Arya M. Sharma, MD

As readers of these pages are well aware, there is currently only one prescription drug (orlistat) for the long-term management of obesity and recent applications for new anti-obesity drugs have run into considerable obstacles with licensing agencies, not least the US Food and Drug Administration.

Thus, despite meeting criteria for efficacy, recent applications for new anti-obesity drugs were rejected due to safety concerns, the argument being that the expected widespread use of these drugs warrants higher standards of safety than for drugs in other therapeutic areas.

This may well be true, if anti-obesity drugs are indeed to be made freely available (even on prescription) and are to be used without careful consideration of the risk/benefit ratio.

As discussed previously in the context of bariatric surgery, the risk side of this equation is not limited to the ‘risk’ associated with treatment but must also include the ‘risk’ associated with not treating the condition.

As our recent papers on the Edmonton Obesity Staging System (EOSS) clearly show, the ‘risk’ associated with obesity varies considerably - from virtually no risk for patients with EOSS 0 to very significant and immediate risk for patients with EOSS 3.

So, while even the smallest treatment-related risk may be unacceptable for treating obesity in a patient with EOSS 0, a higher level of risk would be certainly be acceptable with higher EOSS stages - or in other words - the greater the risk of not treating, the greater the acceptable risk for a potentially beneficially treatment (which, for e.g., is why cancer warrants the considerable treatment risks of chemotherapy).

Of course, this would mean that rather than having to prove the safety and efficacy of anti-obesity drugs in anyone with a BMI over 30, it may be enough to show that these drugs are effective and safe enough to warrant their use in people with higher EOSS stages (2/3) - the greater the risk of the target population, the greater the acceptable risk of treatment.

This should not be difficult. We know that many of the conditions that patients with EOSS 2/3 present with can be ameliorated even with rather modest weight loss. For these patients, the 5-10% sustainable reductions in body weight that can be achieved with anti-obesity drugs, can have very significant health benefits, which would easily outweigh and justify a reasonable risk of adverse effects - a risk that may be unacceptable in patients presenting with EOSS 0/1 obesity.

This, of course means rewriting some of the approval criteria and guidelines for anti-obesity drugs, but also requires redesigning pharmacological trials to focus on the high-risk EOSS 2/3 patients rather than on the low-risk EOSS 0/1 patients.

I certainly look forward to following how this discussion evolves and wether or not we can indeed find a way out of the current impasse of anti-obesity drug development and approval.

AMS
Toronto, 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

As not everyone may have a chance during the week to read every post, here’s a roundup of last week’s posts:

• Setting The Stage For Obesity Staging
• Health Is Not Measured In Pounds
• Will Losing Weight Make You Sick?
• Should Causality Matter In The Edmonton Staging System?
• Can The Edmonton Obesity Staging System Better Guide Indications For Bariatric Surgery?

Have a great Sunday! (or what’s left of it)

AMS
Toronto, Ontario

You can now also follow me and post your comments on Facebook

One of the questions many readers and colleagues have asked, is whether or not the issue of ‘causality’ should matter in the the context of the Edmonton Obesity Staging System (EOSS).

In other words, should only conditions be counted that are ‘causally’ linked to obesity or is it enough that these conditional are merely more common in people with excess weight.

There are essentially two important but distinct aspects to this question that ultimately relate to how EOSS is to be used in clinical practice.

If the primary purpose of EOSS is to identify patients who would benefit from weight loss, then, yes, it matters whether or not the co-morbidities considered, are ‘causally’ related to obesity and can be reversed or ameliorated by reducing and sustaining a lower body weight.

However, if the primary purpose of EOSS is merely to identify obese patients, who are at high risk and need to be prioritized within the health care system in order to receive the appropriate care for their conditions (irrespective of whether or not this ‘care’ involves weight loss or just better management of their comorbidities), then the question of causality is really irrelevant.

Thus, in the first case, one would only count ‘comorbidities’ that are actually ‘causally’ related to excess weight - an example being sleep apnea. There is ample evidence that weight loss reduces symptoms of sleep apnea (while weight gain makes it worse) and so sleep apnea would count as an important comorbidity that can be addressed by obesity treatment.

In the second case, it does not actually matter if the comorbidity is in any way related to excess weight. All that really matters, is whether this comorbidity is present or not. An e.g. would be depression, which, while not caused by obesity and not likely to improve with weight loss (it may sometimes even get worse), may, when present, help identify obese patients, who do have a higher risk of premature death.

