Abstract

The benefits of physical activity in reducing cardiovascular disease (CVD) are thought to be mediated through changes in blood lipids, insulin sensitivity, and thrombogenic factors. Few studies have addressed the effects of both long-term physical activity and inactivity on these factors. The authors assessed associations between long-term leisure-time physical activity, television watching, and biomarkers of CVD risk among 468 healthy male health professionals. Prior to blood collection in 1993–1994, physical activity and television watching were assessed biennially from 1986 to 1994 by a questionnaire. Physical activity was expressed as metabolic equivalents-hours per week. Multivariate linear regression analyses showed that metabolic equivalents-hours in 1994 were significantly associated with high density lipoprotein cholesterol (HDL cholesterol) (positively) and with leptin and C-peptide (inversely). The average number of hours of television watching assessed in 1994 was significantly positively associated with low density lipoprotein cholesterol and significantly inversely associated with HDL cholesterol and apolipoprotein A1. Average hours of television watching per week assessed in 1988–1994 was positively associated with leptin levels (p < 0.01). The associations of television watching and vigorous activity with leptin and HDL cholesterol were independent of each other. In conclusion, physical activity and television watching were significantly associated with several biochemical markers of obesity and CVD risk.

Physical activity is known to lower the risk of cardiovascular disease (CVD) (13). The biological mechanisms for this effect are not completely understood, but probably include beneficial changes in blood pressure, obesity, blood lipids, insulin resistance, and thrombogenic factors (48). Many previous epidemiologic studies have used a single average measurement of physical activity (9, 10) rather than incorporating multiple measures over time. Repeated measurements can minimize random within-person variation in physical activity and provide a better representation of true overall level of activity. Sedentary behaviors, as represented by television watching, have been linked to obesity (1113), but few studies have examined the independent effects of physical activity and inactivity on CVD risk factors (14). In this study, we examine the relations of physical activity and sedentary behavior to biomarkers of obesity and CVD risk.

MATERIALS AND METHODS

Study sample

The Health Professionals Follow-up Study is a prospective cohort study of 51,529 US male health professionals (dentists, optometrists, pharmacists, podiatrists, osteopaths, and veterinarians), which was designed to investigate dietary etiologies of heart disease and cancer (15). The men were aged 40–75 years at baseline in 1986. Health information and disease status were assessed biennially by a self- administered questionnaire. The 468 men included in this analysis are a subset of the 18,225 male health professionals who provided blood samples between 1993 and 1994. This subset was selected randomly among men with various types of reported alcohol consumption patterns (e.g., light, binge, abstain, etc.) to examine the associations of these patterns with biomarkers of cardiovascular disease and diabetes in a separate study (16). Compared with men who did not provide blood samples, the men who provided samples were somewhat younger, but were similar with respect to smoking, physical activity, and diet. There is little difference in various characteristics between the 468 men included in this study and the remaining 17,757 men who gave blood. As described in detail elsewhere (16), participants had no history of myocardial infarction, angina pectoris, stroke, diabetes mellitus, intermittent claudication, gastric or duodenal ulcers, gallbladder disease, liver disease, or cancers (except nonmelanoma skin cancer) prior to 1994. The time line for data collection is given in figure 1. The average follow-up rate for the cohort was 92 percent of the total possible person-years through 1994.

FIGURE 1.

An outline of the time frame for the assessment of physical activity (PA), television watching (TV), diet, and blood draw, Health Professionals Follow-up Study, 1986–1994.

FIGURE 1.

An outline of the time frame for the assessment of physical activity (PA), television watching (TV), diet, and blood draw, Health Professionals Follow-up Study, 1986–1994.

Blood collection and assessment of biomarkers

Blood samples were returned to our laboratory on ice packs in insulated containers via overnight courier. More than 95 percent of all the blood samples received arrived within 24 hours of blood drawing. They were centrifuged, aliquoted, and stored at −150°C. Upon arrival, fewer than 15 percent were slightly hemolyzed, fewer than 3 percent were moderately hemolyzed, fewer than 1 percent were lipemic, and fewer than 0.5 percent were not cool.

Total plasma cholesterol was measured by the esterase-oxidase method, and triglycerides were measured by an enzymatic procedure. High density lipoprotein cholesterol (HDL cholesterol) and low density lipoprotein cholesterol were precipitated by the addition of phosphotungstic acid and magnesium ions. Apolipoprotein A1 (ApoA1) measurement was based on a nephelometric assay. Lipoprotein A was assayed with enzyme-linked immunosorbent assay. Insulin and C-peptide were measured by radioimmunoassay (Linco Research, St. Charles, Missouri). This assay allows accurate assessment with little or no proinsulin and C-peptide cross-reactivity; the coefficient of variation (CV) was less than 10 percent. Hemoglobin A1c (HbA1c) was measured with turbiometric immunoinhibition in red cells by using a Hitachi 911 analyzer (Boehringer Mannheim, Indianapolis, Indiana). For HbA1c levels of 5.6 and 9.6 percent, the coefficients of variation were less than 2.5 percent. Fibrinogen was measured by the Clauss method with a CV of 2.6 percent. Leptin was measured by radioimmunoassay method (Linco Research), intraassay CV was 4.3 percent, and interassay CV was 8.3 percent. Serum cholesterol and apoA1 measurements were not available in one individual each, and HbA1c level was unavailable in two individuals.

Assessment of leisure-time physical activity and other variables

We assessed leisure-time physical activity every 2 years between 1986 and 1994 (figure 1) through a series of questions on the specific type of activity and the average total time per week spent on the activity over the previous year. Questions about the number of hours per week spent walking or hiking outdoors; jogging; running; bicycling; swimming laps; playing tennis, squash or racquetball; and rowing or doing calisthenics were included in the 1986 questionnaire. We added heavy outdoor work in 1988 and weightlifting in 1992. Walking pace, categorized as casual (≤2 miles/hour), normal (2–2.9 miles/hour), brisk (3–3.9 miles/hour), or striding (≥4 miles/hour), was also assessed. Two individuals did not provide a physical activity measure in 1988, four did not provide it in 1992, and one did not provide it in 1994. No one had more than one missing activity measure. We calculated total weekly energy expenditure from leisure-time physical activities for each individual in metabolic equivalents-hours (MET-hr) (17). One MET-hr is equivalent to sitting quietly for 1 hour. Total MET-hr for vigorous activities, defined as requiring MET-hr values greater than or equal to six, were calculated by using MET-hr from jogging, running, bicycling, swimming laps, tennis, playing squash/racquetball, and doing calisthenics/rowing. Nonvigorous activities (MET-hr < 6) were walking, heavy outdoor work, and weightlifting. The Spearman correlation coefficients between the MET-hr values across time ranged from 0.57 to 0.70. Among a subsample (n = 238) of the Health Professionals Follow-up Study cohort, we assessed the validity and reproducibility of this method by comparing physical activity assessed by the questionnaire with the average of four 1-week activity diaries in 1991 that covered all seasons (18). The correlation between reported overall activity on the diaries and the 1992 questionnaire was 0.41. The correlation between the vigorous activities recorded in the diaries and 1992 questionnaire was 0.58.

