CCBYNC Open access

Physical activity and risk of breast cancer, colon cancer, diabetes, ischemic heart disease, and ischemic stroke events: systematic review and dose-response meta-analysis for the Global Burden of Disease Study 2013

BMJ 2016; 354 doi: (Published 09 August 2016) Cite this as: BMJ 2016;354:i3857

Re: Physical activity and risk of breast cancer, colon cancer, diabetes, ischemic heart disease, and ischemic stroke events: systematic review and dose-response meta-analysis for the Global Burden of Disease Study 2013

Re-setting the physical activity threshold – a flawed and potentially harmful approach?

Widely accepted global recommendations regarding the amount of physical activity (PA) required for chronic disease prevention (150-300 minutes/week; [1,2]) were challenged in a recent meta-analysis, which showed insignificant reductions in risk of NCDs at conventionally recommended levels of PA [3]. The inconsistency is likely explained by the fact that global recommendations are based on evidence for PA in leisure/recreational 'exercise' and transport that are reliably recalled, and are over and above the low level 'moving about' which is part of everyday domestic and most occupational activities. In contrast, the meta-analysis by Kyu et al focused on the health consequences of total physical activity, including occupational and domestic activities that show more tenuous relationships with health outcomes; these activities are unreliably recalled (and significantly over-estimated) and the intensity of these activities is usually low [4].

The new meta-analysis included data from generic surveys, in which PA is reported across various domains; these surveys often demonstrate 2-3 times the prevalence of “sufficient activity”, compared with contemporaneous leisure time physical activity surveys in the same countries [5]. Inclusion of all activities, including occupational and domestic activities, resulted in a range of activity reported in the first five figures from 0-32,500 MET.mins/week [3]. At moderate intensity of 4 METs, the upper end of this range is equivalent to 19.3 hours per day of exercise or activity, which is clearly unrealistic. Indeed, this is more than the weekly energy expenditure of contemporary Pennsylvanian Amish men who live without motorised or electrical assistance [6]. Even in their summary figure 7, Kyu et al indicate an upper limit of 10,000 MET.mins/week, which is approximately 6 hours per day of moderate intensity activity. The pattern of the dose-response relations between PA and risk of cardiovascular disease, diabetes and cancers are similar to those in previous consensus documents [1,7], with the maximal decline in risk observed in moving from inactive to currently recommended levels (around 500-600 MET.mins/week). However, the estimated PA required for significant risk reduction is higher in the Kyu paper, because of the dilution effects of inaccurately reported occupational and domestic activities and over-estimation of their MET values, and because values based on all physical activity below 600 MET.mins/day are used as the reference categories. Since low level activities are likely to have some benefit [8], the amount of activity required for significant risk reduction, in comparison with this reference category, is further inflated.

A recent meta-analysis of leisure time and transport-related PA showed a clear risk reduction in mortality of 30-40% when people achieved the recommended 30-60 minutes of moderate intensity PA per/day, and doing 10-fold more than this conferred no greater benefit [9]. Another recent meta-analysis highlighted the importance of the upper range of current PA recommendations (ie 60 minutes per day) to fully offset the health consequences of prolonged occupational/transit sitting time [10]. Both these meta-analyses used METs as a common metric of PA intensity.

In the editorial that accompanied the Kyu paper, Autier and Pizot (12) claim the use of METs for measuring PA is meaningless, because they do not 'tell us whether risk reductions would be different with short duration intense activity or longer duration light activity'. While this is true, METs remain important as the unit of exposure that has provided information on the volume of PA required to mitigate the health consequences of inactivity, as evidenced through hundreds of cohort studies. When METS accrued through self-reported occupational and domestic activity are included, there is likely to be significant measurement error, and if estimates of total daily are required, these should be measured objectively.

The conclusion that PA needs to be several times higher than the levels recommended in current guidelines (30-60 minutes per day, above activities of daily living) for meaningful reduction in the risk of NCDs is misleading. Population studies that have objectively assessed total PA with accelerometers in relation to health outcomes have consistently supported the current PA recommended levels, further suggesting methodological limitations in the Kyu et al study [13].

