Does microalbuminuria predict cardiovascular events in nondiabetic men with treated hypertension?

•

ORIGINAL CONTRIBUTIONS

AJH 1995; 8:337-342

Does Microalbuminuria Predict Cardiovascular Events in Nondiabetic Men Vvith Treated Hypertension? Stefan Agewall, John Wikstrand, Susanne Ljungman, Hans Herlitz, ond Bjorn Fagerberg, on behalf of the Risk Factor Intervention Study Group

The aim of this study was to investigate the predictive value of microalbumlnuria (overnight urinary albumin excretion rate 17 to 100 rng/12 h) as a risk factor for future major cardiovascular events in nondiabetic patients with treated hypertension in a prospective study with follow-up time of 3.3 years. Overnight urinary albumin excretion was measured in 345 nondiabetic treated hypertensive men, aged 50 to 72 years, either with a serum cholesterol of ~6.5 mmol/L or smokers, or both. CardiovascuJar morbidity was closely recorded during the follow-up period. At entry, microalbuminuria was found in 84 patients (24.3%) and 12 patients had macroalbuminuria (3.5%). During the follow-up perle..~ there were no differences in new cardiovascular events between patients with microalbuminuria and those with normoalbuminuria, However, an increase in the risk of future major cardiovascular events oc-

curred in patients with urinary albumin excretion above 100 mg/12 h (macroalbuminuria). In a Cox regression analysis urinary albumin excretion was net associated with the incidence of future major cardiovascular events unless a more detailed approach was used, showing that this was the case for urinary albumin excretion above 100 mg/12 h (macroalbuminuria). Calculations with an alternative definition of rnicroalbuminuria and mortality as end-point did not change the principal result. In conclusion, microalbuminuria does not seem to be a predictor of future mortality and cardiovascular morbidity in nondiabetic men with treated hypertension and at high risk of coronary heart disease. However, macroalbuminuria was associated with future major cardiovascular events in this group of patients. Am J Hypertension 1995; 8:337-342 Hypertension, cardiovascular disease, prediction, microalbuminuria. KEY WORDS:

roteinuria in diabetic and hypertensive pop-

P

Received June 6, 1994. Accepted November 21, 1994. ulationsis associated with an excess risk of Fromthe Departmentof Medicine (SL, HH), Departmentof Nephrology OW), Wallenberg Laboratory for Cardiovascular Remorbidity and mortality from cardiovascular search, Sahlgrenska University Hospital, Goteborg University, disease.r" During the last years, the results G6teborg, Sweden. The Risk Factor Intervention Study group also includes Ove K. of several studies have suggested that the excretion of Andersson, Antje Berglund, Marianne Hartford, Tomas Hedner, small amounts of urinary albumin, microalbuminBengt Persson, Madis Suurkula, Bengt Widgren, and Marian uria, predicts mortality in both insulin-dependent Wysocki. This study was supported by grants from the Swedish Medical and non-insulin-dependent diabetes mellitus't" and Research Council (B92-19X-09937-01A, B93-19X-09937-02B, B94· in the general population.v'" In patients with pri19X-D9937-03A) and the Swedish Heart and Lung Foundation. mary hypertension the clinical relevance of microalAddress correspondence and reprint requests to Dr. StefanAgewall, Department of Medicine, G6teborg University, Sahlgrenska buminuria is less clear. However, we have in a preUniversity Hospital, S-413 45 Gothenburg, Sweden. viously reported cross-sectional study found a weak

© 1995 by the American

Journal ofHypertension, Ltd.

0895-7061/95/$9.50 0895-7061<94100246-8

338 AGEWALL ET AL

association between cardiovascular disease and microalbuminuria in treated hypertensive men.'? The mechanism of the link between urinary albumin excretion and cardiovascular disease is incompletely understood and it is still an open question whether management of the huge hypertensive population can be improved by monitoring microalbuminuria in a manner similar to that in diabetes care. Against this background the aim of this study was to investigate the predictive value of microalbuminuria as a risk factor for future major cardiovascular disease events in a group of nondiabetic men with treated hypertension who previously were examined in a crosssectional study regarding microalbumirurria.!' METHODS Study Population As previously described, 508 male patients with treated primary hypertension were included in a risk factor intervention study. 12 Most of the patients were recruited after having participated in either of two studies, which were based on screening of the general population.P'" The inclusion criteria for the risk factor intervention study were, apart from treated hypertension and male sex, age between 50 and 72 years and one or more of the following: hypercholesterolemia (serum cholesterol ~6.5 mmol/L), tobacco smoking (one or more cigarettes per day), or diabetes mellitus. Exclusion criteria were unwillingness to participate, and malignant or other serious chronic disease. These men are representative of high-risk hypertensives in Gothenburg since the majority (over 90%) were recruited earlier by screening of a random third of all men in their respective age groups in Gothenburg." Of these 508 patients it was possible to obtain an adequate overnight urine collection and subsequent determinationof urinary albumin excretion in 452 patients. In the present analysis 13 subjects with renal disease and 94 subjects with diabetes mellitus, defined as fasting blood glucose 2:6.7 mmoIlL at two separate occasion.P were excluded. Hence, 345 nondiabetic hypertensive patients were included in the present analysis.

