Family Practice Vol. 16, No. 6, 580-585
© Oxford University Press 1999
Albuminuria, cardiovascular morbidity and mortality in diabetic and non-diabetic subjects in a rural general practice
ak
Jak
i
aa
Medical Center Metlika, Cesta bratstva in enotnosti 73, 8330 Metlika, Slovenia and
a A Stampar School of Public Health, Zagreb, Croatia.
Mlaak B, Jaki and Vuleti S. Albuminuria, cardiovascular morbidity and mortality in diabetic and non-diabetic subjects in a rural general practice. Family Practice 1999; 16: 580–585.
Received 17 December 1998; Revised 7 June 1999; Accepted 25 June 1999.
 |
Abstract |
|---|
 Top Abstract IntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
Objectives. The aim of this study was to estimate the frequency of albuminuria in patients with and without diabetes mellitus, answering the question: ‘How relevant for general practice are epidemiological findings that albuminuria is a significant risk indicator for the development of cardiovascular syndromes?’.
Method. The relationship between albuminuria and vascular disease was studied in two groups of subjects: one group consisted of 138 patients with diabetes mellitus and the other of 160 patients without diabetes, randomly selected from a stratified sample comparable with known diabetics by age, sex and profession. The groups were examined in the same way and mortality was followed over 5 years (1994–1998).
Results. Albuminuria was significantly more often present in diabetics (P < 0.0001). Peripheral arterial occlusive disease was diagnosed in 27.3% of patients with diabetes mellitus, and in 8.8% of the control group (P < 0.001). In the diabetic group 40 (29%) had microalbuminuria, but only 20 (12.5%) of the other group had the condition. In the presence of albuminuria, the incidence of arterial occlusive disease was 53.6%, but in those without it was only 9.8%.
Conclusions. Our study has demonstrated that albuminuria is frequently present with diabetes mellitus, peripheral arterial occlusive disease, hypertension, coronary heart disease, hyperlipidaemia and in cases with increased concentration of blood fibrinogen. The albuminuria was significantly more frequent in those who died in the observed 5-year period after first examination. Although albuminuria indicates higher risks of cardiovascular morbidity and mortality in diabetics and in non-diabetics, it is necessary to study further how useful is this evidence for general practice.
Keywords. Albuminuria, cardiovascular diseases, diabetes mellitus, evidence, risk factors.
|
Introduction |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
Several longitudinal studies have shown that albuminuria (albumin excretion of >30 mg/24 h) is associated with the increased risk of cardiovascular morbidity and mortality in patients with insulin-dependent (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM), and also in non-diabetic individuals. Recent studies have demonstrated the higher incidence of complications in diabetics with microalbuminuria as well as their higher mortality.1,2
The aim of our study was to ascertain albuminuria and other usual risk factors for cardiovascular disease in groups with and without diabetes mellitus and to estimate the usefulness of albuminuria findings for monitoring and care in groups of chronic diseases.
|
Studied population |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
The investigation was carried out in Metlika county (Slovenia) with 8300 residents, 60% of whom live in rural areas. It is a wine-producing hilly area served by a health centre, about 30 km away from the nearest hospital and specialist services. The population are prosperous, literate and well nourished. Traditionally, these people were very physically active but recent decades have seen a trend of diminishing activity. Chronic diseases are increasing and becoming the most prominent among health problems.
Two groups of patients registered in the local health centre were studied. The first group included 138 diabetic patients. All registered diabetics in the area (168) received written invitations, and 138 reported for examination. Among the 138 diabetics, 18 (13%) had insulin-dependent type I diabetes. The control group was recruited from 189 people without known diabetes, chosen randomly from a sample stratified by sex, age, place of living and occupation to correspond to the group of diabetics.
Of the 168 registered diabetic patients invited to the health centre, 156 (93%) attended, as did 171 (90%) of non-diabetics. A person was included in the control group, when fasting blood glucose has not exceeded 5.6 mmol/1.
