Friday, October 30, 2015

My Blog Has A New Home!

The Un-American Diet Blog has graduated to the big time and become a . . .

Featured Column on Science 2.0!!!

Which shows some of the content on this past blog, reconsidered and improved, as well as loads of new stuff. Most of the new posts are longer and better.

Now with 30% more SCIENCE!!

I will keep the Un-American Diet open as it still read by many and I don't want to lose friends or create dead google searches. But new stuff, from 2015 on, will all be published at Science 2.0.

This science blogging and news article host publishes on a broad range of topics in all science disciplines. Come check us out.

Thank you for reading and please come find all the new stuff!!


Monday, May 18, 2015

NuVal nutrition scoring: a practical review.

I first noticed the NuVal scores in my HyVee grocery store about four years ago. My first reaction was the same as how I feel about government: I'm in full support of universal rules for long as I get to make the rules. No grocery store dietitian is gonna tell me how to eat!

I began to look into it a little more deeply when I found out that the scoring system was developed by David L. Katz, who does real work in obesity research and whose perspective has always made good sense to me. The system of scoring is described in not-quite-enough detail on the NuVal website, but there are several links to peer reviewed research that examine how this system of  food labeling, based on one simple score that takes into account fat, protein, sugar, vitamins, salt, etc.. can aid consumers looking to improve their shopping choices. I am purposely not going to argue about the research, because this is supposed to be a practical test of how to use the score. Since the purpose of NuVal is to simplify shopping for health, no research should be required. In broad terms, NuVal is a ratio, with good things like protein, vitamins, fiber on the top, divided by “bad” things like sugar, salt, saturated fats on the bottom. The result is a score ranging from 1-100. From Coke to Spinach.

I do the shopping for a family of six. The two adults are mid-forties, normal weight, don't require medications for cholesterol or blood pressure and are still able to exercise as much as desired. The four children are all healthy and happy, tend toward the skinny side of normal, not prone to infections, score well on those goofy fitness tests they do in gym class. I figured, if someone was trying to promote some sort of universal healthy food guidelines for families, my family can't be too far from the end product they are envisioning. I’m not saying we’re Olympians or anything, but we are healthy people. If food is what drives our health and fitness and NuVal can sort out "good" from "bad" food choices, our grocery cart should be a pretty good test of the NuVal system’s logic.

Note the arrogance here. A reasonable person would probably check their cart against the healthy score developed by the team of experts, which has been validated in peer reviewed studies. My thought was exactly the opposite: NuVal ought to reflect our good health. If it’s going to be worth something, it should reflect that my grocery cart fuels an active healthy family. Either the food I buy for my family is what’s needed for health, or my family is so resilient that eating poor food choices can't compromise health. Or perhaps NuVal is inaccurate.

So here are the scores for today's shopping cart:

38-Grass fed beef
57-Organic chicken breast, free range, no antibiotics
8-Krusteaz buttermilk pancake mix (just add water!)
28-Nature valley chewy protein bars
6-Pop tarts, unfrosted
93-Country Fare quick oats
23-Corn flakes
25-Life cereal
38-Quaker oats and honey granola
8-Extra virgin olive oil
29-Skippy peanut butter
64-Pep farm ancient grains
24-Pep farm oatmeal farmhouse bread
26-Pep farm cinnamon raisin bread
25-Athenos Feta Cheese
23-Oscar Mayer Bacon
16-Philly cream cheese
25-Cheddar sticks
33-Org Valley lactose free milk
91-Silk original soy milk
Note that fat free milk is the only WIC approved choice. Hope that's the healthiest.
82-Org valley cow's milk, whole...Skim scores 100, but the kids say it’s watery and I agree
23-Fage Greek yogurt, 99 if fat free, we don’t buy fat free
23-for Noosa, with nutrition listed 11g fat (5 are sat) 30g carb, 14g pro, 30mg chol, 110 sodium, 28 sugars, 1g fiber. I’m at a loss for what’s wrong with this yogurt…sat fat?
47-Simply Orange, much better than a Welch’s concord grape drink which scored 2…
1-Coke, say it ain’t so!!!
3-Chips Ahoy
100-carrots, tomatoes, spinach, iceberg lettuce, bell pepper
94-for bananas, blueberries, raspberries, pineapple
32-crisco canola oil, 2 for butter

