OUP user menu

Association of physical activity volume and hypercholesterolemia in US adults

J.R. Churilla, T.M. Johnson, E.A. Zippel
DOI: http://dx.doi.org/10.1093/qjmed/hcs231 333-340 First published online: 19 December 2012

Abstract

Background: Several studies illustrate the favorable association between physical activity (PA) and cholesterol levels. There is a paucity of data examining the PA patterns of individuals with and without hypercholesterolemia (HC).

Aim: To examine self-reported moderate and vigorous PA (MVPA) patterns using the most recent PA guidelines among US adults with and without HC.

Design: Cross-sectional study utilizing a secondary data analysis approach.

Methods: We used data from the 2009 Behavioral Risk Factor Surveillance System (BRFSS). PA categories were based on the 2008 Department of Health and Human Services (DHHS) guidelines.

Results: The age-adjusted prevalence of self-reported HC in US adults was 34%. When stratified by gender, the age-adjusted prevalence of HC was found to be significantly higher in men (36.2%; 95% CI 35.6, 36.8) compared with women (31.8%; 95% CI 31.3, 32.3). The age-adjusted prevalence of meeting the DHHS PA recommendation was 59.1% among participants reporting HC and 68.3% among participants not reporting HC (P < 0.05). Following adjustment for demographics and health history, the odds ratio for meeting the DHHS PA recommendation among participants with HC compared with those without HC was 0.86 (95% CI 0.83, 0.89).

Conclusions: Although a large proportion of adults reporting HC report engaging in a volume of MVPA necessary to meet national guidelines, their odds of meeting these guidelines and their MVPA volume may be significantly lower than adults who did not report HC.

Introduction

Cardiovascular disease (CVD) remains the leading cause of morbidity and mortality in the USA. It has been postulated that 75% of CVDs can be attributed to modifiable risk factors.1 The World Health Organization estimates that hypercholesterolemia (HC) attributes to more than half of coronary heart disease (CHD) cases worldwide annually.2 Average total cholesterol (TC) levels have remained unchanged worldwide but increased use of pharmacological and nutritional interventions have resulted in declines in North America, Australia, Asia, Western, Central and Eastern Europe.3 TC is classified as high when levels are ≥240 mg/dl, with desirable levels being <200 mg/dl. According to data from 2005 to 2008 National Health and Nutrition Examination Survey (NHANES) ∼16% of US adults had a cholesterol level ≥240 mg/dl.4 Furthermore, data from 1999 to 2006 NHANES illustrated that 8% of US adults have undiagnosed HC.5 In addition to pharmacological and nutritional therapies for HC, engaging in PA has also showed promise in favorably impacting blood lipid values.6–16

Healthcare providers should be aware of current PA trends and the benefits of exercise for CVD risk factors, such as HC. Regular PA has been shown to reduce TC levels and provide health benefits in healthy individuals and those at risk for CVD.7,8,10–12,14–16 Current PA recommendations for all Americans according to the US Department of Health and Human Services (DHHS), is to engage in at least 150 min of moderate intensity PA (MIPA) or 75 min of vigorous intensity PA (VIPA) per week or a combination of both to meet the PA recommendations. In addition, 2 or more days per week of resistance training is recommended.17 Previous research in the area of PA intensity has illustrated that beneficial changes in TC have been shown to be population specific.8,9,13 Furthermore, higher volumes of PA and greater levels of cardiorespiratory fitness have shown to produce beneficial reductions in TC.16 It is important to differentiate the benefits of regular PA and level of physical fitness when examining TC reduction.6 Aims of our study were (i) to investigate if American adults with self-reported HC are engaging in a level of MIPA or VIPA that meets or exceeds the new DHHS PA guidelines and (ii) to compare these PA patterns with the PA patterns of US adults without self-reported HC in large, nationally representative, state-based cohorts of adults.

Methods

Sample

The data for this study came from the 2009 Behavioral Risk Factor Surveillance System (BRFSS). Data from 322 465 respondents were collected, of which we obtained complete data for 253 570 participants. The BRFSS is a state-based, continuous, random-digit-dialed telephone survey conducted among the non-institutionalized US civilian population 18 years and older in all 50 states, the District of Columbia and three US territories. BRFSS obtains data on demographics, current and evolving health issues and health behaviors of US adults. Detailed information on sampling methodology for BRFSS has been previously described.18 The BRFSS data have been shown to provide valid and reliable estimates when examining other nationally representative US survey data.19 The Institutional Review Board of the University of North Florida approved the use of the 2009 BRFSS data.

