The Relation of Physical Activity to Risk for Symptomatic Gallstone Disease in Men

Methods

Study Sample

The Health Professionals Follow-up Study is a longitudinal study involving 51 529 male dentists, veterinarians, pharmacists, optometrists, osteopathic physicians, and podiatrists in the United States who were 40 to 75 years of age in 1986. The participants completed a detailed mailed questionnaire in 1986 about their history of professionally diagnosed medical conditions, medication use, diet, weight, height, smoking habits, alcohol consumption, and physical activity patterns. Follow-up questionnaires were sent every 2 years to update information on potential risk factors and to ascertain the occurrence of newly diagnosed illnesses, including gallstone disease. The follow-up rate for each questionnaire was approximately 94% of total possible person-years through 1994. At baseline, we excluded a priori 2096 men who reported having had cholecystectomy or a diagnosis of gallstone disease before 1986, 231 men who provided no information on physical activity, 1829 men who had had cancer (other than non-melanoma skin cancer) before 1986, 1227 men with an average daily energy intake outside of the range of 3.35 to 17.6 MJ (800 to 4200 kcal), 282 men who left 70 or more food items blank on the food frequency questionnaire, and 51 men with other missing information. After exclusions, the sample for analysis comprised 45 813 men who were followed from 1986 to 1994.

Assessment of Physical Activity

The baseline questionnaire contained a section on primarily nonoccupational physical activities during the previous year. This questionnaire has been shown to categorize men by level of physical activity with reasonable reliability and validity [23]. Participants reported the average weekly time that they spent doing each of eight activities: jogging (slower than 10 minutes per mile), running (10 minutes per mile or faster), bicycling (including that done on stationary machines), lap swimming, tennis, squash or racquetball, calisthenics or rowing, and walking or hiking outdoors. We also inquired about usual walking pace and the number of flights of stairs climbed daily.

Using a compendium of physical activities [24], we calculated a weekly score for each activity by multiplying the reported duration of the specific activity by its energy expenditure requirements, expressed as metabolic equivalents (METs). One MET is defined as the energy expended by sitting quietly and is equivalent to an oxygen uptake of 3.5 mL/kg of body weight per minute for an adult weighing 70 kg. Body weight was not included in the derivation of energy expenditure of physical activity to avoid confounding the MET variable by body weight. Categories that included more than one activity (for example, squash and racquetball) were assigned an average MET value. A total weekly MET-hour score was calculated by summing the weekly expenditures from all activities.

Activities were further classified as vigorous or nonvigorous: Activities that required 6 or more METs per hour were considered vigorous, and those that required less than 6 METs per hour were considered nonvigorous. In addition, a weekly physical inactivity score was developed on the basis of the average weekly time spent watching television, which was reported in the 1988 follow-up questionnaire. For specific activities, we had to divide men into tertiles to provide enough men in each group; this was done before any modeling was performed. To compare younger men with older men, we chose 65 years of age as a cut-off point because it was the current retirement age and had been used previously in our cohort [25, 26].

Other Variables

At baseline, we assessed dietary and alcohol intake by using a 131-item semiquantitative food frequency questionnaire that had been shown to provide useful data on diet in this cohort [27]. For each food, a commonly used unit or portion size was specified, and participants were asked how often, on average during the previous year, they had consumed that amount of that food. Nutrient intakes were computed by multiplying the frequency of consumption of each unit of food by its nutrient content according to food composition tables from the U.S. Department of Agriculture [28]. To control for energy intake, all nutrients were adjusted for total energy intake by using the residual method [29]. Body mass index was calculated in metric units from participants' reports of their own height in inches and weight in pounds.

Identification of Gallstone Disease

At baseline and in each follow-up questionnaire, participants reported whether they had had cholecystectomy or had received a diagnosis of gallstones from a physician. In the follow-up questionnaires, the men were asked whether their gallstone disease was symptomatic and whether the diagnosis had been confirmed by radiography or ultrasonography. To verify the self-reports of diagnosed but unremoved gallstones and cholecystectomy, we reviewed a sample of 441 medical records of men who reported having gallstones or cholecystectomy; of these, 99% (all but 5) confirmed the diagnosis. Moreover, all of the self-reported symptoms and all but one of the self-reported diagnostic procedures were confirmed by medical record review.

