- Email: [email protected]
Microalbuminuria and associated cardiovascular risk factors in the community
71
93 ( 1992) 71-8 1 0 1992 Elsevier Scientific Publishers Ireland, Ltd. All rights reserved 0021-9150/92/$05.00
Atherosclerosis,
Printed and Published in Ireland
ATHERO 04784
Microalbuminuria
and associated cardiovascular risk factors in the community
Peter H. Winocoura’d, Jane O.E. Harlandayd, Jill P. Millar”, Michael F. Lakerb’d and K. George M.M.. Albertia>d The Departments of ‘Medicine. bCIinical Biochemistry, and ‘Primary Health Care and ‘Human Metabolism and Diabetes Research Centre, The Medical School, University of Newcastle upon Tyne. Framlington Place, Newcastle upon Tyne. NE2 4HH (UK.)
(Received 26 July, 1991) (Revised, received 24 December, 1991) (Accepted 30 December 1991)
Summary The prevalence of microalbuminuria and relationship to cardiovascular risk factors was examined in a cross-sectional community survey of cardiovascular risk factors. Microalbuminuria (when classified as albumin concentration 720 pg/ml) was present in 6.3% of subjects but in conjunction with an albumin/creatinine ratio > 3.5 in only 2.2%. Diastolic blood pressure, prevalence of abnormal electrocardiographs, and to a lesser extent systolic blood pressure and fibrinogen concentration, were greater in those with albuminuria concentrations >20 &ml. The strongest positive univariate correlates of albumin/ creatinine ratios in those with detectable albuminuria were age, fibrinogen, blood pressure, total- and low density lipoprotein- (LDL) cholesterol, apo B and alcohol intake, whereas fasting insulin and insulin resistance were inversely correlated. Multiple regression analysis revealed that age, gender, systolic blood pressure and insulin resistance independently accounted for 37% of the variability in albumin/creatinine ratios. When those 10 subjects with microalbuminuria and albumin/creatinine ratios 73.5 were matched with 20 with normoalbuminuria for age, gender and body mass index, the microalbuminuric subjects had significantly lower LDL cholesteroYapo B ratios and a tendency to lower high density lipoprotein (HDL) cholesterol and HDL cholesterohapo Al ratios. Microalbuminuria is uncommon in the general population, and is related to ageing, blood pressure and other vascular risk factors. It may reflect the presence of established cardiovascular disease.
Key words: Microalbuminuria; Lipoproteins
Cardiovascular
disease; Diabetes; Hypertension;
Elderly; Fibrinogen;
Introduction Correspondence to: Dr Peter H. Winocour, Lecturer in Medicine, Freeman Hospital, Freeman Road, High Heaton, Newcastle upon Tyne, NE7 7DN, U.K.
Proteinuria detected by conventional semiquantitative test-strip testing predominantly reflects
12 urinary albumin excretion which is usually the consequence of renal glomerular (or less commonly tubular) disease [ 11. Proteinuria in diabetic and hypertensive populations is associated with an excess mortality from cardiovascular disease [2-51. More recently it has been suggested that so-called microalbuminuria measured by sensitive radiochemical techniques, and by definition not detected by the usual test strips, might also be associated independently with cardiovascular disease [6,7]. Urinary albumin excretion may be modified by posture or exercise, or increased in the presence of urinary infection or parenchymal renal disease [ 11. Intra-individual day-to-day variability of albumin excretion [8] has also restricted assessment of an independent role for albuminuria as a risk factor for cardiovascular disease (CVD), but can be minimised by the use of timed samples, or correction for urine creatinine. The prevalence of microalbuminuria in healthy Caucasian adults is not known, and the suggestion that in diabetic and hypertensive populations albuminuria may be associated with vascular risk factors [9-141 such as smoking, blood pressure, lipids, blood glucose, insulin, fibrinogen and ageing has yet to be explored in the general population. These questions have been addressed in the course of a community based CVD risk factor screening programme. Subjjts
and Metlmds
Ethical approval for the study was obtained from the Joint Ethics Committee of the University of Newcastle upon Tyne and the Newcastle Health Authority. Seven hundred self-selected patients from one urban group general practice responded to an advertisement inviting adults aged 25-65 to take part in a cardiovascular risk factor health screen. This report is based on 447 individuals in whom urinary albumin was estimated. Patients attended during the period September 1988-March 1990. Following an overnight fast, venous blood was drawn for measurement of plasma glucose, glycated haemoglobin (HbA,), fibrinogen, and serum concentrations of insulin, cholesterol, triglycerides, apolipoproteins B and Al. In addi-
tion plasma glucose and serum insulin were measured 1 and 2 h following ingestion of a 75-g oral glucose monohydrate load. High density lipoprotein (HDL) cholesterol was also measured 2 h after ingestion of the glucose load. Venous plasma glucose (separated immediately after sampling) was measured on a continuous flow autoanalyser (Technicon AA2, Technicon Instruments, Basingstoke, U.K.; inter-assay coefficient of variation (CV) l-3%). HbA, was measured by electroendosmosis; reference range 5.0-8.0%. Serum insulin was measured by radioimmunoassay, a modified version of the method of Soeldner and Slone [15], using an in-house antibody, inter-assay CV 6.8-7.5%. Insulin resistance (IR) was calculated from fasting blood glucose and serum insulin concentrations using the computer-solved homeostasis model assessment (HOMA) method described by Matthews et al. [16]: IR = fasting insulin/22.5e-In fastins glucose. Serum cholesterol was measured by a cholesterol oxidase method on a Cobas Bio centrifugal fast analyser (Roche Products Limited, Welwyn Garden City, U.K.) using a commercial kit (number 725242) supplied by BCL &ewes, U.K.). Between batch CVs ranged between 1.5% to 2.2% for cholesterol concentrations ranging from 3.9 to 10.3 mm01 1-l. Serum triglycerides were measured by a lipase-glycerol kinase method without correction for free glycerol on the Cobas Bio centrifugal analyser using a commercial kit (both supplied by Roche Products Limited, Welwyn Garden City, U.K.). Between batch CV was 3.0% to 3.2% with triglyceride standards ranging from 2.31 to 3.47 mm01 1-l. HDL was isolated after precipitation of apo B-containing lipoproteins with heparin and manganese. Cholesterol within HDL was measured by the enzymatic method previously described. The between batch CV at a mean HDL cholesterol of 0.97 mm01 1-l was 3.6O/o.Low density lipoprotein (LDL) cholesterol was calculated by the Friedewald formula [17] (total serum cholesterol (HDL cholesterol + fasting serum triglycerides/ 2.2) mm01 1-l). The apoproteins B and Al were measured by immunonephelometry using a Behring laser nephelometer, and Behring standards and antisera (Hoechst U.K. Ltd., Hounslow, U.K.). Between batch CV for a mean apo B of 0.88 g 1-l was 4.9% and for a mean apo Al of 0.92 g 1-l was 5.6%
73
Plasma fibrinogen was measured by the Clauss technique (between batch coefficient of variation 5%) [18]. Body mass index (BMI) was calculated as Quetelet’s index: BMI = weight (kg)/height2 (m2). Alcohol and tobacco consumption, and patterns of exercise were ascertained by direct enquiry, and the presence or absence of vascular disease assessed by a resting 12 lead electrocardiogram and the WHO Rose questionnaire. ECGs were coded by the Minnesota coding system, and the following codes compatible with ischaemic heart disease were designated abnormal: l-l, l-2, 4-1, 5-1, 8-1, 8-3. Urinary albumin estimation
Each individual brought a specimen of urine passed on waking on the morning of attendance. Samples were tested for glucose, blood, protein, and nitrite (Multistix, Boehringer, U.K.), and a mid-stream sample was sent for microbiological evaluation when positive for any one of the latter three, or when urinary tract infection was suspected from the history. A repeat early morning urine sample was sent in those cases following treatment of the urinary infection or if women were menstruating at the time of the initial collection, and a further sample was sent for albumin and creatinine measurement after detection of a sterile mid-stream urine sample. Initially, albumin and creatinine concentrations were measured by immunonephelometric (Behring antisera and reagents, Behring Diagnostics, Hoechst, Middlesex, U.K.) and Jaffe methods, respectively. The lower limit of sensitivity of albumin concentration estimation by this method is 12 &ml and the between batch variability for urine with a mean albumin concentration of 38.0 &ml was 6%. A repeat timed overnight urine collection was carried out in those 10 cases whose urinary albumin concentration was greater than 20 pg/rnl and in addition whose albumin/creatinine ratio was greater than 3.5, and in 20 normoalbuminuric individuals who were matched sequentially for sex, age (to within 5 years) and body mass index (to within 2 kg/m2 units). All normoalbuminuric cases initially had urinary albumin concentrations close to the limit of sensitivity of the assay (12-14 &ml). The first 2 normoalbuminuric subjects who
fulfilled these criteria were selected from the list of normoalbuminuric patients. On this occasion urine albumin was quantitated by a more sensitive radioimmunoassay method (sensitivity 0.5 &ml, between batch variability 5%) [19]. Intraindividual day-to-day variability in overnight albumin excretion has been estimated at 8-40% by our laboratory [20] and by others [21]. Data analysis
Information was collected on an Amstrad PC4535 and analysed using the statistical package for the social sciences (SPSS). Data are presented as mean (S.E., range). Differences between those with urinary albumin concentrations above or below 20 &ml, and between the normoalbuminuric subjects above or below the detection limit of the assay were, compared by Student’s t-test, as were those with microalbuminuria when compared with the matched normoalbuminuric control group. After establishing that those with normoalbuminuria (< 20 &ml) above and below the detection limit of the assay were comparable with respect to demographic and biological variables, univariate associations in those 74 individuals with urinary albumin concentrations above the detection limit of the assay were examined by Pearson’s correlation coefficient. Stepwise multiple regression analyses were also employed in this group to examine the independent associations between albuminuria and the following variables: cigarette and alcohol consumption, documented hypertension and antihypertensive therapy (including diuretics), fasting and post-glucose challenge plasma glucose and serum insulin levels, age, a first degree relative with hypertension and or diabetes, blood pressure, HbA,, total-, LDL- and HDL-cholesterol, serum triglycerides, apo Al, apo B, fibrinogen and haemoglobin concentrations, and the ratios of LDL/HDL and HDL cholesterol/ape A 1. Results
Of the 447 individuals on whom data were available, there were 209 men and 238 women, aged from 25 to 70 years (mean (S.E.) 47.3 (0.5) years)), average body mass index 25.1 (0.2) kg/m’. Mean values of other measured variables were as follows: alcohol consumption 9 (1) units/week, fasting and
74 2-h plasma glucose 5.2 (0.1) and 5.0 (0.1) mmol/l, respectively, HbAi 6.2 (O.l)O/o,fasting serum insulin 8.8 (0.4) munits/l, total serum and LDL cholesterol 5.9 (0.1) and 3.9 (0.1) mmol/l, respectively, total serum triglycerides 1.45 (0.04) mmolil, HDL cholesterol 1.41 (0.02) mmol/l, apo Al 1.5 (0.01) g/l, apo B 1.3 (0.02) g/l, fibrinogen 3.1 (0.04) g/l, haemoglobin 14.2 (0.1) g/dl, blood pressure (systolic/diastolic) 121 (1)/76 (1) mmHg. The 2-h HDL cholesterol value correlated closely with fasting values (rS = 0.97) in 100 subjects whose fasting HDL cholesterol values ranged between 0.7 and 2.0 mm01 1-i (mean value 1.5 mm01 1-l). There were 68 cigarette smokers whose average consumption was 14.6 (0.9)/day. The sample of 447 was comparable to the original group of 700 with respect to the prevalence of vascular risk factors and apparent vascular disease: diabetes (1.3%
vs. 1.4%, respectively), hypertension (systolic/diastolic > 160 and/or 95 mmHg respectively, or current antihypertensive medication) in 10.3% vs. lO.O%, current smokers 15.2% vs. 15.0%, hypercholesterolaemia (>6.5 mmol/l) in 32% vs. 27%, and historical evidence of cardiovascular disease from the WHO Rose questionnaire in 5.1% and 4.3%, respectively. All but 10 of the 447 were Caucasian, 99 had previously been told by a doctor they had hypertension, of whom 27 received antihypertensive medication, including 17 on diuretic therapy. Aspirin was taken on average more than once a week by 16 individuals. Angina was reported by 18 cases, intermittent claudication by 4, and 5 had a history of previous myocardial infarction. There was a first degree relative with hypertension in 146 cases and of diabetes in 52. Impaired glucose tolerance was present in 11
TABLE 1 URINE ALBUMIN EXCRETION AND CARDIOVASCULAR
RISK FACTORS IN 477 MALE AND FEMALE
Figures are mean (S.E.), range.
