Microalbuminuria in urban Zimbabwean women - Nature

Journal of Human Hypertension (2000) 14, 587–593  2000 Macmillan Publishers Ltd All rights reserved 0950-9240/00 $15.00 www.nature.com/jhh

ORIGINAL ARTICLE

Microalbuminuria in urban Zimbabwean women K-K Hwang1, LJ Scott2, J Chifamba3, J Mufunda3, WS Spielman1 and HV Sparks1 1

Department of Physiology, Michigan State University, East Lansing, MI, USA; 2Section on Genetics and Epidemiology, Joslin Diabetes Center, Boston, MA, USA; 3Department of Physiology, University of Zimbabwe, Harare, Zimbabwe

The prevalence of microalbuminuria (MAU) in African populations has not been reported, nor has the relationship between MAU and hypertension been reported for these populations. We collected spot urine samples from 370 women, 25 years and older as a part of a population-based, cross-sectional blood pressure survey in an urban community in Zimbabwe and analysed the samples for albumin and ␤2-microglobulin. The ageadjusted prevalence of hypertension was 30% for women 25 years and older in this community. After excluding the samples with hematuria (11%), the prevalence of MAU (3.0 ⭐ albumin-to-creatinine ratio (ACR, mg/mmol) ⬍25.0) in the study population was 9%. When age-adjusted to the population in the community, the prevalence was 8% among women 25 years and older. The prevalence of MAU was substantially higher in

hypertensive (HT) than in normotensive (NT) women (16% vs 4%, P ⬍ 0.001). A significantly higher level of log ACR in HT was found in each age group except the youngest age group (age 25–34). In age-adjusted multiple regression, percent fat mass was negatively associated with log ACR (␤ ⴝ ⴚ1.18, 95% CI (ⴚ0.23, ⴚ2.21), P ⴝ 0.02). In a similar regression analysis, higher log ␤8microglobulin-to-creatinine ratio was very strongly associated with higher log ACR (␤ ⴝ 0.34, 95% CI (0.25, 0.43), P ⬍ 0.0001) and significantly associated with lower percent fat mass (␤ ⴝ ⴚ1.02, 95% CI (ⴚ0.25, ⴚ1.8), P ⴝ 0.01). These results suggest that MAU is frequently caused by hypertension, but that other diseases may contribute to its presence. Journal of Human Hypertension (2000) 14, 587–593

Keywords: microalbuminuria; women; black population; Africa

Introduction The level of hypertension in Africa is increasing with increasing urbanisation, and it is particularly high in urban Southern Africa.1–5 In a recent population-based, cross-sectional survey of an urban community in Zimbabwe, we found that the prevalences of hypertension in people 25 years and older were 30% for women and 21% for men when ageadjusted to the population in the community.6 Available data also suggest that hypertensionrelated diseases including stroke, renal failure and cardiac failure are becoming major public health problems in many regions of Africa7–10 and thus may disproportionately affect the urban Zimbabwean population. Microalbuminuria (MAU), a moderately elevated level of urinary albumin excretion, can be caused by increased blood pressure, and may be the result of glomerular dysfunction. The presence of MAU signals an increased risk of the presence or development of cardiovascular and kidney complications.11–13 In many studies including populationbased studies of essential hypertensives, MAU is Correspondence: Jacob Mufunda, Department of Physiology, University of Zimbabwe, P.O. Box MP167, Harare, Zimbabwe. E-mail: physio얀icon.co.zw Received 1 March 2000; revised 10 May 2000; accepted 13 May 2000

positively associated with insulin resistance, hyperlipidaemia, salt sensitivity, obesity, smoking and alcohol consumption.14–20 In addition, MAU is a marker of early intrarenal vascular dysfunction in lean essential hypertensives.21,22 It is also associated with the inflammatory processes that occur in a variety of diseases including trauma and sepsis.23 An elevated level of ␤2-microglobulin in the urine is also associated with a variety of diseases and may be caused by disruption of tubular function rather than glomerular function. Spot urine samples are a useful measure in community studies because they can be collected without timing of urine collection or concerns for completeness of collection. The intra-individual variability in urine concentration can be reduced by the use of albumin-to-creatinine ratio (ACR).24,25 A single sample can predict MAU with 92% sensitivity and 90% specificity in non-diabetic subjects25 and is a useful marker for cardiovascular and renal risks.26,27 Prevalences of MAU in population-based studies range from 2 to 28%, and the levels are higher in older and non-white populations.24,28,29 The levels of MAU in people with hypertension range from 0 to 40%.30 However, the level of MAU in communities in urban Africa is not known, nor has the relationship between MAU and hypertension been examined in an African community. In the present

Microalbuminuria in urban Zimbabwean women K-K Hwang et al

588

study, we determined the ACR and ␤2-microglobulin-to-creatinine ratio (␤2MCR) in spot urine samples collected from a community-based study of 370 women. The aim of our study was to determine the prevalence of microalbuminuria in a populationbased sample of Zimbabwean women and to determine the relationship of systolic blood pressure to the level of ACR. In addition, we examined the relationship between ACR or ␤2MCR and age body mass index (BMI), percent fat mass and urinary Na/K ratio, a crude marker of sodium intake.