The weekly number of hours watching television and video-cassette recorders was assessed biennially starting in 1988. The 1988 questionnaire included six response categories, ranging from 1–2 hours/week to 40 or more hours/week. Subsequent questionnaires contained 13 categories ranging from zero to 40 hours/week.

Cigarette smoking and body weight was also assessed biennially. We calculated body mass index (BMI), a measurement of obesity, for each 2-year cycle as weight (kg)/height2 (m2). Using a validated semiquantitative food frequency questionnaire (19, 20), we calculated average daily intakes of saturated and polyunsaturated fat and dietary fiber intake. Alcohol consumption was measured in 1986, 1990, and 1994.

Statistical analysis

We examined associations between physical activity, sedentary lifestyle, and biomarkers by using multivariate linear regression with a robust variance estimate (21). This variance estimator allows for valid inference without the normal distribution assumption in the dependent variable. Because tests for nonlinearity using spline regression were not statistically significant, we conducted the analysis with linear regression. We used two sets of independent variables and covariates in our study. In the cross-sectional regression analysis that used the 1994 weekly MET-hr values as the independent variable, we adjusted for covariates reported in 1994. In the long-term prospective analysis based on average weekly MET-hr values obtained between 1986 and 1994, we averaged the values of the covariates over this period of time (except for smoking, in which the status in 1994 was used). We calculated Pearson correlations by using log-transformed biomarker values to improve normality. We have also run Spearman correlations, and results were similar.

To control for potential confounding, we adjusted the models for age (continuous variable), alcohol intake (categorized as nondrinkers and 0.1–10, 10.1–20, and >20 g alcohol/day), total dietary fiber (quintiles), saturated fat (quintiles), polyunsaturated fat (quintiles), and smoking status (never smokers, past smokers, current smokers of ≤14 cigarettes/day, and current smokers of >14 cigarettes/day). BMI was included only in secondary analyses because it might be an intermediate variable in the biological pathway between physical activity and CVD biomarkers.

From the 468 men selected for this study, we excluded two who had missing values for at least one of the biomarkers. The 197 individuals who reported food consumption within 6 hours of blood draw were excluded from the analysis of insulin, C-peptide, and triglycerides. All p values are two-tailed.

RESULTS

Means and standard deviations of biomarkers and lifestyle characteristics by quintiles of average available weekly MET-hr are shown in table 1 The median of the first quintile represents exercise intensity equivalent to 2 hours of walking per week at a moderate pace, and the median of the fifth quintile represents the equivalent of 20 hours of walking per week. Men with a higher level of exercise had a lower BMI, ate less saturated fat, and were less likely to smoke. They also had lower plasma levels of leptin, triglycerides, fasting insulin, and C-peptide and a higher HDL cholesterol when compared with those with lower levels of activity.

TABLE 1.

Means (and standard deviations) of diet, lifestyle, biomarkers of obesity, and CVD* risk by quintiles of average available weekly MET-hr* among 466 male health professionals, Health Professionals Follow-up Study, 1986–1994

Characteristics Quintiles of average weekly MET-hr (1986–1994) (median MET-hr)
 
1 (7.5) 2 (16.9) 3 (27.3) 4 (40.3) 5 (60.6) 
Age (years) 61 (8) 60 (8) 61 (9) 61 (8) 59 (9) 
Average BMI* 25.7 (3.1) 25.3 (3.0) 25.0 (2.3) 24.5 (2.3) 24.7 (2.6) 
% current smokers 6.5 7.4 6.4 4.3 4.3 
Alcohol intake/week (g) (1994) 28.9 (41.1) 23.3 (28.2) 26.2 (22.1) 29.2 (38.1) 29.8 (32.5) 
Total fat (g) (kcal adjusted)      
 (1994) 66.9 (13.4) 66.2 (14.9) 64.9 (12.7) 64.2 (13.7) 62.7 (14.0) 
Saturated fat (g) (kcal      
 adjusted) (1994) 22.0 (5.6) 21.4 (6.0) 20.8 (5.4) 20.7 (5.7) 19.9 (6.0) 
Cholesterol (mg/dl) 252.9 (44.0) 244.7 (45.9) 250.4 (51.6) 240.9 (41.1) 254.4 (43.2) 
LDL cholesterol* (mg/dl) 154.2 (36.7) 145.2 (34.0) 151.6 (42.0) 146.2 (34.8) 151.5 (36.3) 
HDL cholesterol* (mg/dl) 60.0 (15.8) 56.7 (14.7) 61.5 (17.3) 64.5 (41.3) 64.3 (19.0) 
Triglycerides (mg/dl) 167.7 (124.1) 179.2 (133.8) 167.7 (98.1) 160.8 (105.5) 159.7 (152.2) 
ApoA1* (mg/dl) 153.2 (28.0) 151.9 (22.4) 152.8 (25.1) 149.8 (25.2) 157.7 (28.7) 
Lp(a)* (mg/dl) 32.7 (40.6) 27.9 (34.1) 34.0 (42.4) 39.2 (48.0) 36.2 (36.7) 
Leptin (ng/ml) 8.8 (6.9) 6.9 (4.3) 6.4 (3.1) 6.1 (3.7) 5.4 (3.8) 
Fibrinogen (mg/dl) 239.4 (34.5) 234.6 (33.1) 236.8 (30.6) 237.4 (34.5) 234.6 (28.7) 
Insulin (μU/ml) 13.9 (7.1) 15.1 (15.4) 12.2 (5.0) 10.0 (3.2) 11.8 (5.5) 
C-peptide (ng/dl) 2.7 (1.6) 2.8 (1.9) 3.0 (2.2) 2.3 (1.6) 2.4 (1.6) 
HbA1c* (%) 5.8 (0.8) 5.7 (0.7) 5.8 (0.7) 5.8 (0.8) 5.7 (0.6) 
Characteristics Quintiles of average weekly MET-hr (1986–1994) (median MET-hr)
 