AIH-APRlL 1995-VOL. 8, NO.4, PART 1

clinical practice. The study took place at our hypertension outpatient department. The study was approved by the ethical committee of the faculty of medicine, Gothenburg University. Measurements After having given informed consent to participate in the study, the patientswere interviewed by a physician who used case records to fill in a standard record form covering documented cardiovascular events, syrn~toms of cardiovascular disease, and diabetes mellitus. Subsequently, allrecordforms were also reviewed by a second physician. The patients were asked to answer a questionnaire regarding smokinghabits and the current medication was documented. Blood pressure was measured by especially trained nurses, using a mercury sphygmomanometer with a cuff of appropriate size. Diastolic pressure was determined as Korotkoff phase V. The patients rested in the recumbent position for 5 min and the mean of two recordings was used. Heart rate was measured by palpation of the radial pulse. Body weight, body mass index, and waist:hip circumference ratio were measured according to recommended principles. ll,12 Venous blood was drawn after an overnight fast and after 5 min of rest in the recumbent position for determination of blood glucose and serum concentrations of total cholesterol high density lipoprotein cholesterol, triglycerides, urate, and creatinine using established methods.11,12 These measurements and all examinations described above were repeated biannually. At entry the patients were asked to bring an overnight (12-h) urine sample for determination of urinary albumin excretion (Albumin RIA, Pharmacia Diagnostics, Uppsala, Sweden, detection limit 0.4 mg/L and a coefficient of variation of ~10%). The subjects were told to avoid coff e and heavy exercise on the evening of the urine collection. Given that the urinary albumin excretion during the day exceeds the excretion at night by a factor of 1.3 to 2,16 the definition of microalbuminuria of 30 to 300 mg/24 h will roughly correspond to an upper cut-off point of 100 mg during the 12-h overnight sampling period. Normoalbuminuria was classified as urinary albumin excretion below 17 mg/12 hi microalbuminuria as 17 to 100 mgl 12 h, and macroalbuminuria as over 100 m~/12 h, according to previously described principles. 1 In an alternative analysis, microalbuminuria, defined as urinary albumin excretion of 20 to 200 JLg/min,7,9 was also used. The amount of microalbuminuria was not available during the treatment period and consequently did not influence the treatment of the patients.

Study Protocol After the initial examination the patients were followed biannually. In the present report we show results after a mean follow-up time of 3.3 years. In the underlying study, patients were randomizedeithertoa moreintensive multiple riskfactor treatment program, aimedat improving the riskfactor profile, or to usual care. 12 The former program was based on information meetings concerning healthier dietary habits and smoking cessation." If necessary, lipidlowering drug therapy couldbe instituted. 12 There was Endpoint Registration Cardiovascular morbidity no difference betweenthe groups regarding antihyper- was closely followed during the follow-up period. Fatensive drug treatment, which was given according to tal and nonfatal myocardial infarction, fatal and non-

