Antihypertensive and LDL cholesterol-lowering medications use were also inversely related with walking distance in both sexes, and in women even when adjusted for BMI. Compared with women who walked < 5 km·wk−1, those who walked > 5 km·wk−1 were significantly less likely to use antihypertensive or LDL cholesterol medications, with significant incremental reductions in odds for both medications with each 10 km/wk increment in distance up to 25 km·wk−1. Adjustment for BMI reduced the coefficient for the women's log odds for antihypertensive medications by nearly three fourths and for LDL cholesterol lowering by more than one half; however, both associations remained significant when adjusted for BMI (Table 3). Figure 1 shows that compared with women who walked < 5 km·wk−1, when adjusted for BMI those who walked > 15 km·wk−1 generally had significantly lower odds for antihypertensive and LDL cholesterol-lowering medications.
Table 3 shows that the length of the longest weekly walk was significantly associated with reduced log odds for the use of all three medications in both sexes, even when adjusted for total distance per week. Adjustment for BMI eliminated the lower use of antidiabetics and antihypertensives with longest distance for women, but not for men. Moreover, the inverse relationship between men's log-odds for antidiabetic, antihypertensive, and LDL-lowering medication use versus longest walking distance remained significant when adjusted simultaneously for both BMI and average weekly distance.
Figure 2 shows that including a walk of at least 4-6 km each week was associated with significantly lower odds for antidiabetic medication use in men and women and antihypertensive use in women, regardless of total weekly distance. The odds for antihypertensive and LDL cholesterol-lowering medications were 28% and 33% lower for women whose longest walk was 6-8 km than for those who never exceeded 4 km, respectively. Among women, longer walks were not associated with further reductions in these medications, whereas in men lower use was most pronounced when walks of eight or more kilometers were included. BMI accounted for the lower medication use with longer walks by women but not by men.
The lower odds for men's diabetic medication use in association with the length of the longest walk were particularly robust to adjustments for total weekly distances and BMI values. Men who included at least one walk > 10 km had 56% lower odds for antidiabetic medication use, 39% lower odds for antihypertensive medication use, and 26% lower odds LDL cholesterol-lowering medication use when adjusted for both BMI and total distance walked per week.
The analyses of Table 3 suggest that the strongest predictor of medication use was self-reported walking speed. Speed was significantly associated with reduced antidiabetic, antihypertensive, and LDL cholesterol-lowering medication use, independently of both BMI and total distance. Adjustment for weekly walking distance generally reduced the logistic regression coefficient for speed by about 15% in women and by 10% or less in men. In contrast, adjusting the coefficient for weekly walking distance for walking speed reduced the logistic regression coefficients for diabetic medication use by 33% in women and 39% in men, and it reduced those for antihypertensive medication use by 26% and 37% in women and men, respectively. Although the inverse association between medication use and walking speed remained statistically significant when adjusted for BMI, the effects were considerably weakened.
Figure 3 shows that the odds ratios for greater speed were only minimally affected by adjustments for walking distance. The odds for medication use showed similar trends in men and women for each 0.3-m·s−1 increment in walking speed from 1.2 through 1.8 m·s−1. The graphs illustrate the greater impact of BMI adjustment on the odds in women than in men. However, with few exceptions, the lower odds ratio for faster men and women remained significant.
The frequency of walks taken per week was largely unrelated to medication use (analyses not displayed).
These data demonstrate pronounced dose-response relationships of walking distance and intensity with the prevalence of antidiabetic, antihypertensive, and LDL cholesterol-lowering medications. High cholesterol, hypertension, and diabetes are all strong predictors of cardiovascular disease, and their inverse relationship with physical activity may partly mediate the reduction in cardiovascular disease risk associated with being more physically active.
Our analyses suggest that the odds for diabetes were inversely related to walking distance through at least 15 km·wk−1 in both men and women, and that the odds for hypertension and high cholesterol in women were inversely related to distance through at least 25 km·wk−1. The BMI-adjusted associations for diabetes shown in Figure 1 are consistent with reductions in adjusted incidence reported from cohort studies in women and men and controlled clinical trials (2,4,8,11,17,19). Our observations that the inverse relationships for women's diabetes are greater than in men, and that the proportions of these relationships attributable to BMI is also greater in women than in men, has been suggested by cohorts followed prospectively (8,17,19) and by comparisons between the Nurses' and Physicians' Health Studies (4). The lower prevalence of LDL cholesterol-lowering medication use with walking distance reported here agrees with meta-analyses of 948 subjects in 22 randomized controlled trials that showed that walking decreased non-HDL cholesterol by about 4% (−0.15 mM) (13). Meta-analyses by Leon and Sanchez (16) also concluded that exercise training, including both moderate and vigorous prescriptions, decreased LDL cholesterol by about 5%.
