to characterize diabetes-related advanced chronic kidney disease in the South area of the Gran Canaria Island. According to our usual practice, patients are referred to this medical facility when they present chronic kidney disease stages 4 or 5, as defined by the National Kidney Foundation’s Kidney Disease Outcomes and Quality Improvement Initiative (KDOQI).13 In all cases, an etiologic diagnosis for kidney disease is assigned by consensus between two nephrologists (C.G.-C. and F.B.). Up to February 2013, 107 of the 170 participants in the CERCA Study had a diagnosis of type 2 diabetes. Twenty-two of them were excluded for the present investigation because kidney disease was attributed to other conditions apart from diabetes. Other seven patients were excluded because a measurement of urinary albumin-to-creatine ratio (UACR) was not available. The final sample was composed of 78 patients. The study has been approved by the Ethical Committee of the Hospital Universitario Insular, and all subjects signed informed consent before participation.

Survey procedure

After an overnight fasting, participants provided a urine specimen and underwent blood extraction. They fulfilled a survey questionnaire covering sociodemographic information, lifestyle behaviors and personal and familial medical history, including current use of drugs. Subjects were classified as current smokers or non-smokers, the latter including past smokers. Physical activity was measured by asking about work-related and leisure time activities, using the questions designed for the 2006 Spanish National Health Survey.14 Subsequently, weight, height, and waist circumference were measured. Blood pressure was taken twice, in a sitting position, using a manual sphygmomanometer and the mean of the two readings was recorded for analysis.

Laboratory investigations

Hemoglobin was measured using a LH750 automated processor (Beckman Coulter, Brea, CA). Urinary albumin and creatinine, used to calculate UACR, as well as serum levels of creatinine, glucose, sodium, potassium, calcium, phosphate, albumin, total cholesterol, high-density lipoprotein (HDL) cholesterol, apolipoprotein B (apoB), triglycerides, uric acid, liver enzymes, creatine kinase and C-reactive protein were determined using a AU2700 platform (Beckman Coulter, Brea, CA). ApoB determinations were standardized to the WHO-IFCC International Reference Materials for apoB (SP3-07). Low-density lipoprotein (LDL) cholesterol was calculated by the Friedewald formula and non-HDL cholesterol was calculated as the difference between total cholesterol and HDL cholesterol. GFR was calculated using both the MDRD4 (Modification of Diet of Renal Disease) formula: 186 × (creatinine)^−1.154× (age)^−0.203× (0.7402 if female)15 and the Cockcroft--Gault formula: [(140 − age) × weight × (0.85 if female)]/(72 × creatinine).16 Glycated hemoglobin (HbA_1c) was determined using high performance liquid chromatography with an HA-8140 analyzer (Menarini Diagnostics-Arkray, Kyoto, Japan). Chemiluminescent immunoassays were used for measure of PTH (UniCel DxI 800 Access Immunoassay System, Beckman Coulter, Brea, CA) and 25OHD3 (Liaison 25, DiaSorin, Italy).

Statistical analyses

For this study, each subject was described as being without abnormal urine protein excretion (i.e., no microalbuminuria or proteinuria) or as having microalbuminuria or proteinuria. Microalbuminuria was defined as an UACR ≥30 and <300 mg/g. Proteinuria was defined as an UACR ≥300 mg/g. Physical activity was recorded into two categories: active (those who referred to regularly practice any physical activity during leisure time and those who spent more work time walking than sitting or standing) and sedentary (the rest). The existence of established cardiovascular disease and chronic microvascular complications of diabetes was settled on the basis of available medical charts and information obtained from the survey questionnaire. For each study group, frequencies were compared using chi-square or Fisher’s exact test, and medians were compared with the Wilcoxon’s signed rank test. Variables significant at p value <0.1 level were included in a multiple logistic regression analysis in order to investigate independent effects of factors associated with normoalbuminuria. The best subset selection, using complete enumeration algorithm and Akaike’s Information Criteria, was used for selection of variables.

Results

A total of 17 subjects (21.8%) had normoalbuminuria, 16 (20.5%) had microalbuminuria, and 45 (57.7%) had proteinuria. Table 1 provides information on demographic data, clinical profile and biochemical studies of groups with normoalbuminuria and either microalbuminuria or proteinuria. Individuals with normoalbuminuria were predominantly women, none of them was smoker and had a lower prevalence of somatic polyneuropathy than subjects with microalbuminuria or proteinuria. Additionally, they presented greater measures of body mass index and waist circumference, higher values of total cholesterol and LDL cholesterol, and lower values of HbA_1c, and serum creatinine than subjects with microalbuminuria or proteinuria. Multivariate logistic regression analysis demonstrated that female sex (positively) and HbA_1c and diabetic polyneuropathy (negatively) were independently associated with the absence of albuminuria (Table 2).

Table 1. Clinical and biochemical characteristics of the studied sample according to levels of urinary albumin to creatinine ratio.

	Total	Normoalbuminuria	Microalbuminuria/proteinuria
	n = 78	n = 17	n = 61	p
Sex (women) (%)	43.6	77.5	34.4	0.002
Age (years)	72.0 (63.0, 77.0)	73.0 (63.0, 77.0)	72.0 (63.0, 77.0)	n.s.
BMI (kg/m^2)	30.7 (28.2, 34.3)	33.7 (31.6, 37.5)	30.3 (27.4, 32.9)	0.005
Waist (cm)	110 (101, 119)	117 (109, 120)	109 (99, 117)	0.037
Diabetes duration (years)	17 (10, 24)	21 (15, 27)	17 (10, 22)	n.s.
Current smoking (%)	17.9	0	23.0	0.032
Sedentary activity (%)	39.7	29.4	42.6	n.s.
Retinopathy (%)	59.0	52.9	60.7	n.s.
Proliferative retinopathy (%)	47.4	29.4	52.5	0.079
Polyneuropathy (%)	44.9	23.5	50.8	0.045
Coronary heart disease (%)	33.3	47.1	29.5	n.s.
Ischemic stroke (%)	15.4	11.8	16.4	n.s.
Hypertension (%)	100	100	100	n.s.
SBP (mmHg)	145 (135, 159)	140 (131, 149)	145 (135, 160)	n.s.
DBP (mmHg)	80.0 (71.5, 88.0)	77.5 (70.0, 80.0	80.0 (72.2, 88.0)	n.s.
Hemoglobin (g/dL)	11.5 (10.3, 13.3)	11.4 (10.3, 12.9)	11.5 (10.2, 13.5)	n.s.
Total cholesterol (mg/dL)	156 (130, 183)	177 (163, 191)	148 (128, 180)	0.055
HDL cholesterol (mg/dL)	40.3 (34.7, 47.7)	40.6 (34.9, 55.1)	39.9 (34.3, 47.4)	n.s.
LDL cholesterol (mg/dL)	81.2 (59.3, 99.9)	90.4 (84.3, 109)	75.1 (57.5, 91.5)	0.035
Triglycerides (mg/dL)	150 (122, 218)	147 (124, 190)	156 (121, 218)	n.s.
Apolipoprotein B (mg/dL)	82 (65, 98)	77 (72, 101)	86 (64, 98)	n.s.
Glucose (mg/dL)	136 (109, 185)	129 (105, 178)	142 (119, 193)	n.s.
HbA1c (%)	7.5 (6.7, 8.4)	6.7 (5.8, 7.5)	7.6 (6.8, 8.4)	0.057
GPT (U/L)	16.0 (11.6, 22.2)	13.4 (10.9, 22.2)	16.2 (11.6, 22.2)	n.s.
Creatinine (mg/dL)	2.56 (2.24, 3.04)	2.07 (1.99, 2.60)	2.66 (2.39, 3.12)	0.023
GFR MDRD4 (mL/min/1.73 m^2)	24.0 (19.3, 27.1)	25.2 (19.3, 25.8)	23.9 (19.3, 27.2)	n.s.
GFR Cockcroft--Gault (mL/min)	29.0 (21.6, 36.5)	29.4 (24.0, 35.4)	28.6 (21.0, 36.6)	n.s.
Uric acid (mg/dL)	6.5 (5.7, 7.4)	7.1 (6.3, 8.8)	6.4 (5.5, 7.3)	0.064
Albumin (g/dL)	4.0 (3.6, 4.2)	4.1 (3.8, 4.3)	3.9 (3.6, 4.1)	n.s.
C-reactive protein (mg/L)	0.35 (0.17, 0.90)	0.25 (0.17, 0.69)	0.37 (0.17, 0.92)	n.s.
PTH (pg/mL)	145 (86, 202)	127 (98, 212)	148 (85, 201)	n.s.
25-OH-D (ng/mL)	13.2 (7.5, 22.4)	13.3 (7.6, 23.6)	12.6 (7.3, 22.2)	n.s.
Insulin treatment (%)	73.1	70.6	73.8	n.s.
ACEI/ARB treatment (%)	59.0	52.9	60.7	n.s.
N° antihypertensive drugs	3 (3, 4)	3 (2, 5)	3 (3, 4)	n.s.
Statin treatment (%)	74.4	76.5	73.8	n.s.
Antiagregant treatment (%)	70.5	58.8	73.8	n.s.