A straight average of these numbers yields a NuVal score of 63. What are we to make of this? Well, you can judge a lot for yourself just by looking at the list and come to your own conclusions about what’s healthy and what’s junk. Of course Coke and PopTarts aren't recommended foods and they are going to cancel out a couple of the positive choices (Truth be told, those two go in only one big male mouth in the family). I don't know what a "good" score is supposed to be, but a 63 probably isn’t it. If you remove the “bonus points” we get for buying the fruits and vegetables, the score is in the 30s for the rest of the cart. This may more accurately reflect how my family is actually eating, as the fruits and veggies likely account for less than 10% of what we actually consume, calorically.

The idea of simply scoring the whole grocery cart is riddled with problems that are not the fault of NuVal. In fact, they don’t really say that we are supposed to average our cart, that’s just something I thought of. Nevertheless, promoting a numbers-based system has got to result in using those numbers in some sort of logical fashion, including trying to find and improve one’s average. So I think the following limitations need consideration: These are the foods that the family needed this week. Many items we eat are still in the fridge or the cabinet, so aren’t counted. The foods I bought will be consumed at different rates. For instance, the milk will be gone in three days, the canola oil will last for months, half of the spinach will be thrown away by Thursday (it’s the truth, you know how slimy it gets). There is no weighting for how much of each food one would consume, or how they will be eaten in combination. We don’t consume individual foods, usually, but meals. In this system, there’s no way to pair the foods we bought.

Some problems are inherent to NuVal itself: I notice that there are very specific biases in the scoring, generally reflecting traditional concepts like “fat is bad, saturated especially,” “added sugar is different from sugar in raw forms,” “avoid salt” and “cholesterol will kill you.” Fruits and vegetables (and fat free milk for some reason) are the only foods to score 100. So, the easiest way to move your cart to a better score would be to buy more of these foods. But even with an extra refrigerator to keep bushels of veggies fresh for the week, a diet scoring anywhere near 100 would have a serious problem: not enough calories to sustain life (and tremendous flatulence for some of us, but that’s a topic for another day). The scores seem to rest very heavily on avoiding "bad" things (saturated fats, cholesterol, simple sugars) without much emphasis on putting in “good” things. In fact, aside from vitamins and fiber, I can’t actually figure out what the good things are in this system. We talk so much about avoiding obesity and diabetes that health professionals sometimes seem to forget a person actually needs energy in the food she eats, particularly if she is a growing kid on a soccer team. Vitamins are great, but they don’t give you energy. The 50 calories contained in a snack of two large carrot sticks isn't going to carry you through the second half. In fact, the caloric content of the food probably needs to be one of the main parameters in the numerator if we are trying to make a ratio of the needed attributes of food.

(I forgot to check diet Coke.  Does it score 100, like iceberg lettuce, because it doesn't have anything "bad" for us in it?)

I've had my resting metabolic rate tested. I've been measured, pinched, bio-impeded and Dexa-scanned to find my body fat percent. I've logged on a score of diet trackers and input my data into a host of online tools to get recommendations. So I am pretty confident that I need around 2400 calories a day to live the way I do. It would be less healthy for me to substitute low calorie options for what I eat. What combination of foods in the NuVal system could provide 2400 calories and give a score anywhere near the "A" range of 90 and above? Just trying to trade up on my sugars, the 280 calories I take in from coke per day would require three cups of blueberries (at a cost of about $12). You might say, "That's a good idea, you should do that. You'd be healthier." But that's not considering that I'm already eating some blueberries. They are part of the other 2100 calories I'm taking in. I don't want, nor does anyone need, three cups of berries a day. Just what, exactly, are we supposed to run on, in this system? If you were to eat only “A” foods scoring 90 and above, you would have a fairly limited diet. You’d be a vegetarian, for one thing. Skim-lacto, no-ovo, vegetarian. Nothing wrong with that, I guess, but it’s not a universal scoring system for guiding families at HyVee. More importantly, I don’t think it’s what the NuVal experts are trying to achieve.

As I looked about the grocery store at the scoring system, I saw some really strange things. Smucker’s Red Raspberry preserves are given the same score as Coke: 1. That may be fair, since they are both fueled by high fructose corn syrup, but the seeds in the preserves imply that they did start with some actual fruit, so is that really worth nothing? Apart from this, Smucker’s Sugar Free Red Raspberry preserves score a whopping 84! Splenda preserves are more nutritious than whole milk? Are you sure? We should eat this instead of an egg? At ten calories per serving, what am I getting for that 84?

To give Dr. Katz and his collaborators some credit, they seem to understand that there are limitations and that the scoring system will need frequent updates. The Q and A section of their website addresses some of these concerns and generally argues that: "Yeah, it's not perfect, but you gotta start somewhere and the food labels are pretty worthless." I think that's true. No one knows anymore whether it's trans fats or saturated fats or MUFAs or PUFAs or omega 3, 6, or 9s that are good or bad for us. We do need something simpler and NuVal is on the right track. I like the idea of being able to compare different versions of, say, Raisin Bran, with a simple score. But I believe the scores themselves need some work.

Here are my suggestions to make this a system that could really guide a father shopping for his family:

1. Realize that humans must actually take in adequate calories and include a factor that is based on “healthiest per unit energy.”

2. Nothing but Coke should score near one. I will still buy it, that’s what’s awesome about free will. But I will understand Coke to be a “one” item only when preserves are scored about 20. If you just throw all things with sugar in the same bucket, you won’t help me.

3.  Remember that we aren’t all 60 year old men trying to avoid a second heart attack. The science for controlling salt, cholesterol, fat (in children especially), even simple sugar, is not sound enough to dominate the scoring system.

4. Perhaps score only within categories? Best peanut butter? Best cut of meat? Best breakfast grain?

5. Reconsider the emphasis on included vitamins, so that Hershey’s syrup can’t get upgraded to a 10 for including Calcium.

6. Review the biases of the working group. Expert agreement is not the same as scientific consensus.

All of these criticisms aside, I really believe that NuVal is onto something big. It is an easier system (and more up to date) than the pyramid or the plate promoted by USDA. Since they are out ahead of other scoring systems, the NuVal team will likely be the best positioned to incorporate new science. They are not conflicted with lobbying from the cattleman’s association and the other political issues that plague government teams trying to do similar work. Nor do they fall prey to unscientific concerns whether things are “organic” or “GMO-free.” For this they deserve a lot of credit. I think the idea of scoring from 1 to 100 is a good one; it’s easy for anyone to understand and does lend itself to comparison and “trading up.” But that trading up can’t be based on substituting Splenda.

Saturday, February 28, 2015

Book Review: Eat Bacon, Don't Jog by Grant Petersen

Grant Petersen is just a guy. He's not a doctor, a dietitian or a nutritionist (although I must confess to being personally confused as to what that last term means...and I think I might be one), but he has put together a concise, accurate, extremely readable guide to weight loss. I picked it up in a hipster bookshop in Minneapolis, when I was, this will shock you, unable to get back home due to snow. My initial reaction was, "Damn it! Awesome Title. Why couldn't I have thought of that?" My second reaction was, "isn't he that bike dude?"

As it happens, he's not really "just a guy."  This is Grant Petersen's second book, the first being Just Ride, which is an adult bike riding manifesto. Peterson is the founder of Rivendell Bicycle Works, which manufactures and sells mostly beautiful, old-school, steel frames meant to be ridden by people who have gotten over the delusion that riding a bike requires lycra jerseys and electrolyte solution (don't get me wrong, I'm still one of those lycra assholes, but I totally understand his message). What really comes through in Eat Bacon, Don't Jog, is that Petersen has been doing his homework on current obesity science. The title suggested this and the book confirms it. In a breezy but authoritative tone, he uses the same format as Just Ride (1-2 page reflections) to cover a lot of ground. He supports the (now becoming dominant) school of thought that obesity is caused by:

1. High carbohydrates in the diet, which cause,
2. High insulin levels, which
3. Drive storage of energy into fat cells

He therefore outlines a reasonable low carbohydrate diet to reverse this process, with suggestions for substitutions and good quick rationales for each of his recommendations. In the process, he does a lot of debunking of standard wisdom, without including perplexing, dull reviews of the scientific literature (for that, please see my other blog posts). For example, #45: If you can't jog yourself skinny, why are the world's best distance runners skinny? In examining this question in three quick paragraphs, he explains how varying sports select for the body types that work in that sport, rather than the reverse. He then briefly points out one science-y fact: endurance training causes breakdown of muscle protein for making needed sugar, giving serious runners that Kenyan look. Very brief, two facts and an insight per page, next subject.