Unfortunately the public and the media are easily misled, as many media reports have now reported on the Kyu paper, indicating people “need between two and five times as much exercise as was previously recommended for health”. Potential adverse consequences of these media reports will be to: [i] de-motivate people who are trying to increase their PA, [ii] provide an unachievable target for most people (fewer than 1% of adults reach the amount recommended by Kyu et al [14] [iii] confuse policymakers and [iv] obfuscate the work of PA advocates. Kyu describes nine potential sources of error in the results, but do not mention the key issue of differential measurement error across self-reported domains of PA. Public debate to prevent these unintended consequences is essential.

Adrian Bauman, University of Sydney
Wendy Brown, University of Queensland
Ronald Plotnikoff, University of Newcastle, Australia


[1] World Health Organization. Global recommendations on Physical activity for health. 2010. WHO, Geneva {accessed August 2016]
[2] Australian Physical activity guidelines 2014.[ accessed August 2016].
[3] Kyu H, Bachman V, Alexander L, et al. Physical activity and risks of breast cancer, colon cancer, diabetes, ischemic heart disease, and ischemic stroke events: systematic review and dose-response meta-analysis for the Global Burden of Disease Study 2013. BMJ 2016;354:i3857.
[4] Peeters GMEE, van Gellecum Y, van Uffelen JGZ, Burton NW, Brown WJ. (2014). Contribution of house and garden work to the association between physical activity and wellbeing in younger, mid-age and older women. British Journal of Sports Medicine, 48(12), 996-1001.
[5] Macniven R, Bauman A, Abouzeid M. A review of population-based prevalence studies of physical activity in adults in the Asia-Pacific region. BMC Public Health. 2012 Jan 17;12(1):1.
[6] Bassett DR, Schneider PL, Huntington GE. Physical activity in an Old Order Amish community. Medicine and Science in Sports and Exercise. 2004; 36(1):79-85.
[7] Haskell, W.L., Lee, I.M., Pate, R.R., Powell, K.E., Blair, S.N., Franklin, B.A., Macera, C.A., Heath, G.W., Thompson, P.D. and Bauman, A., 2007. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation, 116(9), 1081.
[8] Hupin D, Roche F, Gremeaux V, Chatard JC, Oriol M, Gaspoz JM, Barthélémy JC, Edouard P. Even a low-dose of moderate-to-vigorous physical activity reduces mortality by 22% in adults aged≥ 60 years: a systematic review and meta-analysis. British journal of sports medicine. 2015 Oct 1;49(19):1262-7.
[9] Arem H, Moore SC, Patel A, Hartge P, de Gonzalez AB, Visvanathan K, Campbell PT, Freedman M, Weiderpass E, Adami HO, Linet MS. Leisure time physical activity and mortality: a detailed pooled analysis of the dose-response relationship. JAMA internal medicine. 2015 Jun 1;175(6):959-67.
[10] Ekelund, U., Steene-Johannessen, J., Brown, W.J., Fagerland, M.W., Owen, N., Powell, K.E., Bauman, A., Lee, I.M. Does physical activity attenuate, or even eliminate, the detrimental association of sitting time with mortality? A harmonised meta-analysis of data from more than 1 million men and women. The Lancet. Jul 27, 2016. pii: S0140-6736(16)30370-1
[11] Autier P and Pizio C. Meaningless METS: studying the link between physical activity and health. BMJ 2016;354:i4200 doi:10.1136/bmj.i4200
[12] Fishman, EI., Steeves JA, Zipunnikov V, Koster A, Berrigan D, Harris TA, Murphy R. (2016). Association between Objectively Measured Physical Activity and Mortality in NHANES. Medicine and Science in Sports and Exercise 48(7): 1303-1311
[13] ABS NHS Australian health survey 2014-2015 [unpublished re-analysis of weighted CURF data, August 2016]

Competing interests: No competing interests

15 August 2016
Adrian Bauman
Sesquicentenary Professor of Public Health
Professor Wendy Brown, University of Queensland, Brisbane, Australia; Professor Ron Plotnikoff, University of Newcastle, Australia
Sydney University
Charles Perkins Centre and School of Public Health, Sydney University, level 6 CPC Building D17, Sydney University 2006 NSW Australia,
Click to like: