Intensity-Weighted Physical Activity Volume and Risk of All-Cause and Cardiovascular Mortality: Does the Use of Absolute or Corrected Intensity Matter?

in Journal of Physical Activity and Health

Background: Previous epidemiological studies examining the association between physical activity (PA) and mortality risk have measured absolute PA intensity using standard resting metabolic rate reference values that fail to consider individual differences. This study compared the risk of all-cause and cardiovascular mortality between absolute and corrected estimates of PA volume. Methods: 49,982 adults aged ≥40 years who participated in the Health Survey for England and Scottish Health Survey in 1994–2008 were included in our study. PA was classified as absolute or corrected metabolic equivalent (MET)-hours per week, taking participant’s weight, height, age, and sex into account. Cox regression models were used to examine the association between absolute and corrected PA volumes and all-cause and cardiovascular mortality. Results: The authors found no difference in the association between levels of PA and risk of all-cause and cardiovascular mortality for absolute and corrected MET-hours per week, although there was a consistent decrease in mortality risk with increasing PA. There was no difference in mortality when analyses were stratified by sex, age, and body mass index. Conclusions: The association between PA volume and risk of mortality was similar regardless of whether PA volume was estimated using absolute or corrected METs. There is no empirical justification against the use of absolute METs to estimate PA volume from questionnaires.

Engaging in regular physical activity (PA) improves cardiometabolic health and reduces the risk of chronic disease and mortality.1,2 Over recent decades, PA guidelines have evolved from a primary focus on aerobic exercise to a multimodal approach incorporating strength-promoting exercise and other activities (eg, balance and flexibility training) as an adjunct to aerobic exercise.1,3 Performing regular bouts of vigorous-intensity PA is important for optimal health benefits.4

A common measure of PA intensity involves metabolic equivalents (METs), defined as multiples of resting metabolic rate (RMR).5 PA intensity is classified as light (1.6 to <3.0 METs), moderate (3.0 to <6.0 METs), and vigorous (≥6.0 METs).5 The standard 1-MET reference value for RMR is 3.5 mL·O2·kg−1·min−1 and most likely traces back to a medical textbook from 1890 that published the data of a single 70-kg, 40-year-old man.6,7 Multiple studies that have reevaluated RMR in large heterogeneous samples report substantially lower mean RMR of 2.51 to 3.30,811 questioning the use of 3.5 mL·O2·kg−1·min−1 as the standard 1-MET reference.

Generally, absolute METs measure the total amount of energy expended during PA, without considering individual characteristics,1 whereas METs corrected for RMR take into account variation in individual characteristics, such as age, gender, height, and weight when calculating the amount of energy needed to complete a given PA task.9 The Compendium of Physical Activities provides a comprehensive list of absolute METs that determines intensity-weighted PA volume.5 However, the use of absolute METs may underestimate the true effort needed by overweight and obese, older, and female persons to perform PA, the subgroups of the population that are the least likely to meet PA guidelines.9,11 This can lead to considerable misclassification of PA intensity as individuals from these subgroups may engage in more vigorous PA than estimated from absolute METs.9,11 Corrected METs may offer a more accurate estimate of intensity-weighted PA volume.

Epidemiological studies examining links between PA and mortality risk have predominantly used absolute measures of PA intensity,1215 which may lead to PA level misclassification and potential attenuation of the observed associations. Although one study considered the relative intensity of PA using a generic question about perceived effort during sports, they did not measure PA volume.16 To our knowledge, no study has compared mortality risk estimates derived from absolute and corrected METs. The aim of this study was to compare these 2 intensity-weighted PA volume estimation methods (absolute vs corrected METs) on the direction and magnitude of PA–mortality associations using an established pooled data set of population cohorts.15

Methods

Participants

We used data collected from the Health Survey for England (HSE) and the Scottish Health Survey (SHS). The HSE and SHS are household-based population surveillance studies in which a multistage, stratified probability design was used to select households that are representative of the target populations of England and Scotland, respectively.17,18 The HSE surveys were conducted in 1994, 1997, 1998, 1999, 2003, 2004, 2006, and 2008, and the SHS in 1995, 1998, and 2003. Interview response rates ranged from 58% to 71% for the HSE17 and 60% to 81% for the SHS.18 All participants provided written informed consent; but, for this study, only adults aged ≥40 years were included. Ethical approval was granted by relevant local research ethics committees.

