a Department of Community Medicine, University of Cambridge, Cambridge CB2 2SR, UK.
b Department of Mathematics, Hong Kong University of Science and Technology, Hong Kong.
c Clinical Epidemiology Unit, University of Manchester Medical School,M13 9PT, UK.
d Medical Research Council Biostatistics Unit, Cambridge CB2 2SR, UK.
Reprint requests to: Dr NJ Wareham, Department of Community Medicine, Institute of Public Health, University of Cambridge, Robinson Way, Cambridge CB2 2SR, UK. E-mail: njw1004{at}medschl.cam.ac.uk
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Methods In a cross-sectional study of 775 people (4570 years) participating in a continuing population-based cohort study, energy expenditure was assessed by 4 days of heart rate monitoring with individual calibration of the relationship between heart rate and energy expenditure, a method validated against doubly-labelled water and whole body calorimetry. Cardio-respiratory fitness was assessed in a sub-maximal test. To adjust for measurement error in the assessment of usual energy expenditure and fitness, 190 subjects repeated both tests on three further occasions at 4-monthly intervals.
Results A highly significant linear trend in blood pressure was found across quintiles of the physical activity level, the ratio of total energy expenditure to basal metabolic rate. The differences in the mean systolic/diastolic blood pressure between the top and bottom quintile was 6.3/4.4 mmHg in men and 10.7/5.9 mmHg in women. These effects were independent of obesity and cardio-respiratory fitness. Correction for measurement error suggests that the true underlying relationship between usual energy expenditure and blood pressure is stronger still.
Conclusions These findings are compatible with a strong association between usual energy expenditure and blood pressure and support public health strategies aimed at increasing overall energy expenditure.
Keywords Energy expenditure, physical activity, exercise, fitness, blood pressure, hypertension, population study
Accepted 5 January 2000
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The aim of our study was to quantitate the cross-sectional relationship between habitual energy expenditure and blood pressure by using objective and quantitative methods. The gold standard techniques for measuring energy expenditure are doubly labelled water and whole body calorimetry.9 However, the cost of the former and the subject restrictions required of the latter, generally preclude their use in populations.10 More practical methods in free-living individuals include heart rate monitoring with individual calibration of the relationship between heart rate and energy expenditure. This method's validity has been demonstrated by comparison with the gold standard methods.11,12 We used this method in a population-based study to demonstrate that the metabolic syndrome is closely linked to the level of habitual energy expenditure.13,14 In that study, a sub-group underwent repeated measures, so that the regression dilution bias due to measurement error and inherent variability in energy expenditure could be estimated. However, the approach used was univariate. Biostatistical techniques exist for considering the situation where several exposures are measured with error15 and for modelling the sensitivity of the effect size estimates to changes in underlying assumptions.16 These methods are employed here.
![]() |
Patients and Methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Assessment of resting and exercise oxygen consumptionheart rate relationship
Following the measurement of anthropometry and blood pressure, a standard protocol for individually calibrating heart rate and energy expenditure was used.13,14 This method relies on the computation for each individual of resting energy expenditure (REE), and the slope and intercept of the regression line describing the linear relationship between heart rate and energy expenditure during exercise. The final parameter that is measured is the Flex heart rate, which is the level used in the method to distinguish between resting and activity. When heart rate data is collected over the following 4 days, energy expenditure when the heart rate is lower than Flex is assumed to equal REE and when it is above Flex, then it is predicted by the simple linear regression computed during the calibration. In this study, oxygen consumption, energy expenditure and heart rate were measured at rest with the subject lying prone and then seated. Subjects bicycled on a cycle ergometer at several different workloads to provide the slope and the intercept of the line relating energy expenditure to heart rate. Each subject cycled at 50 revolutions per minute and the workload was progressively increased from 0 Watts (W), through 37.5 W, 75 W and 125 W in stages each lasting 5 minutes. At each workload three separate readings were made of heart rate, minute volume and expired air oxygen concentration. Mean resting energy expenditure was taken as the average of the lying and sitting values. Flex heart rate was calculated as the mean of the highest resting pulse rate and the lowest on exercise. Finally the slope and intercept of the least squares regression line of the exercise points were calculated. The VO2max was measured from the linear regression as predicted oxygen consumption at maximal heart rate (220-age) and is expressed in the results per unit body weight. The volunteers wore the heart rate monitor (Polar Electro, Finland) continuously during the waking hours over the 4 days following their visit to the clinic. No account was taken of whether these days included only weekdays, or a combination of weekdays and weekend days. In a previous study, we demonstrated that the mean paired difference between individual recordings of week day and weekend day energy expenditure was 0.0008 (P = 0.98).13 Thus, there was no evidence that the 4-day sampling strategy would introduce bias.
Heart rate readings were directly downloaded into a computer via a serial interface and the individual calibration data were used to predict minute energy expenditure for each person. Sleeping energy expenditure was calculated as 95% of basal metabolic rate (BMR) where this was derived from published prediction equations.19,20 A physical activity level (PAL), which is the ratio of total energy expenditure to BMR, was computed for each day and averaged over the 4-day period.
