Plasma lipoproteins play a major role in the aetiology of atherosclerosis and CVDReference Heiss, Johnson, Reiland, Davis and Tyroler(1, Reference Schaefer2), which continue to be the leading cause of morbidity and mortality in the industrialised worldReference Thom, Haase and Rosamond(3).
It has been well documented that a variety of personal characteristics and environmental factors influence the composition of plasma lipids and lipoproteins, including age, gender and the associated hormonal changes in women, genetics, BMI, body weight, body composition, alcohol consumption, smoking behaviour and medication useReference Heiss, Johnson, Reiland, Davis and Tyroler(1, Reference Hall, Collins, Csemiczky and Landgren4–Reference Arquer, Elosua, Covas, Molina and Marrugat10).
In addition, it has been shown that physical activity can affect the plasma lipid profile. Most studies indicate that regular exercise produces favourable changes in plasma lipids and lipoproteinsReference Ashton, Nanchahal and Wood(11–Reference Pescatello, Murphy and Costanzo21).
Evidence from epidemiological and training studies indicates that those who are physically active exhibit higher levels of HDL-cholesterol (HDL-C)Reference Twisk, Kemper, Mellenbergh, van Mechelen and Post(6, Reference Young, Haskell, Jatulis and Fortmann7, Reference Marti, Suter, Riesen, Tschopp, Wanner and Gutzwiller18, Reference Bijnen, Feskens, Caspersen, Giampaoli, Nissinen, Menotti, Mosterd and Kromhout22–Reference Barengo, Kastarinen, Lakka, Nissinen and Tuomilehto30), lower levels of TAG and a lower ratio of total cholesterol (TC) to HDL-C than those who are less activeReference Ashton, Nanchahal and Wood(11, Reference Durstine, Grandjean, Cox and Thompson12, Reference Kraus, Houmard and Duscha14–Reference Ma, Liu and Ling16, Reference Wei, Macera, Hornung and Blair19, Reference Pescatello, Murphy and Costanzo21, Reference Lippi, Schena, Salvagno, Montagnana, Ballestrieri and Guidi31, Reference Kobayashi, Murase, Asano, Nohara, Kawashiri, Inazu, Yamagishi and Mabuchi32). Durstine et al.Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33) reviewed the literature and concluded that regular exercise can raise HDL-C levels by 2–8 mg/dl and lower TAG by 5–38 mg/dl in men and women.
The preceding results indicate that physical activity can positively affect the lipid profile. However, some investigators have failed to find a significant association between physical activity and favourable changes in the lipid profileReference Danielson, Cauley and Rohay(34–Reference Ring-Dimitriou, von Duvillard, Paulweber, Stadlmann, Lemura, Peak and Mueller36). These discrepancies probably reflect the differences in the literature concerning intensity of the physical activity or energy expenditure due to physical activity. Many investigators agree that an exercise threshold needs to be met before favourable changes in HDL-C can occurReference Hardman(13, Reference Marrugat, Elosua, Covas, Molina and Rubies-Prat17, Reference Kokkinos and Fernhall24, Reference Kodama, Tanaka and Saito27, Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose33). For most individuals, this threshold is associated with a physical activity-related energy expenditure of 5·02 MJ (1200 kcal)Reference Durstine, Grandjean, Cox and Thompson(12) or 8 MJ or moreReference Hardman(13) per week, or an intensity of 29·3 kJ/min (7 kcal/min)Reference Marrugat, Elosua, Covas, Molina and Rubies-Prat(17), or 5–6 METs (metabolic energy equivalent tasks)Reference Lakka and Salonen(15).
Concerning the effect of exercise on TC and LDL-cholesterol (LDL-C) there is inconclusive evidence in both epidemiological and training studies. A careful evaluation of the literature indicates that exercise training seldom alters TC and LDL-CReference Durstine, Grandjean, Cox and Thompson(12, Reference Kraus, Houmard and Duscha14). However, there are a few exceptions. In a meta-analysis of randomised controlled trials, it was concluded that walking results in decreases in LDL-CReference Kelley, Kelley and Tran(37).
Obviously it is not completely clear whether in a normal population it should be advised to augment energy expenditure by high-intensity activities (sport) or by being more physically active in general, in order to obtain a healthier lipid profile. Moreover, relatively few studies have included women and their conclusions are less consistent than in men.
Therefore, the purpose of the current study was to investigate the lipid profile in a general population of Flemish males and females differentiated according to sports-related energy expenditure, energy expenditure during active leisure time and total energy expenditure.