Simplicity, brevity and breezy style are hard to keep going if one wants to remain true to scientific accuracy and occasionally the book suffers from the drive to make points quickly, such as when certain foods are described as "evil" or to be "avoided at all costs." But his tone is such that the reader is aware of the hyperbole and unlikely to be misguided. If I have one criticism of the book it would be simply that it falls into the "pretty fit guy gets a little squishy, loses some weight and decides to tell the world about it" genre of diet books. It is a genre-mostly supported by doctors who had a tiny tummy at one point in their lives. There's no hint here that he's grappled with many people trying to actually eat 20 grams of carbs or less for any period of time, or studied what it's like for someone who weighs a few hundred extra pounds. I have and I've seen that the low carb philosophy, while fairly sound scientifically, is more difficult for those who really need it: the very heavy diabetic individuals whose bodies don't handle sugar well. Patients who can follow it are the exception not the rule. But that's the doctor in me speaking and as he points out, "how helpful has your doctor been?"

This will be my go-to book recommendation for anyone who asks me what the current thinking is on weight loss. Grant Petersen has summarized it neatly, in good style, in as tight a package as you could ever hope to find.

Thursday, January 22, 2015

Diet Soda Changes Gut Bacteria

Diet Soda Changes Gut Bacteria

Many people, including, I'm guessing, most weight loss doctors, have an instinct that diet sodas are a bad idea, but it's been hard to find research that would back this up. While it's true that diet soda and obesity are associated, it seems reasonable to suspect that people turn toward diet drinks because they are getting heavier, rather thinking the weight gain is in any way due to the soda. A previous myth was that the brain-gut-pancreas system couldn't distinguish artificial sweeteners from the real thing, so that drinking diet soda would cause insulin spikes, that over time would lead to obesity. That's been disproved. However, a scientifically sound reason to avoid diet soda has been found.

In this thoughtful series of experiments reported in Nature last fall, the researchers show that diet drinks may sabotage weight loss efforts and hasten diabetes onset through encouraging the growth of unwanted species of bacteria.  Suez and colleagues studied whether saccharin, sucralose and aspartame would alter the composition of the bacteria that live in 10 week old, lean mice.  They compared the artificial sweetener group to controls drinking water, water + sucrose and water + glucose. The mice drinking water, sucrose or glucose had similar glucose tolerance curves falling in the normal range. They continued to handle blood sugar well, so would be low risk for diabetes. All three of the artificial sweetener groups developed glucose intolerance within weeks. When they combined the saccharin feeding with a high fat diet, they produced obese, insulin resistant mice. They went on to examine the composition of microbiota in the feces of the mice and found that the artificial sweetener groups had overgrowth of unnatural bacterial species of many types. They then demonstrated that they were able to transfer the glucose intolerance from the sweetener group to the healthy lean group by fecal transplantation. The researchers then went on to treat the mice with antibiotics (ciprofloxacin and metronidazole) and found that the changes in tolerance to glucose were reversible with antibiotics.

This is a remarkable series of findings which strings together several proposals:

            1. Artificial sweeteners cause glucose intolerance
            2. The glucose intolerance is mediated by gut bacteria
            3. Glucose intolerance is "infectious" through fecal transfer
            4. It is reversible by antibiotic treatment.

So which comes first, obesity or diet soda drinking? This study didn't directly address that, but it gives a biologically plausible explanation for ill effects of artificial sweeteners on our metabolism. With regard to obesity specifically, the experiments have been fairly convincing that bacteria play a role (for example: V.K. Ridaura et al., Science 341, 6 September 2013)  In mice, there is evidence that fulfills roughly all of Koch's postulates: fecal transplants from lean and obese humans will produce lean and obese mice with differing intestinal flora. If you take a lean mouse and have him live with the obese mice, he will not grow heavy, but if you take an obese mouse and put him to live in the cage with the lean mice, he will lose weight. This bodes well for the pressing problem of mouse obesity.