Data Analysis

We computed person-time of follow-up for each participant beginning at the month of return of the 1986 questionnaire and ending at the minimum of the month of cholecystectomy, the month of diagnosis of newly symptomatic gallstones, death, or the end of the study period. Men with asymptomatic gallstone disease, men whose gallstone diagnosis was not based on ultrasonography or radiography, and men with diagnosed cancer (except nonmelanoma skin cancer) were excluded from subsequent follow-up; thus, the eligible sample at risk comprised only those who remained free of gallstone disease and cancer at the beginning of each 2-year follow-up interval.

For multivariate analyses, pooled logistic regression (which accounts for varying time to the outcome event [30] and is asymptotically equivalent to Cox regression modeling [31]) was used to calculate the incidence rate ratios of symptomatic gallstone disease. In this paper, relative risk is equivalent to the incidence rate ratio, given the low incidence of gallstone disease within each follow-up interval. The basic model included age (updated every 2 years); history of diabetes mellitus; smoking; and intake of cholesterol-lowering drugs, thiazide diuretics, non-steroidal anti-inflammatory drugs, alcohol, dietary fiber, and carbohydrates at baseline. The tests for linear trend were done by modeling the specific exposure as a continuous variable in the multivariate logistic model. In addition, we used restricted cubic splines to determine whether physical activity was linearly associated with gallstone disease [32].

Because body mass index could be an intermediate step in the causal pathway linking physical activity to decreased risk for gallstone disease, this variable was not included in our primary analyses. However, we included body mass index in additional models to assess the effect of physical activity on gallstone disease independent of its effect on body weight. In all models that contained body mass index, we also included a variable for body mass index at 21 years of age to account for weight change from 21 years of age to 1986.

In a subanalysis, we calculated the fraction of cases of gallstone disease that could potentially have been prevented [33] if persons had accumulated the currently recommended amount of physical activity: approximately 25 METs or more per week [34]. To determine whether age modified the association between physical activity and gallstone disease, an interaction term between physical activity expressed as a continuous variable and two levels of age (<65 years of age and ≥ 65 years of age) was included in the multivariate model. In a further subanalysis, we corrected the relative risk estimate of physical activity for measurement error by using data from our validation study [23]. This procedure provides an estimate of the association unattenuated by the effects of measurement error [35]. Relative risks are presented with 95% CIs, and reported P values are two-tailed. All analyses were performed with SAS software, release 6.12 (SAS Institute, Cary, North Carolina).

Role of Funding Sources

The funding sources for this study had no role in collection, interpretation, or analysis of the data or in the decision to submit the report for publication.

Results

During 324 263 person-years of follow-up, we documented 828 cases of self-reported symptomatic gallstones, 661 of which required cholecystectomy. We first examined how physical activity related to potential confounders of risk for gallstone disease (Table 1). Men who were more physically active tended to be leaner, to have a lower waist-to-hip ratio, to have a smaller waist circumference, and to smoke less. Physical activity was also inversely associated with a history of diabetes and regular use of diuretics and aspirin, but it was not related to the use of cholesterol-lowering drugs. Men with a history of gallstones before 1986 (excluded from all further analyses) were less physically active (multivariate-adjusted METs per week, 10.8) than were men without such a history (multivariate-adjusted METs per week, 12.9; P < 0.001).

Total physical activity was inversely related to risk for symptomatic gallstone disease. Men in the highest quintile of physical activity had a multivariate relative risk of 0.63 (95% CI, 0.51 to 0.79) compared with men in the lowest quintile (Table 2). To examine the benefit of physical activity independent of its effect on body weight, we adjusted for body mass index at 21 years of age and for body mass index in 1986. The inverse association became slightly weaker (multivariate relative risk, 0.72 [CI, 0.58 to 0.90] for men in the highest quintile compared with men in the lowest quintile). The cubic spline analyses indicated that the association between physical activity and gallstone disease was linear. When we entered METs per week for physical activity as a continuous variable into a multivariate model without body mass index, an increase in physical activity of 25 METs per week (≥ 30 minutes of moderate-intensity, endurance-type exercise on 5 days of the week) was associated with a relative risk of 0.81 (CI, 0.73 to 0.89). After correction for measurement error in the assessment of physical activity, the relative risk was 0.64 (CI, 0.45 to 0.92).

To examine the possibility that symptoms of latent gallstone disease could have induced a reduction in physical activity, biasing our results, we excluded the first 2-year follow-up period. The relation between total physical activity and gallstone disease was essentially unchanged (relative risk, 0.65 [CI, 0.49 to 0.86] for men in the highest quintile compared with men in the lowest quintile). Additional exclusion of respondents who did not have a routine medical check-up between 1986 and 1988 also had a negligible effect (relative risk for men in the highest quintile compared with men in the lowest quintile, 0.62 [CI, 0.45 to 0.85]). An analysis of the prevented fraction indicated that 34% of potential cases of gallstones could be prevented by expending at least 25 METs per week.