Hypercholesterolemia

HC was assessed by asking the participant if they had ever been told by a doctor, nurse or other healthcare professional that they had high cholesterol. The respondents’ choices for answers were (i) yes, (ii) no, (iii) don’t know/not sure and (iv) refused. The 2009 BRFSS participants were asked if their cholesterol was high but levels were not quantified.20 Anyone who answered ‘yes’ to the HC question was classified as having HC and anyone who answered ‘no’ was classified as not having HC. Those who responded ‘don't know/not sure’ or who refused to answer the question were set to missing. The 2009 BRFSS did not ask questions regarding cholesterol medications.

Physical activity

In 2009, respondents were asked if during a typical week they engaged in any MIPA, VIPA or total PA (TOPA) independent of their duties performed at work. They were then asked how often (how many days) and for what duration of time they engaged MIPA, VIPA or TOPA. Respondents were asked how many days during a typical week they engaged in MIPA, which included things like brisk walking, cycling, domestic activities, yard work or any other PAs for at least 10 min in duration and which resulted in ‘some’ increase in breathing or heart rate. They were then asked how many days during a typical week they engaged in VIPA, which included things like running, aerobics, heavy yard work or any other PAs for at least 10 min in duration and which resulted in ‘large’ increases in breathing or heart rate. If they reported 10 min or more of MIPA or VIPA, respondents were asked to provide the total time they spent performing the activities.

Utilizing the 2008 DHHS PA guidelines,17 respondents were considered physically active if they reported a minimum of 150 weekly minutes of MIPA, a minimum of 75 weekly minutes of VIPA or an aggregate of MIPA and VIPA × 2. Respondents were classified into four categories: (i) highly active (HACT) >300 min of weekly MIPA (meets DHHS recommendations); (ii) moderately active (MACT) 150–300 weekly minutes of MIPA or 75–150 weekly minutes of VIPA (meets DHHS recommendations); (iii) low activity (LACT) represented a TOPA <150 weekly minutes of PA or the equivalent (does not meet recommendations) and (iv) inactive (INACT), which were respondents reporting no PA beyond activities of daily living. The following example illustrates the aggregated PA time for someone who reported 30 min of MIPA three times per week and 10 min of VIPA three times per week. This example illustrates how someone reporting PA of varying intensities during 1 week can meet the new recommendations despite not engaging in enough MIPA or VIPA independently. Embedded Image

Statistical analysis

We used SAS21 to recode variables of interest and SUDAAN22 to conduct our analysis. SUDAAN incorporates sampling weights within the context of the sampling design inherent to the BRFSS.20 We calculated age-adjusted prevalence estimates for HC and meeting the 2008 DHHS PA recommendations by demographic variables using PROC DESCRIPT with the direct method based on the 2000 projected US population aged ≥18 years. PROC CROSSTAB was utilized in SUDAAN for calculating descriptive statistics for categorical variables. We used logistic regression using PROC RLOGIST to calculate the odds ratios (ORs) for participants with self-reported HC meeting the DHHS PA recommendation. PROC RLOGIST is a standard procedure in SUDAAN that fits logistic regression models to complex survey and other cluster-sampled data (e.g. BRFSS and NHANES). We dichotomized the four levels of PA: INACT and LACT were collapsed (does not meet DHHS PA recommendations) and MACT and HACT were collapsed (meets DHHS PA recommendations). With logistic regression, our analysis controlled for age, sex, body mass index (BMI), race, smoking, education, diabetes, hypertension and CVD.

Results

HC prevalence and risk

According to the 2009 BRFSS, the age-adjusted prevalence of self-reported HC in US adults was 34%. When stratified by gender, the age-adjusted prevalence of HC was found to be significantly higher in men (36.2%; 95% CI 35.6, 36.8) compared with women (31.8%; 95% CI 31.3, 32.3). Women were found to be 10% less likely to report having HC (OR 0.90; 95% CI 0.86, 0.93). The prevalence of HC was positively associated with age and BMI (P < 0.05 for both). HC was more likely to be present in those who reported being overweight (OR 1.44; 95% CI 1.38, 1.50), obese (OR 1.54; 95% CI 1.46, 1.62), having CVD (OR 1.86; 95% CI 1.76, 1.96) or diabetes (OR 1.65; 95% CI 1.56, 1.75). In addition, those reporting having hypertension were found to be two times more likely to report HC (OR 2.01; 95% CI 1.93, 2.09).