Mortality Outcomes

The primary outcome was risk of all-cause and cardiovascular disease (CVD) mortality. Surviving participants were censored on December 31, 2009 (SHS) and March 31, 2011 (HSE). Primary cause of death was diagnosed and coded according to the International Classification of Diseases and Related Health Problems, Ninth Revision (ICD-9) or Tenth Revision (ICD-10). CVD deaths were recorded using the ICD-9 codes: 390.0 to 459.9 and ICD-10 codes: I01 to I99. Length of survival was calculated based on the time to death or censoring at the end of the study period. Participants with physician-diagnosed CVD or cancer at baseline, and/or who died during the first 24 months of follow-up, were excluded in order to reduce the possibility of reverse causation.

PA Assessment

Nonoccupational PA was assessed using an established questionnaire19 that asked participants to report the frequency and duration of their participation in 3 categories of PA in the 4 weeks prior to the interview: domestic PA (eg, housework and gardening), walking, and sports and exercise. We calculated the absolute MET-hours per week of each participant by multiplying the absolute MET per activity with the number of hours the activity was performed per week, using the standard 1-MET reference of 3.5 mL·O2·kg−1·min−1. Corrected MET-hours per week was calculated using the same method, except we used the corrected 1-MET reference from the Harris–Benedict prediction equation for RMR20 rather than the standard 1-MET reference. Previous work reported that the Harris–Benedict equation is a valid measure of energy expenditure with a minimal difference of 391 kcal/d compared with gas exchange values using indirect calorimetry.21

CorrectedMETintensity=mean absoluteMETper activity×(3.5/Harris–Benedict predictedRMR),
where11 Harris–Benedict predicted RMR = 10 × weight (kg) + 6.25 × height (cm) − 5 × age (years) + 5 for males and 10 × weight (kg) + 6.25 × height (cm) − 5 × age (years) − 161 for females.20

We grouped participants by MET-hours per week into 4 categories based on recent PA guidelines as follows1: inactive (0 MET-h/wk), insufficiently active (<7.5 MET-h/wk, excluding 0 MET-h/wk), sufficiently active (7.5 to <15 MET-h/wk), and highly active (≥15 MET-h/wk). Domestic-related PA was excluded from calculations of total MET-hours per week because its recall is often imprecise due to its incidental nature and, generally, it is not associated with CVD mortality22 or risk factors.23 In a sensitivity analysis, participants were grouped into inactive (0 MET-h/wk) or tertiles of PA according to their total MET-hours per week. Participants whose total MET-hours per week were >5 SDs above the mean were treated as outliers and excluded from all analyses, similar to previous research.24

Covariates

Trained interviewers measured height and weight using standard protocols17,18 to calculate body mass index (BMI). In stratified analyses, we classified a BMI score <25 as indicative of underweight or normal weight, 25 to <30 as overweight, and ≥30 as obese. Additional survey items assessed age, sex, presence of long-standing physical or mental illness, and age-finished full-time education. Smoking habits were categorized as current smoker or noncurrent smoker, and alcohol consumption was classified as drinking alcohol less than, or greater than or equal to, 5 times per week. Scores from the 12-item General Health Questionnaire were used to assess psychological distress with a General Health Questionnaire score ≥4 indicating psychological distress.25

Statistical Analysis

Physical activity levels and participant characteristics were described in relation to absolute and corrected METs. We used Cox proportional hazards regression models to examine associations between PA and risk of all-cause and CVD mortality using absolute and corrected METs. Model 1 was partially adjusted for age and sex. Model 2 was fully adjusted, including age, sex, BMI, long-standing illness, smoking habits, alcohol consumption, psychological distress, and education level. We also ran additional fully adjusted models stratified by sex, age, and BMI. We generated receiver operating characteristic (ROC) curves with corresponding areas under the curve to examine whether corrected MET-hours per week is a better predictor of mortality outcomes than absolute MET-hours per week. We examined the proportional hazards assumption through Kaplan–Meier plots, and no apparent violations were evident. Statistical analyses were conducted using the Statistical Package for the Social Sciences for Windows (version 22.0; IBM, Chicago, IL).

