Bottom line: This post contains two emails from me to a med student friend who recently read Gary Taubes’ Good Calories, Bad Calories (GCBC).
A nutritionally-aware med student friend of mine recently read GCBC and seems to have been as pleasantly surprised as I was when I first read it. However, I’ve learned a lot since then. This post has a few emails I’ve sent her regarding the idea of low carbohydrate dieting. I thought they were sufficiently informative to warrant a blog post. If you disagree, I’d love to hear why in the comments section below. The emails will be minimally edited, so I apologize in advance if parts don’t make sense because they are out of context.
I really enjoyed GCBC, and it taught me a ton that has been useful. More than anything, it encouraged me to be skeptical of medical dogma. Not sure if you remember my questions regarding the unique mortality advantage of statins (with their pleiotropic effects) vs any other cholesterol lowering med during CVPR, or numerous questions regarding Ancel Keys and his disturbingly authoritative methods of establishing much of our current medical recommendations… but these thoughts were in large part set in motion by Taubes.
That said, Taubes is an engineer / journalist by training, and it shows. His critiques on some of our dogmatic beliefs are well-founded. His attacks on methodologically flawed research are astute (who know the cholesterol hypothesis was mostly founded on rabbits – herbivores?). However, many of his propositions and alternative hypotheses reveal his utter lack of formal education in biological sciences. ”Cellular starvation”? Really? He inappropriately applies in vitro data (kinda hypocritical). He fails to mention that HSL activity is significantly impeded by not just postprandial insulin levels, but by basal insulin levels.
The most fundamental problem that Taubes tries to ignore or deny is that the calorie balance “hypothesis” (sloth / gluttony) is absolutely, undeniably true. People cannot make fat out of nothing, people cannot create heat or kinetic energy out of nothing. Up to 99% of massive oral fat loads (thousands of calories) are absorbed with no problem and minimal stool losses, and urinary / respiratory ketone losses in physiologic ketosis (not talking DKA) are only a hundred or so kcal per day. That means that weight loss on low carb diets (LC) is still a matter of input vs output. Interestingly, most of the outpatient research regarding calorie balance and weight loss shows a huge advantage to LC vs LF dieting, at the same reported calorie intake. This difference generally disappears in inpatient or metabolic ward studies, where input and output are meticulously tracked through scales / bomb calorimetry / indirect calorimetry and respiratory quotient, or even doubly labelled water.
What this means to me is that for some reason people are less able to accurately track their calorie intake on a LC vs LF diet. At least that’s the only good explanation until recently. And, to tell the truth, it kinda makes sense. Some of the literature shows a_spontaneous_ reduction of about 30% of total kcal on a LC diet, with no reported increase in hunger. People that are effectively 30% less hungry may also be less likely to “fudge” their reporting. And it’s well known that people suck at estimating their intake – even registered dietitians aren’t very good (1).
The alternative is that there really is some kind of “metabolic advantage” to LC dieting. Several authors promote this idea (for some reason I find tons of COI in nutritional literature. It’s curious how often Googling the authors of a pro-LC article turns up someone that is selling a LC dieting book), but until recently the fine-tuned machinery of the metabolic ward has been unable to demonstrate a statistically significant difference in the metabolic rates of LC vs LF dieters when their diets are standardized for protein content (remember that “futile cycling” / gluconeogenesis / lack of storage depot makes protein about twice as metabolically expensive as fat or carb, ~30% vs ~10–15% “thermic effect of feeding” TEF, but it’s totally possible to do low fat high carb high protein dieting, so LC and high protein are not necessarily synonymous). So if researchers have been unable to detect a difference in metabolic rate with LC vs LF isocaloric isonitrogenous dieting, that means that either A) “metabolic advantage” doesn’t work except to sell books or B) our instruments are not sensitive enough to detect the difference. Given the high sensitivity of our instruments, if even if B is true, the difference may not be enough to be meaningful. But the reason I keep saying “until recently” is that there is a new paper out that suggests B may be the case (2). Granted this paper seems to have a pro-LC bias, and it has not been replicated in similar work, but who knows. If it is replicated, we might finally have a better idea why LC seems to work better than LF in all of the studies using free-living individuals (assuming that the metabolic ward studies are too short in duration to manifest the difference as statistically significant discrepancies in weight loss).
Wow, I sure can rant. What are your thoughts? Was there anything in particular that surprised you about the book? I’m sorely disappointed that he used a bibliography without in-text citations; it makes verifying or refuting his claims very inefficient. A quick google for “taubes criticism” turns up interesting results. If I remember correctly, Taubes even got a response from Bray himself, who’s like the godfather of obesity research in the US (3).
1. Champagne, C. M., Bray, G. A., Kurtz, A. A., Monteiro, J. B. R., Tucker, E., Volaufova, J., & Delany, J. P. (2002). Energy intake and energy expenditure: a controlled study comparing dietitians and non-dietitians. Journal of the American Dietetic Association, 102(10), 1428–1432.