Most of the inverse risk for gallstone disease was explained by vigorous physical activity. After we adjusted for multiple confounding factors, including nonvigorous activity, each increase in vigorous activity of 25 METs per week was associated with a relative risk of 0.77 (CI, 0.67 to 0.88). In contrast, after we adjusted for multiple confounding factors, including vigorous activity, each increase in nonvigorous activity of 25 METs per week was associated with a relative risk of 0.87 (CI, 0.73 to 1.03).

We next examined the effect of age on the association of physical activity with symptomatic gallstone disease. In men younger than 65 years of age (n = 524), the multivariate-adjusted relative risk for total physical activity when extreme quintiles were compared was 0.58 (CI, 0.44 to 0.78). For men 65 years of age or older (n = 304), it was 0.75 (CI, 0.52 to 1.09) (Table 3). Similarly, for vigorous activity, the inverse association was stronger in younger men (relative risk when extreme categories were compared, 0.52 [CI, 0.38 to 0.72]) than in older men (relative risk when extreme categories were compared, 0.81 [CI, 0.52 to 1.26]). However, results of the formal test for age interaction were not statistically significant (P > 0.2). Moreover, declining physical activity did not account for the strong relation between increasing age and increasing risk for symptomatic gallstone disease.

We investigated the relation of various specific leisure-time activities to risk for symptomatic gallstone disease by comparing the second and third tertiles with the lowest tertile (Table 4). In men younger than 65 years of age, significant inverse associations for the highest tertile compared with the lowest tertile were seen for jogging and running combined, racquet sports, and brisk walking pace. For men 65 years of age or older, we found significant inverse associations when comparing extreme categories for racquet sports and brisk walking pace after we adjusted for multiple risk factors.

We also examined the relation between sedentary behavior and risk for gallstone disease (Table 5). Men who watched more than 40 hours of television per week had a multivariate-adjusted relative risk of 2.68 (CI, 1.38 to 5.18) compared with men who watched less than 6 hours of television per week. A particularly strong association was seen for men older than 65 years of age, although the results of a formal test for interaction with age were not statistically significant (P > 0.2).

Discussion

Our data support the hypothesis that physical activity decreases the incidence of newly symptomatic gallstone disease in men. An average of 2 to 3 hours of moderate running per week seemed to reduce the risk for this disease by 20% and by as much as 40% after adjustment for imperfect measurement of physical activity. Most specific leisure-time activities showed a trend toward decreased risk, although jogging and running combined, racquet sports, and brisk walking pace were the only activities or factors that were statistically significant. This suggests that the overall level of physical activity rather than any particular form of exercise is the main determinant of risk.

After we adjusted for body mass index and weight change, the inverse associations were attenuated but still evident. We conclude that physical activity may contribute to the prevention of symptomatic gallstone disease beyond its effect on weight control.

When men younger than 65 years of age and men 65 years of age or older were analyzed separately, vigorous activity was related to a reduced risk in younger men but not in older men. The most probable explanation for this is that older men do not exercise as intensely as younger men, with the exception of a brisk walking pace, which may be a vigorous physical activity for elderly persons. In addition, older men had a stronger positive association between physical inactivity and symptomatic gallstone disease than did younger men. This finding indicates that in younger men, the main contributor to decreased risk is an increased level of vigorous activity, whereas in older men, a reduction in the time spent in sedentary activities may be a more important factor in decreasing risk.

Although some previous studies did not detect an association between physical activity and gallstones, possibly because their physical activity assessment methods were limited or their study samples were small [6-14], our findings are consistent with a study of men of Japanese ancestry in Hawaii that reported that physical activity was associated with a relative risk of 0.7 (CI, 0.6 to 1.0) [15]. One study in women [16] found that physical activity was inversely related to gallstone prevalence. A cross-sectional study [19] found an inverse association in diabetic Pima Indian men but not women, reportedly because of low variability in physical activity among the women. In a cross-sectional survey [17], persons with low work activity had a higher prevalence of cholesterol gallstones. Similarly, one case–control study [18] found an inverse relation with amount of manual labor. Another [20] found an inverse association with time spent walking.