In his meta-analysis of 35 randomized controlled trials, Fagard (7) estimates that training decreases systolic and diastolic blood pressure by 3.4 and 2.4 mm Hg, respectively. Moreover, he notes that the reductions were greater in hypertensive (7.4 and 5.8 mm Hg, respectively) than normotensive subjects (2.6 and 1.8 mm Hg, respectively), suggesting that exercise might have a greater affect on reducing the prevalence of hypertension than on lowering mean blood pressure for the population (7). Whereas Fagard's meta-analyses does not suggest a clear dose-response relationship between reductions in blood pressure and the frequency, duration, or intensity of the intervention (7), Figures 1-3 show that the prevalence of antihypertensive medication use declined proportionately with volume, length of longest walk, and intensity. The trials examined by Fagard, which ranged from 4 to 52 wk, suggest that longer exercise trials were less effective in lowering systolic blood pressure than were shorter trials (7). It is noteworthy, therefore, that the lower prevalence in antihypertensive medication use shown here pertain to men and women who had walked for exercise an average of 8 and 12 yr, respectively.
Excess body weight, particularly intraabdominal adiposity, increases the risks for diabetes, hypertension, and elevated cholesterol. Elsewhere, we have shown that BMI and body circumferences decline in association with walking quantities and intensities (27,28). Declines in women's BMI and regional adiposity with walking distance were shown to be convex (27). Correspondingly, in this paper we show that incremental decreases in the prevalence of women's use of antidiabetic and LDL-lowering medications also diminished above 35 km·wk−1 (Fig. 1).
Our analyses were repeated with and without BMI as a covariate to assess quantitatively the extent to which BMI explained the association of medication use with walking (Table 3). Adjustment for BMI accounted for one half to two thirds of the inverse association with antidiabetic medication use, three fourths to four fifths of the inverse association with antihypertensive medication use, and about one half of the inverse association with LDL cholesterol-lowering medication use with walking distance (Table 3). Although substantial, these proportions are expected to underestimate the true percentages attributable to the inherent errors in assessing adiposity, including both the approximate nature of BMI for estimating the physiologically relevant fat and the imprecision in obtaining heights and weights by self-report.
The significant contributions of BMI to the dose-response relationships reported here do not necessarily negate the importance of exercise. The change in the coefficient for walking distance when BMI is added to the model may reflect in part the mediating effects of BMI; that is, exercise attenuates age-related weight gain, thereby reducing the risks for diabetes, hypertension, and high cholesterol. In Western societies, men and women usually gain weight with age (29). We have shown that long-term runners experience less weight gain over time in proportion to their weekly distance run (29), but they are subject to accelerated weight gain when they quit running (30). Although we have not shown that walking also attenuates age-related weight gain, walking has been shown to help maintain diet-induced weight loss (12), and weight maintenance is the basis for the Institute of Medicine's recommendation of walking briskly for 60 instead of 30 min·d−1 (12).
There are also the confounding effects of BMI attributable to self-selection. Specifically, the distance relationships shown in Figures 1 and Table 3 would become irrelevant if lean men and women (i.e., those at low risk for diabetes, hypertension, and high cholesterol) simply choose to walk farther. Walking distance and medication use would then be secondarily related through preexercise adiposity, but not through cause and effect. We have reported that 40% of the decline in BMI with walking distance in women and 17% of the decline in men are attributable to self-selection-that is, theses declines are attributable to initially leaner sedentary men and women choosing to walk farther each week when they begin exercising (28). Thus, particularly in women, some of the association between walking distance and medication use may be attributable to leaner women (i.e., those who are at lower risk for disease) choosing to walk farther.
We also have shown that the odds for antidiabetic, antihypertensive, and LDL cholesterol-lowering medications were inversely related to walking intensity, independently of weekly distance (Table 3, Fig. 3). Walking may be performed at light (2-3 METs) or moderate absolute intensities (3-4.5 METs) (1). Incidence rates for diabetes, hypertension, and metabolic syndrome have been demonstrated to relate prospectively to cardiorespiratory fitness (3,15,26), and high-intensity walking produces significantly greater increases in fitness than does low-intensity walking (14).
Adjustment for adiposity accounted for only one quarter of the men's decrease in the prevalence of antidiabetic and antihypertensive medication use, and one quarter of the decrease in the prevalence of LDL cholesterol-lowering medication use with walking intensity. The corresponding proportions in women were substantially greater than in men, between one half and two thirds. These proportions are less likely to be attributable to the mediating effect of BMI than to the effects of self-selection. Specifically, we have reported that self-selection accounts for more than 70% of the decline in BMI with walking intensity (28). Yet, in both women and men, Figure 3 shows that when adjusted for BMI, the odds ratios for medication use continued to be significantly inversely related to walking speed. Thus, there seems to be a residual association of walking intensity on diabetes, hypertension, and high cholesterol (or the converse), independent of adiposity.