Methods
Subjects
The subjects included in the present study originate from the Policy Research Centre Sport, Physical Activity and Health. This Policy Research Centre was established to investigate physical fitness, physical activity and sports participation in relation to health in the Flemish population. In order to fulfil this aim, almost 6000 subjects between 18 and 75 years of age were tested in their home town (Theme 1), of which almost 2000 subjects also completed an extensive test battery in the central laboratory of the Policy Research Centre (Theme 2).
For the present study, the data of subjects originating from Theme 2 were used. Of the total sample of almost 2000 subjects, 1632 had complete data on blood lipids, dietary intakes and physical activity habits. It was decided to exclude students, pensioners older than 75 years of age and those who took cholesterol-lowering medication. Thus, a total of 1170 subjects (672 men, mean age 46·30 (sd 10·67) years and 498 women, mean age 44·38 (sd 8·98) years) were included in the present study.
Assessment of physical activity
Participation in sport and physical activity was investigated using the Flemish Physical Activity Computerised Questionnaire (FPACQ), which is extensively described by Matton et al.(38).
In this questionnaire subjects were asked to select a maximum of three of their most important sports out of a list of 196 specific sports. To calculate the hours of health-related sports participation, the sum was made of the hours per week they spent on these three sports. This value could amount to a maximum of 30 h. Sports participation was considered as health-related sports participation if the associated MET valueReference Ainsworth, Haskell and Whitt(39) of the sport was at least 4·5, 4·0 and 3·5 respectively for subjects younger than 35, between 35 and 49 and older than 50 years of age. For each sport, the MET value was multiplied by the time spent on this sport. The sum of these three multiplications results in the total energy expenditure during health-related sports participation and was called EESPORT (MET × h/week).
Furthermore, subjects were asked about time spent on active transportation (walking and cycling) in leisure time and time spent on light, moderate and vigorous household and garden activities. For each active leisure-time activity, the MET value was multiplied by the time spent on this activity. The sum of these multiplications plus the energy expenditure during health-related sports participation resulted in EEALT (MET × h/week), an indication of the overall energy expenditure of active leisure-time activities.
Finally, total energy expenditure, EETOTAL (MET × h/week), was calculated by summing the energy expenditure of all activities during a usual week.
Assessment of dietary intake
Dietary habits and energy intake were assessed by means of the 3 d (two weekdays and one weekend day) food record method. The subjects were asked to write down all the food and drinks they consumed during those three days. The detailed list of dietary products was processed with the BECEL computer program version 5·0 (Hartog-Union and Van den Bergh, Rotterdam, The Netherlands). To control for confounding factors the relationship between dietary intake and lipid profile was analysed by correlation and regression analyses (results not shown). Based on these analyses it was decided to use alcohol consumption as covariate.
Lipid analysis
Fasting blood samples (after an overnight fast) were taken for the measurement of TC, TAG, HDL-C, LDL-C and TC:HDL-C ratio. The blood lipids and lipoproteins were measured on an Olympus AU5400 analyser (Olympus Diagnostica, Hamburg, Germany). The CV between days were 2·0 % (at 7·02 mmol/l) for TC, 1·4 % (at 2·07 mmol/l) for TAG, 3·9 % (at 2·04 mmol/l) for HDL-C and 3·4 % (at 4·02 mmol/l) for LDL-C.
Statistical analysis
The subjects were grouped by sex, age (<45 years, ≥45 to <55 years and ≥55 years), EESPORT, EEALT and EETOTAL (<25th percentile v. >75th percentile). Descriptive statistics (mean, sd, P25 and P75) were calculated for the anthropometric characteristics, physical activity variables and the serum lipid and lipoprotein values. Group differences were analysed by analyses of covariance with BMI and alcohol consumption as covariates since prior analyses (correlation and regression analyses) showed the importance of these factors (results not shown).
All statistical analyses were performed using the Statistical Analysis Systems statistical software package version 9·1 (SAS Institute, Cary, NC, USA). Statistical significance was set at α = 0·05.
Results
Descriptive statistics of the anthropometric characteristics, physical activity-related variables, and the serum lipid and lipoprotein values are presented in Tables 1 and 2 for men and women respectively.
Table 1 Descriptive statistics (mean, sd, 25th percentile (P25) and 75th percentile (P75)) of the anthropometric characteristics, physical activity variables and serum lipid and lipoprotein values among men
EESPORT, energy expenditure during health-related sports participation; EEALT, energy expenditure during active leisure time; EETOTAL, total energy expenditure; TC, total cholesterol; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol.
