Will Losing Weight Make You Sick?

One of the intriguing (some would say frustrating) aspects of analysing large datasets is that this often raises more questions than it answers.

This is certainly the case with one of the studies on the Edmonton Obesity Staging System, published in Applied Physiology, Nutrition and Metabolism this week.

The paper looked at data from the Aerobics Center Longitudinal Study, a cohort of over 29,000 participants who attended the Cooper Clinic (Dallas, TX) for periodic self- or physician referred medical examinations between 1987 and 2001.

Of these, 5,453 men and 771 women both the met the BMI criteria for obesity and had enough information available to allow EOSS grading.

As reported before, irrespective of BMI, EOSS stage 0 and 1 participants had no significant impact on their risk of mortality over 16 years, which, however, was higher in EOSS stage 2 and stage 3 participants.

So, if not body weight (or BMI), what exactly were some of the characteristics of individuals with higher EOSS stages?

It turns out that apart from (as one may expect) the fact that individuals with lower EOSS stages reported eating more fruit and vegetables and had higher cardiorespiratory fitness (as an indicator that they were clearly more physically active), they were also less likely to have a history of weight cycling.

Indeed, lower EOSS stages were associated both with less lifetime weight loss as well as fewer (or no) episodes of prior weight loss.

This certainly poses the question, whether dieting or losing weight in fact increases the long-term risk of health problems and one can only wonder if the folks with higher EOSS scores would be better off had they never lost weight before.

Now, obviously, this association (as all associations) does not prove causality. It could well be that people who already have health problems may be more likely to engage in (or remember) previous weight loss attempts.

It may also be that worth noting that people who tend to engage in weight loss are the ones who often have significant body image and body dissatisfaction issues, as well as a generally higher prevalence of psychiatric illnesses than people who do not diet or lose weight. As psychiatric and mental health are part of the EOSS criteria, it may well be that this alone accounts for the association of yo-yo dieting and elevated EOSS stages.

Or, as we discuss in the paper:

“…for the vast majority of obese individuals, lifestyle-based weight loss is not maintained over the long term (Wing et al. 1995). This is particularly concerning, given that weight cycling is associated with greater weight gain over time (Van Wye et al. 2007) and potentially worse health outcomes, compared with individuals who may have maintained a stable body weight (Blair et al. 1993; Wannamethee et al. 2002). Although we observed that greater reported weight loss was associated with worse EOSS scores, it is unclear whether individuals with more severe EOSS staging had attempted to lose more weight because of their poor health, or whether they had poorer health because they had weight cycled. Furthermore, it is unclear whether obese individuals without existing comorbidities will develop metabolic abnormalities if they remain at a stable BMI…“

These are all intriguing questions for which we currently simply have no definitive answers.

However, it is certainly clear from this study that there are a significant number of people, who meet the BMI criteria of obesity, but do not appear to have any of the health problems that most overweight and obese folks tend to have. It is certainly unclear whether or not these individuals will experience any health benefits from attempting to lower their body weights, given that most people, who lose weight, will simply put it back on.

This is by no means implies that it is now “OK to be fat“, as some media has chosen to report on this study. At best, it means that for some people, it may well be OK to be fat, but these people certainly become rarer at higher ranges of BMI.

Whether those, who do have health problems are better served by interventions primarily focussing on changing health behaviours whilst promoting positive outlooks and size acceptance than by interventions primarily focussed on reducing their weight, will certainly remain a topic of debate for some time to come.

Nevertheless, the results of our study certainly add several shades of grey to the usual black-or-white discussions about the impact of body weight on health and do raise questions about simply recommending weight loss to anyone, who happens to meet the current BMI criteria for obesity.

AMS
Edmonton, Alberta