Subjects and methods Study site and sampling procedures Marondera is the capital of Mashonaland East Province in Zimbabwe, and it has an estimated population of 19 366 people.31 The study was conducted from July to October 1995, in Dombotombo township of Marondera.6 Women 25 years and older were sampled according to the ‘probability proportionate to size’ cluster sampling method used in the International Collaborative Study on Hypertension in Blacks (ICSHIB).32 The method involved random selection of clusters (all the houses on a named street) from all possible clusters in an area. Approximately, 100 women in each age group (25–34, 35– 44, 45–54, 55 years and older) who had lived in the community for more than 3 years were recruited. At the time of the interview, containers for a spot urine sample were provided, and the participant was asked to come to a nearby blood pressure (BP) measurement site by the next day, with the urine sample. Sampling procedures, BP and anthropometric measurement protocols were previously described in detail.6 Blood pressure and anthropometric measurements The BP measurement protocol was adapted from ICSHIB.33 BP was measured using Omron HEM713C automated oscillometric monitors. In two independent studies, the automated BP readings were found to be comparable to those taken manually.34,35 BP was measured three times separated by 1-min rest periods, and the average of the last two readings was used for the analysis. Hypertension was defined as systolic BP (SBP) ⭓140 and/or diastolic BP (DBP) ⭓90 and/or currently taking an antihypertensive medication. Weight, height, waist and hip circumference were measured as previously described, and BMI and waist-to-hip ratio (WHR) were calculated as previously described.32 Overweight and obesity were defined as BMI ⭓27.3 and ⭓32.3 kg/m2, respectively.32 Bioelectrical impedance analysis was performed using an instrument from the RJL Bioelectric Systems (Clinton Township Michigan State, USA), and percent fat mass was calculated using the equations of Kuschner and Schoeller.36,37 Urinary measurements Aliquots of the spot urine samples were frozen and stored at ⫺80°C for the averages of 17, 18 and 21 Journal of Human Hypertension

months before the assays for albumin, creatinine and ␤2-microglobulin (␤2M), respectively. All assays were performed in a blinded fashion. Albumin and ␤2M concentrations were measured by radioimmunoassay (Diagnostic Products Corporation, LA, CA, USA), and creatinine concentration was measured by Jaffe´ reaction (Sigma, St Louis, MO, USA). The coefficients of variation for albumin, creatinine and ␤2M assays were 7%, 3% and 6%, respectively. MAU was defined using a cut-off point of ACR ⭓3.0 mg/mmol. Macroalbuminuria was defined as ACR ⭓25.0. ␤2MCR was calculated as a measure of tubular dysfunction. The abnormal excretion of ␤2M was defined using a cut-off point of ␤2MCR ⭓34.0 ␮g/mmol (300 ␮g/g). Eighty four percent of the samples with ␤2MCR ⭓34.0 ␮g/mmol had ␤2M concentration ⬎290 ␮g/L, which is the cut-off point proposed by the American Conference of Governmental Industrial Hygienists Inc.38 Urine samples were tested for hematuria (sensitivity, 0.15–0.62 mg/L hemoglobin), proteinuria (sensitivity, 150–300 mg/L albumin), glucose (sensitivity, 750–1250 mg/L) and ketone (sensitivity, 50–100 mg/L) using dipsticks (Bayer, Tarrytown, NY, USA). None of the samples showed a positive result for glucose or ketone. Since ␤2M has been shown to be unstable at pH ⬍6,39 the samples were also tested for pH using dipsticks (Bayer), and only 4% of the samples was pH ⬍6. Urinary sodium-to-potassium (Na/K) ratio was calculated from concentrations of the two ions measured by flame photometry. Exclusions From the 391 female participants, 20 urine samples were not available for the assays. One sample was excluded because of a positive result in a urine pregnancy test. Thirty-nine samples with positive results on hemoglobin dipsticks were excluded from the analysis. Of the remaining samples, two samples (ACR ⫽ 178 and 44 mg/mmol or albumin ⫽ 486 and 95 mg/L, respectively) had macroalbuminuria (ACR ⭓25.0) and were excluded. Therefore, 329 samples were available for the subsequent analyses for MAU. The ACR was equal to 0 in eight samples. For these samples the log-transformed value was set to the value of the lowest observation. For the multiple linear regression models, three people with missing percent fat mass were excluded. In addition, the person with the lowest percent fat mass and the people with the three highest Na/K ratio were excluded because they were highly influential (in their presence, more highly negative and significant coefficients were observed for the regression of log ACR). Statistical analysis Data were double entered into Epi-Info (version 6.04a), and non-matching entries were corrected. SAS (version 6.12) was used for all of the analyses except for the general additive models and multiple linear regression which were performed in S-Plus (StatSci Division of MathSoft, Seattle, WA, USA, Version 4.5).