Table 2 Descriptive statistics (mean, sd, 25th percentile (P25) and 75th percentile (P75)) of the anthropometric characteristics, physical activity variables and serum lipid and lipoprotein values among women
EESPORT, energy expenditure during health-related sports participation; EEALT, energy expenditure during active leisure time; EETOTAL, total energy expenditure; TC, total cholesterol; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol.
Analyses of covariance in the younger age group differentiated according to sex and level of energy expenditure (EESPORT, EEALT or EETOTAL) revealed significant differences in lipid profile, all in favour of the most active groups (Tables 3 and 4).
Table 3 Comparison of serum lipid and lipoprotein values between men being part of the lowest and highest quartile of energy spent during health-related sports participation, energy spent during active leisure time and total energy spent during a week: analyses of covariance with BMI and alcohol consumption as covariates
EESPORT, energy expenditure during health-related sports participation; EEALT, energy expenditure during active leisure time; EETOTAL, total energy expenditure; LSMean, least-squares mean; TC, total cholesterol; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol.
Table 4 Comparison of serum lipid and lipoprotein values between women being part of the lowest and highest quartile of energy spent during health-related sports participation, energy spent during active leisure time and total energy spent during a week: analyses of covariance with BMI and alcohol consumption as covariates
EESPORT, energy expenditure during health-related sports participation; EEALT, energy expenditure during active leisure time; EETOTAL, total energy expenditure; LSMean, least-squares mean; TC, total cholesterol; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol.
Concerning the comparison of young men with different levels of EESPORT (<P25 v. >P75), significant results were found for HDL-C (1·37 (sd 0·30) mmol/l v. 1·49 (sd 0·30) mmol/l) and TC:HDL-C (3·96 (sd 0·88) v. 3·54 (sd 0·88); Table 3). For the young women, significant results were observed for HDL-C (1·65 (sd 0·35) mmol/l v. 1·76 (sd 0·34) mmol/l), LDL-C (3·09 (sd 0·78) mmol/l v. 2·76 (sd 0·78) mmol/l) and TC:HDL-C (3·29 (sd 0·78) v. 2·95 (sd 0·71); Table 4).
Differentiation according to EEALT and EETOTAL in young men revealed significant differences for TAG (respectively 1·20 (sd 0·49) mmol/l v. 1·04 (sd 0·49) mmol/l and 1·22 (sd 0·52) mmol/l v. 1·04 (sd 0·51) mmol/l), HDL-C (respectively 1·30 (sd 0·25) mmol/l v. 1·50 (sd 0·25) mmol/l and 1·31 (sd 0·27) mmol/l v. 1·48 (sd 0·27) mmol/l) and TC:HDL-C (respectively 4·09 (sd 0·79) v. 3·53 (sd 0·79) and 4·07 (sd 0·88) v. 3·58 (sd 0·88); Table 3). In young women no significant differences could be observed.
In the middle-aged group (≥45 to <55 years) differentiated according to sex and level of energy expenditure (EESPORT, EEALT, EETOTAL), no significant differences were observed except for HDL-C in men grouped by EESPORT (1·39 (sd 0·30) mmol/l v. 1·53 (sd 0·29) mmol/l; Table 3).
Comparison of the older age group (≥55 years) differentiated according to sex and level of energy expenditure (EESPORT, EEALT, EETOTAL) did not reveal any significant difference except for TAG in men grouped by EEALT (1·51 (sd 0·55) mmol/l v. 1·23 (sd 0·55) mmol/l; Table 3).
Discussion
There is consolidated evidence that physical activity exerts beneficial effects on lipid profiles. Former studies focused on the importance of the intensity of physical activity whereas more recently focus has shifted towards energy expenditure. Durstine et al.Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33) concluded in their review that the effects of exercise on lipid and lipoprotein levels can be initiated at low training volumes and will continue in a dose–response fashion with increasing training volume. However, results are not equivocal in different age groups. Moreover, studies focusing on women often show conflicting results and are rather sparse.
Obviously debate continues regarding the effect of physical activity. The purpose of the present study was to investigate the importance of energy expenditure on the lipid profile in a general population of Flemish males and females of different age groups. More specifically, it was investigated whether energy expenditure in general or sports-related energy expenditure contributes to a healthier lipid profile.
