Obesity has been shown to be positively

Transcription

1 508 Body Fat and Its Distribution in Relation to Casual and Ambulatory Blood Pressure Linda M. Gerber, Peter L. Schnall, and Thomas G. Pickering This study was undertaken to evaluate the associations of body fat and its distribution with casual and ambulatory blood pressure in nonobese men. One hundred and thirty-five normotensive or mildly hypertensive (but untreated) men employed at three work sites were studied. Casual blood pressure was measured at the work site at initial screening and on a second occasion by a nurse. Ambulatory blood pressure was measured noninvasively for 24 hours on a workday and analyzed as work, home, and sleep blood pressure measurements. Anthropometric measurements included height, weight, and waist and hip circumferences. Blood pressure was highest while at work; home blood pressure was higher than screening blood pressure or nurse blood pressure, and sleep blood pressure was lowest. Weight and both waist and hip circumferences (but not their ratio) were all significantly correlated with screening, nurse, and sleep blood pressures but not with work or home blood pressures. Stepwise regression analysis showed that waist circumference was the best overall predictor of blood pressure. We suggest that in situations where blood pressure is the dependent variable, correlations with other variables may be closest for "basal" measures of blood pressure and may be obscured by the effects of daily activities on blood pressure. (Hypertension 1990;15: ) Obesity has been shown to be positively related to essential hypertension. 1-4 Considerable evidence also exists to support an association between the distribution of body fat and blood pressure. Measures of centrally located or upper body fat predominance have been shown to be positively related to levels of both systolic and diastolic blood pressure. 5 " 11 Few studies, however, have examined the parameters of weight, body fat distribution, and blood pressure together, especially in a nonobese population. Hypertension and blood pressure level in general have methodological problems in measurement. Blood pressure is influenced by the instrument and the individual taking the measure, the subject's physical activity and behavior, and the location and setting in which the blood pressure is obtained Ambulatory blood pressure monitoring offers an alternative measurement to casual blood pressure. Ambulatory From the Department of Medicine, Cardiovascular Center, The New York Hospital-Cornell University Medical College, New York, New York. Supported in part by grant HL from the National Heart, Lung, and Blood Institute. Presented in part at the 58th Annual Meeting of the American Association of Physical Anthropologists in San Diego, California, April 6-8, Address for correspondence: Linda M. Gerber, PhD, Cardiovascular Center, The New York Hospital-Cornell University Medical College, 525 East 68th Street, Starr-4, New York, NY Received September 7, 1989; accepted in revised form January 22, monitoring has been reported to be more reliable because of the absence of observer error, the increased number of readings, and the ability to measure blood pressure during usual activities of the subject There are no published data, however, on the relation between ambulatory blood pressure and indexes of overall obesity and fat distribution. The present study examines the relation of measures of body mass and fat distribution with two measures of casual blood pressure and three of ambulatory blood pressure in a nonobese male population. Methods Subjects were participants in a case-control study of working men employed at eight New York City work sites. These subjects are part of a study evaluating the effects of exposure to psychosocial, biological, and anthropometric factors on casual and ambulatory blood pressure. After comprehensive screening of 3,223 male employees, all eligible research subjects with screening diastolic blood pressure more than 85 mm Hg and a random sample (one in eight) of eligible research subjects with screening diastolic blood pressure 85 mm Hg or less were studied. Of 1,466 eligible subjects, 288 were recruited into the study. One hundred thirty-five subjects at the last three sites had weight, height, and waist and hip circumferences measured and constitute the sample for this study. After the research protocol and all attendant risks were