Results

The core analytic sample included 49,982 adults from England and Scotland aged ≥40 years and was predominately white (93.5%). Baseline characteristics are presented in Table 1. Over an average follow-up time of 9.1 (4.5) years (corresponding to 452,335 person-years), 5227 all-cause and 1513 CVD deaths were recorded. When shifting from absolute to corrected METs, 6.2% fewer participants were classified as insufficiently active, 1.1% fewer participants were classified as sufficiently active, and 7.3% more participants were classified as highly active. Supplementary Figure 1 (available online) presents the distribution of the absolute and corrected MET-hours per week in the analytic sample.

Table 1

Baseline Characteristics of Participantsa

Physical activity levelb
Absolute MET-h/wkCorrected MET-h/wk
CharacteristicInactive (n = 10,162)Insufficiently active (n = 17,169)Sufficiently active (n = 8747)Highly active (n = 13,904)Inactive (n = 10,163)Insufficiently active (n = 14,086)Sufficiently active (n = 8179)Highly active (n = 17,554)
Age,c y59.3 (12.7)56.5 (11.7)55.7 (11.1)54.6 (10.6)59.3 (12.7)56.3 (11.7)55.6 (11.2)55.2 (10.8)
Female sex5882 (57.9)10,000 (58.2)5042 (57.6)6801 (48.9)5883 (57.9)7956 (56.5)4705 (57.5)9181 (52.3)
White ethnicity9499 (93.5)15,958 (93.0)8158 (93.3)13,099 (94.3)9500 (93.5)13,071 (92.9)7633 (93.4)16,510 (94.2)
Body mass indexc,d28.0 (5.3)27.6 (4.8)27.1 (4.4)26.7 (4.2)28.0 (5.3)27.4 (4.8)27.2 (4.5)27.0 (4.3)
Long-standing illness5982 (58.9)8443 (49.2)3855 (44.1)5627 (40.5)5980 (58.8)6930 (49.2)3665 (44.8)7332 (41.8)
Current smoker3095 (30.5)3976 (23.2)1728 (19.8)2438 (17.5)3094 (30.4)3368 (23.9)1729 (21.1)3046 (17.4)
Frequent alcohol consumption (≥5 d/wk)1888 (18.6)3725 (21.7)1905 (21.8)3444 (24.8)1887 (18.6)3110 (22.1)1757 (21.5)4208 (24.0)
Psychological distress (GHQ score ≥4)1870 (18.4)2297 (13.4)945 (10.8)1317 (9.5)1871 (18.4)1908 (13.5)967 (11.8)1683 (9.6)
Finished full-time education at age ≥19 y920 (9.1)2716 (15.8)1635 (18.7)3022 (21.7)921 (9.1)2218 (15.7)1484 (18.1)3670 (20.9)
Total MET-h/wkeN/A3.0 (3.5)10.5 (3.6)25.6 (19.2)N/A3.3 (3.5)10.9 (3.7)29.3 (23.1)

Abbreviations: GHQ, General Health Questionnaire; MET, metabolic equivalent; N/A, not applicable; IQR, interquartile range.

aSample includes 11 cohorts of 49,982 adults aged ≥40 years who responded to the Health Survey for England or Scottish Health Survey. bPhysical activity levels were classified into 4 categories: inactive (0 MET-h/wk); insufficiently active (<7.5 MET-h/wk, but not inactive); sufficiently active (7.5 to <15 MET-h/wk); and highly active (≥15 MET-h/wk). cValues are expressed as mean (SD). dBody mass index = weight (in kilograms)/height (in meters squared). eValues are expressed as median (IQR).

Tables 2 and 3 show the partially and fully adjusted multivariate associations between levels of PA and all-cause and CVD mortality risk, respectively. In both models, using absolute and corrected METs, compared with the inactive participants, participation in any amount of PA was associated with a significantly reduced risk of all-cause and CVD mortality, although there was a slight attenuation in HR estimates for fully adjusted models. The pattern of risk reduction in all-cause and CVD mortality between the 4 PA levels did not differ between the use of absolute or corrected METs. In a sensitivity analysis where we replicated these Cox regression models using insufficiently active as the reference group, no appreciable differences were found (data not shown).