The Nurses' Health Study showed that walking intensity predicted lower incident diabetes during follow-up (11). The inverse association in diabetes with walking intensity reported here (Table 3, Fig. 3) and by the Nurses' Health Study may be unexpected, given reports that equivalent energy expenditures by moderately intense (3-6 METs) and vigorously intense (>6 METs) physical activities produce comparable reductions in incident diabetes (11), and nonvigorous and vigorous activity correspond to equivalent improvements in insulin sensitivity (18). Although the Nurses' Health Study also reports that the risk reduction was significant independent of the time spent walking each week (11), this was not exactly a test of whether walking volume and intensity were independently predictive of diabetes risk. Walking volume in their study is reported as MET-hours, which is the product of time and pace. They show that diabetes risk declined with MET-hours of activity, and if this were true, irrespective of walking intensity, then both time and pace effects would be expected to be significant.
In our study, walking volume is reported as distance, and the analyses of Table 3 and Figure 3 show that walking intensity is associated with lower prevalence of diabetes, hypertension, and high cholesterol, independently of walking volume (i.e., distance that is comparable to MET-hours or kilocalories). Individual differences in walking speed and fitness that are not attributable to exercise dose might reflect inherent genetic dissimilarities. Walking speed in older female twins has been attributed to significant genetic (20%), common environment (26%), and individual environmental effects (54%) (21). Genetic factors are estimated to account for approximately 40% of the variance in V˙O2max in family sets (20). Fifty percent heritability was reported for V˙O2max in the sedentary state (5) and ΔV˙O2max in response to training (6) in the HERTIAGE family study.
The longest usual weekly walk was a better discriminator of medication status than the total cumulative distance per week, particularly in men. The usual length of the longest regular walk remained significant when adjusted for total weekly distance. This suggests a possible disadvantage of accumulating only short bouts of exercise compared with including some bouts of longer duration. Prior studies of fractionization of exercise have primarily used fitness and body weight as outcomes, and although most report no significant differences between fractionized versus continuous exercise (9), they may have limited statistical power to detect differences.
The recruitment of walkers through Walking Magazine subscription lists and walking events is likely to have generated a more health conscious and active sample then the population at large. This strategy was pursued in order to target higher doses of walking than represented in other population studies. This may limit the generalizability of these results. However, we believe that the biological processes relating walking to hypertension, hypercholesterolemia, and diabetes would not dramatically differ between this sample and a less selected population.
Our use of self-reported medications for diabetes, hypertension, and high cholesterol may provide greater specificity for these maladies than self-reported physician diagnoses; however, medication use may also provide less sensitivity than physician diagnosis, because not all patients will initiate and remain compliant with pharmacological treatment. We also recognize that the dose-response relationships described herein are cross-sectional rather than prospective. An inherent limitation of cross-sectional analyses is the uncertainty of whether walking preceded the disease, or whether the converse occurred. For example, diabetes may cause peripheral neuropathy or vascular disease, which could diminish walking duration or intensity attributable to muscular discomfort, weakness, or cramping. Walking intensity or duration would be the consequence rather than the antecedent of diabetes, and the decline in prevalence with intensity or duration would overestimate the effect of walking on diabetes risk. Alternatively, walking is prescribed for type 2 diabetes mellitus, peripheral neuropathy, peripheral vascular disease, hypertension, and high cholesterol, and this fact may cause the decline in diabetes prevalence to be underestimated. Despite these possible biases, the decline in the prevalence of diabetes in intensity and distance as shown in Figures 1-3 are entirely consistent with prospective data reported by others, leading us to surmise that at least some of the declines in antidiabetic, antihypertension, and LDL-lowering medications are attributable to walking distance and intensity.
Supported in part by grants HL-45652, HL-072110, and DK066738 from the National Heart Lung and Blood Institute, and conducted at the Ernest Orlando Lawrence Berkeley National Laboratory (Department of Energy DE-AC03-76SF00098 to the University of California). The author has no conflicts of interest to report, other than to use these data to secure additional grant support.
1. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc
. 2000;32(9 suppl):S498-504.
2. Asikainen TM, Kukkonen-Harjula K, Miilunpalo S. Exercise
for health for early postmenopausal women: a systematic review of randomised controlled trials. Sports Med
3. Barlow CE, LaMonte MJ, Fitzgerald SJ, Kampert JB, Perrin JL, Blair SN. Cardiorespiratory fitness is an independent predictor of hypertension incidence among initially normotensive healthy women. Am J Epidemiol
4. Bassuk SS, Manson JE. Epidemiological evidence for the role of physical activity in reducing risk of type 2 diabetes and cardiovascular disease. J Appl Physiol
5. Bouchard C, Daw EW, Rice T, et al. Familial resemblance for V˙O2max
in the sedentary state: the HERITAGE family study. Med Sci Sports Exerc
6. Bouchard C, An P, Rice T, et al. Familial aggregation of V˙O2max
response to exercise
training: results from the HERITAGE Family Study. J Appl Physiol
7. Fagard AH. Exercise
characteristics and the blood pressure response to dynamic physical training. Med Sci Sports Exerc
. 2001; 33(6 suppl):S484-92.