2 Gerber et al Body Fat Distribution and Blood Pressure. 509 reviewed, informed consent was obtained from each participant at the time of recruitment. To be included in the study, subjects had to be between 30 and 60 years of age, employed more than 30 hours per week for at least 3 years in their current job before the onset of high blood pressure, educated in the United States and able to read English, have a body mass index not greater than 30% above ideal, and have no second job of 15 or more hours per week. Subjects were excluded if they had a history of cardiovascular disease or systolic blood pressure greater than 160 mm Hg or diastolic blood pressure greater than 105 mm Hg at screening. Casual screening blood pressures were determined at the worksite on a workday for all 3,223 men. Blood pressure was measured by trained nurses and technicians using a standardized protocol based on American Heart Association criteria. Three blood pressure measurements were taken in a sitting position, and the average of the last two was used as the estimate of blood pressure. For all subjects, a standard size blood pressure cuff was used in determining blood pressure. Using the American Heart Association protocol as described above, a second set of blood pressure measurements was taken by a nurse-practitioner at the work site, usually within 4 weeks of the screening, either in a room in a medical clinic or in office space converted into an examining room. Subjects then wore a Spacelabs 5200 ambulatory blood pressure monitor (Hillsboro, Oregon) for 24 hours beginning at the start of a work day. The monitor was attached at the subject's work site and calibrated by comparison of five successive systolic and diastolic readings against simultaneously determined casual readings taken with a mercury column in which each had to be within 5 mm Hg to be acceptable. The timer on the monitor was set to take readings at 15-minute intervals during the day and at 30-minute intervals during the hours of sleep, and the subject was instructed to proceed through a normal workday. Each time the monitor took a reading during waking hours, the subject was asked to remain as motionless as possible and then to record his activity, location, position, and mood in a diary. Averages could then be calculated for work, home, and sleep blood pressure measurements. Subjects taking medication for hypertension (n=10) were titrated off treatment under medical supervision and wore the ambulatory monitor after all medication was stopped for 3 weeks. Body mass index was calculated according to the formula weight (kg)/height (m) 2. Waist and hip circumferences were measured by one observer using a steel tape. While the subject was standing, the waist was measured at its minimum with the abdomen relaxed and the tape held behind the subject with one edge in the horizontal plane through the center of the umbilicus. The tape was then wrapped carefully around the patient's torso and used as an aid in marking the horizontal plane on TABLE 1. Characteristics of Study Population Characteristic Age (yr) Weight (kg) Height (cm) Body mass index (kg/m 2 ) Waist circumference (cm) Hip circumference (cm) Waist/hip ratio n Mean±SD 43.2± ± ± ± Range the sides and back. The hip circumference was measured at its maximum, with the tape held at the top of the patient's hipbone and then wrapped carefully around the torso. 17 A measure of physical exertion was derived from an item on a self-reported questionnaire. The item, scored from 1 through 4, from strongly disagree to strongly agree, was "My job requires lots of physical effort." Pairwise correlation coefficients were calculated for all independent variables with each other and with the different blood pressure measures. Multiple linear regression analyses were performed using a stepwise variable selection procedure. Two-tailed probability levels for statistical significance tests are reported. Because three different worksites were included in the analysis, site was examined as a potential confounding variable. Regression analyses were performed with and without site as an independent variable. As site did not alter the results, the findings are presented without site. Results Table 1 describes the characteristics of the study population. The men were between 30 and 60 years old with a mean age of approximately 43 years. The mean body mass index was 26, equivalent to roughly the 50th percentile for the US population. Mean waist and hip circumferences were very similar resulting in a mean waist/hip ratio close to 1. Ambulatory, screening, and nurse blood pressure measurements of the study population are presented in Table 2. Of note is the finding that mean ambulatory work blood pressures were the highest, followed by home, and then by sleep blood pressures. This is the case for both systolic and diastolic blood pressures. Mean screening and nurse blood pressure measurements were substantially lower than either work or home ambulatory blood pressures. In addition, the nurse systolic blood pressure measurement was even lower than the screening blood pressure, although for diastolic blood pressure the two were very similar. The correlation coefficients between the anthropometric variables, age, and physical exertion and ambulatory, screening, and nurse systolic and diastolic blood pressure measurements are shown in Table 3. Age was consistently and significantly