Age group-adjusted means of selective variables for normotensive (NT) and hypertensive (HT) or normoalbuminuric and microalbuminuric groups were compared using linear regression. In each age group, age-specific means of log ACR and log ␤2MCR for NT and HT were compared using Student’s t-test. Age adjustment of the levels of selected variables to the underlying Zimbabwean population in Marondera was performed as previously described.6 Relationships between SBP and selective variables were assessed using Spearman correlations. Differences in the age group-adjusted prevalences of categorical variables between normoalbuminuric and microalbuminuric groups were compared using logistic regression. The presence of a non-linear relationship between log ACR and SBP was assessed by a general additive model. The model replaces the usual assumption of linearity with a flexible smooth term, in this case the loess option, to fit a smooth curve to the data.40 The best non-linear fit was determined by the Akaikes Information Criterion. The following variables were also included as linear terms in some models for log ACR: age, percent fat mass or BMI, Na/K ratio and alcohol consumption. These variables, plus a linear term for SBP and/or log ACR, were included in models of log ␤2MCR.

Results Anthropometric and demographic characteristics of the study population, stratified by 10-year-age group are presented in Table 1. The prevalence of hypertension in the study population was 41%. The level of hypertension was lowest, 15%, in the age group 25–34 and highest, 73%, in the age group 55 and older. When adjusted for the age of the community as a whole, the prevalence of hypertension was 30% among women 25 years and older. SBP and DBP increased progressively in older age groups, and BMI appeared to rise and then plateau in the age group 35– 44.6 Table 2 shows the age-specific findings for hematuria, proteinuria, MAU and an abnormal ␤2M excretion in NT and HT women. The overall prevalence of hematuria was 11%. It was most prevalent in the age group 25–34 (19%) and least prevalent in the age group 45–54 (5%). With the exception of the

oldest age group (55 and older), the prevalence of hematuria was higher among women who were hypertensive than those who were normotensive. Overall, the women with hematuria were significantly younger than those without hematuria (P ⫽ 0.05), but their BMI and SBP did not differ significantly from women without hematuria. When age-adjusted to the population in the community, the overall prevalence of hematuria was 14%. After excluding the samples with hematuria, 331 urine samples were available for further analyses. Five percent of women had dipstick-positive proteinuria, and the prevalence was 5% when age-adjusted to the population in the community Albumin and creatinine were measured in the 331 urine samples which did not contain hemoglobin. Two of these samples had ACR in excess of 25.0 mg/mmol and were excluded. The other samples which were positive for dipstick proteinuria had ACR ⬍25.0. The overall prevalence of MAU in the study population was 9%. The proportion of women with MAU increased with age, and the proportions in the age groups 25–34, 35– 45, 45–54 and 55 and older were 5, 8, 10 and 14%, respectively (P ⫽ 0.32, Mantel-Haenszel test for trend, P ⫽ 0.06). The prevalence of MAU was significantly higher in HT than in NT women (16% vs 4%, P ⬍ 0.001). Except the age group 25–34, the prevalence of MAU was also higher for HT than for NT women in each age group (Table 2). When ageadjusted to the population in the community, the prevalence of MAU was 8%. The 95th percentile of ACR was 5.4 mg/mmol (48 mg/g). The age-adjusted mean of log ACR was significantly higher in HT than in NT women (mean ± standard error, ⫺0.1 ± 0.05 vs ⫺0.5 ± 0.04, P ⬍ 0.001), and a significantly higher mean of log ACR for HT women was also found in each age group except the youngest age group (Figure 1). The prevalence of an abnormal ␤2M excretion (␤2MCR ⭓34.0 ␮g/mmol creatinine) was 9%, and it was most prevalent in the age group 55 and older (17%) (Table 2). In overall, the 95th percentile of ␤2MCR was 94.4 ␮g/mmol (834 ␮g/g). When the levels of log ␤2MCR for NT and HT women were compared, there was no significant difference in either the age group-adjusted mean or the age-specific means (data not shown).