THE PREDICTIVE VAf.UE OFMICROALe4.:MINURlA 339

AJH-APRIL 1995-VOL. 8, NO.4, PART 1

fatal stroke, sudden death, and other causes of cardiovascular mortality (ie, ruptured aortic aneurysm) were registered according to established definitions.P Any of the above-mentioned diagnoses, alone or in combination, were used as a new major cardiovascular event in the analysis. Only one, the first event, was used in each patient. All these events were independently coded by two doctors who did not know to which group the patients had been allocated. In case of disagreement, a final decision was reached after discussion with a third colleague. Statistical Methods Results are presented as means and standard deviations. Unpaired t tests were used. to compare continuous variables. Categorical. variables were analyzed by using the l test or Fisher's exact test. The relationship between urinary albumin excretion at entry of the study and the incidence rate of new major cardiovascular events during follow-up was analyzed with a Cox regression analysis. As urinary albumin excretion showed a markedly skewed distribution, log transformation was done before calculating the Cox regression analyses. A second Cox regression analysis was performed where the hazard function of new major caniiovascular event was estimated as a continuous and piecewise linear function of urinary albumin excretion. Two possible breakpoints of the function were chosen, at 17 mgl12 hand 100mg/12 h, corresponding to our definition of microalbuminuria. This was done to obtain a more detailed picture of the relationship between urinary albumin excretion and cardiovascular risk. A P < .05 on two-sided tests was regarded as significant.

RESULTS The characteristics of the 345 patients at entry are shown in Table 1. Microalbuminuria was found in 84 patients (24.3%) and 12 patients had macroalbuminuria (3.5%). Patients with microalbumlnuria were slightly older and had higher serum levels of creatinine compared with norrnoalbuminuric patients (n = 249). In comparison with the normoalbuminuric group the small group of patients with macroalbuminuria had higher systolic blood pressure and higher concentrations of blood glucose, serum creatinine, and serum urate. Treatment with diuretics was more common in the microalbuminuric group compared with the normoalbuminuric group. Calcium antagonists were more frequently used in the macroalbuminuric group compared with the other two groups (Table2). Only four patients, three in the normoalbuminuric group and one in the microalbuminuric group, were on single therapy with a calcium antagonist. At the end of the follow-up period no difference in respect to the use of antihypertensive drugs or lipid-lowering drugs between the normoalbuminuric and the microalbuminuric groups was found (data not shown). Net changes in blood pressure, serum cholesterol, blood glucose, and proportion of patients quitting smoking as well as participation in the intervention program did not differ between the three category groups of urinary albumin excretion during the follow-up period (data not shown). The incidence rates of clinicalendpoints during the follow-up are given in Table 3. There were no differences in any of the clinical event rates between patients with microalbuminuria and those with nor-

TABLE 1. INITIAL CHARACTERISTICS OF PATIENTS Normoalbuminuria «17 mgl12 h; n 249)

Microalbuminuria (17-100 mgl12 h; n 84)

66.0 (4.9)* 81.4 (12.5) 26.4 (3.5) 0.97 (0.05)

67.6 (4.0) 81.7 (13.3) 27.3 (3.7) 0.97 (0.05)

=

Age (years) Weight (kg) Body mass index (kg/m2 ) Waist-to-hip ratio Blood pressure (mm Hg) Systolic Diastolic Heart rate (beats/min) Serum total cholesterol (mmoVL) Serum high density lipoprotein cholesterol (mmollL) Serum triglycerides (mmollL) Blood glucose (mmol/L) Serum creatinine (umol/L) Serum urate (mmollL)

=

(18)t

156

91 (8) 62 (11) 6.9 (1.0)

91

152

1.3 (0.3) 1.8 (0.9) 4.9 (0.7)t 99 (16)*t 356 (84)+

Macroalbuminuria (>100 mgl12 h; n = 12) 67.1 81.8 27.3 0.99

(3.9) (9.4) (2.8) (0.04)

(19) (9) 64 (11) 6.8 (1.1)

165 94 64 7.1

(30) (12) (16) (0.8)

1.3 (0.4) 1.9 (1.1)

1.2 2.2 5.4 125 426

(0.3) (1.0) (1.3) (48)

5.0 (0.7) 103 (18):1: 373 (89):1:

(42)

Values are means (SD). *Normoalbuminuric group v micrcalbuminuric group, P < .05; tnormoalbuminuric group v mucroalbuminutic group, P < .05; imicroalbuminuric group v macrealbuminuric group, P < .05.

AfH-APRIL 1995-VOL. 8, NO.4, PART 1

340 AGEWALL ET AL

TABLE 2. INITIAL PREVALENCE OF CARDIOVASCULAR DISEASE, PHARMACOLOGIC TREATMENT, AND SMOKING HABITS

----

Prevalence of: Previous myocardial infarction or stroke, number (%) Smokers, number (%) Treatment with: Diuretics, number (%) ~.Blockers, number (%) Angiotensin converting enzyme inhibitors, number (%) Calcium antagonists, number (%) Hydralazine, number (%)

Normoalbumlnuria «17 mgl12 hi n ::: 249)

Mlcroalbuminuria (17-100 mgl12 hi n ::: 84)