Number (M/F) Age (years) Body mass index (kg/m2) Alcohol (units/week) Smokers (%) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Fasting blood glucose (mmobl) 1 h blood glucose (mmoiil) 2 h blood glucose (mmoliI) HbA, (%) Fasting insulin (munitsil) 1 h insulin (munitsil) 2 h insulin (munitsil) Insulin resistance Total serum cholesterol (mmobl) Total serum triglycerides (mmol/l) LDL cholesterol (mmolil) HDL cholesterol (mmol/l) APO Al W) APO B W) Fibrinogen (gil) LDL cholesterol/ape B HDL cholesteroYapo Al LDLiHDL cholesterol
Normoalbuminuric (< 12-20 &ml)
Microalbuminuric (>20 &ml)
419 (198221) 47 (l), 25-70 25.1 (0.2) 17.7-45.4 9 (1) O-70 15.3 121 (1) 82-189 76 (I), 54-105 5.2 (O.l), 3.0-11.5 6.6 (O.l), 2.3-19.5 5.0 (O.l), 1.7~20.8 6.2 (0. l), 4.0-9.8 8.8 (0.4), 1.0-46.8 67.1 (2.0), 9.0-254.0 34.4 (1.5), 1.5-280.2 0.23 (O.Ol), 0.02-2.09 5.9 (0. l), 2.7- 10.3 1.5 (0.1) 0.4-6.8 3.9 (O.l), 1.0-7.9 1.4(0. l), 0.6-3.0 1.5 (O.l), 0.6-2.5 1.3 (0.1). 0.4-2.3 3.1 (0.1) 0.7-7.2 1.17 (0.01) 0.38-2.18 0.36 (O.Ol), 0.13-0.75 3.0 (0. I), 0.6-7. I
28 (11:17) 49 (2), 25-65 25.3 (0.7), 19.0-32.0 12 (3). O-80 14.3 127 (3) 102-174’ 81 (2). 63-102** 5.3 (0.1). 4.1-6.1 7.0 (0.4), 3.1-I 1.9 5.6 (0.3) 2.7-10.5 6.4 (0.2) 4.8-9.8 8.7 (1.2), 0.5-26.3 75.8 (6.Q 26.9-174.9 47.0 (9.0), 6.5-260.0 0.23 (0.03) 0.01-0.72 5.9 (0.2) 4.0-8.2 1.2 (0. I), 0.6-2.5 3.9 (0.2) 1.9-6.5 1.4 (0.1) 0.6-2.2 1.6 (0.1) 0.8-2.1 1.3 (0.1). 0.8-1.9 3.4 (0.2) 2.2-5.5, 1.14 (0.02), 0.89-1.30 0.36 (O.Ol), 0.26-0.48 3.0 (0.4), 1.1-10.8
Differences between groups *P < 0.1, **P < 0.05.
SUBJECTS
75 (2.5%) and diabetes in 6 (1.3%) subjects. All were negative for protein on Multistix testing except 3 subjects (2 showed a trace, and in one case 0.1 g/l of proteinuria). Three hundred and seventy-three individuals had albuminuria below the detection limit of the assay, 46 had normoalbuminuria (12-20 &ml) and 28 (6.3%) had microalbuminuria (> 20 &nl). Those with normoalbuminuria above or below the detection limit of the assay were comparable with respect to gender ratio, age, body mass index, consumption of tobacco and alcohol, and biological variables (data not shown). We therefore compared those 28 with microalbuminuria with the remaining 419. Microalbuminuria was associated with significantly higher diastolic blood pressure, and a tendency to somewhat higher levels of systolic blood pressure and plasma fibrinogen. Other biochemical measurements were comparable (Table 1). When
only those 46 with detectable normoalbuminuria were compared with those with microalbuminuria, similar differences in blood pressure and fibrinogen were recorded (data not shown). The prevalence of abnormal ECGs was greater in those with microalbuminuria >20 j&ml (P < 0.05), but there was no clear difference in the prevalence of abnormal glucose tolerance, or use of antihypertensive agents (Table 2). Significant (P < = 0.01) univariate correlates of albuminuria concentrations and the albumin/ creatinine ratios are shown in Table 3. The strongest associations of the albuminkreatinine ratios were with age, systolic blood pressure, total serum and LDL cholesterol, fibrinogen, alcohol, and apo B. Urinary albumin concentrations were correlated with age, systolic and diastolic blood pressure, fibrinogen, serum triglycerides, and apo Al.
TABLE 2 PREVALENCE OF DIABETES, HYPERTENSION, VASCULAR DISEASE AND ANTIHYPERTENSIVE SUBJECTS WITH NORMAL URINARY ALBUMIN CONCENTRATIONS OR MICROALBUMINURIA Normoalbuminuric (< 12-20 &lll) (?I = 419) Previous diagnosis of diabetes (% (number)) Impaired glucose tolerance (% (number)) Diabetes mellitus (% (number)) Previous diagnosis of hypertension (% (number)) Current use of antihypertensive agents (% (number)) Current diuretic use (% (number)) Anginaa (% (number)) Abnormal ECG* (% (number)) Myocardial infarctiona (% (number)) Prevalence of 1st degree relatives with hypertension at any age (% (number)) Prevalence of 1st degree relatives with CVA at any age (% (number)) Prevalence of 1st degree relatives with diabetes mellitus at any age (% (number)) BData generated from WHO Rose Questionnaire. Difference between groups *P < 0.05.
Microalbuminuria ( > 20 &ml) (n = 28)
0.2 (1)
0
2.1 (9)
7.1 (2)
1.4 (6) 22.0 (92)
0 (0) 25.0 (7)
6.2 (26)
3.6 (1)
4.1 (17) 4.1 (17) 2.4 (10) 1.0 (4) 32.5 (136)
0 (0) 3.6 (1) 17.9 (5)* 3.6 (1) 35.6 (IO)
22.7 (95)
25.0 (7)
11.7 (49)
10.7 (3)
(0)
TREATMENT IN
76 TABLE 3 UNIVARIATE PEARSON’S CORRELATION COEFFICIENT (r), OF ALBUMINURIA MEASURES (IN 74 SUBJECTS WITH DETECTABLE ALBUMINURIA) WITH OTHER VARIABLES Correlates
Albumin/ creatinine ratio
Urinary albumin concentration
Age LDL cholesterol Fibrinogen Systolic blood pressure Diastolic blood pressure Total serum cholesterol Apo B Alcohol Triglycerides Fasting blood glucose Fasting insulin Insulin resistance Apo Al HDL cholesterol
0.43*** 0.25* 0.24* 0.24* 0.23* 0.23’ 0.20* -0.20* -0.06 0.12 -0.25* -0.24* 0.13 0.08
0.23* 0.04 0.25* 0.21’ 0.222 -0.03 0.02 0.10 0.20* 0.04 -0.15 -0.14 0.20’ 0.06
‘P < 0.05, ***p < 0.001.