1 (7.5) 2 (16.9) 3 (27.3) 4 (40.3) 5 (60.6) 
Age (years) 61 (8) 60 (8) 61 (9) 61 (8) 59 (9) 
Average BMI* 25.7 (3.1) 25.3 (3.0) 25.0 (2.3) 24.5 (2.3) 24.7 (2.6) 
% current smokers 6.5 7.4 6.4 4.3 4.3 
Alcohol intake/week (g) (1994) 28.9 (41.1) 23.3 (28.2) 26.2 (22.1) 29.2 (38.1) 29.8 (32.5) 
Total fat (g) (kcal adjusted)      
 (1994) 66.9 (13.4) 66.2 (14.9) 64.9 (12.7) 64.2 (13.7) 62.7 (14.0) 
Saturated fat (g) (kcal      
 adjusted) (1994) 22.0 (5.6) 21.4 (6.0) 20.8 (5.4) 20.7 (5.7) 19.9 (6.0) 
Cholesterol (mg/dl) 252.9 (44.0) 244.7 (45.9) 250.4 (51.6) 240.9 (41.1) 254.4 (43.2) 
LDL cholesterol* (mg/dl) 154.2 (36.7) 145.2 (34.0) 151.6 (42.0) 146.2 (34.8) 151.5 (36.3) 
HDL cholesterol* (mg/dl) 60.0 (15.8) 56.7 (14.7) 61.5 (17.3) 64.5 (41.3) 64.3 (19.0) 
Triglycerides (mg/dl) 167.7 (124.1) 179.2 (133.8) 167.7 (98.1) 160.8 (105.5) 159.7 (152.2) 
ApoA1* (mg/dl) 153.2 (28.0) 151.9 (22.4) 152.8 (25.1) 149.8 (25.2) 157.7 (28.7) 
Lp(a)* (mg/dl) 32.7 (40.6) 27.9 (34.1) 34.0 (42.4) 39.2 (48.0) 36.2 (36.7) 
Leptin (ng/ml) 8.8 (6.9) 6.9 (4.3) 6.4 (3.1) 6.1 (3.7) 5.4 (3.8) 
Fibrinogen (mg/dl) 239.4 (34.5) 234.6 (33.1) 236.8 (30.6) 237.4 (34.5) 234.6 (28.7) 
Insulin (μU/ml) 13.9 (7.1) 15.1 (15.4) 12.2 (5.0) 10.0 (3.2) 11.8 (5.5) 
C-peptide (ng/dl) 2.7 (1.6) 2.8 (1.9) 3.0 (2.2) 2.3 (1.6) 2.4 (1.6) 
HbA1c* (%) 5.8 (0.8) 5.7 (0.7) 5.8 (0.7) 5.8 (0.8) 5.7 (0.6) 
*

CVD, cardiovascular disease; MET-hr, metabolic equivalents-hours; BMI, body mass index; LDL cholesterol, low density lipoprotein cholesterol; HDL cholesterol, high density lipoprotein cholesterol; ApoA1, apolipoprotein A1; Lp(a), lipoprotein A; HbA1c, hemoglobin A1c.

Two outliers were excluded.

After adjustment for age, we found significant inverse correlations between average physical activity (from 1986 to 1994) and BMI, leptin, insulin, and C-peptide, and a positive correlation with HDL cholesterol (table 2). Conversely, hours of television watching was positively associated with leptin, C-peptide, and BMI.

TABLE 2.

Partial Pearson correlation coefficient matrix of log biomarkers, BMI, and log average weekly MET-hr from 1986, 1988, 1990, 1992, and 1994, Health Professionals Follow-up Study, 1986–1994

 Average television hours BMI Fibrinogen (mg/dl) TG (mg/dl) Cholesterol (mg/dl) HDL Cholesterol (mg/dl) LDL Cholesterol (mg/dl) C-peptide (ng/dl) HbA1c (%) Leptin (ng/ml) ApoA1 (mg/dl) Lp(a) (mg/dl) Insulin (μU/ml) 
MET-hr −0.052 −0.11* 0.0064 −0.046 −0.017 0.080 −0.046 −0.096* −0.071 −0.22*** 0.040 0.018 −0.15* 
Average television hours  0.13** 0.035 0.047 0.050 −0.056 0.047 0.12* −0.00013 0.15** −0.075 −0.031 −0.094 
BMI   0.14** 0.19*** −0.021 −0.27*** −0.0036 0.34*** 0.24*** 0.59*** −0.14** −0.015 0.45*** 
Fibrinogen    −0.033 0.0058 −0.11* 0.061 0.089 0.17** 0.13** −0.13** 0.039 0.051 
TG     0.35*** −0.39 0.11** 0.33*** 0.33 0.17** −0.15** −0.083 0.27*** 
Cholesterol      0.23* 0.82*** −0.0031 0.012 0.097* 0.22*** 0.16** 0.068 
HDL cholesterol       0.12* −0.23*** −0.096* −0.20*** 0.59*** 0.072 −0.22** 
LDL cholesterol        −0.034 −0.0051 0.12** −0.0084 0.21*** 0.047 
C-peptide         0.17** 0.48*** −0.1** −0.09* 0.77*** 
HbA1c          0.19*** 0.029 −0.050 0.14* 
Leptin           −1.10* 0.039 0.052 
ApoA1            0.055 −0.13* 
Lp(a)             −0.054 
 Average television hours BMI Fibrinogen (mg/dl) TG (mg/dl) Cholesterol (mg/dl) HDL Cholesterol (mg/dl) LDL Cholesterol (mg/dl) C-peptide (ng/dl) HbA1c (%) Leptin (ng/ml) ApoA1 (mg/dl) Lp(a) (mg/dl) Insulin (μU/ml) 
MET-hr −0.052 −0.11* 0.0064 −0.046 −0.017 0.080 −0.046 −0.096* −0.071 −0.22*** 0.040 0.018 −0.15* 
Average television hours  0.13** 0.035 0.047 0.050 −0.056 0.047 0.12* −0.00013 0.15** −0.075 −0.031 −0.094 
BMI   0.14** 0.19*** −0.021 −0.27*** −0.0036 0.34*** 0.24*** 0.59*** −0.14** −0.015 0.45*** 
Fibrinogen    −0.033 0.0058 −0.11* 0.061 0.089 0.17** 0.13** −0.13** 0.039 0.051 
TG     0.35*** −0.39 0.11** 0.33*** 0.33 0.17** −0.15** −0.083 0.27*** 
Cholesterol      0.23* 0.82*** −0.0031 0.012 0.097* 0.22*** 0.16** 0.068 
HDL cholesterol       0.12* −0.23*** −0.096* −0.20*** 0.59*** 0.072 −0.22** 
LDL cholesterol        −0.034 −0.0051 0.12** −0.0084 0.21*** 0.047 
C-peptide         0.17** 0.48*** −0.1** −0.09* 0.77*** 
HbA1c          0.19*** 0.029 −0.050 0.14* 
Leptin           −1.10* 0.039 0.052 
ApoA1            0.055 −0.13* 
Lp(a)             −0.054 
*

p < 0.05

**

p < 0.01

***

p < 0.001

Adjusted for age.