Age and gender

Compared with participants reporting HC, greater levels of HACT and MACT and lower levels of LACT and INACT were found among those not reporting HC across all age categories (P < 0.001) except in the youngest age category (P = 0.223) (Table 1).

View this table:
Table 1

Age-adjusted prevalence of PA patterns among US adults with and without HC

CharacteristicHCNo HC
n = 111 507n = 142 063
HACTMACTLACTINACTHACTMACTLACTINACTP-value
% (SE)% (SE)% (SE)% (SE) % (SE)% (SE)% (SE)% (SE)
Age
    18–2458.8 (4.04)14.5 (2.59)18.4 (3.54)8.2 (2.05)63.5 (1.40)17.8 (1.11)12.8 (0.98)5.9 (0.64)0.223
    25–3443.6 (1.59)22.2 (1.30)25.1 (1.45)9.1 (0.93)52.4 (0.76)21.1 (0.63)19.1 (0.57)7.4 (0.41)<0.001
    35–4442.9 (0.88)22.2 (0.73)23.4 (0.72)11.4 (0.60)48.5 (0.54)22.7 (0.45)21.0 (0.43)7.8 (0.32)<0.001
    45–6439.3 (0.39)21.8 (0.32)25.1 (0.33)13.8 (0.27)46.2 (0.34)21.9 (0.28)22.2 (029)9.7 (0.21)<0.001
    ≥6531.9 (0.37)20.2 (0.31)26.6 (0.36)21.3 (0.32)34.9 (0.41)20.0 (0.31)23.7 (0.39)21.3 (0.32)<0.001
Gender
    Male44.2 (0.45)21.0 (0.36)21.8 (0.38)13.0 (0.29)54.0 (0.44)19.6 (0.35)17.2 (0.32)9.2 (0.24)<0.001
    Female32.7 (0.35)21.5 (0.29)28.6 (0.32)17.3 (0.27)42.3 (0.31)22.7 (0.27)23.8 (0.27)11.2 (0.20)<0.001
BMI (kg/m2)
    <2542.7 (0.57)21.6 (0.47)22.7 (0.50)13.0 (0.37)52.4 (0.44)20.9 (0.35)18.4 (0.33)8.3 (0.23)<0.001
    25 to <3041.6 (0.47)22.0 (0.37)23.9 (0.39)12.6 (0.30)49.5 (0.44)21.9 (0.36)19.6 (0.33)9.0 (0.23)<0.001
    ≥3032.5 (0.50)20.2 (0.41)28.2 (0.44)19.1 (0.38)39.1 (0.54)21.1 (0.45)25.6 (0.44)14.2 (0.35)<0.001
Race
    nH-White38.9 (0.29)22.0 (0.24)25.2 (0.25)14.0 (0.19)49.2 (0.28)21.4 (0.22)20.6 (0.21)8.8 (0.13)<0.001
    nH-Black33.2 (1.09)19.0 (0.89)27.0 (0.95)20.8 (0.82)41.1 (0.96)20.1 (0.90)23.9 (0.82)15.0 (0.63)<0.001
    Hispanic40.1 (1.34)18.7 (1.03)23.7 (1.19)17.6 (0.98)46.3 (1.09)20.1 (0.85)19.1 (0.81)14.5 (0.76)<0.001
    Other37.4 (1.46)20.3 (1.11)25.2 (1.30)17.0 (1.07)44.5 (1.18)22.9 (0.96)21.5 (0.98)11.1 (0.71)<0.001
Smoking
    Smoker38.1 (0.75)18.8 (0.59)24.9 (0.62)18.3 (0.54)50.3 (0.70)18.2 (0.52)19.5 (0.51)12.0 (0.41)<0.001
    Former smoker38.0 (0.46)21.2 (0.37)24.9 (0.40)15.9 (0.32)47.4 (0.47)20.9 (0.38)20.5 (0.36)11.2 (0.28)<0.001
    Never smoked38.9 (0.42)22.0 (0.35)25.5 (0.37)13.6 (0.28)47.2 (0.35)22.2 (0.29)21.1 (0.28)9.39 (0.20)<0.001
Education
    <High school33.1 (1.08)15.4 (0.71)24.5 (0.89)27.3 (0.86)37.4 (1.15)15.8 (0.92)21.7 (0.95)25.0 (0.97)0.018
    High school/GED35.4 (0.52)19.5 (0.42)26.7 (0.47)18.5 (0.39)44.9 (0.56)20.0 (0.45)21.6 (0.42)13.6 (0.34)<0.001
    >High school40.6 (0.36)22.9 (0.30)24.6 (0.31)11.8 (0.23)49.7 (0.31)22.4 (0.25)20.4 (0.24)7.5 (0.15)<0.001
Diabetes
    Yes27.9 (0.59)19.3 (0.52)27.9 (0.57)25.0 (0.57)31.6 (0.85)21.0 (0.88)24.9 (0.72)22.5 (0.74)<0.001
    No40.7 (0.32)21.7 (0.26)24.6 (0.28)13.1 (0.21)48.8 (0.28)21.3 (0.22)20.5 (0.22)9.4 (0.15)<0.001
Hypertension
    Yes34.0 (0.37)20.3 (0.30)26.8 (0.33)18.9 (0.28)38.2 (0.47)21.4 (0.40)24.3 (0.39)16.1 (0.33)<0.001
    No42.9 (0.44)22.2 (0.36)23.5 (0.38)11.3 (0.28)50.7 (0.31)21.2 (0.25)19.7 (0.24)8.4 (0.17)<0.001
CVD
    Yes30.0 (0.89)18.6 (0.48)25.8 (0.54)25.6 (0.54)33.7 (0.89)19.2 (0.79)23.0 (0.74)24.2 (0.81)<0.001
    No40.0 (0.32)21.7 (0.26)25.1 (0.28)13.2 (0.21)48.4 (0.27)21.4 (0.22)20.1 (0.21)9.5 (0.15)<0.001
  • HACT: highly active—the equivalent of >300 weekly minutes of MIPA; MACT: medium activity—engaging in ≥150–300 weekly minutes of MIPA or 75–150 weekly minutes of VIPA; LACT: low activity—engaging in <150 weekly minutes of MIPA or 75 min of VIPA or combined equivalent; INACT: no PA beyond daily living tasks; nH: non-Hispanic.