Both groups had been examined systematically in the same way during one month of winter 1993/1994.
Of the 156 diabetic patients, 151 (97%) submitted a urine sample, as did 167 (98%) of non-diabetics. Further exclusions from the analysis were made on the basis of haematuria and evidence of urinary tract infection (13 diabetics and 7 non-diabetic patients). The remaining 138 diabetic and 160 non-diabetics form the basic group for the present analysis. They were followed-up during 1994–1998.
The group of diabetics had an age range of 22–88 years; the average age was 62, and it consisted of 60 males and 78 females. All of these were individually treated by diet, tablets and insulin when necessary and glycosylated haemoglobin tested every 6 months. They were regularly monitored (on average once every 2 months) for urine and blood glucose, as well as other clinical findings. The age range of the control group was also 22–88 years; the average age was 63 years, and the group comprised 69 males and 91 females. They had not had regular treatment and were examined occasionally according to their demands.
|
Examination methods |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
The examination methods were standardized according to recently published recommendations.2,3 Here are described the most relevant measurements:
Urinary albumin concentration was assessed in a morning on-spot urine sample. A urinary infection was ruled out by a leukocyte count and a urinary sediment analysis. The immunochemical nephelometric method was used on the BNII Boehringer apparatus, using the antisera and standards provided by the manufacturer. Results were classified as micro- and macro-albuminuria. Microalbuminuria was defined as a albumin/creatinin ratio of 2.5–25 mg/mmol (corresponding to the urine albumin excretion rate (UAER) of 30–300 mg/24 h). An albumin:creatinine ratio >25 mmol/l (UAER >300 mg/ 24 h) was interpreted as macroalbuminuria or proteinuria. In cases of positive results, the persistence was measured. The persistent albuminuria result, used in final analysis, was accepted when two of three different samples taken 2–6 weeks apart were in the described ranges.
Glycosylated haemoglobin (HbA1C) was essayed by a latex immunoagglutination inhibition method (DCA 2000 analyser Bayer corporation ElkhartI). The nominal range was 4.2–6.5 %. Measurements have been repeated in diabetics at regular intervals of 6 months.
Coronary heart disease (CHD) was defined in cases of history of confirmed myocardial infarction, recorded angina pectoris or ECG findings. The resting 12-lead ECG was recorded and analysed according to the Minnesota code defining ischaemia, by codes 1.1, 1.2 and 7.1.
Peripheral arterial occlusive disease (PAOD) in the lower extremities was diagnosed by measuring systolic blood pressure using the ultrasound Doppler detector; systolic pressure in the upper arm was measured using the standard mercury sphygmomanometer (Riva-Rocci), while the pressure in the artery dorsalis pedis and the artery tibialis posterior were measured using the ultrasound Doppler detector during rest and 1 minute after exercise after Mahler.4 The standard measure of exercise was established as 30 elevations per minute, the heel being elevated by at least 5 cm. For the pathological value, PAOD, we have taken the ankle arm pressure index <0.9, 1 minute after exercise.5
Hypertension was defined as a positive history of antihypertensive therapy and/or casual blood pressure measured in the sitting position on three different occasions, with 1 result of systolic blood pressure of >160 mmHg and or diastolic blood pressure of >95 mmHg.
Lipids and fibrinogen were determined by standard enzymatic methods (cholesterol CHOD-PAP method; triglycerides GPO-PAP) using the Boehringer-Mannheim standards. Hypercholesterolaemia was considered to be present if cholesterol was over 5.2 mmol/l, and hypertriglyceridaemia if plasma levels were above 2.3 mmol/l. Fibrinogen levels of 3.0 g/l and above were considered elevated.
Alcoholic beverage users were defined as those who reported consuming one or more drink units per week. Smokers were defined as those currently smoking—one or more cigarettes per day. Weight and height were measured in indoor clothing and without shoes, and the body mass index calculated. Obesity was defined by the Quetelet index (normal 2.2–2.4). Physical activity was determined according to known factors: their profession, work they had to perform, the maximal physical activity of an individual and the activity of their usual tasks. Professions were classified as heavy, medium or light in physical activity, and we identified all persons accordingly.