BMI, body mass index; MET-hr, metabolic equivalents-hours; TG, triglycerides; HDL cholesterol, high density lipoprotein cholesterol; LDL cholesterol, low density lipoprotein cholesterol; HbA1c, hemoglobin A1c; ApoA1, apolipoprotein A1; Lp(a), lipoprotein A.

Because many other lifestyle factors may confound the association between physical activity and these biomarkers, we conducted multivariate analyses, adjusting for age, alcohol intake, smoking, total fiber, saturated fat, and polyunsaturated fat intake (table 3). We observed a significant positive association between physical activity and HDL cholesterol and inverse associations with leptin and C-peptide. Each 20 MET-hr increment per week (equivalent to about an additional hour of walking per day) in 1994 was associated with a 1.4 mg/dl (95 percent CI: 0.2, 2.6) higher level of HDL cholesterol (p < 0.05), a 0.6 ng/ml (95 percent CI: −0.4, −0.8) lower level of leptin (p < 0.0001), and 0.1 ng/dl (95 percent CI: −0.2, −0.002) lower level of C-peptide (p < 0.05). These associations were strengthened when average physical activity from 1986 to 1994 was used. The association for HDL cholesterol was slightly weaker without adjustment for average hours of television watching (1.2 mg/dl higher for every 20 MET-hr/week higher level of activity). The associations with leptin and C-peptide were not affected by adjustment for average hours of TV watching (data not shown). While the inverse association between physical activity in 1994 and fasting insulin was not statistically significant, the longer-term measure of physical activity (average of 1986–1994) was significantly (p < 0.05) associated with fasting insulin. Each increment of 20 MET-hr per week was associated with 0.7 µg/ml (95 percent CI: 0.01, 1.4) lower fasting insulin levels.

TABLE 3.

Multivariate linear regression coefficients and standard errors for the association between physical activity (in 20 MET-hr increments per week) and CVD and obesity biomarkers among 466 male health professionals aged 48–83 years, Health Professionals Follow-up Study, 1986–1994

Biomarkers (dependent variables) Physical activity measures (independent variables)
 
1994 MET-hr 1994 MET-hr also adjusted for BMI Average MET-hr (1986–1994) Average MET-hr also adjusted for BMI Average vigorous MET-hr (1986–1994) Average vigorous MET-hr also adjusted for BMI 
Cholesterol (mg/dl) 5.1 (3.4) 5.9 (3.3) 0.1 (1.7) 0.1 (1.6) 1.8 (2.1) 1.8 (2.1) 
LDL cholesterol (mg/dl) −0.1 (1.1) −0.1 (1.1) −0.2 (1.4) −0.1 (1.4) 1.5 (1.8) 1.7 (1.8) 
HDL cholesterol (mg/dl) 1.4 (0.6)* 1.2 (0.5)* 2.4 (0.8)** 2.0 (0.7)** 3.2 (1.3)* 2.6 (1.2)* 
Triglycerides (mg/dl)§ −4.5 (4.3) −3.1 (4.5) −4.8 (7.5) −1.3 (7.6) −3.0 (15.4) −0.9 (9.9) 
ApoA1 (mg/dl) 0.3 (0.7) 0.1 (0.9) 1.2 (1.1) 0.9 (1.0) 2.1 (1.5) 1.6 (1.5) 
Lp(a) (mg/dl) 0.5 (1.0) 0.5 (1.0) −0.03 (1.5) 0.1 (1.5) 2.2 (2.2) 2.5 (2.2) 
Leptin (ng/ml) −0.6 (0.1)** −0.5 (0.08)*** −0.9 (0.2)*** −0.6 (0.1)*** −0.7 (0.2)* −0.4 (0.2)** 
Fibrinogen (mg/dl) 0.3 (0.9) 0.5 (0.6) 0.9 (1.3) 1.3 (1.0) −1.9 (1.4) −1.5 (1.4) 
Insulin (μU/ml)§ −0.5 (0.3) −0.3 (0.3) −0.7 (0.4)* −0.3 (0.3) −0.4 (0.4) 0.03 (0.4) 
C-peptide (ng/dl)§ −0.1 (0.05)* −0.08 (0.04)* −0.12 (0.05)** −0.09 (0.04)* −0.1 (0.06)* −0.06 (0.05) 
HbA1c (%) 0.02 (0.02) 0.02 (0.02) −0.02 (0.03) −0.0005 (0.03) −0.07 (0.04) −0.05 (0.03) 
Biomarkers (dependent variables) Physical activity measures (independent variables)
 
1994 MET-hr 1994 MET-hr also adjusted for BMI Average MET-hr (1986–1994) Average MET-hr also adjusted for BMI Average vigorous MET-hr (1986–1994) Average vigorous MET-hr also adjusted for BMI 
Cholesterol (mg/dl) 5.1 (3.4) 5.9 (3.3) 0.1 (1.7) 0.1 (1.6) 1.8 (2.1) 1.8 (2.1) 
LDL cholesterol (mg/dl) −0.1 (1.1) −0.1 (1.1) −0.2 (1.4) −0.1 (1.4) 1.5 (1.8) 1.7 (1.8) 
HDL cholesterol (mg/dl) 1.4 (0.6)* 1.2 (0.5)* 2.4 (0.8)** 2.0 (0.7)** 3.2 (1.3)* 2.6 (1.2)* 
Triglycerides (mg/dl)§ −4.5 (4.3) −3.1 (4.5) −4.8 (7.5) −1.3 (7.6) −3.0 (15.4) −0.9 (9.9) 
ApoA1 (mg/dl) 0.3 (0.7) 0.1 (0.9) 1.2 (1.1) 0.9 (1.0) 2.1 (1.5) 1.6 (1.5) 
Lp(a) (mg/dl) 0.5 (1.0) 0.5 (1.0) −0.03 (1.5) 0.1 (1.5) 2.2 (2.2) 2.5 (2.2) 
Leptin (ng/ml) −0.6 (0.1)** −0.5 (0.08)*** −0.9 (0.2)*** −0.6 (0.1)*** −0.7 (0.2)* −0.4 (0.2)** 
Fibrinogen (mg/dl) 0.3 (0.9) 0.5 (0.6) 0.9 (1.3) 1.3 (1.0) −1.9 (1.4) −1.5 (1.4) 
Insulin (μU/ml)§ −0.5 (0.3) −0.3 (0.3) −0.7 (0.4)* −0.3 (0.3) −0.4 (0.4) 0.03 (0.4) 
C-peptide (ng/dl)§ −0.1 (0.05)* −0.08 (0.04)* −0.12 (0.05)** −0.09 (0.04)* −0.1 (0.06)* −0.06 (0.05) 
HbA1c (%) 0.02 (0.02) 0.02 (0.02) −0.02 (0.03) −0.0005 (0.03) −0.07 (0.04) −0.05 (0.03) 
*

p < 0.05

**

p < 0.01

***

p < 0.001

MET-hr, metabolic equivalents-hours; CVD, cardiovascular disease; BMI, body mass index; LDL cholesterol, low density lipoprotein cholesterol; HDL cholesterol, high density lipoprotein cholesterol; ApoA1, apolipoprotein A1; Lp(a), lipoprotein A; HbA1c, hemoglobin A1c.

