Arya M. Sharma, MD

Yesterday, at the 29th Scientific Meeting of The Obesity Society, I spoke at a session on the history of obesity, hosted by George Bray and David Haslam, author of Fat, Gluttony and Sloth: Obesity in Literature, Art, and Medicine.

In my talk, I discussed the history of the Quetelet Index, later named the Body Mass Index by Ancel Keys; I spoke about its limitations in clinical practice and presented the Edmonton Obesity Staging System (EOSS) and our latest data showing that this staging system is a far better predictor of mortality than BMI (or waist circumference).

Although it is far too early to know whether or not EOSS (or a variation thereof) will eventually be more widely used in clinical practice, research and policy, I could not but sense that the audience fully appreciated the need for a clinical obesity staging system, which characterizes how ’sick’ rather than simply how ‘big’ patients are. Clearly, such a system is but a natural evolution in our thinking about obesity, especially its treatment.

Click here to download a short powerpoint presentation on the Edmonton Obesity Staging System.

AMS
Orlando, FL

Over the last several days I have been examining various aspects and implications of our recent publications showing that the Edmonton Obesity Staging System (EOSS) does a far better job of predicting mortality than does BMI (in fact BMI does almost nothing in this regard).

Not only does EOSS make intuitive sense to clinicians and most patients (especially the ones who are at EOSS 0) it is also a better way to individualize patient management strategies.

But, despite these two publications in three independent samples that included over 20,000 participants, many important questions remain to be addressed:

it is not clear whether all comorbidities should receive the same weight for defining the EOSS stage - for e.g. should chest pain due to reflux disease count the same as chest pain due to ischemic heart disease (probably not)?

What is the natural history of EOSS stage progression? Or in other words, how long does it take for patients to move from Stage 0 to Stage 1 or from Stage 2 to Stage 3? Are there really patients, who never progress? Are there predictors of progression? If yes, can this progression be delayed or prevented?

What does it take to reverse Stages and does reversing the obesity Stage improve prognosis (it probably does)?

How do cost-effectiveness and risk-benefit ratios of obesity treatment for patients look at different EOSS stages? I am guessing that both increase at higher stages, but is this really the case?

Can we develop a simplified version of EOSS (EOSS-lite?) that only counts certain comorbidities or only acknowledges certain dimensions of quality of life?

Is EOSS a concept that health professionals, decision makers, and funders are ready to adopt and will it improve practice and outcomes?

These are all questions that future research will need to address, some of this work is already underway, but I’d be happy to hear from potential collaborators or people wanting to do some of this research on their own.

If nothing else, I at least hope that the EOSS discussion has opened a whole new way of thinking about clinical assessment and definition of obesity and will find its way into clinical care pathways and management guidelines.

From everything I hear, this is already beginning to happen.

AMS
Edmonton, Alberta

Padwal RS, Pajewski NM, Allison DB, & Sharma AM (2011). Using the Edmonton obesity staging system to predict mortality in a population-representative cohort of people with overweight and obesity. CMAJ : Canadian Medical Association journal = journal de l’Association medicale canadienne PMID: 21844111

Kuk JL, Ardern CI, Church TS, Sharma AM, Padwal R, Sui X, & Blair SN (2011). Edmonton Obesity Staging System: association with weight history and mortality risk. Applied physiology, nutrition, and metabolism = Physiologie appliquee, nutrition et metabolisme PMID: 21838602

Continuing our discussion on the implications of the recent publication of our two studies showing that the Edmonton Obesity Staging System (EOSS), which actually measures how ’sick’ patients are rather than just how ‘big’ patients are, is a better predictor of mortality than BMI, I’d like to briefly discuss a potential implications for public health messaging around obesity and health promotion.

In contrast to what some media commentators have chosen to imply (’It is now OK to be fat’), EOSS actually takes a very differentiated view of the complexity of the relationship between excess weight and illness.

In fact, one key learning for public health messaging from our EOSS papers, would clearly be to re-evaluate the entire notion of a ‘healthy weight’, an outdated and misleading concept that harks back to the rather simplistic idea of ‘ideal’ weight.

Indeed, we previously criticized the notion of ‘ideal’ weight in an article published in 2009 in SOARD, in which we lamented the use of ‘Excess Weight Loss’ in the bariatric surgical literature.