The 5-year mortality rates (January 1994–December 1998) for the subjects with albuminuria in the diabetic and control groups were compared with the rate for the subjects without albuminuria. In analysing causes of death, we included data from hospital autopsies and death certificates.
|
Statistical analysis |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
We used the chi-square test and factorial analysis. Having a great number of variables in the samples, and also assuming their interdependence, we applied the dimensionality reduction method, factor analysis under the varimax component model, to each set of measured variables. We applied Kaiser's criterion as the method of factor extraction. The statistical analysis was accomplished using the programme of Statistical Advanced System (SAS/STAT), edition 6.03.
|
Results |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
The results are presented in the Tables 1–4
. The first three tables are descriptive. In Table 1 the differences between the diabetic group and control group are described regarding their risk indicators and albuminuria findings.
|
|
|
|
Differences between the diabetic and control groups
There was no significant difference between the diabetic and control groups in smoking, alcohol drinking, physical inactivity or obesity. The difference was significant for hyperlipidaemia (P < 0.05), hypertension, coronary heart disease and fibrinogenaemia (P < 0.01), and highly significant for PAOD (P < 0.001).
In the diabetic group, 40 (29%) had microalbuminuria, compared with 20 (12.5%) in the control group. Macroalbuminuria was found in 16 (11.6%) of diabetics, but only in 5 (3.1%) of controls. This difference in the incidence of micro- and macroalbuminuria is highly significant (P < 0.0001). Among patients with type I diabetes, 9 (50%) had albuminuria, compared with only 47 (39.2%) of patients with type II (NIDDM).
Albuminuria, diabetes and vascular characteristics
Among the 38 diabetics with PAOD, 17 (44.7%) had micro- and 13 (34.2%) had macroalbuminuria. The difference to those diabetics without PAOD is highly statistically significant (P < 0.0001). Of 14 patients with PAOD in the control group, 5 (35.7%) had microalbuminuria and 3 (21.4%) proteinuria, which is significant when compared with those without PAOD (P < 0.01).
Albuminuria is significantly more frequent in patients with coronary heart disease (P < 0.0001), hyperlipidaemia (P < 0.01), hypertension (P < 0.001) and hyperfibrinogenaemia (P < 0.01).
The results obtained during the follow-up period
The average glycosylated haemoglobin (HgbA1C) value decreased from 8.2% in June 1994 to 7.5% in December 1997 (P < 0.001). The glycosylated haemoglobin value was higher in the patients with persistent microalbuminuria than in the patients without microalbuminuria (9.1 versus 7.9%; P < 0.001).
In Table 2
are shown main data on those who died during the 5-year observation period.
In Table 3 the relationship between albuminuria and death during 5 years' observation in the diabetic and in the control group is presented.
Mortality increased markedly with increasing urinary albumin concentration in diabetics (P < 0.001). The number of subjects with albuminuria, which is necessary for statistical analysis, was too small in the control group. Patients with the excretion of >300 mg of albumin per litre (macroalbuminuria) had a significantly higher death rate (57.1%, 12 deaths in 21 patients with macroalbuminuria). During the same period, the death rate in the subgroup without albuminuria was 3.7% (8 cases in 217).
However, the significant difference is found between diabetics and controls only in the group of patients with macroalbuminuria.
It is important to observe that not one patient in the group died of uraemia. The predominating causes of death were cardiovascular disease.
In Table 4, the results of multivariate analysis are presented.
The studied variables were reduced to four factors describing the main characteristics in the sample. The variable in an individual factor is important when its value is 0.5 or more. Four factors were clearly defined. Factor 1 (F1) is described by proteinuria, microalbuminuria and having PAOD. This factor could be an indicator of PAOD linked with albuminuria. Factor 2 (F2) is described by age and physical inactivity; factor 3 (F3) by hypertension, hyperlipidaemia and obesity; and factor 4 (F4) by smoking and hyperfibrinogenaemia. We can name the first factor the ‘albuminuria’ factor, the second the ‘age’ factor, the third ‘hypertension-hyperlipidaemia’, and the fourth ‘smoking and hyperfibrinogenaemia’.