589

Table 1 Demographic and anthropometric characteristics of women in the study population, Marondera, Zimbabwe, 1995 Age group Number SBP (mm Hg) DBP (mm Hg) BMI (kg/m2) WHR Hypertensive (%) Overweight (%) Obese (%) Drinks alcohol (%) Uses tobacco (%)

Total

25–34

35– 44

45–54

55 and older

370 130 (24) 82 (15) 26.3 (5.4) 0.82 (0.06) 41 36 14 17 13

95 119 (15) 76 (13) 24.8 (4.1) 0.80 (0.05) 15 23 4 18 2

100 128 (22) 82 (16) 27.8 (6.0) 0.82 (0.05) 32 46 20 18 10

87 132 (25) 82 (14) 26.4 (4.9) 0.83 (0.06) 49 37 14 20 22

88 144 (25) 88 (16) 26.1 (6.1) 0.83 (0.06) 73 40 16 14 19

Values represent means and the standard deviation. BMI, body mass index; WHR, waist-to-hip ratio. Overweight and obesity were defined as BMI ⭓27.3 and ⭓32.3, respectively.32 Journal of Human Hypertension


590

Table 2 Age-specific prevalences (%) of abnormal urinary protein measures in normotensive and hypertensive women, Marondera, Zimbabwe, 1995 Age group

No.

Hematuria Proteinuriaa Abnormal ␤2MCRb MAU (ACR ⭓3.0)†

370 331 331 329

Total

25–34

35– 44

45–54

55 and older

NT

HT

NT

HT

NT

HT

NT

HT

NT

HT

10 4 8 4

11 7 11 16

16 4 9 6

36 0 11 0

9 3 6 3

13 11 7 19

2 5 5 5

7 5 5 15

8 0 18 0

8 7 17 19

Hematuria and proteinuria were determined by dipstick methods. a Urine samples with hematuria were excluded. b Urine samples with hematuria and two samples with ACR ⭓25.0 mg/mmol were excluded. Abnormal ␤2MCR was defined as ␤2MCR ⭓34.0. NT, normotensive; HT, hypertensive; ␤2MCR, ␤2-microglobulin-to-creatinine ratio (␮g/mmol); MAU, microalbuminuria; ACR, albuminto-creatinine ratio (␮g/mmol).

Figure 1 Distributions of the values of ACR in normotensive and hypertensive women, Marondera, Zimbabwe, 1995. Horizontal lines represent the 10th, 25th, 50th, 75th and 90th percentiles of ACR. Urine samples with hematuria and two samples with ACR ⭓25.0 were excluded. Asterisks denote significant differences in the means of log ACR between normotensive (NT) and hypertensive (HT) women in age groups 35– 44 (P ⬍ 0.001), 45–54 (P = 0.009) and age group 55 and older (P ⬍ 0.001). ACR, albumin-to-creatinine ratio (mg/mmol).

Table 3 shows Spearman correlation coefficients of SBP with ACR and ␤2MCR in each age group. Except for the age group 25–35, ACR was positively correlated with SBP. ACR was also significantly correlated with DBP in age group 35– 44 (r ⫽ 0.24, P ⫽ 0.01). In the overall group, ACR was not correlated with age (r ⫽ 0.02, P ⫽ 0.77), BMI (r ⫽ ⫺0.05, Table 3 Spearman correlation coefficients of systolic blood pressure with urinary measures in women, Marondera, Zimbabwe, 1995 Age group Number ACR ␤2MCR

Total

25–34

329 76 0.18*** −0.12 0.03 −0.13

35– 44 89 0.32** −0.07

45–54 55 and older 83 0.19† 0.11

81 0.31** 0.16

Urine samples with hematuria and two samples with ACR ⭓25.0 mg/mmol were excluded. ACR, albumin-to-creatinine ratio (mg/mmol); ␤2MCR, ␤2microglobulin-to-creatinine ratio (␮g/mmol). † P ⬍ 0.1; *P ⬍ 0.05; **P ⬍ 0.01, ***P ⬍ 0.001. Journal of Human Hypertension

P ⫽ 0.42), waist-to-hip ratio (r ⫽ 0.07, P ⫽ 0.24), percent fat mass (r ⫽ ⫺0.09, P ⫽ 0.11) or Na/K ratio (r ⫽ ⫺0.09, P ⫽ 0.12). When the age-adjusted means of selective variables for normoalbuminuric (n ⫽ 299) and microalbuminuric groups (n ⫽ 30) were compared, the levels of SBP and log ␤2MCR were significantly higher in the microalbuminuric group, and the prevalences of hypertension and alcohol consumption were also higher in microalbuminuric group (Table 4). The relationship between SBP and log ACR was further explored using a general additive model which allows for relaxation of the assumption of linearity. There were both linear (P ⫽ 0.002) and nonlinear (P ⫽ 0.02) components to the relationship between log ACR and SBP. The level of log ACR increased between a SBP of approximately 125 to 150 mm Hg and then remained constant at higher SBP. In age-adjusted multiple regression, the Table 4 Differences in the age group-adjusted means and prevalences (%) of selective variables in women with normoalbuminuria and microalbuminuria, Marondera, Zimbabwe, 1995 Normoalbuminuria Number Log ␤2MCR SBP (mm Hg) DBP (mm Hg) BMI (kg/m2) WHR Percent fat mass Na/K ratio Hypertensive (%) Overweight (%) Obese (%) Drinks alcohol (%) Uses tobacco (%)