The results of this study indicate that young women (<45 years) of the highest quartile of energy expenditure during health-related sports participation have healthier values for HDL-C, LDL-C and TC:HDL-C than their inactive counterparts. It was expected to find differences regarding HDL-CReference Twisk, Kemper, Mellenbergh, van Mechelen and Post(6, Reference Young, Haskell, Jatulis and Fortmann7, Reference Kokkinos and Fernhall24, Reference Leon and Sanchez25, Reference Kodama, Tanaka and Saito27–Reference Fransson, Alfredsson, de Faire, Knutsson and Westerholm29, Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose33, Reference Skoumas, Pitsavos, Panagiotakos, Chrysohoou, Zeimbekis, Papaioannou, Toutouza, Toutouzas and Stefanadis40), but finding differences in LDL-C was rather surprising since there is little support in the cross-sectional literature for significant differences in LDL-C between active and inactive groups. Even in training studies significant changes in LDL-C are generally not observed. However, there are a few exceptions: most frequently in studies with exercise training programmes in which participants expended more than 5·02 MJ/week (1200 kcal/week)Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33). Analysing the energy expenditure due to health-related sports participation of the women younger than 45 years of age, it was observed that none of the subjects of the group <P25 and seventy-six subjects of the seventy-eight subjects of the group >P75 had an EESPORT above 4·18 MJ/week (1000 kcal/week). Of these seventy-six subjects, sixty-four women had an EESPORT above 5·02 MJ/week (1200 kcal/week). It is clear that the energy expenditure due to sports participation in the most active young women of the present study is comparable with the energy expenditure in training studies in which significant effects of physical activity on the lipid profile were observed.
When the women younger than 45 years of age were differentiated according to energy spent during active leisure time or overall energy expenditure during a normal week, significant differences were no longer observed. Weller and CoreyReference Weller and Corey(41) showed that physical activity was inversely associated with risk of death in women and pointed out that the contribution of non-leisure (household chores) energy expenditure represented, on average, 82 % of women’s total activity. They argued that when studying the effect of physical activity on CVD in women non-leisure energy expenditure should be taken into account. The results of the present study suggest that sports-related energy expenditure, rather than active leisure time-related energy expenditure in which household activities are included, results in a healthy lipid profile. This conclusion is supported by the findings of O’Connor et al.Reference O’Connor, Hennekens, Willett, Goldhaber, Paffenbarger, Breslow, Lee and Buring(42), who stated that moderate to vigorous sporting activity was directly related to high HDL-C whereas total energy expenditure was uncorrelated with blood lipids. To illustrate: in the present study, less than half of the young women who belonged to the group >P75 differentiated according to energy spent during active leisure time had an EESPORT above 5·02 MJ/week (1200 kcal/week). So it seems that a minimum intensity threshold has to be met.
In young men differentiated according to energy expenditure during health-related sports participation significant results were observed for HDL-C and TC:HDL-C. No significant results were observed for TAG. This finding was unexpected since it was argued by Durstine et al.Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33) that both HDL-C and TAG levels can be changed by similar training volumes. They concluded that training volumes that elicit energy expenditures ≥5·02 MJ/week (≥1200 kcal/week) are associated with elevations in HDL-C levels and reduced TAG levels. When the men younger than 45 years of age were differentiated according to EEALT or EETOTAL, significant differences were observed for HDL-C, TC:HDL-C and also for TAG.
Analysing EESPORT in the group >P75 differentiated according to EESPORT, it was observed that all seventy-eight subjects had values exceeding the threshold of 5·02 MJ/week (1200 kcal/week). When differentiating according to EEALT or EETOTAL most of the subjects belonging to the group >P75 still had EESPORT values above 5·02 MJ/week (1200 kcal/week; this in contrast to the young women). However, when looking at the mean and sd of energy spent during active leisure time of the group <P25 differentiated according to EEALT, it becomes clear that these subjects were much less active [8·35 (sd 3·09) MJ/week (1996 (sd 739) kcal/week)] than the subjects belonging to the group <P25 differentiated according to EESPORT [14·69 (sd 9·47) MJ/week (3511 (sd 2263) kcal/week)]. For the group >P75, when differentiating according to EEALT higher values were observed for EEALT than for EESPORT when differentiating according to EESPORT [43·59 (sd 15·67) MJ/week (10 419 (sd 3745) kcal/week), with a minimum value of 24·33 MJ/week (5815 kcal/week) v. 39·05 (sd 17·98) MJ/week (9333 (sd 4297) kcal/week), with a minimum value of 15·89 MJ/week (3798 kcal/week)]. In contrast to the young women it seems that differentiating young men according to EEALT or EETOTAL results in more contrasting groups than when differentiating according to EESPORT.