Only the third factor (F3: ‘hypertension/hyperlipidaemia factor’) had a different distribution in the group of diabetics in comparison to controls. Both factors 1 and 2 (F1 and F2: ‘albuminuria’ and ‘age’ factors) had different distributions in the groups of patients with and without PAOD.
In summary, the studied examinees could be described by four clear-cut interpretable factors, showing that measured variables do not appear as isolated, but in an aggregation form.
|
Discussion |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
A comparison with published findings
In the several longitudinal studies, established risk factors for cardiovascular diseases, such as smoking, hypertension, hyperlipidaemia and hyperfibrinogenaemia, were higher in diabetics with persistent albuminuria than in those of comparable age and sex without albuminuria. In our cross-sectional study, we found an association between albuminuria and cardiovascular risk factors and in our results this was higher than in some other prospective studies.2,3 Several studies have revealed the incidence of cardiovascular accidents to be 1.5–3 times higher in diabetics than in the general population.2,3,6 In our study, the prevalence of PAOD and albuminuria in the diabetic patients was approximately three times higher than in the control group.
Microalbuminuria or proteinuria, whether in diabetics or controls, was significantly correlated with the prevalence of PAOD and CHD. The results of the factor analysis demonstrate that both microalbuminuria and proteinuria are indicators for PAOD. Even non-diabetics with albuminuria in our control group have a higher frequency of PAOD and CHD. This is in accord with other studies which have shown that persistent macroalbuminuria in diabetics is an indicator not only for nephropathy, but also for proliferative retinopathy and cardiovascular disease.3
Several studies of type I diabetics (IDDM) have shown that microalbuminuria foretells the development of clinical diabetic nephropathy after an approximately 10-year interval. Microalbuminuria also occurs in type II diabetes (NIDDM), but in this setting it predicts cardiovascular complications more commonly than end-stage renal disease.
Diabetic nephropathy in patients with IDDM is associated with the high risk of early death, the relative mortality among patients with nephropathy being 40 times greater than among those without it.7 NIDDM subjects with albuminuria also have an increased mortality rate, especially from cardiovascular disease. Our results confirm the importance of this risk in diabetics: 11 of 16 patients with proteinuria (nephropathy) (68.8%) died, as compared with 7 of the 122 patients without nephropathy (5.7 %).
Opportunities for medical intervention
There have been substantial changes in the treatment of diabetes over the past 10 years. Self-monitoring by measurement of urinary glucose and, more recently, blood glucose, has become routine; together with better education and adjustment of treatment to the patient's lifestyle, this has led to improved metabolic control. In order to reduce the incidence of diabetes, morbidity and mortality, a co-ordinated and continuous programme that has an impact on all population groups is required and this was implemented in our case. All persons, without economic barriers, have had access to health services. Besides, there was an active follow-up procedure. The decrease of the prevalence of diabetic nephropathy, as manifested by persistent proteinuria, is possibly the result of improved glycaemic control. According to the Bojestig et al. study, the decrease of persistent albuminuria from 30 to 8.9% is also described as a probable result of improved glycaemic control.7
Effective treatment of coexistent hypertension seems to reduce renal and cardiovascular mortality among patients with albuminuria.8 The prevalence of macroalbuminuria in our diabetics was 16 (11.6%), but only in 5 (3.1%) of controls.
Treatment with antihypertensive drugs reduced microalbuminuria in patients with diabetes. The decrease in the prevalence of persistent albuminuria and delay of uraemia in our study was most probably related to antihypertensive treatment, because 77.5% patients with diabetic nephropathy received such therapy.
What is the meaning of albuminuria in general practice?