3 510 Hypertension Vol 15, No 5, May 1990 TABLE 2. Ambulatory, Study Population Measurements Ambulatory blood pressure (mm Hg) Systolic 125 Diastolic 125 blood pressure (mm Hg) Systolic Diastolic " blood pressure (mm Hg) Systolic Diastolic, and Blood Pressure of n Mean+SD 130.1± ± ± ± ± ± ± ±9.1 Range related to all measures of blood pressure except sleep systolic pressure. The anthropometric variables, on the other hand, did not show as straightforward a pattern. Among measures of ambulatory systolic blood pressure, sleep blood pressure had the strongest correlation with anthropometric parameters. Weight, height, and waist and hip circumferences all had significant correlations with sleep systolic blood pressure and weaker associations with work and home blood pressure measurements. For the most part, these anthropometric variables, except for height, were even more strongly related to nurse and screening blood pressure measurements when compared with sleep blood pressure; nurse blood pressure had the strongest relation. Physical exertion was not significantly related to any systolic blood pressure measure. With diastolic ambulatory blood pressure, age again had the most consistent correlation across all measures. However, physical exertion had a stronger and negative correlation with work and home blood pressure measurements. blood pressure was also significantly and negatively associated with physical exertion, whereas no relation existed between physical exertion and screening and nurse blood pressure. Among ambulatory diastolic pressures, all relations with anthropometric variables were strongest for sleep blood pressures. Of these, only waist and hip circumferences reached significance. As with systolic blood pressure, the correlations between diastolic blood pressure and anthropometric variables were more often significant and greater for the nurse and the screening blood pressures than the ambulatory measures. Again, the nurse blood pressure had even stronger relations than the screening blood pressure. As expected, among the anthropometric variables, there were many significant associations (Table 4). Weight and height and waist, hip, and body mass index were all significantly correlated. Height was also significantly related to waist and hip circumferences. Waist, in addition, had very high correlation coefficients with weight (r=0.87), hip (r=0.94), and body mass index (r=0.83). A multivariate procedure was used to quantify the relative contribution of each variable to blood pressure level when examined in the presence of the others. The six anthropometric variables, age, and physical exertion were analyzed with a stepwise multiple regression model. The results for ambulatory systolic and diastolic blood pressure are included in Table 5. For both work and home blood pressures, only age was found to be significantly predictive. The percent of variance accounted for by age was approximately 8 and 6% for work and home, respectively. Waist circumference was found to be predictive of sleep blood pressure, explaining 7% of the variance. For ambulatory diastolic blood pressure, physical exertion and waist circumference and physical exertion and age were significant contributors to the fraction of the explained variance for work and home blood pressures, respectively. diastolic blood pressure was dependent in a first step on age, in a second step on physical exertion, and in a third step TABLE 3. Correlation Coefficients Between Anthropometric Variables, Age, and Physical Exertion, and Ambulatory,, and Blood Pressure Systolic blood pressure Diastolic blood pressure Ambulatory Ambulatory Variable Weight Height Body mass index Waist circumference Hip circumference Waist/hip ratio Age Physical exertion *p< t * " 0.24' 0.26' 0.22' * 0.29t 0.30t 0.41f t 0.28t 0.37t 0.36t 0.38t * -0.36t * * 0.21* 0.29f -0.24* 0.23* 0.27t 0.32t 0.32t * It t 0.40t 0.36t t

4 Gerber et al Body Fat Distribution and Blood Pressure 511 TABLE 4. Variable Intrasubject Correlations of Anthropometric Measurements and Age Weight Weight Height BMI Waist Hip Waist/hip ratio Age BMI, body mass index. *p<0.01. tp<0.05. Height 0.56* BMI 0.84* Correlation coefficients Waist 0.87* 0.35* 0.83* Hip 0.86* 0.38* 0.78* 0.94* Waist/hip ratio 0.20t * 0.36* Age ' 0.28* on waist circumference. The addition of waist circumference to work blood pressure contributed nearly 5% to the explained variance. Waist circumference contributed to explaining an additional 4% to sleep blood pressure variance. Table 6 shows the results of regression analysis of screening and nurse blood pressure. Systolic blood pressure was dependent in a first step on age and in a second step on hip circumference. Both weight and waist circumference were significant after adjusting for age and could have been substituted for hip circumference. Age and hip circumference together explained nearly 22% of the variation of screening systolic