Body mass index

We found the prevalence of INACT to be significantly greater among all three BMI categories for those who reported having HC compared with their counterparts who did not report HC (P < 0.001 for all). In addition, independent of self-reported HC status, those in the desirable (<25 kg/m2) and overweight (25 to <30 kg/m2) BMI categories were found to have a significantly greater prevalence of HACT than those in the obese (≥30 kg/m2) BMI category (P < 0.05).

Race

When examining race, independent of HC status, we found non-Hispanic Blacks reported significantly lower levels of HACT and significantly higher levels of INACT compared with non-Hispanic Whites and Hispanics (P < 0.05 for both). In addition, the prevalence of INACT was found to be significant lower in non-Hispanic Whites compared with all other race categories independent of HC status (<0.05 for all) (Table 1).

Smoking

Compared with participants who did not report having HC, significantly fewer participants with HC who reported being current or former smokers reported being HACT (P < 0.001 for both). In addition, the prevalence of INACT was found to be significantly greater among current smokers compared with former or never smokers independent of cholesterol status (P < 0.05 for both) (Table 1).

Education

Significantly fewer participants reporting HC and having a high school education or greater reported being in the HACT category compared with their counterparts not reporting HC (P < 0.001) (Table 1). In addition, among participants not reporting HC and having a high school education or greater, the self-reported prevalence of being HACT was greater when compared with those reporting less than a high school education (P < 0.05).

Diabetes

Independent of HC status, the prevalence of being HACT was found to be significantly greater among those reporting no diabetes. Furthermore and also independent of cholesterol status, the prevalence of self-reported inactivity was found to be 2-fold among those with diabetes compared with those without diabetes (P < 0.05 for both) (Table 1).