Repeated measures sub-study
A random group of 190 subjects in the cohort reattended for measurements on a further three occasions at 4-monthly intervals for one year, when height, weight and impedance were also remeasured. The calibration between heart rate and resting and exercise energy expenditure was repeated and the volunteers then underwent 4-day heart rate monitoring.
Statistical analysis
The within- and between-subject mean square and the reliability coefficients for PAL, height, weight, Body mass index (BMI), percentage body fat and VO2max per kg were estimated using the formulae described by Armstrong et al.21 By this method, each of n subjects is measured k times with Xij being the jth measure of subject i.i is the mean of k measurements in subject i, and
= overall mean. The reliability coefficient is given as
= (BMS WMS)/(BMS + (k 1)WMS) where BMS = between subject mean square = Between subject sum of squares/degrees of freedom= k i (
i
)2/(n 1) and WMS = within subject mean square = within subject sum of squares/degrees of freedom =
i
j (xij
i)2/n(k 1). Simple linear regression and multiple regression were undertaken using SAS and the regression coefficients are presented per standard deviation for each variable. The univariate correction was undertaken using the univariate reliability coefficients. The multivariate correction factors for PAL and VO2max per kg were estimated from the variance-covariance matrices using the method described by Wong et al.16 Both the univariate and bivariate correction factors were calculated under the assumption that the errors associated with repeated measures on the same individual were independent. In the absence of other methods for measuring PAL or VO2max per kg, the validity of this assumption cannot be assessed. We investigated the effect of departures from this independence assumption by assuming a range of values from 0.0 to 0.2 for the correlation between the errors in measuring PAL and VO2max per kg.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Even though there is a large literature of the relationship between physical activity and increased blood pressure, it is difficult to make this type of quantitative estimate from existing studies. The longitudinal studies from the University of Pennsylvania Alumni Study clearly demonstrate that self-reported participation in sports and exercise predict the incidence of self-reported hypertension over the subsequent 2231 years of follow-up.2,3 Similar results were obtained in Harvard male alumni in whom the risk of developing hypertension was 1.30 in those with a physical activity index <2000 kcal/week compared to those whose index was >2000.4 In that study, the physical activity index was assessed using a composite of the energy cost of walking, stair climbing and sporting activity.22 Although indicative of an underlying association, the subjective nature of the questionnaire, its limited focus and the lack of validation study against objective measures of energy expenditure, make it unlikely that quantitative estimates based on this data would be accurate.23 Similar criticisms could be made of other epidemiological studies using physical activity questionnaires focusing on leisure time activity.5,24,25
The fact that in some of these studies the effect of physical activity is restricted to those who participate in vigorous activity26 could be interpreted as evidence that exercise needs to be conditioning for it to have a benefit on blood pressure. Indeed in the follow-up study from the Cooper Clinic, those individuals who were in the lowest fitness group had a relative risk for future hypertension of 1.52 (95% CI : 1.082.15) compared to those in the highest fitness group.27 From these data one might conclude that a public health recommendation aimed at reducing blood pressure through physical activity should advocate moderately intense cardio-respiratory fitness-enhancing exercise.28 However, such a conclusion would ignore the fact that an alternative explanation to the failure to find an association between blood pressure and activity at lower levels is the difficulty of accurately assessing overall energy expenditure. These issues can only be resolved in epidemiological studies by employing measurement instruments which can objectively quantitate the different dimensions of activity. Although the current study is cross-sectional, further follow-up will confirm whether the observations seen in this study are reinforced over time.
An alternative to attempting to estimate the likely public health benefit through observational studies is to rely on estimates from intervention studies. However, recent reviews of experimental studies of habitual physical activity and hypertension2931 demonstrate that although physical activity appears to lower blood pressure, the findings need to be interpreted carefully as few studies have avoided all the major pitfalls that cause bias in experimental studies. Thus, although it is intuitively appealing to look to intervention studies to provide an answer to estimating the benefit of increasing physical activity at a population level, the methodological difficulties that these studies present make this more challenging than would at first seem apparent. Therefore, it is likely that observational studies will continue to provide much of the support ive information needed to inform the process of measuring the public health benefit of population changes. Improving the quality of the estimates that are derived from such epidemiological studies requires advances in the assessment and quantitation of exposures such as physical activity and diet, combined with the development of statistical modelling to separate their different aetiological effects.
![]() |
Acknowledgments |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Paffenbarger RS, Thorne MC, Wing AL. Chronic disease in former college students. VIII. Characteristics in youth predisposing to hypertension in later years. Am J Epidemiol 1968;88:2532.[ISI][Medline]
3 Paffenbarger RS, Lung DL, Leung RW, Hyde RT. Physical activity and hypertension: An epidemiological view. Ann Med 1991;23: 31927.[ISI][Medline]
4 Paffenbarger RS, Wing AL, Hyde RT, Jung DL. Physical activity and incidence of hypertension in college alumni. Am J Epidemiol 1983; 117:24557.[Abstract]
5 Folsom AR, Prineas RJ, Kaye SA, Munger RG. Incidence of hypertension and stroke in relation to body fat distribution and other risk factors in older women. Stroke 1990;21:70106.[Abstract]
6 Wareham NJ, Rennie K. The assessment of physical activity in individuals and populations: Why try to be more precise about how physical activity is assessed ? Int J Obesity 1998;22(S2):S3038.[ISI]
7 Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep 1985;100:12631.[ISI][Medline]
8 MacMahon S, Peto R, Cutler J et al. Blood pressure, stroke, and coronary heart disease. Lancet 1990;335:76574.[ISI][Medline]
9 Montoye HJ, Kemper HCG, Saris WHM, Washburn RA (eds). Measuring Physical Activity and Energy Expenditure. Champaign, IL: Human Kinetics, 1996.