Multivariate regression adjusted for age (continuous variable); alcohol (nondrinkers and 0.1–10, 10.1–20, and ≥20 g/day); smoking (never smokers, past smokers, current smokers of ≤14 cigarettes/day, current smokers of >14 cigarettes/day); fiber; saturated and polyunsaturated fats (quintiles); television watching (continuous variable).

§

Includes only 269 men who provided fasting samples.

For several biomarkers, vigorous activity was a better predictor than total activity (table 3). For example, HDL cholesterol levels were more strongly associated with vigorous activity than with total activity. However, the associations between physical activity and total cholesterol, insulin, and C-peptide were not substantially different when we used vigorous activity rather than total activity. Additional adjustment for BMI only somewhat attenuated the associations between the biomarkers and physical activities (table 3). The results for analyses with lipids as outcome variables did not change appreciably after exclusion of those who took cholesterol-lowering medication or after further adjustment for baseline hypertension.

To determine whether measures of a sedentary lifestyle also were important predictors, we examined the associations between hours of television watching and these biomarkers independent of physical activity and other known confounders (table 4). The number of hours of television watching in 1994 was inversely associated with HDL cholesterol (p < 0.01) and ApoA1 (p < 0.05) and positively associated with low density lipoprotein cholesterol (p < 0.05). Average leptin level increased by 1.29 ng/ml (95 percent CI: 0.3, 2.3) for each 14-hour increment per week of television watching between 1988 and 1994.

TABLE 4.

Multivariate linear regression coefficients and standard errors for the association between television watching (14 hours/week increments) and CVD biomarkers among 466 male health professionals aged 48–83 years, Health Professionals Follow-up Study, 1986–1994

Biomarkers 1994 television hours 1994 television hours also adjusted for BMI Average television hours in 1988–1994 Average television hours in 1988–1994 also adjusted for BMI 
Cholesterol (mg/dl) −6.4 (8.8) −8.5 (8.7) 4.0 (4.6) 4.0 (4.6) 
LDL cholesterol (mg/dl) 6.1 (2.9)* 6.1 (2.9)* 5.2 (3.6) 5.0 (3.6) 
HDL cholesterol (mg/dl) −3.9 (1.2)** −3.4 (1.2)** −4.2 (2.3) −3.2 (2.2) 
Triglycerides (mg/dl)§ 23.8 (13.2) 20.9 (13.3) 20.5 (15.2) 17.0 (15.2) 
ApoA1 (mg/dl) −5.3 (2.0)* −4.9 (2.0)* −6.2 (2.6)* −5.5 (2.5) 
Lp(a) (mg/dl) −2.1 (3.0) −2.3 (3.0) 2.4 (4.4) 2.0 (4.4) 
Leptin (ng/ml) 0.6 (0.,4) 0.3 (0.3) 1.3 (0.5)** 0.8 (0.4)* 
Fibrinogen (mg/dl) 2.9 (2.3) 2.5 (2.3) 3.5 (3.0) 2.8 (3.0) 
Insulin (μU/ml)§ 0.7 (0.6) 0.3 (0.6) 0.3 (0.7) −0.3 (0.3) 
C-peptide (ng/dl)§ 0.1 (0.1) 0.06 (0.09) 0.02 (0.1) −0.004 (0.1) 
HbA1c (%) −0.07 (0.05) −0.09 (0.05) −0.07 (0.06) −0.01 (0.06) 
Biomarkers 1994 television hours 1994 television hours also adjusted for BMI Average television hours in 1988–1994 Average television hours in 1988–1994 also adjusted for BMI 
Cholesterol (mg/dl) −6.4 (8.8) −8.5 (8.7) 4.0 (4.6) 4.0 (4.6) 
LDL cholesterol (mg/dl) 6.1 (2.9)* 6.1 (2.9)* 5.2 (3.6) 5.0 (3.6) 
HDL cholesterol (mg/dl) −3.9 (1.2)** −3.4 (1.2)** −4.2 (2.3) −3.2 (2.2) 
Triglycerides (mg/dl)§ 23.8 (13.2) 20.9 (13.3) 20.5 (15.2) 17.0 (15.2) 
ApoA1 (mg/dl) −5.3 (2.0)* −4.9 (2.0)* −6.2 (2.6)* −5.5 (2.5) 
Lp(a) (mg/dl) −2.1 (3.0) −2.3 (3.0) 2.4 (4.4) 2.0 (4.4) 
Leptin (ng/ml) 0.6 (0.,4) 0.3 (0.3) 1.3 (0.5)** 0.8 (0.4)* 
Fibrinogen (mg/dl) 2.9 (2.3) 2.5 (2.3) 3.5 (3.0) 2.8 (3.0) 
Insulin (μU/ml)§ 0.7 (0.6) 0.3 (0.6) 0.3 (0.7) −0.3 (0.3) 
C-peptide (ng/dl)§ 0.1 (0.1) 0.06 (0.09) 0.02 (0.1) −0.004 (0.1) 
HbA1c (%) −0.07 (0.05) −0.09 (0.05) −0.07 (0.06) −0.01 (0.06) 
*

p < 0.05

**

p < 0.01

CVD, cardiovascular disease; BMI, body mass index; LDL cholesterol, low density lipoprotein cholesterol; HDL cholesterol, high density lipoprotein cholesterol; ApoA1, apolipoprotein A1; Lp(a), lipoprotein A; HbA1c, hemoglobin A1c.

Multivariate regression adjusted for age (continuous variable); alcohol (nondrinkers and 0.1–10, 10.1–20, and ≥20 g/day); smoking (never smokers, past smokers, current smokers of ≤14 cigarettes/day, current smokers of >14 cigarettes/day); fiber; saturated and polyunsaturated fats (quintiles); and television watching (continuous variable).

§

Includes only 269 men who provided fasting samples.