299 0.89 (0.03) 129 (1.3) 81 (0.8) 26.3 (0.3) 0.82 (0.01) 0.37 (0.01) 3.6 (0.2) 37 37 14 16 13

Microalbuminuria

1.23 141 87 25.7 0.83 0.35 2.9

30 (0.10) (4.0) (2.6) (1.0) (0.01) (0.02) (0.7) 73 30 10 33 13

P

0.001*** 0.007** 0.054 0.55 0.40 0.29 0.34 0.002** 0.32 0.43 0.02* 0.72

Values represent least squared means and standard error of the mean. Urine samples with hematuria and two samples with ACR ⭓25.0 mg/mmol were excluded. The least squared means of log ACR in normoalbuminuric and microalbuminuric groups are −0.46 ± 0.03 and 0.82 ± 0.09, respectively. ACR, albumin-to-creatinine ratio (mg/mmol); ␤2MCR, ␤2microglobulin-to-creatinine ratio (␮g/mmol); BMI, body mass index; WHR, waist-to-hip ratio.


relationship of SBP and log ACR remained significantly and similarly non-linear (P ⫽ 0.02). Percent fat mass was negatively associated with log ACR (␤ ⫽ ⫺1.18, 95% CI (⫺0.23, ⫺2.21), P ⫽ 0.02), and Na/K ratio was associated with a higher level of log ACR (␤ ⫽ 0.04, 95% CI (⫺0.001, 0.061), P ⫽ 0.04). Alcohol consumption was not related to log ACR (P ⫽ 0.62). The model containing SBP, percent fat mass, Na/K ratio and alcohol consumption accounted for 9.6% of the variation in log ACR. When BMI was substituted for percent fat mass as a measure of adiposity, a negative, but not significant association with log ACR was observed (P ⫽ 0.11). In a multiple linear regression analysis, higher log ␤2MCR was very strongly associated with higher log ACR (␤ ⫽ 0.34, 95% CI (0.25, 0.43), P ⬍ 0.0001). It was also significantly associated with lower percent fat mass (␤ ⫽ ⫺1.02, 95% CI (⫺0.25, ⫺1.8), P ⫽ 0.01) and higher Na/K ratio (␤ ⫽ 0.05, 95% CI (0.02, 0.08), P ⫽ 0.002). Log ␤2MCR was not associated with SBP (P ⫽ 0.92). The effect of SBP on ␤2MCR might be mediated through the same pathway as the one causing an elevation of log ACR. To examine the possibility that the lack of association between SBP and ␤2MCR might be due to the inclusion of log ACR in the model,41 we reran the model without log ACR. In the reduced model, SBP was more strongly, but still not significantly, associated with ␤2MCR (P ⫽ 0.12). Overall, the model, which contained log ACR, SBP, percent fat mass and Na/K ratio accounted for 21% of the variation in log ␤2MCR.

Discussion Microalbuminuria is present in 9% of the urban Zimbabwean women in this study. When ageadjusted to the population in the community, the estimated prevalence of MAU is 8% among women 25 years and older. The prevalence of MAU is substantially higher in the hypertensive compared to normotensive women (16% vs 4%, P ⬍ 0.001). In general, the proportion of Zimbabwean women with MAU, 8%, fell within proportions measured in other population-based studies. Estimates of MAU range from 2 to 28% and are higher among nonEuropean populations than among European populations.28 In a study of young black American women (aged 19–32), the prevalence of MAU was 5.7%.42 The variations in the prevalence of MAU among different populations may be caused by variations in the levels of risk factors.15,16,43 As reported in other studies24,29 the levels of MAU are higher at older ages: the prevalence in the youngest age group (25–34 years) is 5% and in the women 55 and older is 14%. The levels of MAU in different groups of people with hypertension are highly variable.30 In a population-based study of 518 white hypertensive women, the prevalence of MAU was 2%,12 which is considerably lower than the value in our study population. In a small clinic-based study of middleaged men and women, 32% of the black hypertensives had MAU.44 This was higher than the 14% of white hypertensives in the same study. In a study of middle-aged hypertensive Jamaicans, the prevalence