Hypertension

Independent of HC status, those with self-reported hypertension were found to report lower levels of HACT and twice the rate of inactivity compared with those without hypertension (P < 0.05 for both) (Table 1).

Cardiovascular disease

Among those with self-reported CVD and independent of HC status, significantly lower rates of MACT and HACT were reported compared with those without CVD (P < 0.05 for all) (Table 1). Furthermore, independent of HC status and similar to those with hypertension and diabetes, inactivity rates were found to be significantly lower among those reporting no CVD (P < 0.05 for both).

Meeting the DHHS PA recommendations

Overall, 59.1% of participants with HC and 68.3% of participants without HC reported a level of PA that met or exceeded the DHHS recommendations following age-adjustment. In both men and women and across race categories, US adults reporting HC were found to meet the DHHS PA recommendation at significantly lower rates than those not reporting HC (P < 0.01 for all). Furthermore, men with or without self-reported HC reported meeting the DHHS PA recommendations at a greater rate than women, independent of race with the exception of those in the other category (P < 0.05 for all) (Figure 1).

Figure 1

Prevalence of US adults with (light-gray bars) and without (dark-gray bars) self-reported HC that reported meeting the DHHS PA recommendations, 2009 BRFSS. Error bars represent 95% CI.

Hypercholesterolemia and physical activity odds ratios

The unadjusted OR of meeting the DHHS PA recommendations among adults reporting HC compared with those not reporting HC was 0.67 (95% CI 0.64, 0.69). This OR was slightly attenuated, but remained statistically significant following adjustments for age (model 2) and demographic variables including sex, BMI, race, smoking, education, diabetes, hypertension and CVD (model 3) (Table 2).

View this table:
Table 2

ORs for US adults with self-reported HC meeting the current DHHS PA guidelines compared with US adults not reporting HC, 2009 BRFSS

HC statusOR95% CI
Yesa0.670.64, 0.69
No1.00
Yesb0.770.74, 0.79
No1.00
Yesc0.860.83, 0.89
No1.00
  • aUnadjusted model, bage-adjusted model and cadjusted for age, gender, BMI, race, smoking, education, diabetes, hypertension and CVD.

Discussion

There is a paucity of data examining the PA patterns of US adults with HC. Our study utilized data from the 2009 BRFSS, a nationally representative sample, and illustrated that 59.1% of US adults with self-reported HC meet the current DHHS PA recommendations. This is an important finding considering that data from 1999 to 2006 National Health and Nutrition Examination Survey illustrates 69.2% of US adults that have been told by their healthcare provider that they had HC and advised to increase their PA level are complying. Additionally, our findings indicate that US adults with self-reported HC were less likely to report the volume of PA necessary to meet the current DHHS PA recommendation compared with those not reporting HC. Previous epidemiological studies have been conducted utilizing comparable methodology examining hypertension, CHD and diabetes.23–25 Our findings are compelling since previous research supports lipids being favorably impacted by regular PA 6,7,10–12 and higher levels of functional capacity.6 There may be several reasons for this incongruity. The DHHS PA recommendation does not include frequency; therefore, an individual who reports 4 h (240 min) of cycling 1 day per week would meet the DHHS recommendation17 because of the focus on total PA volume (TPAV). This same individual would not meet the PA recommendations put forth by the American College of Sports Medicine and the American Heart Association26 because of the frequency being only 1 day per week. Also, data on PA from various domains, which were not traditionally collected during previous investigations, are now becoming more available due to investigators and federal surveys refining methodology. If we are not going to consider frequency when making PA recommendations, then more studies are needed that will control for TPAV.