Table 2

Associations Between Physical Activity and Risk of All-Cause Mortality Using Absolute and Corrected MET-Hours Per Weeka

Model 1cModel 2d
Physical activity levelbNo. of deathsnHR95% CIHR95% CI
Absolute MET-h/wk
 Inactive185010,1621.00 (reference)N/A1.00 (reference)N/A
 Insufficiently active175117,1690.700.65–0.750.770.72–0.83
 Sufficiently active72887470.630.58–0.690.730.67–0.80
 Highly active89813,9040.540.50–0.580.640.59–0.70
 Total522749,982N/AN/AN/AN/A
Corrected MET-h/wk
 Inactive185110,1631.00 (reference)N/A1.00 (reference)N/A
 Insufficiently active146114,0860.710.67–0.760.780.73–0.84
 Sufficiently active70381790.640.59–0.700.730.67–0.79
 Highly active121217,5540.560.52–0.600.660.61–0.71
 Total522749,982N/AN/AN/AN/A

Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio; MET, metabolic equivalent; N/A, not applicable.

aSample includes 11 cohorts of 49,982 adults aged ≥40 years who responded to the Health Survey for England or Scottish Health Survey. bPhysical activity levels were classified into 4 categories: inactive (0 MET-h/wk); insufficiently active (<7.5 MET-h/wk, but not inactive); sufficiently active (7.5 to <15 MET-h/wk); and highly active (≥15 MET-h/wk). cAdjusted for age and sex. dAdjusted for age, sex, BMI, long-standing illness, alcohol consumption, smoking, psychological distress, and education.

Table 3

Associations Between Physical Activity and Risk of Cardiovascular Mortality Using Absolute and Corrected MET-Hours Per Weeka

Model 1cModel 2d
Physical activity levelbNo. of deathsnHR95% CIHR95% CI
Absolute MET-h/wk
 Inactive54610,1621.00 (reference)N/A1.00 (reference)N/A
 Insufficiently active49417,1690.680.61–0.770.770.68–0.87
 Sufficiently active21787470.660.56–0.770.790.67–0.92
 Highly active25613,9040.550.47–0.640.680.58–0.80
 Total151349,982N/AN/AN/AN/A
Corrected MET-h/wk
 Inactive54710,1631.00 (reference)N/A1.00 (reference)N/A
 Insufficiently active40714,0860.690.60–0.780.770.68–0.88
 Sufficiently active20381790.650.55–0.770.760.64–0.89
 Highly active35617,5540.580.50–0.660.710.62–0.82
 Total151349,982N/AN/AN/AN/A

Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio; N/A, not applicable.

aSample includes 11 cohorts of 49,982 adults aged ≥40 years who responded to the Health Survey for England or Scottish Health Survey. bPhysical activity levels were classified into 4 categories as follows: inactive (0 MET-h/wk), insufficiently active (<7.5 MET-h/wk, but not inactive), sufficiently active (7.5 to <15 MET-h/wk), and highly active (≥15 MET-h/wk). cAdjusted for age and sex. dAdjusted for age, sex, BMI, long-standing illness, alcohol consumption, smoking, psychological distress, and education.

Supplementary Tables 1 and 2 (available online) show the fully adjusted multivariate associations between PA levels and risk of all-cause and CVD mortality, respectively, stratified by sex, age, and BMI. Using corrected METs slightly attenuated the association between PA levels and all-cause mortality in adults aged >60 years and with obesity (BMI ≥ 30). We found modest differences in CVD mortality associations between absolute and corrected METs in certain population subgroups, including highly active females, sufficiently active adults aged 50–59 years, and overweight adults (BMI: 25 to <30) who were sufficiently or highly active. Overall, differences in stratified mortality risk estimates between absolute and corrected METs were not consistent in direction or magnitude. For results from the sensitivity analysis where we grouped participants into tertiles based on MET-hours per week, PA–mortality risks were also similar between absolute and corrected METs (Supplementary Tables 3 and 4 [available online]). The areas under the curve did not differ between absolute and corrected METs, indicating a comparable predictive ability for mortality outcomes (Supplementary Table 5 [available online]).