Notes: Values are expressed as percentages and medians (interquartile range). BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; GFR, glomerular filtration rate; MDRD, modification of diet of renal disease; ACEI, angiotensin-converting enzyme inhibitors; ARB, angiotensin receptor blockers.

Table 2. Multiple logistic regression model of variables associated with normoalbuminuria.

	p	Odds ratio (95% CI)
Female sex	<0.001	21.2 (3.64–123)
HbA1c	0.012	0.557 (0.329–0.941)
Polyneuropathy	0.012	0.154 (0.030–0.795)

Discussion

The traditional view of diabetic nephropathy assumed that the decline of renal function that eventually leads to end-stage renal disease is preceded by a several years lasting phase of slowly increasing urinary albumin excretion. This clinical paradigm was first challenged by the results of the United Kingdom Prospective Diabetes Study (UKPDS), which showed that about 50% of patients with type 2 diabetes who experienced some degree of renal impairment (an estimated creatinine clearance ≤60 mL/min per 1.73 m²) did not develop albuminuria during follow-up time.17 Subsequent investigations on different cohorts of subjects with both type 2 diabetes and chronic kidney disease, the latter defined as a GFR <60 mL/min/1.73 m², have confirmed that urinary albumin excretion is normal in a percentage of cases ranging between 35 and 70%.2–10

In our study, specifically focused on patients with stages 4 or 5 of chronic kidney disease, the prevalence of normal urinary albumin excretion was 21.8%. A lower rate of normoalbuminuria in this kind of patients is compatible with the well-established observation that proteinuria is one of the factors more strongly associated with progression of renal disease. Even so, our prevalence was still lower than those found in the Renal Insufficiency And Cardiovascular Events (RIACE)7 and in the Developing Education on Microalbuminuria for Awareness of Renal and Cardiovascular Risk in Diabetes (DEMAND)8 studies, both of which reported a prevalence of normoalbuminuria of 27% in the subgroups of subjects with GFR <30 mL/min. These differences could be due to the fact that, similarly to other previous multicentric or population-based surveys conducted on patients with type 2 diabetes,2,5,6,9,10 the participants in both the RIACE and DEMAND studies were uniquely assessed based on their reduced GFR, and a part of them could be affected by other non-diabetic renal diseases not characteristically associated with albuminuria. In our study, although renal biopsy was not performed in any case, a diagnostic work-up by experienced nephrologists concluded that diabetes was the main cause of renal disease in all the participants.

