Does Dieting Increase Diabetes Risk?

That's certainly the message that HAES practitioner Linda Bacon wanted her followers to believe. In fact her tweet suggests that "even short periods of calorie restriction" increase diabetes risk.

In case you're not familiar with Dr. Bacon her work challenges the assumptions made about obesity and she has been highly critical of studies linking obesity with morbidity and mortality.

In an interview she gave to Med Journal Watch she explains why she believes not everyone agrees with her conclusions,

"My experience from having worked closely with many obesity researchers who are more conventionally-minded than me is that they are so strongly mired in their assumptions, that they don't look at the evidence."

And now back to diabetes and dieting, the story of which in this case begins back in time during the Dutch famine of World War II which they not so affectionately call the Hongerwinter (hunger winter).

It was September 1944. The Germans blockaded Holland and cut off food and fuel shipments to punish the Dutch people who opposed the Nazi regime. Food stocks dwindled. By the end of November rations amounted to fewer than 1,000 calories a day, and by February, to 580. 4.5 million people suffered, and over 22,000 perished. At the famine's worst daily rations amounted to half of a medium sized potato and 2 slices of bread. To compound matters, fuel was nearly impossible to come by and despite frigid winter temperatures, gas, heat and electricity were turned off. The famine lasted until May 1945.

Can you imagine being a child in Holland during the famine? The suffering and the horror they must have felt are unfathomable and when coupled with the severe and prolonged under-nutrition they experienced perhaps it's no surprise that their cohort have seen increases in risk of a variety of medical conditions including type 2 diabetes.  It's their experiences that Dr. Bacon uses as the source for her diabetes related caution against dieting.

The link Dr. Bacon provided in her assertive tweet about dieting and type 2 diabetes risk led to a a paper entitled, "Famine Exposure in the Young and the Risk of Type 2 Diabetes in Adulthood". In it researchers studied 7,557 Dutch women who endured the Hongerwinter and who on average were 9 years old when it occurred. Subjects were stratified into 3 groups on the basis of their self-reported exposure to famine - none, moderate and severe. The researchers found a small and statistically significant increase in risk to those women who reported themselves as moderately or severely affected by famine and this risk, while slightly attenuated, persisted after controlling for age at start of famine, education, BMI, waist circumference and waist to hip ratio.

The authors also noted that their study was unable to distinguish whether or not this association was related to under-nutrition or to famine related stress and point readers to a Finnish study that found a similar increase in lifetime risk of type 2 diabetes associating with the results of a psychological stress test in childhood war evacuees.

So in the end it would seem that from reading the evidence enduring a moderate to severe famine at a young age increases the risk of developing type 2 diabetes perhaps by way of the impact of under-nutrition, or perhaps by way of the impact of psychological stress, or perhaps by another as yet not elucidated cause.

In no way shape or form does this data suggest that "dieting" increases the risk of type 2 diabetes and in no way shape or form is suffering through a famine where daily rations include 2 small slices of bread and half a potato fairly described as a "short period of calorie restriction".

For Dr. Bacon this seems to be a bit of a pattern - slam conventional researchers for being, "so strongly mired in their assumptions that they don't look at the evidence", and then be so strongly mired in her assumptions that either she herself doesn't look at the evidence, or if she does her confirmation bias is so powerful that she'll happily find a way to present it to fit her narrative (click here for more examples). When I questioned her about this particular example on Twitter, despite her battle cry on Huffington Post of, "Show me the data we demand, and you should, too", she blocked me.

A few days ago dietitian and HAES advocate Julie Rochefort asked me on Twitter what barriers I saw to mobilizing HAES into practice. The main one I see seems to be regularly reflected by Dr. Bacon - knee jerk anger and either the willful manipulation of evidence, or a lack of critical appraisal of data so long as it seems to fit the HAES storyline.

Dr. Bacon is certainly HAES' most visible champion and role model. Responding to criticism with anger and manipulating or simply not critically evaluating HAES friendly data undermines the credibility of HAES as a whole, makes HAES easier for detractors to dismiss, and sets an absolutely terrible example for HAES practitioners and supporters to follow.

Annet van Abeelen, Sjoerd Elias, Patrick Bossuyt, Diederick Grobbee, Yvonne van der Schouw, Tessa Roseboom, & Cuno Uiterwaal (2012). Famine Exposure in the Young and the Risk of Type 2 Diabetes in Adulthood Diabetes DOI: 10.2337/db11-1559

Why HAES May Never go Mainstream

It's not the concept that's for sure.

For those of you who aren't familiar with the acronym HAES, it stands for "Health at Every Size", and it's a principle with which I strongly agree.

According to the official HAES community page HAES,

"acknowledges that good health can best be realized independent from considerations of size. It supports people—of all sizes—in addressing health directly by adopting healthy behaviors."

And truly, I could not agree more in that the words "healthy", and "weight", are not mutually inclusive or exclusive terms.