Discussion

To our knowledge, this study is the first to compare the impact of using absolute versus corrected METs on the direction and magnitude of PA–mortality associations. In accordance with previous studies using comparable large, heterogeneous cohorts, we found that any amount of PA was associated with a reduced risk of all-cause and CVD mortality.14,15 Our results showed that correcting intensity-weighted PA volume (MET-h/wk) for variations in RMR did not materially change PA–mortality associations, compared with using absolute MET-hours per week. We also did not find consistent differences in mortality risk estimates between absolute and corrected METs when stratified by the RMR correction parameters, sex, age, and BMI.

There are several explanations for our results. One possibility is that the use of absolute METs led to little misclassification in this sample compared with the corrected METs. Supplementary Figure 1 (available online) shows a similar frequency distribution of absolute and corrected MET-hours per week, indicating that correcting MET-hours per week using the Harris–Benedict equation led to the minimal reclassification of the sample. In particular, the population subgroups most likely to have PA intensity misrepresented by absolute METs (overweight and obese, older and female persons)9,11 are also more likely to be inactive and may not have therefore been reclassified from correcting MET-hours per week. Furthermore, an absence of differences in mortality rates at the group level does not necessarily imply that there is no impact of using absolute versus corrected MET-hours per week at the individual level. Our findings might simply show that there are no differences at the group level.

Previous research also suggested that the Harris–Benedict equation may overestimate RMR in whites, relative to measured RMR.26,27 This potential overestimation of RMR may have attenuated the magnitude of difference in PA–mortality associations between absolute and corrected METs, possibly explaining, at least in part, the observed null difference. There is evidence that the Mifflin–St Jeor equation is more accurate at predicting RMR than the Harris–Benedict equation27 and offers an alternative prediction equation. Finally, the questionnaire used in our study to estimate PA inquires about bouts lasting for at least 10 to 15 minutes,28 a convention that contradicts the most recent US guidelines, which acknowledge that PA of any duration enhances health.1 Questionnaires cannot capture incidental PA of higher intensity, and the capacity of incidental PA to reach relative vigorous intensity is often underappreciated,29 especially when considering the relative intensity of PA in population subgroups with lower RMR.30 It is possible that estimating levels of incidental PA more accurately (eg, using accelerometers) could have shifted more of the distribution of corrected MET-hours per week and would have contributed to more pronounced differences in PA–mortality associations.

This is the first study, to our knowledge, to compare intensity-weighted PA volume corrected for RMR. Strengths of our study included the large, population-based sample and our comprehensive and data-driven analytic approach. We reduced the possibility of reverse causation by removing early deaths and participants with underlying diseases at baseline. Nonetheless, our study had a number of limitations. First, 93.5% of the sample was white, which limits the generalizability of our results to other racial groups. Second, the data sources HSE and SHS were cross-sectional surveys that only assessed PA at baseline. As a result, it is unknown whether levels of PA changed from baseline to follow-up. Third, PA was self-reported using a questionnaire of mostly leisure-time PA,15 which may have led to recall and social desirability bias.

In summary, we found no evidence of different associations between PA levels and risk of all-cause and CVD mortality between absolute and corrected METs. Future research should reexamine differences in mortality risk estimates derived from absolute METs using different correction methods, such as accelerometer or heart rate data.

Acknowledgments

The authors are thankful for Dr Francisco Schneuer (Charles Perkins Centre, Faculty of Medicine and Health, University of Sydney) who helped with the preparation of the histogram figure. This study was supported by a summer research scholarship awarded to J.M. from the Charles Perkins Centre, University of Sydney as well as a senior research fellowship for Prof E.S. from the National Health and Medical Research Council (NHMRC).

References

  • 1.

    Piercy KL, Troiano RP, Ballard RM, et al. The physical activity guidelines for Americans. JAMA. 2018;320(19):2020–2028. PubMed ID: 30418471 doi:10.1001/jama.2018.14854

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2.

    Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT; Lancet Physical Activity Series Working Group. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet. 2012;380(9838):219–229. PubMed ID: 22818936 doi:10.1016/S0140-6736(12)61031-9

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3.

    Garber CE, Blissmer B, Deschenes MR, et al. American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Med Sci Sports Exerc. 2011;43(7):1334–1359. PubMed ID: 21694556 doi:10.1249/MSS.0b013e318213fefb

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4.