blood pressure. Age and waist circumference contributed significantly to predicting nurse systolic blood pressure. With diastolic blood pressure, waist circumference was the only variable to significantly predict screen- TABLE 5. Stepwise Multiple Regression Analysis of Ambulatory Systolic and Diastolic Blood Pressure Dependent variable Multiple R r 2 t Systolic blood pressure Wnrt wurk Step 1: waist Diastolic blood pressure Step 1: physical exertion Step 2: waist Step 1: physical exertion Step 2: age Step 2: physical exertion Step 3: waist Independent variables included were age, weight, height, body mass index, waist and hip circumferences, waist/hip ratio, and physical exertion. ing blood pressure. Waist circumference and age both contributed significantly to predicting nurse diastolic blood pressures. Discussion These results suggest that, in nonobese hypertensive and normotensive men, the relation of measures of body fat and distribution to level of blood pressure varies depending on the measure of blood pressure examined. The anthropometric variables were all highly intercorrelated as well as highly correlated with age and were also significantly correlated with measures of ambulatory, screening, and nurse blood pressures. The strongest relations among ambulatory systolic and diastolic measures were found between sleep blood pressure and waist and hip circumferences. and nurse systolic and diastolic blood pressure measurements had even higher correlation coefficients with waist and hip circumferences than sleep blood pressures. In addition, weight TABLE 6. Stepwise Multiple Regression Analysis of and Systolic and Diastolic Blood Pressure Dependent variable Systolic blood pressure Step 2: hip* Step 2: waist Diastolic blood pressure Step 1: waist Multiple R r ( Step 1: waist Step 2: age Independent variables included were age; weight, height, body mass index, waist and hip icircumferences, waist/hip ratio, and Dhvsical exertion. 'Weight («=) and waist (r=) were significant after adjusting for age but did not enter into the equation after hip was added.

5 512 Hypertension Vol 15, No 5, May 1990 was significantly related to both systolic and diastolic screening and nurse blood pressure measurements. Stepwise regression analyses showed that the anthropometric variables best predicted, of all the blood pressure measurements, ambulatory sleep, screening, and nurse blood pressures. For sleep blood pressures, waist circumference for systolic and diastolic blood pressures contributed significantly to the fraction of explained variance. The anthropometric parameters found to be predictive of screening systolic blood pressure level included waist circumference, hip circumference, and weight, whereas only waist circumference predicted screening diastolic blood pressure. Waist circumference contributed significantly to predicting both nurse systolic and diastolic blood pressures. One caveat that should be kept in mind is that many of the anthropometric measures were highly correlated with each other (e.g., waist and hip circumferences, r=0.94), which created a problem of multicollinearity in the database. As a consequence, in many of the stepwise regression analyses, hip circumference was virtually indistinguishable from waist circumference in its relation to blood pressure, and weight and body mass index lagged only slightly behind. Two important conclusions emerge from our findings. First, although there was a strong relation between waist circumference and many measures of blood pressure, a relatively weak relation between the waist/hip ratio and blood pressure measures was observed. The former finding was not unexpected; the second, however, was. Results from the Framingham study report a correlation coefficient of 0.27 between waist circumference and systolic blood pressure among men, the strongest relation among all the indexes of obesity examined. 18 Berglund et al 19 also found a significant relation between waist circumference and mean arterial pressure among men. In this study also, blood pressure was found to be more highly correlated to waist circumference than to body weight. 19 The present study confirms these observations and extends their validity as waist circumference was found to be strongly related to ambulatory as well as screening and nurse blood pressures. The weak relation between the waist/hip ratio and blood pressure measures observed in this study was not expected. Larsson et al 20 reported that as the waist/hip ratio increased, so did levels of both systolic and diastolic blood pressure. Although hip circumference was not directly measured in the Health Examination Survey, Gillum 11 derived an index of fat distribution computed as waist circumference divided by the estimated hip circumference. Systolic and diastolic blood pressures were both found to be significantly correlated with this fat distribution index. The waist and hip circumferences