Multiple regression analysis demonstrated that 37% of the variance in the albumin/creatinine ratio could be accounted for by independent influences of age, gender, systolic blood pressure, and insulin resistance (Table 4). Microalbuminuria, in association with an albumin/creatinine ratio >3.5 was present in 10 individuals (9 women) (2.2% prevalence overall). Three were positive (0.1 g/l) for urinary protein on stick testing. Four had previously been told by a doctor that they had hypertension, but only one
was currently receiving treatment (with a beta blocker). None of these received diuretics or aspirin, nor had experienced myocardial infarction, although the only 1 who smoked (1 S/day) suffered from angina and intermittent claudication. One further subject had impaired glucose tolerance, but the rest were normoglycaemic by conventional criteria. Repeat urine albumin estimation on an overnight collection by the more sensitive assay confirmed persistent elevation of both urinary albumin concentration and the albumin/creatinine ratio in 8 cases. However, the overnight albumin excretion rate was greater than 16 &nin in all 10 cases. Average variability in albumin concentration was 13% (range l-20%), whilst that of the albumin/creatinine ratio was 19.7%. When these 10 were compared with the matched group of 20 persistently normoalbuminuric individuals, the major differences between the 2 groups was in the LDL cholesterohapo B ratio and to a lesser extent the l-h post-glucose challenge plasma glucose, the HDL cholesterol concentration, and the HDL cholesterohapo Al ratio (Table 5). In addition, abnormal ECGs were recorded in 50% and 25% of the microalbuminuric and the matched normoalbuminuric groups, respectively (P < 0.1). Dlscnssion We have recorded a low prevalence (6.3%) of microalbuminuria in a community based population who were unselected with respect to vascular disease or the presence of vascular risk factors such as hypertension or diabetes. Although the
TABLE 4 PREDICTOR VARIABLES ALBUMINURIA Predictor variables in model
Age
Gender Systolic blood pressure Insulin resistance Constant
FOR
ALBUMINCREATININE
R2
0.21 0.29 0.34 0.37
RATIOS
IN 74 INDIVIDUALS
Regression coeBicient zt S.E. 0.008 0.210 0.003 -0.169 -0.921
i f f zt f
0.003 0.058 0.002 0.094 0.249
WITH
DETECTABLE
Multiple correlation coefficient
Significance of variable
0.46 0.55 0.58 0.61
0.0014 0.0806 0.0709 0.0772 0.0004
P
77 TABLE 5 CARDIOVASCULAR RISK FACTORS IN 10 SUBJECTS WITH BOTH MICROALBUMINURIA ALBUMINICREATININE RATIO AND IN 20 MATCHED NORMOALBUMINURIC SUBJECTS
AND
RAISED
Results shown as mean (S.E.), range.
Baseline albuminuria (ccg/ml) Baseline albumin/creatinine ratio Repeat overnight albuminuria (pglml) Overnight albumin excretion rate (&mm) Repeat overnight albumin/creatinine ratio Sex ratio (M/F) Age (years) Body mass index (kg/m’) Alcohol (units/week) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Fasting blood glucose (mmolil) 1 h blood glucose (mmolil) 2 h blood glucose (mmolil) HbA, (%) Fasting insulin (munitsfl) 1 h insulin (munitsil) 2 h insulin (munits/l) Insulin resistance Total serum cholesterol (mmolil) Total serum triglycerides (mmol/l) LDL cholesterol (mmol/I) HDL cholesterol (mmol/l) Apo Al (fl) APO B (g/l) Fibrinogen (gil) LDL cholesterohapo B HDL cholesterol/ape A I LDL/HDL cholesterol
Normoalbuminuric group
Microalbuminuric group
< 12.2 (0.1). < 12.0-14.0
39.2 (6.2), 20.0-77.0**** 5.0 (0.3), 3.6-6.4**** 38.2 (7.2), 16.8~90.0**** 32.0 (4.9) 18.0-66.0**** 5.1 (0.2), 3.0-12.0**** I:9 56 (3). 39-65 25.0 (I .O), 20.2-29.6 3 (2) o-15 132 (7) 109-168 83 (3). 69-96 5.3 (0. I), 4.5-5.7 7.3 (0.5). 5.0-9.6* 5.8 (0.6) 3.8-10.5 6.6 (0.5) 5.0-9.9 9.4 (2.5), I .7-26.3 75.9 (14.8), 35.8-174 60.5 (23. I), 9.0-260.0 0.25 (0.07) 0.05-0.72 6.6 (0.3) 5.6-8.2 I .4 (0.2), 0.8-2.5 4.5 (0.4) 2.9-6.5 1.5(0.2), 0.6-2.4’ I .6 (0. I), 0.8-2.0 I.5 (0.1) 1.1-1.9 3.6 (0.3), 2.2-5.2 I.18 (0.03) l.O3-1.30** 0.36 (0.02) 0.29-0.46* 3.73 (0.84). 1.20-10.80
4.1 (O.S), 1.5-8.1 3.5 (0.4), 1.5-8.3 0.7 (O.I), 0.3-1.2 2:18 56 (2), 39-65 24.5 (0.5), 21.0-29.7 5 (2), O-30 124 (5), 84-189 78 (3), 60- IO3 5.1 (0. I), 4.2-6.0 6.7 (0.6). 3.3-13.5 4.7 (O.S), 2.4-12.1 6.4 (0.2), 4.8-7.9 6.4 (0.9), I .4-18.0 66.0 (8.0), 13.6-145.5 30.1 (6.3), 3.3-107.0 0.17 (0.03) 0.04-0.51 6.6 (0.2), 4.9-8.5 1.2 (O.l), 0.6-2.5 4.3 (0.2), 3.2-6.1 1.8 (O.l), 1.1-2.5 1.8 (O.I), 1.3-2.3 1.3(0. l), 0.9-2. I 3.2 (0.1). 2.3-4.6 1.27 (0.03) 1.03-1.54 0.39 (0.01) 0.29-0.46 2.56 (0.16) 1.28-4.06
Differences between groups not significant unless *P < 0.1, **P < 0.05, **** P < 0.001.
population under study were self selected, they were representative of the larger group of 700 subjects in terms of their vascular risk profile. The prevalence of diabetes and hypercholesterolaemia was also very similar to that recorded in previous unselected population studies throughout the U.K. [22,23]. The prevalence of smoking and hypertension in the U.K. in general appears more variable in different regions, but in certain districts throughout the country is broadly the same a’sthat recorded in the present investigation [23,24]. We would not claim that our reported prevalence of microalbuminuria is necessarily totally represen-
tative of the general population, and in the light of our associated findings it is likely that microalbuminuria could be detected more frequently in populations with a higher prevalence of hypertension. The present findings are in accordance with one previous study of 127 adults and children in the U.K.[8], but much lower than the lo-20% recorded in other reports where diabetes and cardiovascular disease was over-represented [6,1 l] or where the study population was considerably older [I 11. Haffner et al. [25] reported a 13% prevalence of microalbuminuria in 3 16 non-diabetic Mexican
78 Americans who were similar in other respects to our population with the exception of the fact that they were on average more obese and hyperinsulinaemic. Racial differences [26], selection bias, and the fact that our urine samples were assayed for albumin within 1 week of collection whereas Haffner et al. [25] processed their samples after an average delay of 12 months, may also have contributed to the discrepancy. Although the vast majority of subjects had albuminuria concentrations below the sensitivity of the routine assay (12 @ml) this test appears adequate for the purposes of screening the population, and differentiating normoalbuminuric from microalbuminuric samples. When prevalence of microalbuminuria was determined from elevations in both albumin/ creatinine ratios (> 3.5) and urinary albumin concentrations (>20 &ml), in order to reduce any bias as a result of errors in collection, only 10 of the 28 with albumin concentrations >20 &ml fultilled this category. Inclusion of false positives amongst the 28 cases in the initial comparison (Table 1) could have minimised any apparent differences in cardiovascular risk factors that were directly attributable to ‘microalbuminuria’. We therefore specifically elected to compare also those 10 with elevated urinary albumin concentration and albumin/creatinine ratios with 20 distinctly normoalbuminuric subjects in order to confirm that the differences in risk factors were comparable in the more selected microalbuminuric group. This procedure also reduced any potential confounding influence of gender, age or body mass. Alterations in lipoprotein composition were suggested, but the relatively small sample size and impact of modest microalbuminuria on other vascular risk factors may be why only nonsignificant differences in blood pressure, fibrinogen and HDL cholesterol were recorded. It is recognised that albuminuria is variable within individuals on a day-to-day basis [8,20,21] but it has previously been suggested that this may be lessened by the use of overnight collections 1271, and minimised further by the use of albumin/ creatinine ratios [27,28]. The extent of miscategorisation of subjects on the basis of single urine collections has been shown to be negligible unless albumin excretion rates exceed 30 kg/mm (equivalent to an albumin/creatinine ratio of >3.5)
[20,27]). In the present study 2 of 10 subjects did not fulfil both criteria for microalbuminuria on an overnight collection although their matched controls were persistently normoalbuminuric. However, the albumin excretion rate was > 16&ml in all 10 cases and this is well outside previously quoted reference ranges for healthy subjects [8,21]. Furthermore those previous studies which demonstrated that microalbuminuria acted as a prognostic marker for diabetic nephropathy and cardiovascular disease were based on albumin excretion rates determined from single overnight collections [6,7,29,30]. The dominant determinant of detectable albuminuria in our study was age, which is in accord with Damsgaard et al. who have recorded the highest prevalence of microalbuminuria (1 j-20%) in healthy subjects whose average age was 67 [111. Yudkin et al. [6] found no correlation between age and albuminuria in an older (average age >60 years) selected group of subjects with a high prevalence of vascular disease, and an inverse association between age and albuminuria has even been recorded and attributed to increased urinary flow rates in older subjects, although both children and adults were included in this latter study [8]. Albumin/creatinine ratios in the present study were also independently determined by gender, which could in part be accounted for by differences in muscle bulk in men and women. Blood pressure is closely associated with albuminuria in diabetic [ 111,hypertensive [6,10] or elderly [6,11] populations and this relationship was apparent in our study, although not in healthy normotensive individuals studied by Watts et al. [8]. Only a small proportion of subjects in our study had hypertension according to WHO criteria, however. Fasting plasma glucose and insulin concentrations were only weak correlates of albuminuria in our population where the prevalence of impaired glucose tolerance/diabetes was similar to that expected from a random population sample in the United Kingdom [22]. Haffner et al. [25] found that insulin was more conspicuously associated with albuminuria than in the present study but it was unclear how common impaired glucose tolerance was in their obese hyperinsulinaemic population. Impaired glucose tolerance and/or diabetes is clearly associated with greater
79 albuminuria than an appropriate matched normoglycaemic group [l 11. Although marked pharmacological hyperinsulinaemia may increase transcapillary albumin escape in vivo [31], a pathophysiological role for insulin in albuminuria remains speculative. We did, however, record an independent contribution from insulin resistance (inverse association) to albumin/creatinine ratios, which would if anything detract from a role for insulin resistance in microalbuminuria. In general it would appear that microalbuminuria is more often a concomitant of established cardiovascular disease or clearly identified risk factors for future cardiovascular disease such as hypertension, diabetes or ageing. It is uncertain whether albuminuria could be an independent predictor of future vascular events in the absence of these other factors and it has recently been reported to accompany acute myocardial infarction [32]. Albuminuria is certainly an independent predictor of subsequent cardiovascular mortality in hypertensive [3], diabetic [4,5] and elderly [7] populations. The relative increase in the number of abnormal ECGs in those with microalbuminuria in our study is in keeping with the concept that albuminuria is a marker of atherosclerosis. The mechanism for such an association remains unclear. Abnormalities of important cardiovascular risk factors such as LDL composition, HDL cholesterol, fibrinogen and blood pressure have been recorded in association with albuminuria in insulin dependent diabetes mditus (IDDM) [12,13,33], and it was of interest that in addition to altered LDL composition, we recorded reduced HDL cholesterol and increased fibrinogen concentrations in the small microalbuminuric with elevated group albumin/creatinine rates as well as linear associations with blood pressure and fibrinogen in those with detectable albuminuria, although clearly such minor disturbances of lipid metabolism and coagulation are unlikely to account in full for the link with coronary heart disease. Microalbuminuria is associated with increased transcapillary escape rates of albumin in diabetes and essential hypertension [34] and it has been suggested that this could reflect a generalised vasculopathy secondary to endothelial damage [35]. The consequence of such a disorder could be
disordered haemostasis and progressive atherosclerosis if endothelial fibrinolytic activity was compromised and if loss of endothelial integrity also led to enhanced transvascular escape of atherogenic macromolecules such as modified LDL cholesterol. Microalbuminuria might also reflect altered anionic charge of the glomerular basement membrane due to loss of heparan sulphate from both glomerular and arterial intimal cells [35], although these attractive hypotheses require much clarification. In our predominantly normoglycaemic population we did not find an independent association between smoking and microalbuminuria, supporting findings of a recent study in IDDM [36]. Microalbuminuria appears uncommon in the absence of established cardiovascular disease or other recognised cardiovascular risk factors. Although widespread screening is therefore not indicated, routine measurement in diabetic and elderly subjects might be justified on the grounds that microalbuminuria appears to be a powerful independent predictor of cardiovascular mortality in these groups who are already at excess risk of cardiovascular disease [6,7,30]. The demonstration of microalbuminuria in diabetic and elderly populations could discriminate those at greatest risk. A case might also be made for screening for microalbuminuria in those with essential hypertension on the grounds that it may be an early marker of general&d vascular disease [34,35] and clinically evident macroproteinuria is an independent risk factor for mortality in essential hypertension [3]. Acknowledgements We are grateful to Mrs. L. Ashworth and colleagues for estimation of insulin and urinary albumin, and to the British Diabetic Association for assistance with these studies. References de Wardener, H.E., Tests of Glomerular Functional Integrity and Proteinuria: The Kidney. An Outline of Normal and Abnormal Function, 5th Edn., Churchill Livingstone, Edinburgh, 1985, pp. 36-55. Kannel, W.B., Stampfer, M.J., Castelli, W.P. and Verter J., The prognostic significance of proteinuria: The Framingham study, Am. Heart J., 108 (1984) 1347.