In any case, our findings confirm that a substantial number of patients with diabetes-related chronic kidney disease, even in advanced stages of renal failure, present normal urinary albumin excretion. Several non-controlled factors could have also led to overestimate the proportion of patients with non-albuminuric renal disease in our population. Undoubtedly, one of them could be the use of drugs blocking the renin--angiotensin--aldosterone system (RAAS), which may achieve remission of albuminuria in a considerable proportion of patients.18,19 Although the percentage of patients receiving RAAS-blocking agents was not significantly different between groups with or without albuminuria, more than a half of our participants with normoalbuminuria were taking angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, and we cannot rule out that, in such cases, micro or even macroalbuminuria had reverted to normoalbuminuria before enrollment in the study. However, because of inhibitors of RAAS have become the standard of care for patients with diabetes and either hypertension or proteinuria, the measure of albuminuria should be nowadays interpreted as the measure of residual albuminuria after treatment with these agents. The number of other antihypertensive treatments was not also different between groups with or without albuminuria.

Factors associated with normal UACR in our population were essentially analogous to those found in studies performed on subjects with earlier stages of kidney failure. This suggests that a stable cluster of patients experience progressive impairment of renal function while their urinary albumin excretion remains normal. Seventy six percent of our patients with normoalbuminuria were women, and female sex was the variable most strongly associated with an UACR <30 mg/g in our multivariate analysis. All previous studies performed on patients with GFR <60 mL/min found that, as compared with those with micro or macroalbuminuria, normoalbuminuric patients were characterized by a significantly higher proportion of women.3–7,9,10 Prospective studies following patients with type 2 diabetes have also shown that male sex is a risk factor for development of albuminuria, while female sex is a risk factor for renal impairment.17,20 On the whole, it can be accepted that the non-albuminuric form of type 2 diabetes-related chronic kidney disease is characteristically associated with female sex.

Serum creatinine was significantly lower in non-albuminuric subjects, very probably as a result of the greater proportion of women in this group of patients. In fact, creatinine was excluded from our multivariate regression model by the introduction of the variable gender, and GFR, estimated by either the MDRD4 or the Cockcroft–Gault equations, did not differ among groups with or without albuminuria. In this regard, the GFR estimates provided by the Cockcroft–Gault formula were higher than those obtained by the MDRD4 formula. This finding is probably explained by the fact that most of our patients were obese, and the Cockcroft–Gault equation, which uses body weight as a key variable, tends to overestimate GFR in subjects with obesity.21

In addition to the effect of gender, our study revealed that normal UACR was inversely related to HbA_1c and diabetic polyneuropathy. The negative relationship between albuminuria and HbA_1c is not surprising, as both cross-sectional6,7 and prospective17,20 studies have observed that HbA_1c is an independent correlate of albuminuria, but not of impaired renal function, thus indicating that non-albuminuric renal disease is probably less related to hyperglycemia and microvascular disease than the classic proteinuric form of kidney disease. Regarding to polyneuropathy, our results should be taken with caution, because we used non-standardized methods for diagnosing non-renal chronic complications of diabetes. Said that, it is noteworthy than polyneuropathy was more related to albuminuria than retinopathy and even than proliferative retinopathy. While retinopathy has been shown to be absent in a significant proportion of patients with diabetic nephropathy,22 concomitant retinopathy is still commonly used as a clinical marker for diagnosis of diabetic renal disease.23

The inverse relationship between UACR and polyneuropathy observed in our study could be in apparent conflict with the results of the UKPDS,17 in which previous neuropathy was a risk factor for incident renal failure, but not for proteinuria. In contrast, in the Japan Diabetes Clinical Data Management Study, neuropathy was independently associated with albuminuria, but not with renal insufficiency.6 In general, several population-based studies have found a positive relationship between diabetic neuropathy and proteinuria in people with type 2 diabetes.24,25

In conclusion, 22% of subjects with chronic kidney disease associated with type 2 diabetes, characteristically women, reach advanced grades of renal failure presenting normal excretion of urinary albumin. Levels of HbA_1c and diabetic polyneuropathy are inversely related to this non-albuminuric form of nephropathy.

Declaration of interest

This study was partly funded by the Spanish Ministry of Science and Innovation (IP 11 1880). The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.