One limitation of our study was that participants were not systematically screened for the presence of gallstones by using ultrasonography or other imaging tests. We restricted our analyses to men with cholecystectomy or diagnostically confirmed but unremoved symptomatic gallstones. Although doing this underascertains the true incidence of gallstones in the study sample because most cases of gallstones are asymptomatic [36], relative risks in follow-up studies are not biased by uniform underascertainment [37]. Moreover, our analysis focused on clinically relevant gallstone disease. We cannot exclude the possibility that some less physically active persons may not have exercised because they had gastrointestinal symptoms and may have consulted physicians more frequently, resulting in a greater diagnosis of gallstones. However, this is not likely to account for our results because the physically active men in our cohort were more likely than the less active men to have routine medical check-ups. In addition, we found no appreciable changes in the association with physical activity when we excluded men with unremoved gallstones, who were presumably less symptomatic and more prone to detection bias. Therefore, detection bias is unlikely to have influenced our results. We presume that some of the cases of symptomatic gallstone disease ascertained during follow-up existed at baseline in an asymptomatic state. However, it seems unlikely that the presence of asymptomatic and undiagnosed gallstones at baseline is related to the men's reporting of physical activity. This is further substantiated by the analysis that excluded the first 2 years of follow-up, in which the findings were virtually unchanged. The composition of the gallstones in our study was unknown to us, but nearly 80% of those in men in the Framingham study were cholesterol or mixed stones [10].

Although information on physical activity was collected entirely by self-report, a recent validation study [23] found that this cohort provides reasonably accurate and reliable information on both physical activity and sedentary behavior. In addition, the physical activity variable tracks fairly well over time, showing a correlation coefficient of 0.57 between MET-hours in 1986 and 1988. Because the data on physical activity were gathered before symptoms of gallstone disease developed, misclassification would tend to weaken rather than strengthen the relation between physical activity and symptomatic gallstone disease.

Physical activity may decrease risk for symptomatic gallstone disease through several metabolic pathways both independently and as a result of its role in maintaining body weight. Serum HDL cholesterol levels, which are increased in persons who exercise regularly [38], are inversely associated with gallstone prevalence [39]. Levels of HDL cholesterol also correlate directly with total biliary bile acids [40] and inversely with bile lithogenicity [41], possibly because HDL cholesterol is preferentially metabolized to bile acids [42]. Serum triglyceride levels, which are inversely associated with physical activity [43], show a positive relation with gallstone occurrence [44]. Triglycerides accumulated in liver cells may stimulate mucin hypersecretion by gallbladder mucosal cells [45]; this condition plays an important role in biliary cholesterol lithogenesis [46].

Because impaired glucose tolerance [47] and elevated insulin levels [48] seem to be risk factors for gallstones and because regular physical activity improves glucose utilization [49], the protective influence of physical activity on gallstone development may be due to the effect of physical activity on insulin resistance and hyperinsulinemia. Insulin enhances cholesterol uptake by the liver by activating hepatic low-density lipoprotein receptors [50]. Insulin also stimulates 3-hydroxy-3-methylglutaryl coenzyme A reductase activity [51] and decreases 7α-hydroxylase activity [52], resulting in increased biliary cholesterol secretion and decreased bile acid secretion. The improvement in glycemic control with exercise training has been shown to be partly independent of the effect of weight loss on insulin sensitivity [53].

Some patients with gallstone disease have reduced intestinal transit [54], and intestinal migrating motor complexes may be involved in this condition [55]. In contrast, physical exercise seems to have a prokinetic effect on the gut [56, 57]. Alterations in intestinal transit may modify the enterohepatic circulation of bile salts, leading to increased colonic absorption of deoxycholic acid [58], which is known to promote cholesterol nucleation [59]. Marathon running increases cholecystokinin levels [60], indicating that certain types of physical activity may increase gallbladder motility.

The protective effect of physical activity on gallstone formation may be mediated by changes in hepatobiliary function. In animal models, long-term moderate exercise increases bile acid excretion [61]. Exercise also improves the age-related decline in bile formation and secretion of bile salts [62].

In conclusion, our data suggest that physical activity may be an important determinant of symptomatic gallstone disease. The apparent protective effect is greatest for vigorous physical activity but can also be achieved through moderate exercise, such as rapid walking or hiking. In addition, physical inactivity may be independently associated with increased risk for symptomatic gallstone disease. Our results are generalizable to U.S. men 40 years of age and older and may not apply directly to other populations. Further research is needed to evaluate our findings in other groups, especially women, who are at greater risk for gallstones.

Drs. Giovannucci and Stampfer: Channing Laboratory, Department of Medicine, Harvard Medical School and Brigham and Women's Hospital, 181 Longwood Avenue, Boston, MA 02115.

Dr. Spiegelman and Mr. Wing: Department of Epidemiology, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115.