Previous research supports regular exercise favorably impacting lipid profiles7 and self-reported PA being favorably associated with less atherogenic lipid profiles.10,11,14,15 Findings from studies of a targeted risk reduction intervention through defined exercise (STRRIDE) show that total exercise volume not intensity can induce favorably changes in lipoproteins.7 Participants in STRRIDE were randomized to one 6-month control group (sedentary) and three 8-month exercise groups. The exercise group characteristics were low-volume moderate intensity [∼1200 kcal/week (800 MET min/week at 40–65% VO2PEAK)]; low-volume vigorous intensity [∼1200 kcal/week (800 MET min/week at 60–75% VO2PEAK)] and high-volume vigorous intensity [∼2200 kcal/week (1314 MET min/week at 60–75% VO2PEAK)]. One MET (metabolic equivalent) represents an estimate of resting metabolism (∼3.5 ml/kg/min). The MET level corresponds with the intensity of PA, the higher the MET level the greater the level of oxygen consumption. Metabolic equivalents are commonly used in exercise epidemiology research to calculate energy expenditure and PA volume to allow investigators to examine potential dose-response relationships. The 22 overweight men and women assigned to the high-volume vigorous intensity group in STRRIDE experienced an ∼10% increase in high-density lipoprotein cholesterol (HDL-C) and a 20% decrease in triglycerides (TGs), two components that contribute to TC and potentially HC. These changes were seen independent of weight loss and with no change in TC levels. This is important because every 1% increase in HDL-C can reduce CHD risk 3% by favorably impacting the TC/HDL-C ratio.27 Hokanson et al.28 reported independent of HDL-C, men and women were 14 and 37% more likely to suffer from CHD, respectively, when their TGs were elevated. Thus, lowering TGs will also favorably impact CHD risk and may lower TC levels. Favorable changes in the lipoprotein panels of all exercise groups were seen in STRRIDE; however, the high-volume vigorous intensity exercise resulted in the greatest changes with TPAV being more important than intensity.

Findings from the ATTICA study, which was a health and nutrition survey, illustrate how various levels and modes of PA impacted CVD risk factors in a randomly selected group of 1514 men and 1528 women living in the region of ATTICA Greece.10,11 The PA data were collected using a translated version of the International Physical Activity Questionnaire (IPAQ).29 Participants reporting no PA were classified as sedentary, those reporting ≥600 MET min/week but <1500 MET min/week were classified as sufficiently active and participants reporting a minimum of 1500 MET min/week of vigorous intensity PA 3 days/week or 3000 MET min/week of aggregated moderate or vigorous intensity PA 7 days/week were classified as highly active aerobic only or combination aerobic and resistance exercise (CARE). Biochemical data from blood samples indicated that the baseline prevalence of HC (TC >200 mg/dl) in men and women participating in the ATTICA study was 46 and 40%, respectively.

Following analysis of the IPAQ data and adjustment for potential confounding factors, men and women classified in the CARE group were found to have TC levels 5 and 6.3% lower than INACT controls, respectively. Men in the CARE group were found to have low-density lipoprotein cholesterol (LDL-C) levels 10% lower than both sufficiently active and sedentary men along with apoprotein-B (apo-B) levels 7.9 and 9.5% lower compared with the highly active aerobic and sedentary men, respectively. Additionally, men in the CARE group also had TGs levels 23% lower than the highly active aerobic exercise only men. Women in the CARE group had LDL-C levels 11.9 and 13.5% lower compared with sedentary and sufficiently active women, respectively. Additionally, women in the CARE group also had apo-B levels 11.8% lower than sufficiently active women. Findings from the ATTICA study suggest that the aggregate effect of combining aerobic type exercise with RE may have a more favorably impact on lipoproteins and lipoprotein subfractions compared with aerobic exercise alone. Our findings indicate a greater proportion of US adults without self-reported HC meet the DHHS PA recommendation compared with those with HC, therefore including RE to an exercise program has the potential to increase TPAV, which may be a key mediating factor when examining exercise, PA and health risks.

These study findings translate into support for research conducted by Williams examining the association of total running volume on lipid levels and other CHD risk factors in female14 and male15 runners. When comparing cross-sectional survey data on running distances with physician provided medical data, significantly lower levels of TC and greater levels of HDL-C were found in female and male runners reporting greater running volumes compared with the lowest running volume groups. Additionally, a significant inverse dose–response relationship was found between running volume and both LDL-C and TGs in males.

Conclusion

In summary, US adults with self-reported HC reported being less physically active than their counterparts who did not report HC. Our findings suggest that although a majority of Americans who report HC do report a level of PA that would meet the current PA recommendations, a substantial minority do not, and are more likely than Americans not reporting HC to not meet the recommendations. Greater TPAV may increase the odds of meeting governmental PA guidelines as well as reduce CHD risk from HC.

Conflict of interest: None declared.

References

View Abstract