reported for Swedish men were 86.0 and 93.7 cm, respectively. The mean waist/hip ratio was 0.93 cm with a range from 0.75 to 1.10 cm. 20 Although hip circumference data among Framingham residents were not obtained, the mean waist circumference among men was reported to be 91.5 cm. 18 In the present study, the mean waist circumference of 89.5 cm falls between the results of the Swedish and Framingham studies. The mean hip circumference is much closer to the mean waist circumference in the current study compared with that in Sweden, and as a result, the mean waist/hip ratio is much closer to 1.0. In addition, the range and standard deviation of the ratio in the current study is smaller, possibly contributing to reducing the correlations with blood pressure compared with other studies. Nevertheless, the mean waist and hip circumferences in the present study had ranges in values similar to the Swedish study. The associations of waist and hip circumferences to blood pressure measures were also strong and comparable to those in the Swedish study. It may be that the waist/hip ratio is just not as important in relation to blood pressure in this sample. Second, within our database, nurse, screening, and sleep blood pressures were all better correlated with body fat and distribution measures than were ambulatory work and home blood pressures. This result was somewhat surprising as, for several reasons, ambulatory blood pressure measurement had been considered superior to casual blood pressure measurement in reliability 15 and in its association with several measures of target organ damage However, it is possible that the more variable blood pressure occurring during daily activities while at work and at home obscured the relation between anthropometric variables and blood pressure in the present study. The situations that gave the best correlations (ambulatory sleep, nurse, and screening blood pressures) were all circumstances in which daily activities were either eliminated or highly controlled. On the assumption that physical activity at work and at home might confound the weightambulatory blood pressure relation, a measure of physical exertion was entered into our regression model. Although physical exertion was highly correlated with ambulatory diastolic blood pressures, it was not correlated with the anthropometric measures, and when entered into the regression analyses, did not substantially alter the findings. Given the subjective nature of this measure of activity, it is possible that other more objective measures may remain as potential confounders. Our findings of the higher correlations of anthropometric measures with nurse, screening, and ambulatory sleep blood pressures may be interpreted as the consequence of the fact that body mass and distribution directly influence basal blood pressure, whereas work and home ambulatory blood pressures are additionally influenced by both activity and psychosocial factors. It is noteworthy that generally similar correlations were obtained with these three measures of basal pressure, which were taken on three separate occa-

6 Gerberetal Body Fat Distribution and Blood Pressure 513 sions. Although these findings may at first appear to be in conflict with the studies showing that daytime or 24-hour ambulatory blood pressure measures correlate better with target organ damage, there are two important differences in the present study. First, in the case of target organ damage, it is assumed to be the prevailing level of blood pressure that is the independent variable, whereas in the case of the anthropometric measurements, blood pressure is assumed to be the dependent variable, so that the daytime fluctuations of pressure may be merely adding "noise" to the measure of blood pressure. Second, our measures of casual pressure are different from the conventional measures of clinic pressure made by a physician in his office. The closest correlations with anthropometric measurements were obtained with the nurse blood pressure measurements, which were made in relaxed circumstances at the work sites and which were lower than both the daytime ambulatory blood pressure and the screening blood pressure measurements. Numerous studies have shown that physician-measured clinic blood pressures are consistently higher than daytime ambulatory blood pressures. 