was 27%.45 Based on the high levels of MAU in black hypertensives in other populations44,45 and the high proportion of hypertension in Zimbabwean women,6 a very high level of MAU might have been expected in the study population, particularly, in the older Zimbabwean women. However, the level of MAU in Zimbabwean women with hypertension in the different age groups ranged from 0 to 19%, proportions that fall in the middle to lower end of the levels reported in other studies (0– 40%),12,44,45 and thus the level of MAU is relatively moderate compared to other populations.46 Comparison of the levels of MAU in different populations are complicated by the use of differing criteria for MAU, ways of measuring albumin excretion, lack of information on age-specific prevalences, and because of variations in the method of BP measurements and the definition of hypertension.15,16,43 In our study, borderline hypertensives were included in the HT group, and a cut-off point of ACR ⭓3.0 mg/mmol in spot urine samples was used to detect MAU, which has been shown to predict MAU with 92% sensitivity and 90% specificity in non-diabetic subjects.25,27 Other investigators used a cut-off point of ACR ⭓3.5 in first morning urine samples and found 4% of MAU in women in the community-based population.24 When the cutoff point of ACR ⭓3.5 was used in our study, only two samples were excluded from the microalbuminuric group. The relationship between log ACR and SBP or MAU and hypertension varied by age. In each age group except the youngest age group (age 25–34), there was a significantly higher level of log ACR in HT compared to NT women, and there was a positive correlation of ACR with SBP. In the age group 25–34, the prevalence of hypertension was relatively low (15%), and none of the hypertensive women showed MAU. Other studies have suggested that there may not be a correlation between BP and albumin excretion in young healthy subjects.47 It is also possible that the relationship between SBP and ACR could be affected by the large proportion (19%) of samples in this age group, which were excluded from the final analysis because of hematuria. Some studies have described no or a weak relationship between SBP or mean arterial pressure and albumin excretion at BP levels below the hypertensive range,48,49 while another found a log-linear relationship.46 We observed an increase in ACR between SBP of 苲125 and 150 but very little additional increase at higher blood pressures. This pattern remained constant after adjustment for potential confounders and was fairly constant over the different age groups (data not shown). We observed an inverse relation between measures of adiposity such as percent fat mass or BMI and log ACR. The relationship between lower percent fat mass and higher log ACR was most pronounced in women between 35–55 years of age. This is in contrast to many studies that have reported a positive association between some measures of adiposity such as BMI and increased albumin excretion,19,20,46 although a Jamaican study did not find an association between ACR and measures of

591

Journal of Human Hypertension


592

obesity such as BMI.45 It was unexpected because we had previously found that higher BMI is strongly related to higher BP in this group of Zimbabwean women.6 One possible explanation for our findings is that creatinine output is correlated with body surface area.50 We found a small but significant correlation between urinary creatinine concentration and BMI (r ⫽ 0.12, P ⫽ 0.03) but did not observe any correlation between ACR and urinary creatinine concentration (r ⫽ ⫺0.04, P ⫽ 0.43). Therefore, it is unlikely that the negative association observed in this study is due to confounding by the amount of creatinine in the urine. It has been suggested that in hypertensives, MAU is the consequence of glomerular dysfunction rather than the result of tubular dysfunction.43,51 In our study, elevated ACR but not ␤2MCR was associated with hypertension. This suggests an increased level of glomerular dysfunction in HT women. ␤2MCR, in contrast, was more strongly associated with a lower percent fat mass than ACR. Despite these differences, there was a very strong relationship between the levels of ACR and ␤2MCR. Twenty-eight percent of the women with MAU showed an abnormal level of ␤2M excretion (␤2MCR ⭓34.0 ␮g/mmol) whereas only 8% of the women without MAU had an abnormal ␤2M excretion. One possible explanation is that another process in addition to hypertension is responsible for the related elevations in ACR and ␤2MCR. The urinary ␤2M values have been found to vary widely among populations. In our study population (without hematuria), the 95th percentile of ␤2MCR was 94.4 ␮g/mmol (834 ␮g/g) compared to 166– 500 ␮g/g in other populations.39,52 Acquired immune deficiency syndrome (AIDS) can increase albumin and ␤2M excretion.53,54 Upper urinary tract infection such as pyelonephritis can also increase the level of ␤2M excretion.53 One limitation of this study is that we did not have an independent measure of urinary tract infection or of other diseases that might cause elevated ACR. In particular, we do not have information on the presence of HIV or of diabetes by a method more sensitive than glucose in the urine. The presence of these diseases may complicate the interpretation of the relationships of SBP, log ACR and measures of obesity in this population. Because we only had a single urine sample from which to determine the ACR, a temporary infection or other stress to the kidney could have caused an abnormally high value. Even in the absence of factors identified with variation in albumin excretion, such as increased physical activity, the level of albumin excretion can vary by 40% from day to day and more variation is induced by the use of ACR.46,56 This variation would most likely have had the effect of attenuating the relationships between urinary albumin and systolic blood pressure and may cause some misclassification of the level of MAU. As with all crosssection studies a causal relationship between elevated SBP and elevated ACR can not be inferred from the data.55 Nine percent of this population has microalbuminuria which indicates the common presence


of diseases that affect renal function, among which are hypertension. The association of MAU and hypertension in this population is another indication that the population may have an increased risk of cardiovascular and renal complications. Further studies and analysis of other known cardiovascular and renal risk factors will be useful in determining the patterns of development of hypertension related diseases in this population.