8. Haapanen N, Miilunpalo S, Vuori I, Oja P, Pasanen M. Association of leisure time physical activity with the risk of coronary heart disease, hypertension and diabetes in middle-aged men and women. Int J Epidemiol
9. Hardman AE. Issues of fractionization of exercise
(short vs long bouts). Med Sci Sports Exerc
. 2001;33(6 suppl):S421-7.
10. Hogan P, Dall T, Nikolov P. Economic costs of diabetes in the US in 2002. Diabetes Care
11. Hu FB, Sigal RJ, Rich-Edwards JW, et al. Walking compared with vigorous physical activity and risk of type 2 diabetes in women: a prospective study. JAMA
12. Institute of Medicine. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients)
. Washington (DC): The National Academies Press; 2002.
13. Kelley GA, Kelley KS, Tran ZV. Walking and non-HDL-C in adults: a meta-analysis of randomized controlled trials. Prev Cardiol
14. King AC, Haskell WL, Young DR, Oka RK, Stefanick ML. Long-term effects of varying intensities and formats of physical activity on participation rates, fitness and lipoproteins in men and women age 50 to 65 years. Circulation
15. LaMonte MJ, Barlow CE, Jurca R, Kampert JB, Church TS, Blair SN. Cardiorespiratory fitness is inversely associated with the incidence of metabolic syndrome: a prospective study of men and women. Circulation
16. Leon AS, Sanchez OA. Response of blood lipids to exercise
training alone or combined with dietary intervention. Med Sci Sports Exerc
. 2001;33(6 Suppl.):S502-15.
17. Lipton RB, Liao Y, Cao G, Cooper RS, McGee D. Determinants of incident non-insulin-dependent diabetes mellitus among blacks and whites in a national sample. The NHANES I Epidemiologic follow-up study. Am J Epidemiol
18. Mayer-Davis EJ, D'Agostino R Jr., Karter AJ, et al. Intensity and amount of physical activity in relation to insulin sensitivity: the Insulin Resistance Atherosclerosis Study. JAMA
19. Meisinger C, Thorand B, Schneider A, Stieber J, Doring A, Lowel H. Sex differences in risk factors for incident type 2 diabetes mellitus: the MONICA Augsburg cohort study. Arch Intern Med
20. Montoye HJ, Gayle R. Familial relationships in maximal oxygen uptake. Hum Biol
21. Pajala S, Era P, Koskenvuo M, et al. Contribution of genetic and environmental factors to individual differences in maximal walking speed with and without second task in older women. J Gerontol A Biol Sci Med Sci
22. Pate RR, Pratt M, Blair SN, et al. Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention
and the American College of Sports Medicine. JAMA
23. Schoenborn CA, Barnes PM. Leisure-Time Physical Activity Among Adults: United States, 1997-98. Advance Data From Vital and Health Statistics; no. 325
. Hyattsville (MD): National Center for Health Statistics; 2002.
24. Tully MA, Cupples ME, Hart ND, et al. Randomised controlled trial of home-based walking programmes at and below current recommended levels of exercise
in sedentary adults. J Epidemiol Community Health
25. U.S. Department of Health and Human Services. Physical Activity and Health: A Report of the Surgeon General.
Atlanta (GA): U.S. Department of Health and Human Services, Centers for Disease Control and Prevention
, National Center for Chronic Disease Prevention
and Health Promotion; 1996.
26. Wei M, Gibbons LW, Mitchell TL, Kampert JB, Lee CD, Blair SN. The association between cardiorespiratory fitness and impaired fasting glucose and type 2 diabetes mellitus in men. Ann Intern Med
27. Williams PT. Nonlinear relationships between weekly walking distance and adiposity in 27,596 women. Med Sci Sports Exerc
28. Williams PT. Self-selection contributes significantly to the lower adiposity of faster, longer-distanced, male and female walkers. Int J Obes
29. Williams PT, Thompson PD. Dose-dependent effects of training and detraining on weight in 6406 runners during 7.4 years. Obesity (Silver Spring)
30. Williams PT, Wood PD. The effects of changing exercise
levels on weight and age-related weight gain. Int J Obes
Keywords:©2008The American College of Sports Medicine
PREVENTION; EPIDEMIOLOGY; POPULATION; AGING; EXERCISE