In humans, the prospect of fecal transplant is being tried experimentally and when it becomes accepted, I will try to be one of the first to open a walk-in Eat S*#T! clinic on a busy urban street. Until that time, we are stuck trying to guess which dietary factors to manipulate in order to change the bacteria from "obesity prone" to "obesity resistant" colonies. This paper lends support to the idea that bacteria affect our weight and glucose handling and that it may be worth actively searching out certain microorganisms to aid our metabolism.

Still, It's a long way from there to an obesity cure. I've always been sort of amazed at people's willingness to ingest yogurt because it has active cultures of bacteria in it. I think a bit of skepticism is warranted by consumer. If it is a good bacteria in the yogurt, good for what? And shouldn't one get a stool sample to see what one's levels are before trying to add an organism to the mix? What would be the right dose? Can we just hope that the bacteria will duke it out and that some magical balance will occur with more "good" bacteria than "bad"? Are you sure that you are in need of more L. delbrueckii subsp. bulgaricus in your diet? Do you trust the yogurt makers to decide that?

Don't get me wrong, I like yogurt, but not for the bacteria included.

If there were to be clinics which specialized in fecal transplant, before and after tests would likely be routine, so maybe over time, we would find the answers to the above questions. For now, we will have to wait for some human trials to reproduce the mouse proof of concept studies.

Wednesday, January 21, 2015

Can Sugar Free Red Bull Aid in Weight Loss?

People seem to have one of two opinions regarding Red Bull: either you hate it, because it tastes like someone added carbonated water to a fistful of SweetTarts, or you love it, because it tastes like someone added carbonated water to a fistful of SweetTarts. I myself fall in the second camp, but don't drink Red Bull regularly because it sends my heart rate up enough that I fear it contains something illegal, or at least dangerous. The non-caffeine stimulants in the drink are what give me pause: taurine, glucuronolactone and B vitamins. These are what distinguishes the "energy drink" beverages sector from the soda sector. The presumed increase in energy from the alternative stimulants is their appeal. That combined with the fact that people base jump with Red Bull parachutes, so the brand is undoubtedly "cool."

With regard to diet drinks, the hypothesis is that taking in caffeine without any sugar should lead to weight loss. While normal high sugar sodas contain enough calories to more than counterbalance the benefits of the caffeine, sugar free sodas should, theoretically, provide a net calorie burn. Caffeine increases metabolic rate, so a calorie free caffeinated beverage should help weight loss (It doesn't, but that's discussed in the next post). What we are discussing here is whether Red Bull's other ingredients do anything in addition to the caffeine.

The authors took healthy young men who were daily caffeine drinkers and gave them either sugar free Red Bull or the same amount of caffeine in simple water. This was designed to test whether there was anything else, besides caffeine, changing metabolic rate. What they found is that the Red Bull drinkers and the water + caffeine drinkers experienced an equivalent rise in resting energy expenditure of 4 percent that lasted about two hours. There was no difference between the two groups, so the researchers concluded that the taurine, B vitamins and glucuronolactone did not have an effect on energy. 

Red Bull, Monster, Rock Star and 5 Hour Energy are all utilizing caffeine to give an energy boost, or even a "rush" that we will view positively or negatively based on our view of how it feels to be artificially sped up.  They are different from sodas not due to the fact that they contain a collection of other purported stimulants, like Taurine or Guarana, but simply because they contain the equivalent of three or four sodas in each serving. Instead of a soda, you are getting something that tastes sweet but has the caffeine of a double espresso.

Getting back to weight loss, are sugar free energy drinks a good idea? The 4% increase in resting energy expenditure seems promising, but we need to recall that our resting energy accounts for roughly 60% of total energy. If we raise that 60% by 4% for two hours, we would likely not make a dent in the overall daily numbers. If we were to drink the Red Bull every two hours while awake, a 4% change in energy would certainly be significant, but then we'd run into an unpleasant side effect profile that most coffee drinkers know well: irritability, lack of focus, racing heart, impulsive online purchases (that may just be me) poor sleep, etc. 