    Batacan R Jr, Duncan M, Dalbo VJ, Tucker PS, Fenning AS. Effects of high-intensity interval training on cardiometabolic health: a systematic review and meta-analysis of intervention studies. Br J Sports Med. 2016;51(6):494–503. PubMed ID: 27797726 doi:10.1136/bjsports-2015-095841

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5.

    Ainsworth BE, Haskell WL, Herrmann SD, et al. 2011 Compendium of Physical Activities: a second update of codes and MET values. Med Sci Sports Exerc. 2011;43(8):1575–1581. PubMed ID: 21681120 doi:10.1249/MSS.0b013e31821ece12

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6.

    Howley ET. You asked for it question authority. ACSM’s Health Fit J. 2000;4(1):6, 40.

  • 7.

    Wasserman K, Hansen JE, Sue DY, Stringer WW, Whipp BJ. Principles of exercise testing and interpretation: including pathophysiology and clinical applications. Med Sci Sports Exerc. 2005;37(7):1249. doi:10.1249/01.mss.0000172593.20181.14

    • Search Google Scholar
    • Export Citation
  • 8.

    Wilms B, Ernst B, Thurnheer M, Weisser B, Schultes B. Correction factors for the calculation of metabolic equivalents (MET) in overweight to extremely obese subjects. Int J Obes. 2014;38(11):1383–1387. doi:10.1038/ijo.2014.22

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9.

    Byrne NM, Hills AP, Hunter GR, Weinsier RL, Schutz Y. Metabolic equivalent: one size does not fit all. J Appl Physiol. 2005;99(3):1112–1119. doi:10.1152/japplphysiol.00023.2004

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10.

    Kwan M, Woo J, Kwok T. The standard oxygen consumption value equivalent to one metabolic equivalent (3.5 ml/min/kg) is not appropriate for elderly people. Int J Food Sci Nutr. 2004;55(3):179–182. PubMed ID: 15223593 doi:10.1080/09637480410001725201

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11.

    Kozey S, Lyden K, Staudenmayer J, Freedson P. Errors in MET estimates of physical activities using 3.5 ml × kg(−1) × min(−1) as the baseline oxygen consumption. J Phys Act Health. 2010;7(4):508–516. PubMed ID: 20683093 doi:10.1123/jpah.7.4.508

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Lear SA, Hu W, Rangarajan S, et al. The effect of physical activity on mortality and cardiovascular disease in 130 000 people from 17 high-income, middle-income, and low-income countries: the PURE study. Lancet. 2017;390(10113):2643–2654. PubMed ID: 28943267 doi:10.1016/S0140-6736(17)31634-3

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Oja P, Kelly P, Pedisic Z, et al. Associations of specific types of sports and exercise with all-cause and cardiovascular-disease mortality: a cohort study of 80 306 British adults. Br J Sports Med. 2017;51(10):812–817. PubMed ID: 27895075 doi:10.1136/bjsports-2016-096822

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14.

    Shiroma EJ, Sesso HD, Moorthy MV, Buring JE, Lee IM. Do moderate-intensity and vigorous-intensity physical activities reduce mortality rates to the same extent? J Am Heart Assoc. 2014;3(5):e000802. PubMed ID: 25326527 doi:10.1161/JAHA.114.000802

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 15.

    O’Donovan G, Lee IM, Hamer M, Stamatakis E. Association of “weekend warrior” and other leisure time physical activity patterns with risks for all-cause, cardiovascular disease, and cancer mortality. JAMA Intern Med. 2017;177(3):335–342. doi:10.1001/jamainternmed.2016.8014

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16.

    Lee IM, Sesso HD, Oguma Y, Paffenbarger RS Jr. Relative intensity of physical activity and risk of coronary heart disease. Circulation. 2003;107(8):1110–1116. PubMed ID: 12615787 doi:10.1161/01.CIR.0000052626.63602.58

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 17.

    Mindell J, Biddulph JP, Hirani V, et al. Cohort profile: the health survey for England. Int J Epidemiol. 2012;41(6):1585–1593. PubMed ID: 22253315 doi:10.1093/ije/dyr199

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 18.