As we noted in that article:

…“ideal weight” is synonymous with the term “desirable weight,” first introduced in 1943 by the Metropolitan Life Insurance Company (MLIC) in their standard height-weight tables for men and women. These tables were determined from actuarial data indicating the lowest mortality risk related to a range of weights for a given height in the studied population. Recognizing that the relationship between mortality and body weight is anything but simple, the MLIC life tables also introduced the concept of frame size, a concept that was hard to understand and implement and was therefore largely ignored. In fact, the current common use of “ideal weight” generally refers to the medium frame category, regardless of the patient’s actual size or form. Nevertheless, it is important to recognize that for a 170 cm (5’ 7”) woman, depending on her frame size, the MLIC “ideal weight” can range from 56 kg for a small frame to 74 kg for a large frame, or a difference of 18 kg. This translates into a body mass index (BMI) range of 19.3–25.6 kg/m2, or more than 6 BMI points.”

Thus, not only was this notion of ‘ideal’ weight based on an unrepresentative actuarial sample but there was actually a weight range of about 18 kg (or about 40 lbs) over which a person’s weight for the same height could be considered ‘ideal’.

Our recent EOSS data shows that for some people (albeit a minority at higher BMI levels), this range of ‘ideal’ weight based on mortality can indeed be even larger.

Thus, recognizing that weight (or BMI) is indeed such a crude (some would say irrelevant) measure of ‘health’, the entire concept of ‘healthy weights’ may need to be abandoned in favour of a more differentiated look at the rather complex relationship between health and weight.

At a minimum, public messaging around obesity, may need to make it very clear that ‘health’ can be achieved and maintained across a wide range of weights and that it is probably far more important to focus on health behaviours and other indicators of health than to jump to conclusions about someone’s health based solely on their weight.

Obviously, I fully appreciate that public health messages need to be kept simple but there are certainly arguments to be made that, given the rather loose relationship between health and weight, continuing to promote the notion of ‘healthy weights’ may do more harm than good, as such messaging would simply continue to promote obesity related stereotypes and can potentially set people off on paths of unhealthy weight loss in the firm belief that losing weight is equivalent to improving health.

On the other hand, none of this should distract from the fact that we do have an obesity epidemic and that the vast majority of people with ‘excess’ weight do indeed have relevant comordibities that put them in higher EOSS categories - for these individuals, improving access to evidence-based preventive and treatment resources must be a priority of any healthcare system.

I can certainly see how dealing with this rather complex problem in public health messages and policies will likely lead to a most interesting discussion in the months to come.

AMS
Edmonton, Alberta

Regular readers of these pages will recall our recent paper in PLoS showing increased susceptibility for cardiometabolic risk factors in South Asians at substantially lower body weights than in Caucasians. These findings are consistent with a large body of evidence suggesting that the BMI definition of ‘obesity’ should likely start a lower BMI levels in South Asian and East Asian populations.

In India, not too long ago, redefinition of ‘obesity’ with lower BMI cutoffs, resulted in a major reclassification of the ‘burden of obesity’ in that country.

There is, however, good reason to assume that even at this lower BMI cutoff, the relationship between BMI and actual risk is probably as poor in this population, as it has proven to be in Caucasian samples.

We there think it reasonable to propose that the Edmonton Obesity Staging Stystem, which characterizes obesity stages based on how ’sick’ rather than simply on how ‘big’ a given patient may be, will also prove a better system to determine individual care plans than simply suggesting that everyone with a BMI that crosses a certain threshold to lose weight.

Since our recent papers on EOSS examined a predominantly white population, I very much hope that other investigators with access to South Asian or other ethnic populations will examine the relationship between EOSS and mortality in their subjects.

I would frankly be very surprised if their results turn out to be any different in that EOSS is a much better predictor of individual risk than BMI.

AMS
Edmonton, Alberta

Padwal RS, Pajewski NM, Allison DB, & Sharma AM (2011). Using the Edmonton obesity staging system to predict mortality in a population-representative cohort of people with overweight and obesity. CMAJ : Canadian Medical Association journal = journal de l’Association medicale canadienne PMID: 21844111

Kuk JL, Ardern CI, Church TS, Sharma AM, Padwal R, Sui X, & Blair SN (2011). Edmonton Obesity Staging System: association with weight history and mortality risk. Applied physiology, nutrition, and metabolism = Physiologie appliquee, nutrition et metabolisme PMID: 21838602

Amongst all the enthusiastic accolades that welcomed the publication of the two recent papers showing that the Edmonton Obesity Staging System (EOSS) is superior to BMI in predicting mortality risk, there were some commentators (including readers of these pages), who apparently misinterpreted these findings as ‘a license to be fat’ or even ‘as sending a devastating public health message that fat can be healthy’.

While the lay media and lay readers may be forgiven for this simplistic and cursory interpretation of our findings, more sophisticated readers should actually have noted that nowhere in these papers (or in the original proposal for EOSS) do we ever equate EOSS 0/1 (low risk) with doing nothing.

In fact, we were very careful in pointing out that even patients presenting with EOSS 0/1 should be investigated for underlying drivers of weight gain and should be counselled on prevention of further weight gain (weight maintenance) and, insofar there is indeed any additional room for improvement, to eat healthily and get as much activity as they can enjoy.

The caveat regarding ‘additional room’ was put in to imply that, yes, people in lower EOSS categories already appear to be eating quite healthy and certainly are more physically fit than their higher-risk EOSS counterparts (meaning - they may already be doing the best they can).

To me, investigating someone for the causes of excess weight, counselling on weight maintenance, and scheduling repeat consults (say at 12-18 months) is not ‘doing nothing’. It is also by no means ‘a license to be fat’ (as if being fat - per se is a bad thing).

In fact, the whole point of EOSS is to better identify and focus treatment on high-risk individuals while pretty much leaving lower-risk EOSS folks alone (albeit with minimal investigation and intervention).

But of course, this notion is unlikely to appease the ‘weight-loss-at-all-cost’ enthusiasts, who firmly believe that there is something inherently ‘unhealthy’ about extra fat (when clearly our studies show that there isn’t).

These ‘weight-loss-enthusiasts’ will also make the ‘prevention’ argument, which of course is based on nothing, given that there are indeed no studies showing that obese people who are currently healthy (or which subset of these) will actually progress to developing relevant morbidity, IF THEY CONTINUE TO MAINTAIN THEIR CURRENT WEIGHT AND HEALTHY LIFESTYLES!

As it takes months if not years or even decades to progress from one EOSS stage to the next, and as most health problems in EOSS 1 and 2 (by definition) are reversible, there is really no argument to pre-empt these problems by recommending weight loss to people, who really have no weight problem.

Indeed, if the first rule of medical practice is ‘do no harm’, I would be the first to point out that recommending weight loss to someone, who is essentially healthy, actually does have the potential to do harm.

Firstly, this person could go out and begin engaging in unhealthy weight loss practices just to get those supposedly ‘extra’ pounds off - this practice alone could pose a health risk.

Secondly, losing some weight only to put it right back on (as do 95% of people who try to lose weight) may well have negative physical and emotional consequences (not to mention the negative impact on their wallets).

Thirdly, having engaged in ‘unsuccessful’ weight loss may make this person less motivated and perhaps even less likely to succeed in losing weight, when, at a later time, this may indeed be indicated - crying ‘wolf’ now may simply lead to the real wolf being ignored later.

But here is a final argument that could perhaps appease the ‘weight-loss-at-any-cost’ enthusiasts - we are (sadly perhaps) only talking about a minority of overweight and obese people, to whom this ‘low-risk’ status actually applies.

Thus, in the overweight category, only 15% of the individuals were EOSS 1, a proportion that decreased to only 8-5% in higher BMI classes.

Thus, the supposed ‘licence’ would in any case only apply to 1 in 6 overweight folks or even only 1 in 15-20 folks with a BMI over 30 - the vast majority would not be in this supposed ‘wait and see’ category.

Indeed, amongst individuals with Class II or III obesity, almost 80% of individuals were classified as EOSS 2/3 - this is the group that is very likely to benefit from obesity treatments.

As I’ve said before, EOSS adds important shades of grey to what many consider to be black or white - either you are at a ‘healthy weight’ (whatever that is) or you need to lose weight - whether or not you are actually healthy at your current weight does not appear to enter into their reasoning.

Of course, when there is money to be made in weight loss, and when most of your clients happen to be EOSS 0, our studies suggest that you should perhaps consider changing your business model (or at least have your clients sign a consent form stating that losing weight may not be in their best interest).

But if you are a health professional or decision maker wondering about just how to dedicate your limited resources to those obese patients most likely to benefit, first addressing the needs of those presenting with higher EOSS stages seems a perfectly rational and reasonable argument.

Only if you still believe that health can be measured simply by stepping on a scale, are you likely to continue thinking that EOSS is a disservice to medical care and dangerous to the public.

For the rest of us, hopefully, EOSS will prove a viable strategy to deliver evidence-based health care to where it is needed the most.

AMS
Ottawa, Ontario

Padwal RS, Pajewski NM, Allison DB, & Sharma AM (2011). Using the Edmonton obesity staging system to predict mortality in a population-representative cohort of people with overweight and obesity. CMAJ : Canadian Medical Association journal = journal de l’Association medicale canadienne PMID: 21844111

Kuk JL, Ardern CI, Church TS, Sharma AM, Padwal R, Sui X, & Blair SN (2011). Edmonton Obesity Staging System: association with weight history and mortality risk. Applied physiology, nutrition, and metabolism = Physiologie appliquee, nutrition et metabolisme PMID: 21838602