10 Drury TF (ed.). Assessing Physical Fitness and Physical Activity in Population-based Surveys. Hyattsville, Maryland: National Center for Health Statistics, 1989.
11 Spurr GB, Prentice AM, Murgatroyd PR, Goldberg GR, Reina JC, Christman NT. Energy expenditure from min-by-minute heart-rate recording: comparison with indirect calorimetry. Am J Clin Nutr 1988;48:55259.[Abstract]
12 Ceesay SM, Prentice AM, Day KC, Murgatroyd PR, Goldberg GR, Scott W. The use of heart rate monitoring in the estimation of energy expenditure: a validation study using indirect whole-body calorimetry. Br J Nutr 1989;61:17586.[ISI][Medline]
13 Wareham NJ, Hennings SHJ, Prentice AM, Day NE. Feasibility of heart-rate monitoring to estimate total level and pattern of energy expenditure in a population-based epidemiological study: the Ely young cohort feasibility study 19945. Br J Nutr 1997;78:889900.[ISI][Medline]
14 Wareham NJ, Hennings SJ, Byrne CD, Hales CN, Prentice AM, Day NE. A quantitative analysis of the relationship between habitual energy expenditure and the Metabolic Cardiovascular Syndrome. Br J Nutr 1998;80:23541.[ISI][Medline]
15 Rosner B, Spiegelman D, Willett WC. Correction of logistic regression relative risk estimates and confidence intervals for measurement error: the case of multiple covariates measured with error. Am J Epidemiol 1990;132:73445.[Abstract]
16 Wong MY, Day NE, Wareham NJ. Measurement error in Epidemiology: the design of validation studies II: multivariate situation. Stat Med 1999;18:283145.[ISI][Medline]
17 Williams DRR, Wareham NJ, Brown DC et al. Glucose intolerance in the community; the Isle of Ely Diabetes Project. Diabetic Med 1995: 12:3035.[ISI][Medline]
18 Wareham NJ, Byrne CD, Williams DRR, Day NE, Hales CN. Fasting proinsulin concentrations predict the development of Type 2 diabetes. Diabetes Care 1999;22:26270.[Abstract]
19 Goldberg GR, Prentice AM, Davies HL, Murgatroyd PR. Overnight and basal metabolic rates in men and women. Eur J Clin Nutr 1988;42: 13744.[ISI][Medline]
20 James WPT, Schofield EC. Human Energy Requirements. Oxford: Oxford Medical Publications, 1990.
21 Armstrong BK, White E, Saracci R. Principles of Exposure Measurement in Epidemiology. Oxford: Oxford University Press, 1994.
22 Wolf AM, Hunter DJ, Colditz GA et al. Reproducibility and validity of a self-administered physical activity questionnaire. Int J Epidemiol 1994;23:99199.[Abstract]
23 Rennie KL, Wareham NJ. The validation of physical activity measurement instruments: problems and pitfalls. Public Health Nutr 1998; 1:26571.[Medline]
24 Folsom AR, Caspersen CJ, Taylor HL et al. Leisure time physical activity and its relationship to coronary risk factors in a population-based sample. Am J Epidemiol 1985;121:57079.[Abstract]
25 Reaven PD, Barrett-Connor E, Edelstein S. Relation between leisure-time physical activity and blood pressure in older women. Circulation 1991;83:55965.[Abstract]
26 Paffenbarger RS, Lung DL, Leung RW, Hyde RT. Physical activity and hypertension: An epidemiological view. Ann Med 1991;23:31927.[ISI][Medline]
27 Blair SN, Goodyear NN, Gibbons LW, Cooper KH. Physical fitness and incidence of hypertension in healthy normotensive men and women. JAMA 1984;252:48790.[Abstract]
28 Seals DR, Hagberg JM. The effect of exercise training on human hypertension: a review. Med Sci Sports Exerc 1984;16:20715.[ISI][Medline]
29 Arroll B, Beaglehole R. Does physical activity lower blood pressure: A critical review of the clinical trials. J Clin Epidemiol 1992;45:43947.[ISI][Medline]
30 Kelley G, McClellan P. Antihypertensive effects of aerobic exercise. A brief meta-analytical review of randomised controlled trials. Am J Hypertens 1994;7:11519.[ISI][Medline]
31 Ebrahim S, Smith GD. Lowering blood pressure: a systematic review of sustained effects of non-pharmacological interventions. J Public Health Med 1998;20:44148.[Abstract]