80 3
4
5
6
I
8
9
10
11
12
13
14
15
16
17
18
Bulpitt, C.J., Beilihn, L.J., Clifton, P., Coles, E.C., Dollery, CT., Gear, J.S.S., Harper, G.S., Johnson, B.F. and Munroe-Faure, A.D., Risk factors for death in treated hypertensive patients. Report from the DHSS hypertension care computing project, Lancet, ii (1979) ,l34. Borch-Johnsen, K., Andersen, P.K. and Deckert, T., The effect of proteinuria on relative mortality in Type 1 (insulin dependent) diabetes mellitus, Diabetologia, 28 (1985) 590. Nelson, R.G., Pettitt, D.J., Carraher, M.J., Baird, M.J. and Knowler, W.C., The effect of proteinuria on mortality in NIDDM, Diabetes, 37 (1988) 1499. Yudkin, J.S., Forrest, R.D. and Jackson, C.A., Microalbuminuria as a predictor of vascular disease in non-diabetic subjects. Islington Diabetes Survey, Lancet, ii (1988) 530. Damsgaard, E.M., Froland, A., Jorgensen, O.D. and Mogensen C.E., Microalbuminuria as predictor of increased mortality in elderly people, Br. Med. J., 300 (1990) 297. Watts, G.F., Morris, R.W., Khan, K. and Polak, A., Urinary albumin excretion in healthy adult subjects: reference values and some factors affecting their interpretation, Clin. Chim. Acta, 172 (1988) 191. Dales, L.G., Friedman, G.D., Siegelaub, A.B., Seltzer, CC. and Ury, H.K., Cigarette smoking habits and urine characteristics, Nephron, 20 (1978) 163. Parving, H.-H., Jensen, H., Mogensen, C.E. and Evrin, P.E., Increased urinary albumin excretion rate in benign essential hypertension, Lancet, i, (1974) 1190. Damsgaard, E.M. and Mogensen, C.E., Microalbuminuria in elderly hyperglycaemic patients and controls, Diabetic Med., 3 (1986) 430. Winocour, P.H., Durrington, P.N., Ishola, M., Anderson, D.C. and Cohen, H., Influence of proteinuria on vascular disease, blood pressure and lipoproteins in insulin dependent diabetes mellitus, Br. Med. J., 294 (1987) 1648. Jensen, T., Stender, S. and Deckert, T., Abnormalities in plasma concentrations of lipoproteins and fibrinogen in type 1 (insulin-dependent) diabetic patients with increased urinary albumin excretion, Diabetologia, 31 (1988) 142. Mulhauser, I., Sawicki, P. and Berger, M., Cigarette smoking as a risk factor for macroproteinuria and proliferative retinopathy in type 1 (insulin-dependent) diabetes, Diabetologia, 29 (1986) 500. Soeldner, J.S. and Slone, D., Critical variables in the radioimmunoassay of serum insulin using the double antibody technic, Diabetes, 14 (1965) 771. Matthews, D.R., Hosker, J.P., Rudenski, A.S., Naylor, B.A., Treacher, D.F. and Turner, R.C., Homeostasis model assessment: insulin resistance and B-cell function from fasting plasma glucose and insulin concentrations in man, Diabetologia, 28 (1985) 412. Friedewald, W.T., Levy, R. and Fredrickson, D.S., Estimation of serum low density lipoprotein without use of preparative ultracentrifuge, Chn. Chem., 18 (1972) 499. Clauss, A., Gerinnungsphysiologische schnell methode
19
20 21
22
23
24
25
26
27 28
29
30
31
32 33
zur bestimmung des tibrinogens, Acta Haematol., 17 (1957) 237. Christensen, C.K. and Orskov, H., Rapid screening PEG radioimmunoassay for quantification of pathological microalbuminuria, Diabetic Nephrop., 3 (1984) 92. Marshall, S.M., Albumin Excretion in Diabetes Mellitus. MD Thesis, University of Glasgow, 1990. Cohen, D.L., Close, C.F. and Viberti, G.C., The variability of overnight urinary albumin excretion in insulindependent diabetic and normal subjects, Diabetic Med., 4 (1987) 437. Sharp, CL., Butterfield, W.J.H. and Keen, H., Diabetes survey in Bedford 1962, Proc. Roy. Sot. Med., 57 (1964) 193. Mann, J.I., Lewis, B., Shepherd, J., Winder, A.F., Fenster, S., Rose, L. and Morgan, B., Blood lipid distribution, prevalence and detection in Britain, Br. Med. J., 296 (1988) 1701. Shaper, A.G., Pocock, S.J., Walker, M., Cohem, N.M., Wale, C.J. and Thomson, A.G., British Regional Heart Study: cardiovascular risk factors in middle-aged men in 24 towns, Br. Med. J., 283 (1981) 179. Haffner, SM., Stem, M.P., Gruber, K.K., Hazuda, H.P., Mitchell, B.D. and Patterson, J.K., Microalbuminuria. Potential marker for increased cardiovascular risk factors in nondiabetic subjects? Arteriosclerosis, IO (1990) 727. Haffner, SM., Mitchell, B.D., Pugh, J.A., Stem, M.P., Kozlowski, M.K., Hazuda, H.P., Patterson, J.K. and Klein, R., Proteinuria in Mexican Americans and nonHispanic whites with NIDDM, Diabetes Care, 12 (1989) 530. Hutchinson, A.S. and Paterson, K.R., Collecting urine for microalbumin assay, Diabetic Med., 5 (1988) 527. Price, D.A., Fielding, B.A., Davies, A.G. and Postlethwaite, R.J., Short term variability of urinary albumin excretion in normal and diabetic children, Diabetic Nephrop., 4 (1985) 169. Viberti, G.C., Hill, R.D., Jarrett, R.J., Argyropoulos, A., Mahmud, U. and Keen, H., Microalbuminuria as a predictor of clinical nephropathy in insulin-dependent diabetes mellitus. Lancet, i (1982) 1430. Jarrett, R.J., Viberti, G.C., Argyropoulos, A., Hill, R.D., Mahmud, U. and Murrells, T.J., Microalbuminuria predicts mortality in non-insulin dependent diabetes, Diabetic Med., 1 (1984) 17. Nestler, J.E., Barlascini, C.O., Tetrault, G.A., Fratkin, M.J., Clore, J.N. and Blackard, W.G., Increased transcapillary escape rate of albumin in nondiabetic men in response to hyperinsulinaemia, Diabetes, 39 (1990) 1212. Gosling, P. and Reynolds, T.M., Early diagnosis of acute myocardial infarction (Letter), Br. Med. J., 302 (1991) 53 Winocour, P.H., Durrington, P.N., Bhatnagar, D., Ishola, M., Mackness, M. and Arrol, S., The influence of early diabetic nephropathy on very low density lipoprotein (VLDL), intermediate density lipoprotein (IDL), and low density lipoprotein (LDL) composition, Atherosclerosis, 89 (1991) 49.
81 34
35
Feldt-Rasmussen, B., Increased transcapillary escape rate of albumin in type 1 (insulin dependent) diabetic patients with microalbuminuria, Diabetologia, 29 (1986) 282. Deckert, T., Feldt-Rasmussen, B., Borch-Johnsen, K., Jensen, T. and Kofoed-Enevoldsen, A., Albuminuria reflects widespread vascular damage. The Steno hypothesis, Diabetologia, 32 (1989) 219.
36
Ardron, M., Macfarlane, LA., Martin, P., Walton, C., Day, J., Robinson, C. and Calverly, P., Urinary excretion of albumin, alpha-I-microglobulin, and N-acetyl-B-Dglucosaminidase in relation to smoking habits in diabetic and non-diabetic subjects, J. Diabetic Complic., 3 (1989) 154.