It's worth noting that majority of over the counter weight loss products use caffeine  as the main active ingredient, including many that the FDA has banned due to cardiac side effects. But assuming one is only going to have one or two Red Bulls or other energy drinks per day, I doubt there's any harm. If you like the taste, go ahead. Just don't expect to see a major change in your weight when switching from the sugared to the sugar free version.

Friday, June 27, 2014

Ketosis Does Not Improve Weight Loss

This is part 2 for "Adherence predicts weight loss."

Almost every physician believes that weight loss is about "calories in/calories out." We learned in med school that the ATP cycle can be driven by fat, carbohydrate or protein and that, as far as our mitochondria are concerned, "a calorie is a calorie." As a physician trying to uncover the mysteries underlying weight gain, over the last ten years, I've searched far and wide for other explanations for obesity and found many leads, many interesting side stories and many dead ends. I've learned that the macronutrients (carbs/fats/proteins) are processed differently in the gut, have differential efficiencies for creating energy and may serve different roles in health. I've been inclined to discount the pure mathematical truth I learned in medical school and to espouse a philosophy a little bit outside of pure science: Carbohydrates are the cause of obesity.

For many readers this may seem obvious. For trainers and weight loss doctors, this is a statement of fact. Most assume you can't be lean unless you cut carbs. But for physicians, who read randomized controlled trials, meta-analyses and evidence-based guidelines...the low carb-diet trend is unsound, a fad, not scientifically validated and perhaps, quackery, sold by doctors who want to get famous (Atkins, anyone?). It was with great trepidation that I slowly adopted a "carbs are evil" message in my obesity clinic. A little sheepishly, I began to suggest to patients that their weight problem stemmed not from how much they ate, but what they ate instead. I initially was quite embarrassed to espouse these views as they flew in the face of most randomized trials which showed results that seemed to never back up low carb claims.

When I suggest to patients that they consume a higher protein/lower carb diet, it does, in fact work. It's a good way to eat to lose weight. But we need to ask, is it because you eat fewer calories, or because something changes metabolically to make weight loss easier on this diet. The explanation used by Atkins and most low-carb advocates is that, without carbs, the body must use fat as a source of calories. The brain must have sugar or ketones to function. Protein can't get in there. Generally, it's easier for the body to convert fat to ketones than it is create sugar through disassembling the carbon atoms in protein, so it will go into "ketosis" after a few days after removing carbs from the diet. "Ketosis" is considered the magical land of "fat burning" and is generally the "metabolic switch" that fad diet books talk about.

As a follow up to the previous post on adherence, I now have mathematical support that ketosis does not represent a state of "fat burning" beyond what would be expected from calorie deficit. As I mentioned, we tracked clients on a minimal carb (less than 50g), higher protein (over 100g), low calorie (about 1000) calorie diet and compared their weight loss over various periods with predictions from a purely mathematical model. The model has a slight increase in weight loss for low carb based on the actual weight of glycogen and can account for salt, but the few pounds are a question of water balance, not anything metabolic, like a "faster metabolism." In short, the model doesn't "believe" that ketosis has anything to do weight loss. Weight loss is caused by calorie deficit, period. The model is, basically, a big bummer for weight loss doctors like me, who want to espouse special knowledge.

So, why do I say that ketosis doesn't matter? Because without accounting for any acceleration in fat burning due to a different metabolic process, the model predicted weight loss nearly perfectly for the participants. There was not one patient who had a surprisingly rapid weight loss that would suggest any metabolic magic was causing an increase in fat burning. Using simply "calories in/calories out" math, we were able to predict how the individual bodies would react. The model, knowing nothing of mitochondria, insulin, leptin, ghrelin, CCK, PPY, Agouti-related protein, NPY, hormone sensitive lipase (I could go on) - the model predicted weight loss just using calorie deficit. None of the hormones that interest obesity researchers mattered in this analysis. Knowing calories and some basic facts about the size, age and activity of the clients was enough. Certainly nothing as simple as "turning on ketosis" contributed to the results.

So what are obesity experts talking about?

The researchers aren't trying to figure out "why" we gain weight. They are doing what scientists do. They are working out "how." All those internal hormonal signals certainly are the controllers of our metabolism, but they merely explain "how" it happens. The "why" is still all about calories. The answer, until someone develops a suite of medicines to manipulate the above hormones, will remain calorie restriction.