References 1 Seedat YK, Seedat MA, Nkomo MN. The prevalence of hypertension in the urban Zulu. S Afr Med J 1978; 53: 923–927. 2 Seftel HC, Johnson S, Muller EA. Distribution and biosocial correlations of blood pressure levels in Johannesburg Blacks. S Afr Med J 1979; 57: 313–320. 3 Simmons D et al. Blood pressure and salt intake in Malawi: an urban rural study. J Epidemiol Community Health 1986; 40: 188–192. 4 Siziya S, Marufu T, Matchaba-Hove RB. Relationship of casual blood pressure to smoking, education and occupation in a high density town near Harare, Zimbabwe. Cent Afr J Med 1995; 41: 219–224. 5 Sparks BT et al. Prevalence of hypertension among women in rural Zimbabwe: a comparison of pregnant and non-pregnant women. Cent Afr J Med 1996; 42: 93–97. 6 Mufunda J et al. Correlates of blood pressure in an urban Zimbabwean population and comparison to other populations of African origin. J Hum Hypertens 2000; 14: 65–73. 7 Laville M et al. Epidemiological profile of hypertensive disease and renal risk in black Africa. J Hum Hypertens 1994; 12: 839–843. 8 Mensah GA, Barkey NL, Cooper RS. Spectrum of hypertensive target organ damage in Africa: a review of published studies. J Hum Hypertens 1994; 8: 799–808. 9 Steyn K, Fourie J, Bradshaw D. The impact of chronic diseases of lifestyle and their major risk factors on mortality in South Africa. S Afr Med J 1992; 82: 227– 231. 10 Veriava Y et al. Hypertension as a cause of end-stage renal failure in South Africa. J Hum Hypertens 1990; 4: 379–383. 11 Bigazzi R et al. Microalbuminuria predicts cardiovascular events and renal insufficiency in patients with essential hypertension. J Hypertens 1998; 16: 1325–1333. 12 Jensen JS et al. Microalbuminuria and its relation to cardiovascular disease and risk factors: a populationbased study of 1254 hypertensive individuals. J Hum Hypertens 1997; 11: 727–732. 13 Pontremoli R et al. Prevalence and clinical correlates of microalbuminuria in essential hypertension: the MAGIC study. Hypertension 1997; 30: 1135–1143. 14 Spangler JG, Konen JC. Hypertension, hyperlipidemia, and abnormal obesity and the development of microalbuminuria in patients with non-insulin-dependent diabetes mellitus. J Am Board Fam Pract 1996; 9: 1–6. 15 Bianchi S, Bigazzi R, Campese VM. Microalbuminuria in essential hypertension. J Nephrol 1997; 10: 216– 218. 16 Ruilope LM. Microalbuminuria as risk in essential hypertension. Nephrol Dial Transplant 1997; 12 (Suppl 2): 2–5. 17 Mykkanen L et al. Microalbuminuria is associated with insulin resistance in nondiabetic subjects: the insulin resistance atherosclerosis study. Diabetes 1998; 47: 793–800.


18 Nesovic M et al. Microalbuminuria is associated with salt sensitivity in hypertensive patients. J Hum Hypertens 1996; 10: 573–576. 19 Ribstein J, DuCailar G, Mimran A. Combined renal effects of overweight and hypertension. Hypertension 1995; 26: 610–615. 20 Valensi P et al. Microalbuminuria in obese patients with or without hypertension. Int J Obes 1996; 20: 574 –579. 21 Mimran A, Ribstein J, DuCailar G. Is microalbuminuria a marker of early intrarenal vascular dysfunction in essential hypertension? Hypertension 1994; 23: 1018–1021. 22 Pedrinelli R et al. Microalbuminuria and endothelial dysfunction in essential hypertension. Lancet 1994; 344: 14 –18. 23 Gosling P. Microalbuminuria and cardiovascular risk: a word of caution. J Hum Hypertens 1998; 12: 211– 213. 24 Winocour PH et al. Microalbuminuria and associated cardiovascular risk factors in the community. Atherosclerosis 1992; 93: 71–81. 25 James MA, Fotherby MD, Potter JF. Screening tests for microalbuminuria in non-diabetic elderly subjects and their relation to blood pressure. Clin Sci 1995; 88: 185–190. 26 Gerber LM, Johnston K, Alderman MH. Assessment of a new dipstick test in screening for microalbuminuria in patients with hypertension. Am J Hypertens 1998; 11: 1321–1327. 27 Pontremoli R et al. Prevalence of micro-and macroalbuminuria and their relationship with other cardiovascular risk factors in essential hypertension. Nephrol Dial Transplant 1995; 10 (Suppl 6): 6–9. 28 Metcalf P et al. Albuminuria in people at least 40 years old: effect of obesity, hypertension, and hyperlipidemia. Clin Chem 1992; 38: 1802–1808. 29 Ritz E et al. Proteinuria and hypertension. Kidney Int 1994; 46: S76–S80. 30 Pedrinelli R. Microalbuminuria in hypertension. Nephron 1996; 73: 499–505. 31 Central Statistics Office. Census 1992. Provincial profile, Harare. Central Statistical Office. Harare, Zimbabwe. 1992. 32 Rotimi CN et al. Distribution of anthropometric variables and the prevalence of obesity in populations of west African origin: the International Collaborative Study on Hypertension in Blacks (ICSHIB). Obes Res 1995; 3: S95–S105. 33 Ataman SL et al. Standardization of blood pressure measurement in an international comparative study. J Clin Epidemiol 1996; 49: 869–877. 34 Mufunda J et al. Comparison of the Omron HEM-713C automated blood pressure monitor with a standard ausculatory method using a mercury manometer. Cent Afr J Med 1996; 42: 230–232. 35 Cooper R et al. Evaluation of an electronic blood pressure device for epidemiologic studies. Blood Press Monit 1997; 2: 35– 40. 36 Luke A et al. Relation between body mass index and body fat in black population samples from Nigeria, Jamaica, and the United States. Am J Epidemiol 1997; 145: 620–628. 37 Kushner RF, Schoeller DA. Estimation of total body water by bioelectrical impedance analysis. Am J Clin Nutr 1986; 44: 417– 424.