So while in the first example, EOSS would be used to decide who needs to lose weight, in the second example, EOSS simple serves to identify obese people, who are at highest risk of complications and death.

Apart from the second scenario being the real reason that EOSS was developed, it is also a far more practical approach to using EOSS, because for many comorbidities it may be impossible to answer the ‘chicken or egg’ question or even determine if these are simply two different chickens.

In clinical medicine this phenomenon is referred to as ‘phenocopy’, a term used to describe a case where two distinct and unrelated conditions, present clinically with the same symptoms or ‘phenotype’.

In obesity, this is particularly common, because, while many symptoms may be ‘causally’ related to obesity, these same symptoms may just happen to present in an obese individual but have nothing to do with that patient’s excess weight.

For EOSS, this question would not really matter - whether the knee pain is from an accident or from carrying around the excess weight makes no difference - the only thing that matters is that this is a patient with excess weight AND knee pain and therefore, this patient is an obese patient, who is at higher risk for mortality than an obese patient without knee pain.

Remember, bariatric care, as I define it, is not about losing weight but rather about medical care for the bariatric patient. As resources are limited, all I want EOSS to tell me is, who to see first.

Of course there are ‘normal-weight’ people with knee pain, but they are not my problem. There are also ‘normal-weight’ people with high blood pressure, sleep apnea, diabetes and depression - again, they are not my problem.

As someone, who works in an obesity clinic, my job is to assess obese patients and help them the best I can, whether their treatment requires weight loss or not. If EOSS can help me decide, who to see most urgently, then EOSS has done its job.

Of course, further research is needed to determine whether EOSS actually works well in clinical practice (e.g. outside of a speciality centre) and we may perhaps need to simplify and clarify the criteria. But the principle stands: it is simply not enough to look at BMI to decide who needs (urgent) medical attention and who does not.

AMS
Edmonton, Alberta

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

As anyone watching the news over the past 24 hours will probably have noticed, yesterday saw the release of two large studies looking at whether or not the Edmonton Obesity Staging System is a better predictor of mortality risk than BMI alone - the short answer is “yes”.

The results from these studies was reported by all major media outlets around the globe including CNN, TIME, MSNBC, and virtually all national print, radio, and tv stations. It was also picked up by the blogsphere.

Interestingly, when I first suggested the use of a staging system for obesity in 2008, it was born out of both a practical need and my own medical experience with obese patients.

The practical need was to better determine, who needs to be seen in our obesity program, given the rather limited resources and long waiting times.

My medical experience had long taught me that the commonly used BMI classification of obesity, or even the suggested use of waist circumference were rather blunt instruments in determining which patients needed obesity treatments and which did not.

So, the idea was simply to create a clinical tool that would help us decide, which obese patients required our attention most urgently.

However, as readers will imagine, with all the talk about ‘healthy” weights and ‘benefits’ of weight loss, our proposal was met with considerable scepticism - not about whether or not obese people with obesity related health problems (EOSS 2-4) needed treatment, but rather whether or not obese people who appeared pretty healthy (EOSS 0/1) were indeed at a low risk from their excess weight.

So we looked for large datasets in which we could apply EOSS and compare it to BMI in predicting death.

In one collaboration, on which my colleague Raj Padwal took the lead, assisted by David Allison and Nicholas Pajewski from the University of Birmingham, Alabama, we looked at the impact of EOSS on mortality in two separate sets of the NHANES study - a representative sample of the US population. The results of this analysis were published yesterday in the Canadian Medical Association Journal.

In another collaboration, on which Jennifer Kuk and Chris Ardern (York University, Toronto) took the lead with help from Timothy Church (Pennington Biomedical Research Center, Baton Rouge, LA), and Xuemi Sui and Steven Blair (University of South Carolina), we looked at the impact of EOSS on mortality in the Aerobics Center Longitudinal Study (n = 29 533). The results were released yesterday in the Applied Physiology, Nutrition & Metabolism.

In today’s post, I will not go into details of the studies or begin a lengthy discussion of the findings - suffice it to say, till someone comes up with an even better way to clinically assess the health status of obese patients, to help decide who does and who does not need obesity treatment, the Edmonton Obesity Staging System may be just the tool that clinicians, payers, and patients have been looking for, to help dispel the notion that health can be measure by simply stepping on a scale.

AMS
Edmonton, Alberta