Cart before Horse?

Predicting the exact impact of weight upon mortality has proven to be a tricky business. That obesity is on the rise is universally acknowledged, but in recent years we have seen research studies reach differing conclusions, depending on the populations examined and the measures used. For example, there has been debate over whether the breakpoints used to analyse BMI, the most prominent weight measure are appropriate in all populations. A key issue posited, however, has been that of "reverse causality" - which in this context is usually interpreted as the failure to adjust for weight scores that are considered "healthy" under the BMI scale but in fact derive from illness-related weight-loss. The question is whether some of the uncertainty around the health-effects of BMI — especially for the overweight category (BMI 25-30) — would be clarified by adjusting for reverse causality.

A paper published this month in the Lancet attempted to take a global look at obesity and mortality whilst adjusting for reverse causality. The populations involved were large - more than 10 million participants culled from studies across Europe, North America, Asia and Australia/New Zealand. The steps taken to adjust for reverse causality were, however, similarly substantial. The meta-analysis excluded data from anyone who had ever smoked, anyone with a pre-existing disease, and any deaths that occured within five-years of follow-up. This left just less than 4 million in the final dataset. Adjustments such as these are not universally lauded. Even setting aside methodological challenges they make the dataset researched very different from a general population, and in particular, at older ages the study will be considering a very select bunch indeed, due to the rising prevalance of chronic conditions and shorter survival-times at that lifestage. Nevertheless where researchers seek a targeted measure of the effect of obesity on the the otherwise healthy, such adjustments are emerging as a common way to proceed. The result in this latest paper was unequivocal: all-cause mortality was higher for all of the underweight, overweight and obese categories of the BMI measure - and further, this behaviour was seen across all four continents considered.

It should be noted that similar issues may arise in other areas, such as when modelling the effect of alcohol abstention. Modelling the life-long teetotal in the same category as reformed chronic alcoholics or those who abstain due to a pre-existing medical condition may similarly reduce the clarity of the results. As an example, this mortality study of male doctors in the UK found clear differences between never-drinkers and ex-drinkers, something that may not always have been accounted for in earlier work.

Some studies have debated the relative importance of fitness over fatness, and undoubtedly work in this area will continue. This review from 2014 concluded that overweight-but-fit individuals exhibited similar risk to their normal-weight counterparts. Perhaps crucially, however, while that review made some attempt to account for chronic disease, it took no specific position on reverse causality for the data included. A more recent study from Sweden examined the risk of early death for a population of male swedish military conscripts, and concluded that fitness did not offset the impact of being overweight in that population, finding unfit but normal-weight individuals exhibiting 30% lower rates of all-cause mortality than fit but obese individuals. Few will be surprised if longevity maximisation in the absence of pre-existing disease is eventually proved to require both a healthy BMI and aerobic fitness. Living longer with a chronic condition may yet be found to require a more complex analysis.

References

WHO expert consultation. (2004) Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. The Lancet. Volume 363, No. 9403, p157–163, 10 January 2004 DOI: http://dx.doi.org/10.1016/S0140-6736(03)15268-3

The Global BMI Mortality Collaboration (2016) Body-mass index and all-cause mortality: individual-participant-data meta-analysis of 239 prospective studies in four continents. The Lancet. Published Online: 13 July 2016 DOI: http://dx.doi.org/10.1016/S0140-6736(16)30175-1

Flegal, K.M. et al. (2014) Reverse Causation and Illness-related Weight Loss in Observational Studies of Body Weight and Mortality. Am. J. Epidemiol. (2011) 173 (1): 1-9. doi: 10.1093/aje/kwq341

Doll, R. et al. (2005) Mortality in relation to alcohol consumption: a prospective study among male British doctors. Int. J. Epidemiol. (2005) 34 (1): 199-204. doi: 10.1093/ije/dyh369

Barry, V.W. et.al. (2014) Fitness vs. Fatness on All-Cause Mortality: A Meta-Analysis. Progress in Cardiovascular Diseases Volume 56, Issue 4, Pages 382–390 DOI: http://dx.doi.org/10.1016/j.pcad.2013.09.002

Nordström, P. et.al. (2015) Aerobic fitness in late adolescence and the risk of early death: a prospective cohort study of 1.3 million Swedish men. Int. J. Epidemiol. (2015) doi: 10.1093/ije/dyv321

Antonopoulos, A.S. et.al. (2016) From the BMI paradox to the obesity paradox: the obesity–mortality association in coronary heart disease. Obesity reviews. Early View. DOI: 10.1111/obr.12440

 

Comments

captcha

Find by key-word


RECENT POSTS

The upcoming EU General Data Protection Regulation places focus on ... Read more
Assume we have a random variable, \(X\), with expected value ... Read more
Our new book, Modelling Mortality with Actuarial Applications , describes ... Read more
Gavin Ritchie
Gavin Ritchie is the IT Director of Longevitas