93 ( 1992) 71-8 1 0 1992 Elsevier Scientific Publishers Ireland, Ltd. All rights reserved 0021-9150/92/$05.00
Atherosclerosis,
Printed and Published in Ireland
ATHERO 04784
Microalbuminuria
and associated cardiovascular risk factors in the community
Peter H. Winocoura’d, Jane O.E. Harlandayd, Jill P. Millar”, Michael F. Lakerb’d and K. George M.M.. Albertia>d The Departments of ‘Medicine. bCIinical Biochemistry, and ‘Primary Health Care and ‘Human Metabolism and Diabetes Research Centre, The Medical School, University of Newcastle upon Tyne. Framlington Place, Newcastle upon Tyne. NE2 4HH (UK.)
(Received 26 July, 1991) (Revised, received 24 December, 1991) (Accepted 30 December 1991)
Summary The prevalence of microalbuminuria and relationship to cardiovascular risk factors was examined in a cross-sectional community survey of cardiovascular risk factors. Microalbuminuria (when classified as albumin concentration 720 pg/ml) was present in 6.3% of subjects but in conjunction with an albumin/creatinine ratio > 3.5 in only 2.2%. Diastolic blood pressure, prevalence of abnormal electrocardiographs, and to a lesser extent systolic blood pressure and fibrinogen concentration, were greater in those with albuminuria concentrations >20 &ml. The strongest positive univariate correlates of albumin/ creatinine ratios in those with detectable albuminuria were age, fibrinogen, blood pressure, total- and low density lipoprotein- (LDL) cholesterol, apo B and alcohol intake, whereas fasting insulin and insulin resistance were inversely correlated. Multiple regression analysis revealed that age, gender, systolic blood pressure and insulin resistance independently accounted for 37% of the variability in albumin/creatinine ratios. When those 10 subjects with microalbuminuria and albumin/creatinine ratios 73.5 were matched with 20 with normoalbuminuria for age, gender and body mass index, the microalbuminuric subjects had significantly lower LDL cholesteroYapo B ratios and a tendency to lower high density lipoprotein (HDL) cholesterol and HDL cholesterohapo Al ratios. Microalbuminuria is uncommon in the general population, and is related to ageing, blood pressure and other vascular risk factors. It may reflect the presence of established cardiovascular disease.
Key words: Microalbuminuria; Lipoproteins
Cardiovascular
disease; Diabetes; Hypertension;
Elderly; Fibrinogen;
Introduction Correspondence to: Dr Peter H. Winocour, Lecturer in Medicine, Freeman Hospital, Freeman Road, High Heaton, Newcastle upon Tyne, NE7 7DN, U.K.
Proteinuria detected by conventional semiquantitative test-strip testing predominantly reflects
12 urinary albumin excretion which is usually the consequence of renal glomerular (or less commonly tubular) disease [ 11. Proteinuria in diabetic and hypertensive populations is associated with an excess mortality from cardiovascular disease [2-51. More recently it has been suggested that so-called microalbuminuria measured by sensitive radiochemical techniques, and by definition not detected by the usual test strips, might also be associated independently with cardiovascular disease [6,7]. Urinary albumin excretion may be modified by posture or exercise, or increased in the presence of urinary infection or parenchymal renal disease [ 11. Intra-individual day-to-day variability of albumin excretion [8] has also restricted assessment of an independent role for albuminuria as a risk factor for cardiovascular disease (CVD), but can be minimised by the use of timed samples, or correction for urine creatinine. The prevalence of microalbuminuria in healthy Caucasian adults is not known, and the suggestion that in diabetic and hypertensive populations albuminuria may be associated with vascular risk factors [9-141 such as smoking, blood pressure, lipids, blood glucose, insulin, fibrinogen and ageing has yet to be explored in the general population. These questions have been addressed in the course of a community based CVD risk factor screening programme. Subjjts
and Metlmds
Ethical approval for the study was obtained from the Joint Ethics Committee of the University of Newcastle upon Tyne and the Newcastle Health Authority. Seven hundred self-selected patients from one urban group general practice responded to an advertisement inviting adults aged 25-65 to take part in a cardiovascular risk factor health screen. This report is based on 447 individuals in whom urinary albumin was estimated. Patients attended during the period September 1988-March 1990. Following an overnight fast, venous blood was drawn for measurement of plasma glucose, glycated haemoglobin (HbA,), fibrinogen, and serum concentrations of insulin, cholesterol, triglycerides, apolipoproteins B and Al. In addi-
tion plasma glucose and serum insulin were measured 1 and 2 h following ingestion of a 75-g oral glucose monohydrate load. High density lipoprotein (HDL) cholesterol was also measured 2 h after ingestion of the glucose load. Venous plasma glucose (separated immediately after sampling) was measured on a continuous flow autoanalyser (Technicon AA2, Technicon Instruments, Basingstoke, U.K.; inter-assay coefficient of variation (CV) l-3%). HbA, was measured by electroendosmosis; reference range 5.0-8.0%. Serum insulin was measured by radioimmunoassay, a modified version of the method of Soeldner and Slone [15], using an in-house antibody, inter-assay CV 6.8-7.5%. Insulin resistance (IR) was calculated from fasting blood glucose and serum insulin concentrations using the computer-solved homeostasis model assessment (HOMA) method described by Matthews et al. [16]: IR = fasting insulin/22.5e-In fastins glucose. Serum cholesterol was measured by a cholesterol oxidase method on a Cobas Bio centrifugal fast analyser (Roche Products Limited, Welwyn Garden City, U.K.) using a commercial kit (number 725242) supplied by BCL &ewes, U.K.). Between batch CVs ranged between 1.5% to 2.2% for cholesterol concentrations ranging from 3.9 to 10.3 mm01 1-l. Serum triglycerides were measured by a lipase-glycerol kinase method without correction for free glycerol on the Cobas Bio centrifugal analyser using a commercial kit (both supplied by Roche Products Limited, Welwyn Garden City, U.K.). Between batch CV was 3.0% to 3.2% with triglyceride standards ranging from 2.31 to 3.47 mm01 1-l. HDL was isolated after precipitation of apo B-containing lipoproteins with heparin and manganese. Cholesterol within HDL was measured by the enzymatic method previously described. The between batch CV at a mean HDL cholesterol of 0.97 mm01 1-l was 3.6O/o.Low density lipoprotein (LDL) cholesterol was calculated by the Friedewald formula [17] (total serum cholesterol (HDL cholesterol + fasting serum triglycerides/ 2.2) mm01 1-l). The apoproteins B and Al were measured by immunonephelometry using a Behring laser nephelometer, and Behring standards and antisera (Hoechst U.K. Ltd., Hounslow, U.K.). Between batch CV for a mean apo B of 0.88 g 1-l was 4.9% and for a mean apo Al of 0.92 g 1-l was 5.6%
73
Plasma fibrinogen was measured by the Clauss technique (between batch coefficient of variation 5%) [18]. Body mass index (BMI) was calculated as Quetelet’s index: BMI = weight (kg)/height2 (m2). Alcohol and tobacco consumption, and patterns of exercise were ascertained by direct enquiry, and the presence or absence of vascular disease assessed by a resting 12 lead electrocardiogram and the WHO Rose questionnaire. ECGs were coded by the Minnesota coding system, and the following codes compatible with ischaemic heart disease were designated abnormal: l-l, l-2, 4-1, 5-1, 8-1, 8-3. Urinary albumin estimation
Each individual brought a specimen of urine passed on waking on the morning of attendance. Samples were tested for glucose, blood, protein, and nitrite (Multistix, Boehringer, U.K.), and a mid-stream sample was sent for microbiological evaluation when positive for any one of the latter three, or when urinary tract infection was suspected from the history. A repeat early morning urine sample was sent in those cases following treatment of the urinary infection or if women were menstruating at the time of the initial collection, and a further sample was sent for albumin and creatinine measurement after detection of a sterile mid-stream urine sample. Initially, albumin and creatinine concentrations were measured by immunonephelometric (Behring antisera and reagents, Behring Diagnostics, Hoechst, Middlesex, U.K.) and Jaffe methods, respectively. The lower limit of sensitivity of albumin concentration estimation by this method is 12 &ml and the between batch variability for urine with a mean albumin concentration of 38.0 &ml was 6%. A repeat timed overnight urine collection was carried out in those 10 cases whose urinary albumin concentration was greater than 20 pg/rnl and in addition whose albumin/creatinine ratio was greater than 3.5, and in 20 normoalbuminuric individuals who were matched sequentially for sex, age (to within 5 years) and body mass index (to within 2 kg/m2 units). All normoalbuminuric cases initially had urinary albumin concentrations close to the limit of sensitivity of the assay (12-14 &ml). The first 2 normoalbuminuric subjects who
fulfilled these criteria were selected from the list of normoalbuminuric patients. On this occasion urine albumin was quantitated by a more sensitive radioimmunoassay method (sensitivity 0.5 &ml, between batch variability 5%) [19]. Intraindividual day-to-day variability in overnight albumin excretion has been estimated at 8-40% by our laboratory [20] and by others [21]. Data analysis
Information was collected on an Amstrad PC4535 and analysed using the statistical package for the social sciences (SPSS). Data are presented as mean (S.E., range). Differences between those with urinary albumin concentrations above or below 20 &ml, and between the normoalbuminuric subjects above or below the detection limit of the assay were, compared by Student’s t-test, as were those with microalbuminuria when compared with the matched normoalbuminuric control group. After establishing that those with normoalbuminuria (< 20 &ml) above and below the detection limit of the assay were comparable with respect to demographic and biological variables, univariate associations in those 74 individuals with urinary albumin concentrations above the detection limit of the assay were examined by Pearson’s correlation coefficient. Stepwise multiple regression analyses were also employed in this group to examine the independent associations between albuminuria and the following variables: cigarette and alcohol consumption, documented hypertension and antihypertensive therapy (including diuretics), fasting and post-glucose challenge plasma glucose and serum insulin levels, age, a first degree relative with hypertension and or diabetes, blood pressure, HbA,, total-, LDL- and HDL-cholesterol, serum triglycerides, apo Al, apo B, fibrinogen and haemoglobin concentrations, and the ratios of LDL/HDL and HDL cholesterol/ape A 1. Results
Of the 447 individuals on whom data were available, there were 209 men and 238 women, aged from 25 to 70 years (mean (S.E.) 47.3 (0.5) years)), average body mass index 25.1 (0.2) kg/m’. Mean values of other measured variables were as follows: alcohol consumption 9 (1) units/week, fasting and
74 2-h plasma glucose 5.2 (0.1) and 5.0 (0.1) mmol/l, respectively, HbAi 6.2 (O.l)O/o,fasting serum insulin 8.8 (0.4) munits/l, total serum and LDL cholesterol 5.9 (0.1) and 3.9 (0.1) mmol/l, respectively, total serum triglycerides 1.45 (0.04) mmolil, HDL cholesterol 1.41 (0.02) mmol/l, apo Al 1.5 (0.01) g/l, apo B 1.3 (0.02) g/l, fibrinogen 3.1 (0.04) g/l, haemoglobin 14.2 (0.1) g/dl, blood pressure (systolic/diastolic) 121 (1)/76 (1) mmHg. The 2-h HDL cholesterol value correlated closely with fasting values (rS = 0.97) in 100 subjects whose fasting HDL cholesterol values ranged between 0.7 and 2.0 mm01 1-i (mean value 1.5 mm01 1-l). There were 68 cigarette smokers whose average consumption was 14.6 (0.9)/day. The sample of 447 was comparable to the original group of 700 with respect to the prevalence of vascular risk factors and apparent vascular disease: diabetes (1.3%
vs. 1.4%, respectively), hypertension (systolic/diastolic > 160 and/or 95 mmHg respectively, or current antihypertensive medication) in 10.3% vs. lO.O%, current smokers 15.2% vs. 15.0%, hypercholesterolaemia (>6.5 mmol/l) in 32% vs. 27%, and historical evidence of cardiovascular disease from the WHO Rose questionnaire in 5.1% and 4.3%, respectively. All but 10 of the 447 were Caucasian, 99 had previously been told by a doctor they had hypertension, of whom 27 received antihypertensive medication, including 17 on diuretic therapy. Aspirin was taken on average more than once a week by 16 individuals. Angina was reported by 18 cases, intermittent claudication by 4, and 5 had a history of previous myocardial infarction. There was a first degree relative with hypertension in 146 cases and of diabetes in 52. Impaired glucose tolerance was present in 11
TABLE 1 URINE ALBUMIN EXCRETION AND CARDIOVASCULAR
RISK FACTORS IN 477 MALE AND FEMALE
Figures are mean (S.E.), range.
Number (M/F) Age (years) Body mass index (kg/m2) Alcohol (units/week) Smokers (%) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Fasting blood glucose (mmobl) 1 h blood glucose (mmoiil) 2 h blood glucose (mmoliI) HbA, (%) Fasting insulin (munitsil) 1 h insulin (munitsil) 2 h insulin (munitsil) Insulin resistance Total serum cholesterol (mmobl) Total serum triglycerides (mmol/l) LDL cholesterol (mmolil) HDL cholesterol (mmol/l) APO Al W) APO B W) Fibrinogen (gil) LDL cholesterol/ape B HDL cholesteroYapo Al LDLiHDL cholesterol
Normoalbuminuric (< 12-20 &ml)
Microalbuminuric (>20 &ml)
419 (198221) 47 (l), 25-70 25.1 (0.2) 17.7-45.4 9 (1) O-70 15.3 121 (1) 82-189 76 (I), 54-105 5.2 (O.l), 3.0-11.5 6.6 (O.l), 2.3-19.5 5.0 (O.l), 1.7~20.8 6.2 (0. l), 4.0-9.8 8.8 (0.4), 1.0-46.8 67.1 (2.0), 9.0-254.0 34.4 (1.5), 1.5-280.2 0.23 (O.Ol), 0.02-2.09 5.9 (0. l), 2.7- 10.3 1.5 (0.1) 0.4-6.8 3.9 (O.l), 1.0-7.9 1.4(0. l), 0.6-3.0 1.5 (O.l), 0.6-2.5 1.3 (0.1). 0.4-2.3 3.1 (0.1) 0.7-7.2 1.17 (0.01) 0.38-2.18 0.36 (O.Ol), 0.13-0.75 3.0 (0. I), 0.6-7. I
28 (11:17) 49 (2), 25-65 25.3 (0.7), 19.0-32.0 12 (3). O-80 14.3 127 (3) 102-174’ 81 (2). 63-102** 5.3 (0.1). 4.1-6.1 7.0 (0.4), 3.1-I 1.9 5.6 (0.3) 2.7-10.5 6.4 (0.2) 4.8-9.8 8.7 (1.2), 0.5-26.3 75.8 (6.Q 26.9-174.9 47.0 (9.0), 6.5-260.0 0.23 (0.03) 0.01-0.72 5.9 (0.2) 4.0-8.2 1.2 (0. I), 0.6-2.5 3.9 (0.2) 1.9-6.5 1.4 (0.1) 0.6-2.2 1.6 (0.1) 0.8-2.1 1.3 (0.1). 0.8-1.9 3.4 (0.2) 2.2-5.5, 1.14 (0.02), 0.89-1.30 0.36 (O.Ol), 0.26-0.48 3.0 (0.4), 1.1-10.8
Differences between groups *P < 0.1, **P < 0.05.
SUBJECTS
75 (2.5%) and diabetes in 6 (1.3%) subjects. All were negative for protein on Multistix testing except 3 subjects (2 showed a trace, and in one case 0.1 g/l of proteinuria). Three hundred and seventy-three individuals had albuminuria below the detection limit of the assay, 46 had normoalbuminuria (12-20 &ml) and 28 (6.3%) had microalbuminuria (> 20 &nl). Those with normoalbuminuria above or below the detection limit of the assay were comparable with respect to gender ratio, age, body mass index, consumption of tobacco and alcohol, and biological variables (data not shown). We therefore compared those 28 with microalbuminuria with the remaining 419. Microalbuminuria was associated with significantly higher diastolic blood pressure, and a tendency to somewhat higher levels of systolic blood pressure and plasma fibrinogen. Other biochemical measurements were comparable (Table 1). When
only those 46 with detectable normoalbuminuria were compared with those with microalbuminuria, similar differences in blood pressure and fibrinogen were recorded (data not shown). The prevalence of abnormal ECGs was greater in those with microalbuminuria >20 j&ml (P < 0.05), but there was no clear difference in the prevalence of abnormal glucose tolerance, or use of antihypertensive agents (Table 2). Significant (P < = 0.01) univariate correlates of albuminuria concentrations and the albumin/ creatinine ratios are shown in Table 3. The strongest associations of the albuminkreatinine ratios were with age, systolic blood pressure, total serum and LDL cholesterol, fibrinogen, alcohol, and apo B. Urinary albumin concentrations were correlated with age, systolic and diastolic blood pressure, fibrinogen, serum triglycerides, and apo Al.
TABLE 2 PREVALENCE OF DIABETES, HYPERTENSION, VASCULAR DISEASE AND ANTIHYPERTENSIVE SUBJECTS WITH NORMAL URINARY ALBUMIN CONCENTRATIONS OR MICROALBUMINURIA Normoalbuminuric (< 12-20 &lll) (?I = 419) Previous diagnosis of diabetes (% (number)) Impaired glucose tolerance (% (number)) Diabetes mellitus (% (number)) Previous diagnosis of hypertension (% (number)) Current use of antihypertensive agents (% (number)) Current diuretic use (% (number)) Anginaa (% (number)) Abnormal ECG* (% (number)) Myocardial infarctiona (% (number)) Prevalence of 1st degree relatives with hypertension at any age (% (number)) Prevalence of 1st degree relatives with CVA at any age (% (number)) Prevalence of 1st degree relatives with diabetes mellitus at any age (% (number)) BData generated from WHO Rose Questionnaire. Difference between groups *P < 0.05.
Microalbuminuria ( > 20 &ml) (n = 28)
0.2 (1)
0
2.1 (9)
7.1 (2)
1.4 (6) 22.0 (92)
0 (0) 25.0 (7)
6.2 (26)
3.6 (1)
4.1 (17) 4.1 (17) 2.4 (10) 1.0 (4) 32.5 (136)
0 (0) 3.6 (1) 17.9 (5)* 3.6 (1) 35.6 (IO)
22.7 (95)
25.0 (7)
11.7 (49)
10.7 (3)
(0)
TREATMENT IN
76 TABLE 3 UNIVARIATE PEARSON’S CORRELATION COEFFICIENT (r), OF ALBUMINURIA MEASURES (IN 74 SUBJECTS WITH DETECTABLE ALBUMINURIA) WITH OTHER VARIABLES Correlates
Albumin/ creatinine ratio
Urinary albumin concentration
Age LDL cholesterol Fibrinogen Systolic blood pressure Diastolic blood pressure Total serum cholesterol Apo B Alcohol Triglycerides Fasting blood glucose Fasting insulin Insulin resistance Apo Al HDL cholesterol
0.43*** 0.25* 0.24* 0.24* 0.23* 0.23’ 0.20* -0.20* -0.06 0.12 -0.25* -0.24* 0.13 0.08
0.23* 0.04 0.25* 0.21’ 0.222 -0.03 0.02 0.10 0.20* 0.04 -0.15 -0.14 0.20’ 0.06
‘P < 0.05, ***p < 0.001.
Multiple regression analysis demonstrated that 37% of the variance in the albumin/creatinine ratio could be accounted for by independent influences of age, gender, systolic blood pressure, and insulin resistance (Table 4). Microalbuminuria, in association with an albumin/creatinine ratio >3.5 was present in 10 individuals (9 women) (2.2% prevalence overall). Three were positive (0.1 g/l) for urinary protein on stick testing. Four had previously been told by a doctor that they had hypertension, but only one
was currently receiving treatment (with a beta blocker). None of these received diuretics or aspirin, nor had experienced myocardial infarction, although the only 1 who smoked (1 S/day) suffered from angina and intermittent claudication. One further subject had impaired glucose tolerance, but the rest were normoglycaemic by conventional criteria. Repeat urine albumin estimation on an overnight collection by the more sensitive assay confirmed persistent elevation of both urinary albumin concentration and the albumin/creatinine ratio in 8 cases. However, the overnight albumin excretion rate was greater than 16 &nin in all 10 cases. Average variability in albumin concentration was 13% (range l-20%), whilst that of the albumin/creatinine ratio was 19.7%. When these 10 were compared with the matched group of 20 persistently normoalbuminuric individuals, the major differences between the 2 groups was in the LDL cholesterohapo B ratio and to a lesser extent the l-h post-glucose challenge plasma glucose, the HDL cholesterol concentration, and the HDL cholesterohapo Al ratio (Table 5). In addition, abnormal ECGs were recorded in 50% and 25% of the microalbuminuric and the matched normoalbuminuric groups, respectively (P < 0.1). Dlscnssion We have recorded a low prevalence (6.3%) of microalbuminuria in a community based population who were unselected with respect to vascular disease or the presence of vascular risk factors such as hypertension or diabetes. Although the
TABLE 4 PREDICTOR VARIABLES ALBUMINURIA Predictor variables in model
Age
Gender Systolic blood pressure Insulin resistance Constant
FOR
ALBUMINCREATININE
R2
0.21 0.29 0.34 0.37
RATIOS
IN 74 INDIVIDUALS
Regression coeBicient zt S.E. 0.008 0.210 0.003 -0.169 -0.921
i f f zt f
0.003 0.058 0.002 0.094 0.249
WITH
DETECTABLE
Multiple correlation coefficient
Significance of variable
0.46 0.55 0.58 0.61
0.0014 0.0806 0.0709 0.0772 0.0004
P
77 TABLE 5 CARDIOVASCULAR RISK FACTORS IN 10 SUBJECTS WITH BOTH MICROALBUMINURIA ALBUMINICREATININE RATIO AND IN 20 MATCHED NORMOALBUMINURIC SUBJECTS
AND
RAISED
Results shown as mean (S.E.), range.
Baseline albuminuria (ccg/ml) Baseline albumin/creatinine ratio Repeat overnight albuminuria (pglml) Overnight albumin excretion rate (&mm) Repeat overnight albumin/creatinine ratio Sex ratio (M/F) Age (years) Body mass index (kg/m’) Alcohol (units/week) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Fasting blood glucose (mmolil) 1 h blood glucose (mmolil) 2 h blood glucose (mmolil) HbA, (%) Fasting insulin (munitsfl) 1 h insulin (munitsil) 2 h insulin (munits/l) Insulin resistance Total serum cholesterol (mmolil) Total serum triglycerides (mmol/l) LDL cholesterol (mmol/I) HDL cholesterol (mmol/l) Apo Al (fl) APO B (g/l) Fibrinogen (gil) LDL cholesterohapo B HDL cholesterol/ape A I LDL/HDL cholesterol
Normoalbuminuric group
Microalbuminuric group
< 12.2 (0.1). < 12.0-14.0
39.2 (6.2), 20.0-77.0**** 5.0 (0.3), 3.6-6.4**** 38.2 (7.2), 16.8~90.0**** 32.0 (4.9) 18.0-66.0**** 5.1 (0.2), 3.0-12.0**** I:9 56 (3). 39-65 25.0 (I .O), 20.2-29.6 3 (2) o-15 132 (7) 109-168 83 (3). 69-96 5.3 (0. I), 4.5-5.7 7.3 (0.5). 5.0-9.6* 5.8 (0.6) 3.8-10.5 6.6 (0.5) 5.0-9.9 9.4 (2.5), I .7-26.3 75.9 (14.8), 35.8-174 60.5 (23. I), 9.0-260.0 0.25 (0.07) 0.05-0.72 6.6 (0.3) 5.6-8.2 I .4 (0.2), 0.8-2.5 4.5 (0.4) 2.9-6.5 1.5(0.2), 0.6-2.4’ I .6 (0. I), 0.8-2.0 I.5 (0.1) 1.1-1.9 3.6 (0.3), 2.2-5.2 I.18 (0.03) l.O3-1.30** 0.36 (0.02) 0.29-0.46* 3.73 (0.84). 1.20-10.80
4.1 (O.S), 1.5-8.1 3.5 (0.4), 1.5-8.3 0.7 (O.I), 0.3-1.2 2:18 56 (2), 39-65 24.5 (0.5), 21.0-29.7 5 (2), O-30 124 (5), 84-189 78 (3), 60- IO3 5.1 (0. I), 4.2-6.0 6.7 (0.6). 3.3-13.5 4.7 (O.S), 2.4-12.1 6.4 (0.2), 4.8-7.9 6.4 (0.9), I .4-18.0 66.0 (8.0), 13.6-145.5 30.1 (6.3), 3.3-107.0 0.17 (0.03) 0.04-0.51 6.6 (0.2), 4.9-8.5 1.2 (O.l), 0.6-2.5 4.3 (0.2), 3.2-6.1 1.8 (O.l), 1.1-2.5 1.8 (O.I), 1.3-2.3 1.3(0. l), 0.9-2. I 3.2 (0.1). 2.3-4.6 1.27 (0.03) 1.03-1.54 0.39 (0.01) 0.29-0.46 2.56 (0.16) 1.28-4.06
Differences between groups not significant unless *P < 0.1, **P < 0.05, **** P < 0.001.
population under study were self selected, they were representative of the larger group of 700 subjects in terms of their vascular risk profile. The prevalence of diabetes and hypercholesterolaemia was also very similar to that recorded in previous unselected population studies throughout the U.K. [22,23]. The prevalence of smoking and hypertension in the U.K. in general appears more variable in different regions, but in certain districts throughout the country is broadly the same a’sthat recorded in the present investigation [23,24]. We would not claim that our reported prevalence of microalbuminuria is necessarily totally represen-
tative of the general population, and in the light of our associated findings it is likely that microalbuminuria could be detected more frequently in populations with a higher prevalence of hypertension. The present findings are in accordance with one previous study of 127 adults and children in the U.K.[8], but much lower than the lo-20% recorded in other reports where diabetes and cardiovascular disease was over-represented [6,1 l] or where the study population was considerably older [I 11. Haffner et al. [25] reported a 13% prevalence of microalbuminuria in 3 16 non-diabetic Mexican
78 Americans who were similar in other respects to our population with the exception of the fact that they were on average more obese and hyperinsulinaemic. Racial differences [26], selection bias, and the fact that our urine samples were assayed for albumin within 1 week of collection whereas Haffner et al. [25] processed their samples after an average delay of 12 months, may also have contributed to the discrepancy. Although the vast majority of subjects had albuminuria concentrations below the sensitivity of the routine assay (12 @ml) this test appears adequate for the purposes of screening the population, and differentiating normoalbuminuric from microalbuminuric samples. When prevalence of microalbuminuria was determined from elevations in both albumin/ creatinine ratios (> 3.5) and urinary albumin concentrations (>20 &ml), in order to reduce any bias as a result of errors in collection, only 10 of the 28 with albumin concentrations >20 &ml fultilled this category. Inclusion of false positives amongst the 28 cases in the initial comparison (Table 1) could have minimised any apparent differences in cardiovascular risk factors that were directly attributable to ‘microalbuminuria’. We therefore specifically elected to compare also those 10 with elevated urinary albumin concentration and albumin/creatinine ratios with 20 distinctly normoalbuminuric subjects in order to confirm that the differences in risk factors were comparable in the more selected microalbuminuric group. This procedure also reduced any potential confounding influence of gender, age or body mass. Alterations in lipoprotein composition were suggested, but the relatively small sample size and impact of modest microalbuminuria on other vascular risk factors may be why only nonsignificant differences in blood pressure, fibrinogen and HDL cholesterol were recorded. It is recognised that albuminuria is variable within individuals on a day-to-day basis [8,20,21] but it has previously been suggested that this may be lessened by the use of overnight collections 1271, and minimised further by the use of albumin/ creatinine ratios [27,28]. The extent of miscategorisation of subjects on the basis of single urine collections has been shown to be negligible unless albumin excretion rates exceed 30 kg/mm (equivalent to an albumin/creatinine ratio of >3.5)
[20,27]). In the present study 2 of 10 subjects did not fulfil both criteria for microalbuminuria on an overnight collection although their matched controls were persistently normoalbuminuric. However, the albumin excretion rate was > 16&ml in all 10 cases and this is well outside previously quoted reference ranges for healthy subjects [8,21]. Furthermore those previous studies which demonstrated that microalbuminuria acted as a prognostic marker for diabetic nephropathy and cardiovascular disease were based on albumin excretion rates determined from single overnight collections [6,7,29,30]. The dominant determinant of detectable albuminuria in our study was age, which is in accord with Damsgaard et al. who have recorded the highest prevalence of microalbuminuria (1 j-20%) in healthy subjects whose average age was 67 [111. Yudkin et al. [6] found no correlation between age and albuminuria in an older (average age >60 years) selected group of subjects with a high prevalence of vascular disease, and an inverse association between age and albuminuria has even been recorded and attributed to increased urinary flow rates in older subjects, although both children and adults were included in this latter study [8]. Albumin/creatinine ratios in the present study were also independently determined by gender, which could in part be accounted for by differences in muscle bulk in men and women. Blood pressure is closely associated with albuminuria in diabetic [ 111,hypertensive [6,10] or elderly [6,11] populations and this relationship was apparent in our study, although not in healthy normotensive individuals studied by Watts et al. [8]. Only a small proportion of subjects in our study had hypertension according to WHO criteria, however. Fasting plasma glucose and insulin concentrations were only weak correlates of albuminuria in our population where the prevalence of impaired glucose tolerance/diabetes was similar to that expected from a random population sample in the United Kingdom [22]. Haffner et al. [25] found that insulin was more conspicuously associated with albuminuria than in the present study but it was unclear how common impaired glucose tolerance was in their obese hyperinsulinaemic population. Impaired glucose tolerance and/or diabetes is clearly associated with greater
79 albuminuria than an appropriate matched normoglycaemic group [l 11. Although marked pharmacological hyperinsulinaemia may increase transcapillary albumin escape in vivo [31], a pathophysiological role for insulin in albuminuria remains speculative. We did, however, record an independent contribution from insulin resistance (inverse association) to albumin/creatinine ratios, which would if anything detract from a role for insulin resistance in microalbuminuria. In general it would appear that microalbuminuria is more often a concomitant of established cardiovascular disease or clearly identified risk factors for future cardiovascular disease such as hypertension, diabetes or ageing. It is uncertain whether albuminuria could be an independent predictor of future vascular events in the absence of these other factors and it has recently been reported to accompany acute myocardial infarction [32]. Albuminuria is certainly an independent predictor of subsequent cardiovascular mortality in hypertensive [3], diabetic [4,5] and elderly [7] populations. The relative increase in the number of abnormal ECGs in those with microalbuminuria in our study is in keeping with the concept that albuminuria is a marker of atherosclerosis. The mechanism for such an association remains unclear. Abnormalities of important cardiovascular risk factors such as LDL composition, HDL cholesterol, fibrinogen and blood pressure have been recorded in association with albuminuria in insulin dependent diabetes mditus (IDDM) [12,13,33], and it was of interest that in addition to altered LDL composition, we recorded reduced HDL cholesterol and increased fibrinogen concentrations in the small microalbuminuric with elevated group albumin/creatinine rates as well as linear associations with blood pressure and fibrinogen in those with detectable albuminuria, although clearly such minor disturbances of lipid metabolism and coagulation are unlikely to account in full for the link with coronary heart disease. Microalbuminuria is associated with increased transcapillary escape rates of albumin in diabetes and essential hypertension [34] and it has been suggested that this could reflect a generalised vasculopathy secondary to endothelial damage [35]. The consequence of such a disorder could be
disordered haemostasis and progressive atherosclerosis if endothelial fibrinolytic activity was compromised and if loss of endothelial integrity also led to enhanced transvascular escape of atherogenic macromolecules such as modified LDL cholesterol. Microalbuminuria might also reflect altered anionic charge of the glomerular basement membrane due to loss of heparan sulphate from both glomerular and arterial intimal cells [35], although these attractive hypotheses require much clarification. In our predominantly normoglycaemic population we did not find an independent association between smoking and microalbuminuria, supporting findings of a recent study in IDDM [36]. Microalbuminuria appears uncommon in the absence of established cardiovascular disease or other recognised cardiovascular risk factors. Although widespread screening is therefore not indicated, routine measurement in diabetic and elderly subjects might be justified on the grounds that microalbuminuria appears to be a powerful independent predictor of cardiovascular mortality in these groups who are already at excess risk of cardiovascular disease [6,7,30]. The demonstration of microalbuminuria in diabetic and elderly populations could discriminate those at greatest risk. A case might also be made for screening for microalbuminuria in those with essential hypertension on the grounds that it may be an early marker of general&d vascular disease [34,35] and clinically evident macroproteinuria is an independent risk factor for mortality in essential hypertension [3]. Acknowledgements We are grateful to Mrs. L. Ashworth and colleagues for estimation of insulin and urinary albumin, and to the British Diabetic Association for assistance with these studies. References de Wardener, H.E., Tests of Glomerular Functional Integrity and Proteinuria: The Kidney. An Outline of Normal and Abnormal Function, 5th Edn., Churchill Livingstone, Edinburgh, 1985, pp. 36-55. Kannel, W.B., Stampfer, M.J., Castelli, W.P. and Verter J., The prognostic significance of proteinuria: The Framingham study, Am. Heart J., 108 (1984) 1347.