According to HAES' founder Linda Bacon, one of HAES' tenets is, "Show me the data", and in her recent Huffington Post piece, she says that we should all be demanding the data too and adopting HAES' "more skeptical" mantra.

Again, I could not agree more.

Yet despite readily agreeing that fat has been regularly and unfairly vilified by society and the medical community for decades if not centuries, and despite regularly telling my otherwise healthy overweight and moderately obese patients that their weights aren't likely contributing much if anything to them in the way of medical risk, I struggle with HAES as it would seem to me that they are fighting misinformation with misinformation, and in so doing, weakening and cheapening their incredibly important and valuable message.

Looking at Linda's debut Huffington Post piece here are the 3 things she wants readers to understand are, "known (even if everyone can't accept it yet)",

"-  Stable fat is blown out of proportion as a health risk (even dreaded "tummy fat"), but yo-yoing weights common to dieters do harm health.

-  The "ironclad" notion that obesity leads to early death is wrong: Mortality data show "overweight" people, on average, live longest, and moderately "obese" people have similar longevity to those at weights deemed "normal" and advisable.

-  Life spans have lengthened almost in lockstep with waistlines over the last few decades, which should make you wonder about the supposed deadliness of fat."

If we're talking, "show me the data", then lets talk data.  First, the data on yo-yo diets, otherwise known as weight cycling. Looking at the most recent and robust data, one set from than Cancer Prevention Study II Nutrition Cohort which followed 55,983 men and 66,655 women from 1992-2008, and the other set from the Nurses Health Study which followed 44,882 women from 1972-1994, neither demonstrated any relationship between weight cycling and mortality. Other studies have exonerated weight cycling from increasing the risk of hypertension, and type 2 diabetes, and there's a mixed bag of studies suggesting both protective and causal effects of weight cycling on various forms of cancer. But if we're really talking "show me the data", the only thing very conclusively linked to weight cycling are increased body fat percentages, and while I definitely agree weight cycling is symptomatic of a broken societal approach to weight management, and may well carry with it some harm, the data simply do not currently support a blanket, "harm health" statement.

Next the "ironclad" comment. While it's true that "overweight" has been shown to be protective in the over 65 population, and that "Class I", or "moderate", obesity carries the same risks as "normal" weight in that same population, what Linda omits here, other than the age qualifiers, is the ironclad fact that as weights rise more dramatically than simply "moderate" obesity, so too does risk. And it's not just as weights rise, but also as weight responsive conditions accumulate as is clearly shown by Dr. Sharma's Edmonton Obesity Staging System work which demonstrates that as EOSS stage rises, where EOSS evaluates weight in the context of having or not having weight related co-morbidities or quality of life impacts, so too does mortality.

Lastly we get to the lockstep comment about the last few decades. Here I'm nearly at a loss for words. Is Dr. Bacon honestly suggesting that the very simple fact that our life spans are continuing to lengthen, while at the same time as a society we're gaining weight, is in turn an argument that weight can't possibly be deadly? Isn't the whole point of HAES' existence to combat what HAES sees as correlations not being causal? Couldn't there be dozens, if not hundreds of other explanations for why our life spans are increasing despite our weight gains even if those gains did carry risk? Like for instance the very dramatic improvements in medicine that have occurred over the course of the past few decades?

Fighting misinformation with misinformation, relevant omissions with relevant omissions, and logical fallacies with logical fallacies, is not the way to accredit your movement, and if HAES has any hope of actually penetrating mainstream medicine, something I would dearly love to see happen, they're going to need to hold themselves up to at least the same, if not a higher level of scrutiny to which they hold others. If they don't do so, then their detractors will have an easy time dismissing them as champions of a self-serving, non-evidence based, over-hyped agenda, which ironically is the very same thing of which HAES is accusing mainstream medicine.

"Fat Children Eat Less Than Their Thinner Peers"!?

"Fat Children Eat Less Than Their Thinner Peers". That's the tweet Linda Bacon from HAES fame sent out to her followers at 9:55am yesterday morning.

It certainly fits the HAES narrative that the world's completely backwards in regard to anything and everything weight related.

Sadly it also continues Linda's confusing practice of tweeting bad data.

Linda's HAES platform, whether you agree or disagree with it, rests on the shoulders of her critical analysis of the medical literature on obesity, and her take is that many of the studies with which researchers and clinicians have vilified obesity, were either poorly designed or poorly analyzed.

But yet here's Linda promoting a Medscape news piece on a non-peer reviewed, poster presentation from a pediatric conference, where the findings are readily debatable.

The poster whose findings she was authoritatively retweeting, tracked the dietary recall of 12,316 children between the ages of 1 and 17 years of age.

Tweet and actual data accuracy aside, what the researchers truly reported wasn't that all fat children eat less than their thinner peers, but rather that the caloric intake of overweight and obese kids aged 1-7 exceeded that of their thinner peers but that the pattern "flipped" at age 7.

So what does "flipped" mean?