Macroalbuminuria (>100 mg/12 hi n ::: 12)

27 (10.B) 73 (29.3)

12 (14.3) 29 (34.5)

3 (25.0) 4 (33.3)

119 (47.8)* 184 (73.9)

55 (65.5) 54 (64.3)

6 (50.0) 10 (83.3)

13 (5.2) 29 (11 .6)t 40 (16.1)

0 10 (11.9):1: 10 (11.9)

1 (8.3) 5 (41.7) 0

Valucs arc r.umbers (percentage). *Normoalbumi/lltric group v microa/blllni/luric group, P < .05; fnorllloaibuminuric group v macroa/bllmilluric group, P < .05; tlllicroaibuminuric group v macroa/brll/tilt/lric group, P < .05.

moalbuminuria. However, in comparison with the microalbuminuric group the macroalbuminuric patients had higher incidence rates of csrdiovascular mortality, fatal and nonfatal myocardial infarction, fatal and nonfatal stroke, and major cardiovascular events taken. together. The latter endpoint also occurred more frequently in the macroalbuminuric group than in the group with normoalbuminuria. In the Cox regression analysis the logarithm value of 'urinary albumin excretion was not significantly associated with the incidence of new major cardiovascular events (P = .11). However, in the second Cox regression analysis in which two possible breakpoints of the function was chosen, at 17 mgl12 hand 100 mgl12 h, corresponding to our definition of microalbuminuria, urinary albumin excretion (log transformed) above 100 mg/12 h was significantly associated with an increased risk of new major cardiovascular event (P = .033). To compare the results with those of other studies,

microalbuminuria was also defined as urinary albumin excretion of 20 to 200 ltg/min (Table 4). Mortality during the follow-up period did not differ between patients with and without microalbuminuria according to this definition. DISCUSSION The results of this study showed that in a group of treated hypertensive men at high coronary risk, baseline urinary albumin excretion was only significantly associated with new major cardiovascular events such as myocardial infarction, stroke, and sudden death in patients withurinary albumin excretion above the level that was defined as microalbuminuria. Among hypertensive patients, overt proteinuria is a well-known risk factor for future disease. l -3 The group, defined as having microalbuminuria, did not have an increased incidence of future major cardiovascular events. This was found only among the patients with macroalbuminuria. These results are not

TABLE 3. CLINICAL EVENTS DURING 3.3 YEARS OF FOLLOW-UP OF NONDIABETIC TREATED HYPERTENSIVE MEN p

Myocardial infarction, number (%) Stroke, number (%) Cardiovascular mortality, number (%) Total mortality, number (%) Major cardiovascular event, number (%)

2

Normoalbuminuria «17 mg/12 hi n = 249)

Microalbuminuria (17-100 mg/12 h; n 84)

Macroalbuminuria (>10!} mg/12 hi n = 12)

(x , all groups compared)

15 (6.0) 8 (3.2)

1(1.2)* 1 (1.2)*

4 (25.0) 2 (16.7)

.003 .017

10 (4.0) 13 (5.2)

0* 2 (2.4)

2 (16.7) 2 (16.7)

.009 .09

24 (9.6)+

2 (2.4)*

5 (41.7)

.0001

=

Value:; arc numbers (percentage). >tMicroalburninuric group v macroatbuminuric group, P < .05; tnormcalbuminuric group v macroa/buminuric group, P < .05. Cardiovascular mortality was defined as any ofthe following diagnoses: fatal myocardial infarction, fatal stroke, sudden death, and fatal ruptured aortic aneurysm. Major cardiovascular event was defined as any ofthe diagnoses: nOllfatal myocardial infarction, nOr!fatal stroke, or cardiovascular death.

.

THE PREDICTIVE VALUE OF MICROALBUMINURIA 341

AJH-.4PRIL 1995-VOL. 8, NO.4, PART 1

TABLE 4. PREVALENl:E OF MICROALBUMINURIA ACCORDING TO CONVENTIONAL CRITERIA AND ASSOCIATED RELATIONSHIP TO FUTURE MORTALITY IN COMPARISON WITH PREVIOUS STUDIES Patients (reference) Present study GP practice9* NIDDM 7 Nondiabetics'? NlDDMs NlDIJM4 100M8

Number

345

Age (years) 67 60

167 141 216

56

232

65

44

51 40

63

68

Microalbuminuria Definition

Number (%)

Follow-up time (years)