For medical practitioners, ‘risk factors' indicate necessity for further assessment and a series of decisions, of which not all are well understood and some need further research and explanation.9,10 It is necessary to differentiate actual meaning of found ‘risk factors' in the served population and in the case of a particular patient. Our findings of significant associations between UAER and cardiovascular morbidity in this cross-sectional study confirm albuminuria as an indicator of CHD and PAOD. At the same time, 5-year follow-up has shown predictive power of albuminuria for cardiovascular mortality.
According to these findings, the guidelines for intervention have also to include regular determination of albuminuria as useful additional information for management of people at risk of prevalent chronic diseases like hypertension, coronary heart disease and diabetes.
The primary health care physician is having the greatest potential for systematic delivery of diabetes and preventive measures in cases of prevalent chronic diseases. The present study, however, does not give good evidence for the importance of efficient metabolic control (short period of observation and small number of patients, especially with IDDM) in prevention of nephropathy. However, decreasing HbA1C in the observed period, aggressive antihypertensive treatment of all diabetics with hypertension, might explain present findings such as nearly the equal longevity (average age of deceased was 76.3 years in diabetics and 77 years in controls) and probable postponement of early cardiovascular mortality and uraemia. The (micro)albuminuria might help indicate the risk of these complications.
|
Conclusions |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
The presence of albuminuria indicates a need for both careful follow-up of diabetic complications and improvement in metabolic control. Measurement of albuminuria should be possibly considered as a routine investigation in the management of IDDM and NIDDM. Screening for microalbuminuria in such a population may identify high-risk patients with abnormalities that are potentially treatable, and further efforts in this direction are necessary.
|
References |
|---|
|
TopAbstractIntroductionStudied populationExamination methodsStatistical analysisResultsDiscussionConclusionsReferences |
|---|
1 Damsgaard EM, Froland A, Jorgensen OD, Mogensen CE. Microalbuminuria as predictor of increased mortality in elderly people. Br Med J 1990; 300: 297–300.
2 The World Health Organisation Multinational Study of Vascular Disease in Diabetics. Prevalence of small vessel and large vessel disease in diabetic patients from 14 countries. Diabetologia 1985; 28 (Suppl.): 615–640.
3 Haffner SM, Stern MP, Gruber MKK, Hazuda HP, Mitchell BD, Paterson JK. Microalbuminuria. Potential marker for increased cardiovascular risk factors in nondiabetic subjects. Arteriosclerosis 1990; 107: 727–731.
4 Mahler F. Systolische Druckmessung nach Belastung. In: Kriessmann A, Bollinger A, Keller M (eds). Praxis der Doppler Sonographie. Stuttgart-New York: G Thieme Verlag, 1982: 32–39.
5 Straub H, Ludwig M. Der Doppler Kurs. Münich/Bern/Vienna: W Zuckschwerdt Verlag, 1990: 80–100.
6 Yudkin JS. How can we best prolong life? Benefits of coronary risk factor reduction in non-diabetic and diabetic subjects. Br Med J 1993; 306: 1313–1318.
7
Bojestig M, Arnquist HJ, Hermansson G, Karlberg BE, Ludvigsson J. Declining incidence of nephropathy in insulin-dependent diabetes mellitus. N Engl J Med 1994; 330: 15–18.
8 Borch-Johnsen K, Andersen PK, Deckert T. The effect of proteinuria on relative mortality in Type 1 (Insulin-dependent) diabetes mellitus. Diabetologia 1985; 28: 590–596.[ISI][Medline]
9
Jones R. Published research in primary care: original papers in the US and UK, 1996. Fam Pract 1998; 15: 291–295.
10 Farmer A, Milne R. Evidence based medicine—1. Giving patients an accurate prognosis. Student Br Med J 1995; 3: 100–102.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. A Donker, D. M Fleming, F. G Schellevis, and P. Spreeuwenberg Differences in treatment regimes, consultation frequency and referral patterns of diabetes mellitus in general practice in five European countries Fam. Pract., August 1, 2004; 21(4): 364 - 369. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||