We examined the joint associations of duration of television watching and vigorous physical activity with leptin and HDL cholesterol (table 5). As expected, the most favorable leptin or HDL cholesterol profile was among men with the highest level of vigorous activity (tertile 3) and the lowest level of average hours of television watching (tertile 1). Compared with men with an average of 1.43 MET-hr/week (tertile 1) between 1986 and 1994 and 18.1 hours of television per week (tertile 3), men who had an average of 37.1 MET-hr per week (tertile 3) and 3.4 hours of television watching (tertile 1) had 5.1 ng/ml (95 percent CI: −7.2, −3.0) lower leptin and 12.8 mg/dl (95 percent CI: −2.2, 27.8) higher HDL cholesterol levels. The associations with vigorous activity and television watching were independent of each other. Hence, across the same level of television watching, HDL cholesterol generally increased and leptin tended to decreased with increasing level of vigorous activity. Across similar levels of vigorous activity, more television watching was associated with lower HDL cholesterol levels and higher leptin levels. Additional adjustment for BMI only slightly attenuated the results. Men at the highest level of physical activity and the lowest level of television watching had a 9.7 mg/dl (95 percent CI: −4.4, 23.9) higher HDL cholesterol level and –3.6 ng/ml (95 percent CI: −5.2, −2.1) lower leptin level compared with those at the lowest level of physical activity and the highest level of television watching (data not shown).

TABLE 5.

Multivariate linear regression coefficients for average HDL cholesterol and leptin by tertiles (range) of average weekly hours of television watching and average vigorous activities, Health Professionals Follow-up Study, 1986–1994

Vigorous activities tertile (MET-hr/week) Television-watching tertiles (hours/week)
 
Leptin
 
HDL cholesterol
 
3 (12.9–32.9) 2 (6–12.8) 1 (0.1–5.8) 3 (12.9–32.9) 2 (6–12.8) 1 (0.1–5.8) 
1 (0.0, 4.0) 0 (reference) −2.7* −2.5** 0 (reference) −0.1 6.5 
2 (4.1, 17.9) −3.2** −3.6** −3.8** −5.8* −2.7 3.1 
3 (18.0, 109) −2.3* −4.2** −5.1*** 7.6* 0.7 12.8 
Vigorous activities tertile (MET-hr/week) Television-watching tertiles (hours/week)
 
Leptin
 
HDL cholesterol
 
3 (12.9–32.9) 2 (6–12.8) 1 (0.1–5.8) 3 (12.9–32.9) 2 (6–12.8) 1 (0.1–5.8) 
1 (0.0, 4.0) 0 (reference) −2.7* −2.5** 0 (reference) −0.1 6.5 
2 (4.1, 17.9) −3.2** −3.6** −3.8** −5.8* −2.7 3.1 
3 (18.0, 109) −2.3* −4.2** −5.1*** 7.6* 0.7 12.8 
*

p < 0.05

**

p < 0.01

***

p < 0.001.

HDL cholesterol, high density lipoprotein cholesterol; MET-hr, metabolic equivalents-hours.

Multivariate regression adjusted for age (continuous variable); alcohol (nondrinkers and 0.1–10, 10.1–20, and ≥20 g/day); smoking (never smokers, past smokers, current smokers of ≤14 cigarettes/day, current smokers of >14 cigarettes/day); fiber; saturated and polyunsaturated fats (quintiles). Number of subjects ranged from 43 to 60 in each strata.

DISCUSSION

In this study, we observed independent associations between physical activity and several biomarkers of obesity and CVD risk. Measures of activity that took into account the average activity level over 8 years were somewhat more strongly associated with HDL cholesterol than was a single, cross-sectional measure. However, additional adjustment for BMI reduced the difference. Inactivity, expressed as hours of television watching, was also independently associated with an adverse CVD risk profile.

In our analysis, MET-hr from vigorous activities were generally a stronger predictor of HDL cholesterol and ApoA1 than were MET-hr from all sources. This suggests that some forms of leisure-time physical activity may be more effective than others, or alternatively, that our assessment tool better quantifies vigorous activity (22). In a validation study comparing these physical activity questions administered before and after a 4-week activity diary, correlations between vigorous activity and the diary were higher than for overall activity. Even though physical activity in general is associated with a favorable lipid profile (2326), several studies have shown that a certain level of intensity and duration may be necessary to affect HDL cholesterol level. Marrugat et al. (24) found that participation in physical activities that consume 9.5 kcal/minute and above (e.g., squash, running) was positively associated with HDL cholesterol, while those that consume less than 9.5 kcal/minute were not. However, other studies did not find a difference in exercise intensity and changes in HDL cholesterol (27, 28).

Consistent with previous studies, overall physical activity was inversely associated with C-peptide and insulin levels (2931). In a 7-year follow-up study, Folsom et al. (32) observed an inverse association between average physical activity and average fasting insulin level across the follow-up period. The Zutphen Elderly Study showed an inverse relation between current physical activity and insulin response to a glucose tolerance test (33), but not with fasting C-peptide among 389 men aged 65–84 years of age. Similarly, Siscovick et al. (25) found that elderly subjects who had the highest intensity of exercise exhibited the lowest level of fasting insulin. Physical activity is known to reduce insulin levels through improvements in insulin sensitivity and glucose disposal in muscle tissue (3436).

Previous studies have investigated the role of exercise on leptin (3740). Perusse et al. (40) observed a significant decrease in leptin among men after a 20-week endurance training program; however, the result was only marginally significant after adjustment for the reduction in fat mass. A 12-week aerobic training program without changes in fat mass did not result in a change in leptin level in men (38), but the number of subjects was small. Conversely, others have found lower leptin levels associated with activity, independent of changes in fat mass. Haluzik et al. (37) observed a lower level of leptin among rugby players when compared with sedentary individuals with similar body fat content. Pasman et al. (39) observed a decrease in leptin among obese subjects with increasing hours of exercise independent of percent body fat. In our analysis, the association of physical activity with leptin was only somewhat attenuated after adjustment for BMI. Our previous analyses suggest that leptin levels predicted weight gain in the subsequent 4 years (16).

High plasma fibrinogen is associated with CVD risk, but the acute or long-term effects of physical activity on fibrinogen are not firmly established (41). We found a nonsignificant inverse relation between vigorous activity and fibrinogen. Previous cross-sectional studies have found stronger inverse associations with fibrinogen (25, 42). A significant association of exercise with fibrinogen was found in some exercise training studies but not in others (4346).

The association between television watching and many of the biomarkers may be mediated in part through obesity. Television watching has been linked to obesity and weight gain (1113), and associations of obesity with dyslipidemia, hyperinsulimia, and elevated fibrinogen have been documented (4751). We found that the association between average hours watching television MET-hr, and CVD biomarkers was only somewhat attenuated after adjustment for BMI, suggesting that biologic processes other than weight gain may contribute to the greater CVD risk profile among men with a low level of physical activity and a high level of television viewing.