According to the poster, 9-11 year old kids with overweight and obesity reported consuming 1,988 calories daily, while their thinner peers reported consuming 2,069 (a difference of 4% which I'd venture isn't likely to be a statistically significant one). The study's 15-17 year olds with overweight and obesity reported consuming 2,271 calories daily, while their thinner counterparts reported 2,537 (a difference of 12%).

But can we really trust the dietary recall of children with overweight and obesity?

I'm not trying to be harsh. This world is not kind to overweight and obese children (or adults), and Linda would certainly know better than most of the stigma, bias and bullying those kids likely face on a daily basis - potentially even from their parents, their schools and their physicians. I don't think it would be an even remotely surprising finding that when participating in dietary recall surveys, children with overweight and obesity, especially older children who've had more time to experience hateful weight bias, might be more likely to under report.

So is there data to suggest that's a real possibility? Could these kids be under-reporting by more than the 12% seen in the oldest age group?

Absolutely.

In fact just this past February there was a review paper published in the International Journal of Pediatric Obesity titled, Assessing dietary intake in children and adolescents: Considerations and recommendations for obesity research. Regarding under-estimation, here's what the review paper's authors had to say,

"One of the most robust findings in dietary studies of children and adolescents is the positive association between under reporting and increased body fatness, particularly in adolescents (4,14,15). This is consistent with studies in overweight and obese adults (16). The extent of mis-reporting irrespective of weight status increases with age and has been reported as 14% of energy intake in 6-year-olds (17), 25% in 10-year-olds (18) and 40% (4,19) to 50% (14) in obese adolescents.."

The authors further report that the type of study most likely to suffer from under-reporting is the very type performed for the poster in question,

"Studies characterising under-reporting have focused on total diet assessment methods and in particular, energy intake"

To be clear, I think Linda Bacon's contribution to the field of overweight and obesity research, as well as public policy and attitude, is tremendously important. I just can't rationalize the scientifically critical Linda Bacon with her Twitter persona that seems to just retweet anything that satisfies the HAES narrative, no matter how weak or poorly designed the study (or in this case, the poster) may be.

Sigh.

There's got to be a better way to fight misinformation and statistically indefensible conclusions than the promulgation of misinformation and statistically indefensible conclusions.

Magarey, A., Watson, J., Golley, R., Burrows, T., Sutherland, R., McNaughton, S., Denney-Wilson, E., Campbell, K., & Collins, C. (2011). Assessing dietary intake in children and adolescents: Considerations and recommendations for obesity research International Journal of Pediatric Obesity, 6 (1), 2-11 DOI: 10.3109/17477161003728469

Did you hear the one about kids who eat candy being thinner?

I did.

I heard about it when Linda Bacon from HAES tweeted a link to a press release about it multiple times, calling it "Myth Busting". Knowing that Linda knows how to critically appraise a journal article, I figured it'd be worth reading the actual study.

I was wrong.

The study looked at one solitary day's 24hr. dietary recall collected from 11,182 children between the ages of 2-18 years of age, and then compared candy intake to overweight and obesity status in those same children.

Now dietary recall is known to be fraught with error, especially when it comes to less than healthy foods.

So is there evidence of error here?

Well according to their results, only 30% of children have candy on a daily basis, where candy means a sugar candy or chocolate (more on that in a bit).

That sure sounds like an awfully small number.

And of the kids who actually admitted eating candy, how much were they found to be eating?

One chocolate bar worth for teens aged 14-18 and about 2/3 of a chocolate bar worth for kids aged 2-13.

That sounds like an awfully small number too.

Now maybe kids really don't eat candy any more. Maybe the world's changed more than I've envisioned and only 3 out of 10 children eat candy daily, and do so in rather tempered amounts. And maybe candy's not only not bad for you, but it's good for you, specifically good for you in regard to weight in that this study found that the kids who reported eating candy, were 22 to 26 percent less likely to have overweight or obesity!

Of course the other possibility is that it's just an awful study that doesn't fairly lend itself to any conclusion whatsoever (pro or con).

And while we're at the awful study angle, given that this is a study where the authors' conclusion and the public relations spin is that candy's not bad for you, and apparently is protective against overweight and obesity, I think it's probably also worth asking what wasn't counted as candy?

Cookies, freezies, ice-cream, pudding, fruit roll ups, cake, pie, etc. Just chocolate candy and sugar candy. So what else wouldn't fit? Any other junk food - chips, pretzels etc.

Sigh.

The only conclusion I'm able to fairly draw from this study is that those of us who have any degree of Twitter influence, we really have to hold ourselves to a higher standard of retweeting. It's always tempting to retweet a press release or a blog post about a study that fits within our own confirmation biases, but before we do, we should really feel obligated to first read the actual study and evaluate it just as critically as we would those studies that don't fit neatly within our personal narratives.

E. O'Neil, C., L. Fulgoni Iii, V., & A. Nicklas, T. (2011). Association of candy consumption with body weight measures, other health risk factors for cardiovascular disease, and diet quality in US children and adolescents: NHANES 1999–2004 Food & Nutrition Research, 55 DOI: 10.3402/fnr.v55i0.5794