So, does this mean we just need to reduce calories and forget about trying to find a diet with the right composition? Probably not. Just because a lower carb/higher protein diet doesn't turn on a magical weight loss switch doesn't mean that it isn't helpful for a different reason: this diet makes it easier for many people to eat fewer calories. Getting rid of carbohydrates helps curb cravings. Eating protein makes you feel full. Taken together, those two things can help you eat less. But that's probably the extent of the magic.

This is not a conclusion that makes me happy to realize, nor does it help me advertise my weight loss clinic. Which is why I'm pretty sure it's accurate.

Monday, June 9, 2014

Adherence Predicts Weight Loss, Part 1

I've recently had a very unsettling experience
I've just published, collaborating with Kevin Hall (whom I mentioned in the blog pieces on Obesity Math), an article testing whether simple "calories in/calories out" calculations are useful. The article, "Is Mathematical Modeling Applicable to Obesity Treatment in the Real World?"  is out in the June issue of the journal Obesity. The unsettling aspect of the paper and the analysis of our clients is that when you can assure compliance to a weight loss program, the body acts pretty much like a machine.

We looked at a database of clients participating in a supervised medical weight loss program with meal replacement. This program is terribly strict, supplying most of the food to the clients and asking only that they supplement with two cups of fresh vegetables and a "real" protein source of chicken, lean beef, or fish for dinner. Strict adherence to the diet provides between 900 and 1100 calories. It's a low carb, low fat, low calorie diet, that is, by design, also a higher protein diet...since you cut everything else.

Whether the clients are adhering to the diet is pretty simple: if they lose a ton of weight, they are adherent. This is way too few calories for pretty much everyone, so if your patient comes in having stayed the same for the week, you can be pretty sure there was some deviation from the program. We reviewed records from over 100 patients and found 49 for whom compliance seemed very tight for at least four weeks, based on the clinic notes which review the patients' self assessment on how close to protocol they've been. Then we asked:

If one is extremely compliant to a rigid diet program, does the weight loss match what a mathematical model would predict?

The reason this is even in doubt is that pretty much every mathematical model will over-estimate the weight loss seen in a real, outpatient, free-living subject. Kevin and I wondered if that was due to some inherent flaw in these models, or whether free-living people are simply never compliant with recommendations (actually Kevin probably didn't wonder at all, because he's a scientist, but I'm just a clinician, so I really wondered if there was some mysterious explanation of variable weight loss outcomes). With this very strict regimen and the close follow up, we were able to see how the weight loss of truly compliant patients added up. We plugged patients' height, weight, age and activity level into the NIH body weight simulator and asked the program to predict how individual's weight loss would progress on the reduced diet. We excluded patients who admitted to straying from the diet.

The unsettling part is that the weight loss for patients who were able to be strict with themselves was essentially exactly as predicted. As the paper shows, the correlation between predicted and actual weight loss is remarkably close (R-2 of .816). Basically, for everyone who could maintain this diet (and don't get me wrong, I could never maintain this diet, I've tried and lasted 6 days on my best attempt), the body works in fairly perfect thermodynamic harmony. A calorie, unfortunately, and I say this with great regret, because it breaks my heart so, is actually, a calorie.

Why do I find this unsettling?

Because for me, it's quite a surprise. I've made a career out of weight loss counseling and that involves being able to work with people month after month, whether things are going well or not. I don't demand rigid adherence, but simply a good attitude and the ability to show up for the recheck appointment. I have patients who plateau on weight loss for months and months. I spend a good deal of time speculating on the biological basis of plateaus, since we do in fact, have evidence that many hormonal controls in the body react to counteract weight loss (leptin decreases, ghrelin increases, etc). What the data in the paper tells me is that, regardless of the biology, plateaus and rebound weight gain come through the patient overeating again. When we are consistent, the weight loss will not, cannot, fail. If the data in the study came back all over the map, not in line with the computer model, I would have said, "see, the body is complex, everyone is different, we all need individual solutions relating to our unique biology." But the opposite seems true: the bodies of our subjects behaved almost exactly as expected, with almost no variation.

The question turns from "how do we better understand the biological signalling responsible for weight control?" to simply "how do we get people to follow the rules?"