80 3
4
5
6
I
8
9
10
11
12
13
14
15
16
17
18
Bulpitt, C.J., Beilihn, L.J., Clifton, P., Coles, E.C., Dollery, CT., Gear, J.S.S., Harper, G.S., Johnson, B.F. and Munroe-Faure, A.D., Risk factors for death in treated hypertensive patients. Report from the DHSS hypertension care computing project, Lancet, ii (1979) ,l34. Borch-Johnsen, K., Andersen, P.K. and Deckert, T., The effect of proteinuria on relative mortality in Type 1 (insulin dependent) diabetes mellitus, Diabetologia, 28 (1985) 590. Nelson, R.G., Pettitt, D.J., Carraher, M.J., Baird, M.J. and Knowler, W.C., The effect of proteinuria on mortality in NIDDM, Diabetes, 37 (1988) 1499. Yudkin, J.S., Forrest, R.D. and Jackson, C.A., Microalbuminuria as a predictor of vascular disease in non-diabetic subjects. Islington Diabetes Survey, Lancet, ii (1988) 530. Damsgaard, E.M., Froland, A., Jorgensen, O.D. and Mogensen C.E., Microalbuminuria as predictor of increased mortality in elderly people, Br. Med. J., 300 (1990) 297. Watts, G.F., Morris, R.W., Khan, K. and Polak, A., Urinary albumin excretion in healthy adult subjects: reference values and some factors affecting their interpretation, Clin. Chim. Acta, 172 (1988) 191. Dales, L.G., Friedman, G.D., Siegelaub, A.B., Seltzer, CC. and Ury, H.K., Cigarette smoking habits and urine characteristics, Nephron, 20 (1978) 163. Parving, H.-H., Jensen, H., Mogensen, C.E. and Evrin, P.E., Increased urinary albumin excretion rate in benign essential hypertension, Lancet, i, (1974) 1190. Damsgaard, E.M. and Mogensen, C.E., Microalbuminuria in elderly hyperglycaemic patients and controls, Diabetic Med., 3 (1986) 430. Winocour, P.H., Durrington, P.N., Ishola, M., Anderson, D.C. and Cohen, H., Influence of proteinuria on vascular disease, blood pressure and lipoproteins in insulin dependent diabetes mellitus, Br. Med. J., 294 (1987) 1648. Jensen, T., Stender, S. and Deckert, T., Abnormalities in plasma concentrations of lipoproteins and fibrinogen in type 1 (insulin-dependent) diabetic patients with increased urinary albumin excretion, Diabetologia, 31 (1988) 142. Mulhauser, I., Sawicki, P. and Berger, M., Cigarette smoking as a risk factor for macroproteinuria and proliferative retinopathy in type 1 (insulin-dependent) diabetes, Diabetologia, 29 (1986) 500. Soeldner, J.S. and Slone, D., Critical variables in the radioimmunoassay of serum insulin using the double antibody technic, Diabetes, 14 (1965) 771. Matthews, D.R., Hosker, J.P., Rudenski, A.S., Naylor, B.A., Treacher, D.F. and Turner, R.C., Homeostasis model assessment: insulin resistance and B-cell function from fasting plasma glucose and insulin concentrations in man, Diabetologia, 28 (1985) 412. Friedewald, W.T., Levy, R. and Fredrickson, D.S., Estimation of serum low density lipoprotein without use of preparative ultracentrifuge, Chn. Chem., 18 (1972) 499. Clauss, A., Gerinnungsphysiologische schnell methode
19
20 21
22
23
24
25
26
27 28
29
30
31
32 33
zur bestimmung des tibrinogens, Acta Haematol., 17 (1957) 237. Christensen, C.K. and Orskov, H., Rapid screening PEG radioimmunoassay for quantification of pathological microalbuminuria, Diabetic Nephrop., 3 (1984) 92. Marshall, S.M., Albumin Excretion in Diabetes Mellitus. MD Thesis, University of Glasgow, 1990. Cohen, D.L., Close, C.F. and Viberti, G.C., The variability of overnight urinary albumin excretion in insulindependent diabetic and normal subjects, Diabetic Med., 4 (1987) 437. Sharp, CL., Butterfield, W.J.H. and Keen, H., Diabetes survey in Bedford 1962, Proc. Roy. Sot. Med., 57 (1964) 193. Mann, J.I., Lewis, B., Shepherd, J., Winder, A.F., Fenster, S., Rose, L. and Morgan, B., Blood lipid distribution, prevalence and detection in Britain, Br. Med. J., 296 (1988) 1701. Shaper, A.G., Pocock, S.J., Walker, M., Cohem, N.M., Wale, C.J. and Thomson, A.G., British Regional Heart Study: cardiovascular risk factors in middle-aged men in 24 towns, Br. Med. J., 283 (1981) 179. Haffner, SM., Stem, M.P., Gruber, K.K., Hazuda, H.P., Mitchell, B.D. and Patterson, J.K., Microalbuminuria. Potential marker for increased cardiovascular risk factors in nondiabetic subjects? Arteriosclerosis, IO (1990) 727. Haffner, SM., Mitchell, B.D., Pugh, J.A., Stem, M.P., Kozlowski, M.K., Hazuda, H.P., Patterson, J.K. and Klein, R., Proteinuria in Mexican Americans and nonHispanic whites with NIDDM, Diabetes Care, 12 (1989) 530. Hutchinson, A.S. and Paterson, K.R., Collecting urine for microalbumin assay, Diabetic Med., 5 (1988) 527. Price, D.A., Fielding, B.A., Davies, A.G. and Postlethwaite, R.J., Short term variability of urinary albumin excretion in normal and diabetic children, Diabetic Nephrop., 4 (1985) 169. Viberti, G.C., Hill, R.D., Jarrett, R.J., Argyropoulos, A., Mahmud, U. and Keen, H., Microalbuminuria as a predictor of clinical nephropathy in insulin-dependent diabetes mellitus. Lancet, i (1982) 1430. Jarrett, R.J., Viberti, G.C., Argyropoulos, A., Hill, R.D., Mahmud, U. and Murrells, T.J., Microalbuminuria predicts mortality in non-insulin dependent diabetes, Diabetic Med., 1 (1984) 17. Nestler, J.E., Barlascini, C.O., Tetrault, G.A., Fratkin, M.J., Clore, J.N. and Blackard, W.G., Increased transcapillary escape rate of albumin in nondiabetic men in response to hyperinsulinaemia, Diabetes, 39 (1990) 1212. Gosling, P. and Reynolds, T.M., Early diagnosis of acute myocardial infarction (Letter), Br. Med. J., 302 (1991) 53 Winocour, P.H., Durrington, P.N., Bhatnagar, D., Ishola, M., Mackness, M. and Arrol, S., The influence of early diabetic nephropathy on very low density lipoprotein (VLDL), intermediate density lipoprotein (IDL), and low density lipoprotein (LDL) composition, Atherosclerosis, 89 (1991) 49.
81 34
35
Feldt-Rasmussen, B., Increased transcapillary escape rate of albumin in type 1 (insulin dependent) diabetic patients with microalbuminuria, Diabetologia, 29 (1986) 282. Deckert, T., Feldt-Rasmussen, B., Borch-Johnsen, K., Jensen, T. and Kofoed-Enevoldsen, A., Albuminuria reflects widespread vascular damage. The Steno hypothesis, Diabetologia, 32 (1989) 219.
36
Ardron, M., Macfarlane, LA., Martin, P., Walton, C., Day, J., Robinson, C. and Calverly, P., Urinary excretion of albumin, alpha-I-microglobulin, and N-acetyl-B-Dglucosaminidase in relation to smoking habits in diabetic and non-diabetic subjects, J. Diabetic Complic., 3 (1989) 154.