20-200 fLg/min 20-200 j.Lglmin 20-200 J.tg/min >7.52 J.tg/min 30-140 J.tg/mL 30-240 J.tglmin 30-HO fJ.g/min

94 (22) 18 (11) 36 (26) 108 (50) 76 (33) 7 (16) 8 (13)

3.3 3.4 3.4 6 9 14 23

Mortality during follow-up, n (%) Mic:roalbuminuria

Normoalbuminuria

2 (2) 6 (33)

13 (5) 3 (2) 4 (4)

10 (28) 23 (21) 59 (78) 6 (86)

5 (62)

8(7)

62 (48) 11 (25) 17 (31)

*Rnndomly selected from general practice.

in agreement with studies of diabetic patients, which have shown that microalbuminuria is a predictor of cardiovascular morbidity or mortality. 4-8 Similar findings have been obtained in screened cases from a general practice? with a high prevalence of diabetes mellitus and impaired glucose tolerance (30%) and in a nondiabetic population sample with an arbitrary definition of microalbuminuria.l" The fact that our results do not support the suggestion that microalbuminuria might be a predictor of future morbidity also in hypertensive patients raises several questions. The sampling methods to determine urinary albumin excretion are not uniform in previous studies. Some studies, like the present one, used overnight urine collection,4,7,8 whereas others used morning urine 5,6 or urine collection during 1 to 2 h. 9, l O It is also well-known that the intra- and interindividual variability in urinary albumin excretion is high, necessitating repeated urinary collections or large population samples." In the latter respect our study is one of the largest. None of the studies referred to in Table 4 used multiple collections of urine for measurements of albumin excretion in all participating subjects. 4,5,7-10 The definition of rr.icroalbuminuria and studied endrooints vary when different studies are compared.v 0 However, our results remained principally unchanged when we used the same criteria for microalbuminuria as in other studies or chose mortality as an endpoint. The power of the present study to show a postulated predictive value of microalbuminuria does not seem to be insufficient. First, the total annual mortality rate in the group defined as normoalbuminuric was of the same magnitude as in most of the previous prospective studies regarding the predictive value of microalbuminuria. 4,7-10 These studies have observed an up to 17fold increased mortality in the microalbuminuric group and in the present study the mortality tended to be smaller in the microalbuminuric group compared to the normoalbuminuric group . Second, the duration of the follow-up period was as long as in

other studies. 7,9 Finally, our patient sample is larger than those in almost all other similar studies. 4,5,7-10 Concomitant treatment may confound urinary albumin excretion as a risk factor for later cardiovascular disease. In this study the only obvious difference between the microalbuminuric and normoalbuminuric groups was a more frequent use of diuretics in the former group. This cannot explain why microalbuminuria was not predictive of future morbidity or mortality. In previous studies the proportions of patients with antihypertensive drug treatment among patients with microalbuminuria have ranged from no reported data to 67%.4-10,17 The potentially confounding effects of different antihypertensive agents have not been analyzed or discussed in any of these studies. In the present study calcium antagonists were more frequently used in the group with urinary albumin excretion above 100 mg/12h, a group with a high prevalence of cardiovascular disease at baseline and also with a high incidence of new major cardiovascular events during follow-up. Since no patients in the macroalbuminuric group were on calcium antagonists as a single drug therapy, a probable explanation may be that this class of drugs, for clinical reasons, was prescribed more often for patients with more advanced cardiovascular disease. Renal hemodynamic effects leading to increased urinary albumin excretion secondary to an increase of glomerular capillary pressure may also possibly contribute to this finding. 18 Changes in cardiovascular risk factors could not explain the outcome of future morbidity or mortality, since net changes in blood pressure, serum cholesterol, blood glucose, and proportion of patients quitting smoking and participation in the intervention program did not differ between the three category groups of urinary albumin excretion during the follow-up period. As regards the underlying mechanisms, explaining why urinary albumin excretion above 100 mg/12 h (macroalbuminuria) was an indicator of future major cardiovascular disease, one might speculate that in-