We generally found a stronger association of physical activity with HDL cholesterol for a composite score based on activities from the previous 8 years than on the physical activity levels measured in 1994. The composite score based on repeated measurements may be a more reliable and stable measurement of overall physical activity because measurement error due to within-person variations over time would be reduced. In previous studies, we have demonstrated that repeated dietary measurements yield stronger associations with CHD risk than use of a single baseline measure (52).

Accurate assessment of physical activity has always been a challenge in epidemiologic studies. The 10 leisure-time physical activity questions on our questionnaire may not be adequate to capture all types of exercise, especially seasonal activities such as skiing and hockey. Even if we did not capture all activities, our instrument is effective in ranking individuals by their activity level, as we included the most common leisure-time activities (53, 54). In a validation study of the questionnaire, reasonably good correlations were obtained between the activities assessed by the questionnaire and those assessed by 4 weeks of activity diary (18). Our cohort of health professionals may be particularly aware of their health behaviors, since they have been able to accurately self-report many diet and lifestyle characteristics (1820). Because we only measured biomarker levels in one blood sample drawn, short-term random individual variation may have limited our ability to detect a significant association with some biomarkers.

In conclusion, physical activity is significantly associated with several biochemical markers of obesity and CVD. Long-term leisure-time physical activity tended to be a stronger predictor of HDL cholesterol levels than a single cross-sectional measure. Additionally, hours of television watching was significantly associated with an adverse lipid profile independent of overall physical activity. This study suggests the importance of both leisure-time physical activity and sedentary behaviors in the prevention of CVD. We recommend that further studies of physical activity and chronic disease should include measurements of inactivity to obtain a more complete view of the overall activity profile.

Correspondence to Dr. Teresa Fung, Department of Nutrition, Simmons College, 300 The Fenway, Boston, MA 02115.

Supported by research grants HL35464, CA55075, and AA11181. The work of Dr. Hu is partially supported by a Research Award from the American Diabetes Association.

The authors thank Dr. Meir Stampfer for his invaluable critique of the manuscript.

REFERENCES

1.
Miller TD, Balady GJ, Fletcher GF. Exercise and its role in the prevention and rehabilitation of cardiovascular disease.
Ann Behav Med
 
1997
;
19
:
220
–9.
2.
Havranek EP. Primary prevention of CHD: nine ways to reduce risk.
Am Fam Physician
 
1999
;
59
:
1455
–63.
3.
Wannamethee SG, Shaper AG, Walker M, et al. Lifestyle and 15-year survival free of heart attack, stroke, and diabetes in middle-aged British men.
Arch Intern Med
 
1998
;
158
:
2433
–40.
4.
Buemann B, Tremblay A. Effects of exercise training on abdominal obesity and related metabolic complications.
Sports Med
 
1996
;
21
:
191
–212.
5.
Berg A, Halle M, Franz I, et al. Physical activity and lipoprotein metabolism: epidemiological evidence and clinical trials.
Eur J Med Res
 
1997
;
2
:
259
–64.
6.
Jessup JV, Lowenthal DT, Pollock ML, et al. The effects of endurance exercise training on ambulatory blood pressure in normotensive older adults.
Geriatr Nephrol Urol
 
1998
;
8
:
103
–9.
7.
Young DR, Appel LJ, Jee S, et al. The effects of aerobic exercise and T'ai Chi on blood pressure in older people: results of a randomized trial.
J Am Geriatr Soc
 
1999
;
47
:
277
–84.
8.
Scarabin PY, Aillaud MF, Amouyel P, et al. Associations of fibrinogen, factor VII and PAI-1 with baseline findings among 10,500 male participants in a prospective study of myocardial infarction–the PRIME Stud. Prospective Epidemiological Study of Myocardial Infarction.
Thromb Haemost
 
1998
;
80
:
749
–56.
9.
Adamopoulos PN, Macrilakis K, Papamichael C, et al. Physical activity and relationship with coronary heart disease risk factors.
Acta Cardiol
 
1993
;
48
:
523
–34.
10.
Eaton CB, Lapane KL, Assaf AR, et al. Physical activity, physical fitness, and coronary heart disease risk factors.
Med Sci Sports Exerc
 
1995
;
27
:
340
–6.
11.
Ching PLYH, Willett WC, Rimm EB, et al. Activity level and risk of overweight in Male Health Professionals.
Am J Public Health
 
1996
;
86
:
25
–30.
12.
Coakley EH, Rimm EB, Colditz G, et al. Predictors of weight change in men: results from the Health Professionals Follow-up Study.
Int J Obes
 
1998
;
22
:
89
–96.
13.
Tucker LA, Friedman GM. Television viewing and obesity in adult males.
Am J Public Health
 
1989
;
79
:
516
–18.
14.
Fitzgerald SJ, Kriska AM, Pereira MA, et al. Associations among physical activity, television watching, and obesity in adult Pima Indians.
Med Sci Sports Exerc
 
1997
;
27
:
910
–15.
15.
Rimm EB, Giovannucci EL, Willett WC, et al. A prospective study of alcohol consumption and the risk of coronary disease in men.
Lancet
 
1991
;
338
:
464
–8.
16.
Chu N-F, Spiegelman D, Yu J, et al. Plasma leptin concentrations and 4-year weight gain among US men–the Health Professionals Follow-up Study.
Int J Obes
  (in press).
17.
Ainswort B, Haskell W, Leon A., et al. Compendium of physical activities: classification of energy costs of human physical activities.
Med Sci Sports Exerc
 
1993
;
25
:
71
–80.
18.
Chasan-Taber S, Rimm EB, Stampfer MJ, et al. Reproducibility and validity of a self-administered physical activity questionnaire for male health professionals.
Epidemiology
 
1996
;
7
:
81
–6.
19.
Rimm EB, Giovannucci EL, Stampfer MJ, et al. Reproduci-bility and validity of a expanded self-administered semiquantitative food frequency questionnaire among male health professionals.
Am J Epidemiol
 
1992
;
135
:
1114
–26.
20.
Giovannucci E, Colditz G, Stampfer MJ, et al. The assessment of alcohol consumption by a simple self-administered questionnaire.
Am J Epidemiol
 
1991
;
133
:
810
–17.
21.
White H. A herteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity.
Econometrica
 
1980
;
48
:
817
–38.
22.
Jacobs DR Jr, Ainsworh BE, Hartman TJ, et al. A simultaneous evaluation of 10 commonly used physical activity questionnaires.
Med Sci Sports Exerc
 
1993
;
25
:
81
–91.
23.
Mensink GB, Heerstrass DW, Neppelenbroek SE, et al. Intensity, duration, and frequency of physical activity and coronary risk factors.
Television
 