38 American Conference of Governmental Industrial Hygienists Inc. Cadmium in documentation of the threshold limit values and biological exposure indices: 6th Edn, Vol 3. American Conference of Governmental Industrial Hygienists Inc: Cincinnati, 1991. 39 Bernard AM, Moreau D, Lauwerys R. Comparison of retinol-binding protein and ␤2-microglobulin determination of urine for the early detection of tubular proteinuria. Clin Chim Acta 1982; 126: 1–7. 40 Hastie TJ, Tibshirani RJ. Generalized Additive Models. Chapman and Hall: London, 1990, pp 9–35. 41 Robins J. The control of confounding by intermediate variables. Stat Med 1989; 8: 679–701. 42 Jiang X et al. Microalbuminuria in young adults related to blood pressure in a biracial (black-white) population. The Bogalusa Heart Study. J Hypertens 1994; 7: 794 –800. 43 Mimran A, Ribstein J. Microalbuminuria in essential hypertension. J Hum Hypertens 1996; 10: 657–661. 44 Summerson JH, Bell RA, Konen JC. Racial differences in the prevalence of microalbuminuria in hypertension. Am J Kidney Dis 1995; 26: 577–579. 45 Reid M et al. Microalbuminuria, renal function and waist:hip ratio in black hypertensive Jamaicans. J Hum Hypertens 1998; 12: 221–227. 46 Metcalf PA, Scragg RK Epidemiology of microalbuminuria in the general population. J Diabetes Complications 1994; 8: 157–163. 47 Ljungman S. Microalbuminuria in essential hypertension. Am J Hypertens 1990; 3: 956–960. 48 Gosling P, Beevers DG. Urinary albumin excretion and blood pressure in the general population. Clin Sci 1989; 76: 39– 42. 49 Watts GF et al. Urinary albumin excretion in healthy adult subjects: reference values and some factors affecting their interpretation. Clin Chim Acta 1988; 172: 191–198. 50 Graystone JE. Creatinine excretion during growth. In: Cheek, DB (ed). Human growth. Lea and Febiger Co.: Philadelphia, 1968, pp 182–187. 51 Mogensen CE et al. Microalbuminuria: studies in diabetes, essential hypertension, and renal diseases as compared with the background population. Adv Nephrol 1991; 20: 191–228. 52 Sakurai H et al. Biological monitoring of cadmiumexposed workers. In: Fiserova-Bergerova V, Ogata M (eds). Biological monitoring of exposure to industrial chemicals. American Conference of Governmental Industrial Hygienists Inc: Cincinnati, 1990, pp 105– 111. 53 Kabanda A et al. Low molecular weight proteinuria in human immunodeficiency virus-infected patients. Am J Kidney Dis 1996; 27: 803–808. 54 Luke DR, Sarnoski TP, Dennis S. Incidence of microalbuminuria in ambulatory patients with acquired immunodeficiency syndrome. Clin Nephrol 1992; 38: 69–74. 55 Howie AJ. ‘Benign’ essential hypertension and kidney damage: a histopathologist’s view. J Hum Hypertens 1996; 10: 691–694. 56 Mogensen CE et al. Microalbuminuria and potential confounders. A review and some observations on variability of urinary albumin excretion. Diabetes Care 1995; 18: 572–581.

593