    Gray L, Batty GD, Craig P, et al. Cohort profile: the Scottish health surveys cohort: linkage of study participants to routinely collected records for mortality, hospital discharge, cancer and offspring birth characteristics in three nationwide studies. Int J Epidemiol. 2010;39(2):345–350. PubMed ID: 19349480 doi:10.1093/ije/dyp155

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19.

    Scholes S, Coombs N, Pedisic Z, et al. Age- and sex-specific criterion validity of the health survey for England physical activity and sedentary behavior assessment questionnaire as compared with accelerometry. Am J Epidemiol. 2014;179(12):1493–1502. PubMed ID: 24863551 doi:10.1093/aje/kwu087

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20.

    Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO. A new predictive equation for resting energy expenditure in healthy individuals. Am J Clin Nutr. 1990;51(2):241–247. PubMed ID: 2305711 doi:10.1093/ajcn/51.2.241

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21.

    Lee SH, Kim EK. Accuracy of predictive equations for resting metabolic rates and daily energy expenditures of police officials doing shift work by type of work. Clin Nutr Res. 2012;1(1):66–77. PubMed ID: 23429979 doi:10.7762/cnr.2012.1.1.66

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 22.

    Stamatakis E, Hamer M, Lawlor DA. Physical activity, mortality, and cardiovascular disease: is domestic physical activity beneficial? The Scottish Health Survey—1995, 1998, and 2003. Am J Epidemiol. 2009;169(10):1191–1200. PubMed ID: 19329529 doi:10.1093/aje/kwp042

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23.

    Stamatakis E, Hillsdon M, Primatesta P. Domestic physical activity in relationship to multiple CVD risk factors. Am J Prev Med. 2007;32:320–327.e3. PubMed ID: 17383563 doi:10.1016/j.amepre.2006.12.020

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24.

    Perreault K, Bauman A, Johnson N, Britton A, Rangul V, Stamatakis E. Does physical activity moderate the association between alcohol drinking and all-cause, cancer and cardiovascular diseases mortality? A pooled analysis of eight British population cohorts. Br J Sports Med. 2017;51(8):651–657. PubMed ID: 27581162 doi:10.1136/bjsports-2016-096194

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25.

    Russ TC, Stamatakis E, Hamer M, Starr JM, Kivimäki M, Batty GD. Association between psychological distress and mortality: individual participant pooled analysis of 10 prospective cohort studies. BMJ. 2012;345:e4933. PubMed ID: 22849956 doi:10.1136/bmj.e4933

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26.

    Douglas CC, Lawrence JC, Bush NC, Oster RA, Gower BA, Darnell BE. Ability of the Harris Benedict formula to predict energy requirements differs with weight history and ethnicity. Nutr Res. 2007;27(4):194–199. PubMed ID: 19081830 doi:10.1016/j.nutres.2007.01.016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 27.

    Frankenfield DC. Bias and accuracy of resting metabolic rate equations in non-obese and obese adults. Clin Nutr. 2013;32(6):976–982. PubMed ID: 23631843 doi:10.1016/j.clnu.2013.03.022

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28.

    Stamatakis E, Ekelund U, Wareham NJ. Temporal trends in physical activity in England: the Health Survey for England 1991 to 2004. Prev Med. 2007;45(6):416–423. PubMed ID: 17316777 doi:10.1016/j.ypmed.2006.12.014

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 29.

    Stamatakis E, Johnson NA, Powell L, Hamer M, Rangul V, Holtermann A. Short and sporadic bouts in the 2018 US physical activity guidelines: is high-intensity incidental physical activity the new HIIT? [published online ahead of print February 20, 2019] Br J Sports Med. doi:10.1136/bjsports-2018-100397

    • Search Google Scholar
    • Export Citation
  • 30.

    Tremblay MS, Esliger DW, Tremblay A, Colley R. Incidental movement, lifestyle-embedded activity and sleep: new frontiers in physical activity assessment. Can J Public Health. 2007;98(suppl 2):S208–S217.

    • PubMed
    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Martenstyn, Powell, Nassar, and Stamatakis are with Charles Perkins Centre, Sydney School of Public Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia. Hamer is with the Institute of Sport Exercise & Health, Faculty of Medical Sciences, University College London, United Kingdom.

Stamatakis (emmanuel.stamatakis@sydney.edu.au) is corresponding author.