In the middle-aged (≥45 to <55 years) and older age groups (≥55 years) no significant differences were observed except for HDL-C in middle-aged men grouped by EESPORT and TAG in men older than 55 years of age differentiated according to EEALT. However, several studies have shown the beneficial effect of physical activity on the lipid profile in middle-aged and older adultsReference Schubert, Rogers, Remsberg, Sun, Chumlea, Demerath, Czerwinski, Towne and Siervogel(8, Reference Arquer, Elosua, Covas, Molina and Marrugat10, Reference Lakka and Salonen15, Reference Pescatello, Murphy and Costanzo21, Reference Bijnen, Feskens, Caspersen, Giampaoli, Nissinen, Menotti, Mosterd and Kromhout22, Reference Sunami, Motoyama, Kinoshita, Mizooka, Sueta, Matsunaga, Sasaki, Tanaka and Shindo26, Reference Kelley, Kelley and Tran37, Reference Ainslie, Reilly, Maclaren and Campbell43). It has already been suggested that energy expenditure has to be above a certain thresholdReference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33). Fonong et al.Reference Fonong, Toth, Ades, Katzel, Calles-Escandon and Poehlman(35) reported that short-time exercise training, generating less than 3·77 MJ/week (900 kcal/week) in exercise energy expenditure, fails to influence HDL-C levels in healthy older men and women. In the present study, when participants of the middle-aged and older age groups were differentiated according to EESPORT, none of the subjects belonging to the group <P25 and all of the subjects belonging to the group >P75 had an EESPORT exceeding 5·02 MJ/week (1200 kcal/week). It seems that the threshold suggested by Durstine et al.Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33) does not elicit positive effects on the lipid profile in the present study sample older than 45 years of age.
It can be postulated that also an intensity threshold has to be met to expect significant effects on blood lipids. For most individuals, this threshold is associated with 5–6 METsReference Lakka and Salonen(15, Reference Marrugat, Elosua, Covas, Molina and Rubies-Prat17). Regarding this finding it can be argued that the criterion used to consider sports with an intensity of at least 3·5–4·0 METs (walking) as health-related sports participation in the older and middle-aged groups was rather low. However, former studies have shown that even walking has beneficial effects on the lipid profile and lowers the risk of coronary events, especially in older participantsReference Kelley, Kelley and Tran(37, Reference Ainslie, Reilly, Maclaren and Campbell43, Reference Manson, Hu, Rich-Edwards, Colditz, Stampfer, Willett, Speizer and Hennekens44).
It should be mentioned that also in the present study BMI and alcohol consumption were observed to be important determinants of the lipid profile. It has previously been reported that greater weekly intakes of alcohol were associated with significantly higher HDL-C in both men and womenReference Williams(45). Thus, alcohol is a potentially important confounder and studies which fail to take this factor into account may overestimate the contribution of physical activity to a healthy lipid profile. The same is true for BMI. A recent study showed that high BMI was more strongly related to adverse cardiovascular biomarker levels than physical inactivity. However, within BMI categories, physical activity was generally associated with more favourable cardiovascular biomarker levels than inactivityReference Mora, Lee, Buring and Ridker(46). It is clear that studies focusing on the effect of physical activity on the lipid profile should take these determinants into account.
It can be concluded that young men and women with higher levels of energy expenditure due to sports activities show a better lipid profile than their sedentary counterparts. When differentiating subjects according to EEALT or EETOTAL only for the men could a healthier lipid profile be observed in favour of the most active subjects. It seems that, in the present study, taking household and garden activities into account to differentiate groups obscures differences in lipid profiles in women and amplifies differences in lipid profiles in men. For men and women younger than 45 years of age the conclusion of Durstine et al.Reference Durstine, Grandjean, Davis, Ferguson, Alderson and DuBose(33), that training volumes eliciting energy expenditures ≥5·02 MJ/week (≥1200 kcal/week) contribute to a healthy lipid profile, can be confirmed.
In the middle-aged and older age groups differentiated according to energy expenditure (EESPORT, EEALT or EETOTAL), no differences could be observed in lipid profile except for HDL-C in middle-aged men grouped by EESPORT and TAG in older men grouped by EEALT.
Acknowledgements
The Policy Research Centre Sport, Physical Activity and Health was supported by the Flemish government.
T.S. was involved in development and design of the study and wrote the manuscript. L.V.L. performed the statistical analyses and interpretation and was a significant manuscript reviewer. R.P., W.D. and M.T. contributed to the concept and design of the study. N.D., L.M. and K.W. participated in the data acquisition. J.L. was involved in development and design of the study and provided statistical expertise. All authors assisted in revising the manuscript. None of the authors have any interests that might be interpreted as influencing the research.