15 The reason for this difference may be the "white coat" effect provoked by a physician, which varies in extent from one patient to another, hence explaining the generally poor correlation between clinic pressure and blood pressurerelated variables. At the very least, our findings suggest a need for researchers to include varying methods of measuring both body mass and distribution as well as different measures of blood pressure depending on the independent variable of interest. References 1. Kannel W, Brand N, Skinner JJ Jr, Dawber TR, McNamara PM: Relation of adiposity to blood pressure and development of hypertension: The Framingham Study. Ann Intern Med 1967;67: Chiang BN, Perlman LV, Epstein FH: Overweight and hypertension, a review. Circulation 1969;39: Garrison RJ, Kannel WB, Stokes J, Castelli WP: Incidence and precursors of hypertension in young adults: The Framingham Offspring Study. Prev Med 1987;16: Berglund G, Ljungman S, Hartford M, Wilhelmsen L, Bjorntorp P: Type of obesity and blood pressure. Hypertension 1982;4: Weinsier RL, Norris DJ, Birch R, Bernstein RS, Wang J, Yang MU, Pierson RN Jr, Van Itallie TB: The relative contribution of body fat and fat pattern to blood pressure level. Hypertension 1985;7: Krotkiewski M, Bjorntorp P, Sjostrom L, Smith U: Impact of obesity on metabolism in men and women: Importance of regional adipose tissue distribution. / Clin Invest 1983; 72: Hartz AJ, Rupley DC, Rimm AA: The association of girth measurements with disease in 32,856 women. Am J Epidemiol 1984;119: Kalkhoff RK, Hartz AJ, Rupley D, Kissebah AH, Kelber S: Relationship of body fat distribution to blood pressure, carbohydrate tolerance, and plasma lipids in healthy obese women. J Lab Clin Med 1983;102: Blair D, Habicht J-P, Sims EAH, Sylvester D, Abraham S: Evidence of an increased risk for hypertension with centrally located body fat and the effect of race and sex on this risk. Am J Epidemiol 1984;119: Seidell JC, Bakx JC, DeBoer E, Deurenberg P, Hautvast JGAJ: Fat distribution of overweight persons in relation to morbidity and subjective health. Int J Obes 1985;9: Gillum RF: The association of body fat distribution with hypertension, hypertensive heart disease, coronary heart disease, diabetes and cardiovascular risk factors in men and women aged years. / Chron Dis 1987;40: Gordon T, Sorlie P, Kannel WB: Problems in the assessment of blood pressure: The Framingham Study. Int J Epidemiol 1976;5: Carey RM, Reid RA, Ayers CR, Lunch SS, McLain WL, Vaughan ED: The Charlottesville blood-pressure survey: Value of repeated blood pressure measurements. JAMA 1976; 236: Pickering TG, James GD: Some implications of the differences between home, clinic and ambulatory blood pressure in normotensive and hypertensive patients. J Hypertens 1989;7(suppl 3):S65-S James GD, Pickering TG, Yee LS, Harshfield GA, Riva S, Laragh JH: The reproducibility of average ambulatory, home and clinic pressures. Hypertension 1988,11: Pickering TG: The influence of daily activity on ambulatory blood pressure. Am Heart J 1988;116: Weiner JS, Lourie JA: Human Biology, a Guide to Field Methods. Philadelphia, FA Davis, Stokes J III, Garrison RJ, Kannel WB: The independent contributions of various indices of obesity to the 22-year incidence of coronary heart disease: The Framingham Heart Study, in Vague J, Bjorntorp P, Guy-Grand B, Rebuffe-Scrive M, Vague P (eds): Metabolic Complications of Human Obesities. New York/Amsterdam, Elsevier Science Publishing Co, Inc, 1985, pp Berglund G, Hellstrom A, Hartford M, Ljungman S, Wikstrand J: Adipose tissue distribution in relation to blood pressure in the population, in Vague J, Bjorntorp P, Guy- Grand B, Rebuffe-Scrive M, Vague P (eds): Metabolic Complications of Human Obesities. New York/Amsterdam, Elsevier Science Publishing Co, Inc, 1985, pp Larsson B, Svardsudd K, Welin L, Wikhelmsen L, Bjorntorp P, Tibblin G: Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13 year follow-up of participants in the study of men born in Br Med J 1984;228: Sokolow M, Werdegar D, Kaim HK, Hinman AT: Relationship between level of blood pressure measured casually and by portable recorders and severity of complications in essential hypertension. Circulation 1966;34: Rowlands DB, Glover DR, Ireland MA, McLeay RAB, Stallard TJ, Watson RDS, Littler WA: Assessment of left ventricular mass and its response to antihypertensive treatment. Lancet 1982;l: Drayer JIM, Weber MA, DeYoung JL: BP as a determinant of cardiac left ventricular muscle mass. Arch Int Med 1983; 143: Devereux RB, Pickering TG, Harshfield GA, Kleinert HD, Denby L, Clark L, Pregibon D, Jason MN, Kleiner B, Borer JS, Laragh JH: Left ventricular hypertrophy in patients with hypertension: Importance of blood pressure response to regularly recurring stress. Circulation 1983,68: Gosse P, Campello G, Aouizerate E, Roudaut R, Broustet JP, Dallochio M: Left ventricular hypertrophy in hypertension: Correlation with rest, exercise, and ambulatory systolic blood pressure. J Hypertens 1986;4(suppl 5):S297-S Vermeersch P, Duprez D, Packet L, Clement DL: Left ventricular hypertrophy in mild hypertension: Value of ambulatory recordings. / Hypertens 1987;5(suppl 5):S495-S Opsahl JA, Abraham PA, Halstenson CE, Keane WF: Correlation of office and ambulatory blood pressure measurements with urinary albumin and /V-acetyl-/3-D-glocosaminidase excretion in essential hypertension. Am J Hypertens 1988; 1: KEY WORDS: blood pressure ambulatory blood pressure body fat body fat distribution body weight anthropometry