342

AGEWALt ET At

creased urinary albumin excretion just may be a marker of a more advanced cardiovascular disease." This is a possible explanation since there was a high prevalence of cardiovascular disease at baseline in the macroalbuminuric group in this study. Furthermore, macroalbuminuria may indicate endothelial dysfunction which may promote the atherosclerotic process." The present patients are not representative of the whole hypertensive population, as theywere middleaged and older men with at least one further cardiovascular risk factor apart from hypertension. However, the patients were mostly recruited from representative samples of the male hypertensive population. 13 Our interpretation is that microalbuminuria does not seem to be a predictor of future mortality and cardiovascular morbidity in men with treated hypertension and at high risk of coronary heart disease. These results have been obtained by using a study design that is similar to that used in several studies demonstrating a predictive value of microalbuminuria in patients with diabetes mellitus. However, our results confirm the cu.rrent concept that overt macroalbuminuria is a predictor of future cardiovascular disease in hypertensive patients, demonstrating that the borderline between microalbuminuria and overt albuminuria is relative. ACKNOWLEDGMENTS Anders Oden, PhD, statistician, helped us with the statistical analyses. We are grateful to Ann-Louise Eriksson, Maija-Lisa Holmasaari, and Tuija Kaikkonen at our hypertension outpatient department.

REFERENCES 1. Bulpitt C], Beevers G, Butler A, et al: The survival of

treated hypertensive patients and their causes of death: A report from the DHSS Hypertensive Care Computing Project (DHCCP). J Hypertens 1986;4:9399. 2. Kannel WB, Stampfer MJ, Castelli WP, et al:Theprognostic significance of proteinuria: The Framingham study. Am Heart J 1984;108:1347-1352. 3. Samuelsson 0, Wilhelmsen L, Elmfeldt D, et al: Predictors of cardiovascular morbidity in treated hypertension. Results from the primary preventive trial in Goteborg, Sweden. J Hypertens 1985;3:167-176. 4. [arett RJ, Viberti GC, Argyropoulus A, et al: Microal-

A/H-APRIL 1995-VOL. 8, NO.4, PART 1

buminuria predicts mortality in non-insulin-dependent diabetes. Diabetic Med. 19R4;1:17-19. 5. Mogensen CE: Microalbuminuria predictsclinical proteinuria and earlymortality in matury-onset diabetes. N Engl J Med 1984;310:356-360. 6. Schmitz A, Vaeth M: Microalbuminuria: A major risk factor in non-insulin-dependent diabetes. A 10-year follow-up study of 503 patients. Diabetic Med 1988;5: 126-134. 7. Mattock MB , Morrish NJ, Viberti G, et al: Prospective study of microalbuminuria as predictor of mortality in NIDDM. Diabetes 1992;41:736-741. 8. Messent WC, Elliott TG, Hill RD, et al: Prognostic significance of microalbuminuria in insulin-dependent diabetes mellitus: A twenty-three year follow-up study. Kidney Int 1992;41:836-839. 9. YudkinJS, Forrest RD, Jackson CA: Microalbuminuria as predictor of vascular disease in non-diabetic subjects. Lancet 1988;2:530-533. 10. Damsgaard EM, Freland AI Iergensen 00, et al: Microalbuminuria as predictor of increased mortality in elderly people. BrMed J 1990;300:297-300. 11. Agewall 5, Persson B, Samuelsson 0, et al: Microalbuminuria in treated hypertensive men at high risk of coronary disease. J Hypertens 1993;11:461-469. 12. Agewall 5, Fagerberg H, Berglund G, et al: Multiple cardiovascular risk factor intervention in treated hypertensive men: what can be achieved? Nut! Metab Cardiovasc Dis 1993;3:128-135. 13. Samuelsson 0: Hypertension in middle-aged men: management, morbidity and prognostic factors during long-term hypertensive care. Acta Med Scan 1985; suppI702:1-79. 14. Wikstrand J, Warnhold I, Tuomilehto J, et al: Metoprolol versus thiazide diuretics in hypertension. Morbidity results from the MAPHY study. Hypertension 1991;17:579-588. 15. World Health Organization Stud}' Group. Diabetes Mellitus: WHO Tech Rep Ser 1985; No. 727. 16. Ljungman 5: Microalbuminuria in essential hypertension. AmJ Hype-tens 1990;3:956--960. 17. Stehouwer CDA, NautaJJP, Zeldenrust GC, et al:Urinary albumin excretion, cardiovascular disease and endothelial dysfunction in non-insulin dependent diabetes mellitus. Lancet 1992;340:319-323. 18. Epstein M: Calcium antagonists. Current status and future perspectives. Arch Intern Med 1992;1.'52:15731584. 19. Deckert T, Feldt-Rasmussen B, Borch-Johnsen K, et al: Albuminuria reflects widespread vascular damage. The Steno hypothesis. Diabetologia 1989;32:219-226.