1997
;
29
:
1192
–8.
24.
Marrugat J, Elosua R, Covas M-I, et al. Amount and intensity of physical activity, physical fitness, and serum lipids in men. The MARATHON Investigators.
Am J Epidemiol
 
1996
;
143
:
562
–9.
25.
Siscovick DS, Fried L, Mittelmark M, et al. Exercise intensity and subclinical cardiovascular disease in the elderly: the Cardiovascular Health Study.
Am J Epidemiol
 
1997
;
145
:
977
–86.
26.
Thune I, Njolstad I, Lochen ML, et al. Physical activity improves the metabolic risk profiles in men and women: the Tromso Study.
Arch Intern Med
 
1998
;
158
:
1633
–40.
27.
Fonong T, Toth MJ, Ades PA, et al. Relationship between physical activity and HDL-cholesterol in healthy older men and women: a cross-sectional and exercise intervention study.
Atherosclerosis
 
1996
;
127
:
177
–83.
28.
Visich PS, Goss FL, Gordon PM, et al. Effects of exercise with varying energy expenditure on high-density lipoprotein-cholesterol.
Eur J Appl Physiol
 
1996
;
72
:
242
–8.
29.
Nilsson PM, Lindholm LH, Schersten BF. Life style changes improve insulin resistance in hyperinsulinaemic subjects: a one-year intervention study of hypertensives and normotensives in Dalby.
J Hypertens
 
1992
;
10
:
1071
–8.
30.
Regensteiner JG, Mayer EJ, Shetterly SM, et al. Relationship between habitual physical activity and insulin levels among nondiabetic men and women: San Luis Valley Diabetes Study.
Diabetes Care
 
1991
;
14
:
1066
–74.
31.
Schneider SH, Khachadurian AK, Amorosa LF, et al. Ten-year experience with an exercise-based outpatient life-style modification program in the treatment of diabetes mellitus.
Diabetes Care
 
1992
;
15
:
1800
–10.
32.
Folsom AR, Jacobs DR Jr, Wagenknecht LE, et al. Increase in fasting insulin and glucose over seven years with increasing weight and inactivity of young adults.
Am J Epidemiol
 
1996
;
144
:
235
–46.
33.
Feskens EJM, Loeber JG, Kromhout D. Diet and physical activity as determinants of hyperinsulinemia: the Zutphen Elderly Study.
Am J Epidemiol
 
1994
;
140
:
350
–60.
34.
Tuominen JA, Ebeling P, Koivisto VA. Exercise increases insulin clearance in healthy man and insulin-dependent diabetes mellitus patients.
Clin Physiol
 
1997
;
17
:
19
–30.
35.
Devlin JT. Effects of exercise on insulin sensitivity in humans.
Diabetes Care
 
1992
;
15
:
1690
–3.
36.
Eriksson J, Taimela S, Koivisto VA. Exercise and the metabolic syndrome.
Diabetologia
 
1997
;
40
:
125
–35.
37.
Haluzik M, Boudova L, Nedvidkova J, et al. Lower serum leptin concentrations in rugby players in comparison with healthy non-sporting subjects–relationships to anthropometric and biochemical parameters.
Eur J Appl Physiol Occup Physiol
 
1998
;
79
:
58
–61.
38.
Hickey MS, Houmard JA, Considine RV, et al. Gender-dependent effects of exercise training on serum leptin levels in humans.
Am J Physiol
 
1997
;
272
:
E562
–6.
39.
Pasman WJ, Westerterp-Plantenga MS, Saris WH. The effect of exercise training on leptin levels in obese males.
Am J Physiol
 
1998
;
274
:
E280
–6.
40.
Perusse L, Collier G, Gagnon J, et al. Acute and chronic effects of exercise on leptin levels in humans.
J Appl Physiol
 
1997
;
83
:
5
–10.
41.
Meade TW. Exercise and haemostatic function.
J Cardiovasc Risk
 
1995
;
2
:
323
–9.
42.
El-Sayed MS. Fibrinogen levels and exercise. Is there a relationship?
Sports Med
 
1996
;
21
:
402
–8.
43.
Zanettini R, Bettega D, Agostoni O, et al. Exercise training in mild hypertension: effects on blood pressure, left ventricular mass and coagulation factor VII and fibrinogen.
Cardiology
 
1997
;
88
:
468
–73.
44.
Schuit AJ, Schouten EG, Kluft C, et al. Effect of strenuous exercise on fibrinogen and fibrinolysis in healthy elderly men and women.
Thromb Haemost
 
1997
;
78
:
845
–51.
45.
Ponjee GAE, Janssen EME, Hermans J, et al. Regular physical activity and changes in risk factors for coronary heart disease: a nine months prospective study.
Eur J Clin Chem Clin Biochem
 
1996
;
34
:
477
–83.
46.
Stratton JR, Chandler WL, Schwartz RS, et al. Effects of physical conditioning on fibrinolytic variables and fibrinogen in young and old healthy adults.
Circulation
 
1991
;
83
:
1692
–7.
47.
Krobot K, Hense HW, Cremer P, et al. Determinants of plasma fibrinogen: relation to body weight, waist-to-hip ratio, smoking, alcohol, age and sex: results from the second MONICA Augsburg Survey, 1989–1990.
Arterioscler Thromb
 
1992
;
12
:
780
–8.
48.
McGill JB, Schneider DJ, Arfken CL, et al. Factors responsible for impaired fibrinolysis in obese subjects and NIDDM patients.
Diabetes
 
1994
;
43
:
104
–9.
49.
Everson GA, Goldberg DE, Helmrich SP, et al. Weight gain and the risk of developing insulin resistence syndrome.
Diabetes Care
 
1998
;
21
:
1637
–43.
50.
Fendri S, Roussel B, Lormeau B, et al. Insulin sensitivity, insulin action, and fibrinolysis activity in nondiabetic and diabetic obese subjects.
Metabolism
 
1998
;
47
:
1372
–5.
51.
De Lorenzo F, Mukherjee M, Kadziola Z, et al. Association of overall adiposity rather than body mass index with lipids and procoagulant factors.
Thromb Haemost
 
1998
;
80
:
603
–6.
52.
Hu FB, Stampfer MJ, Manson JE, et al. Dietary fat intake and the risk of coronary heart disease in women.
N Engl J Med
 
1997
;
337
:
1491
–9.
53.
Lee IM, Paffenbarger RS Jr, Hsieh CC. Time trends in physical activity among college alumni, 1962–1988.
Am J Epidemiol
 
1992
;
135
:
915
–25.
54.
Stephen T, Jacobs DRJ, White CC. A descriptive epidemiology of leisure-time physical activity.
Public Health Rep
 
1985
;
100
:
147
–58.