2. Ebbeling, C. B., Swain, J. F., Feldman, H. A., Wong, W. W., Hachey, D. L., Garcia-Lago, E., & Ludwig, D. S. (2012). Effects of dietary composition on energy expenditure during weight-loss maintenance. JAMA: The Journal of the American Medical Association, 307(24), 2627–2634. doi:10.1001/jama.2012.6607
3. Bray, G. (2008). Good calories, bad calories by gary taubes; new york: Aa knopf. Obesity Reviews, 9(3), 251–263.</blockquote>
Enjoyed your email! I agree with most, but there are 2 points to which I’m compelled to make a quick rebuttal. Please keep in mind that I absolutely love this stuff – it’s the reason I tried out for medical school in the first place – and I love learning about it even more. I have not gotten to spend the time to keep up on all the newest literature since starting med school, so I ask that you please point out details that I’ve gotten wrong. On the flip side, people sometimes get offended if I tell them I think they’re wrong. Please don’t – I think you’re brilliant and a fantastic person.
- Carbs / insulin / obesity. Although it does make sense that high carbs -> high insulin -> obesity, this is not true, and we as medical professionals should work to fight this fallacy that permeates the popular nutrition literature. I promise that anyone in primary care will encounter patients (and other providers) that have read this in a pop science mag. Hyperinsulinemia does *not* lead to obesity in the absence of hyperalimentation (though it’s possible that it leads to hyperalimentation, but this is in accordance with and supports “calorie balance hypothesis”). Remember insulin has peripheral effects in HSL but it has profound central effects on satiety (promoting it). Perhaps that’s partly why the highest scoring food on the “satiety index” is a potato (pubmed link — [full text link redacted, try Google —n8henrie Sat Aug 4 19:54:48 MDT 2012]). Points of evidence:
- Protein is super insulinogenic, in many cases as much as CHO (something else they didn’t teach in biochem, right?). The first paper below is one of my all-time favorites (actually one of the first papers I ever found a “full-text” copy of — I used to print them and keep them in a binder haha). Therefore, if insulin were sufficient for high-carb obesity, we should also observe this in high-protein diets (but the opposite is found). Not only is approximately 50% of glutamine (the major amino acid in meat protein) hepatically converted to glucose, but also leucine and TCA cycle intermediates are potent insulin secretagogues.
- Holt, S. H., Miller, J. C., & Petocz, P. (2006). An insulin index of foods: the insulin demand generated by 1000-kJ portions of common foods. The American journal of clinical nutrition, 66, 1264–1276.
- Bao, J., de Jong, V., Atkinson, F., Petocz, P., & Brand-Miller, J. C. (2009). Food insulin index: physiologic basis for predicting insulin demand evoked by composite meals. The American journal of clinical nutrition, 90(4), 986–992. doi:10.3945/ajcn.2009.27720
- MacDonald, M. J., Fahien, L. A., Brown, L. J., Hasan, N. M., Buss, J. D., & Kendrick, M. A. (2005). Perspective: emerging evidence for signaling roles of mitochondrial anaplerotic products in insulin secretion. American journal of physiology Endocrinology and metabolism, 288(1), E1–15. doi:10.1152/ajpendo.00218.2004
- It’s getting somewhat popular to study insulin levels in diet studies. Curiously, it’s hit-and-miss whether or not insulin levels actually decrease during LCHO dieting, and from what I’ve seen this is independent of weight loss.
- Inhibiting insulin secretion doesn’t lead to weight loss. (Mannoheptulose is an interesting sugar found in decent quantaties in avocados that does this — but probably not clinically relevant quantaties).
- Due, A., Flint, A., Eriksen, G., Møller, B., Raben, A., Hansen, J. B., & Astrup, A. (2007). No effect of inhibition of insulin secretion by diazoxide on weight loss in hyperinsulinaemic obese subjects during an 8-week weight-loss diet. Diabetes, obesity & metabolism, 9(4), 566–574. doi:10.1111/j.1463-1326.2006.00645.x
- Johnson, B. F., & Wolff, F. W. (1970). Trial of mannoheptulose in man. Metabolism: clinical and experimental, 19(5), 354–362.
- Viktora, J. K., Johnson, B. F., Penhos, J. C., Rosenberg, C. A., & Wolff, F. W. (1969). Effect of ingested mannoheptulose in animals and man. Metabolism: clinical and experimental, 18(2), 87–102.
- Insulin-mediated inhibition of fat use is considerable at even *fasting* levels of insulin. Keep this in mind when you’re reading unreferenced material about insulin and carbs, because the fanatics will often pull from in vitro studies using zero-insulin states. This is not physiologic, and is almost totally irrelevant to human physiology. If there’s anything I learned from GCBC, it’s that humans are not animals are not in vitro cell studies are not biochemical studies. Insulin levels in fasting humans are nowhere near zero (why else would we pay so much for Lantus?).
- If you haven’t heard of it (I don’t recall this ever coming up in MS), read about acylation stimulating protein (ASP). Ultra-important and potent fat-trapping enzyme. Eating carbs and no fat you block burning endogenous fat, eating fat and no carbs your body just gets better at storing ingested fat.
- High carb diets lead to weight loss just fine — as long as they’re hypocaloric.</ol>
- I didn’t mean to say that LCHO dieting lacks variety! What I meant was that — by definition — limiting anything in a diet limits variety, compared with limiting nothing. And I find that strict carb restriction is more limiting than strict fat restriction, partly because low fat dieting has been en vogue for so long that there is a “low fat” option for almost everything. Low carb is definitely coming along, but I don’t think it’s quite there yet. If we were to play a game, where we made a list of foods, traded lists, and the other had to respond to each item with a trace-carb or trace-fat alternative… I think low fat would “take the cake,” so to speak (sorry, had to).
- I think this matters for many people that try to diet by “reducing portions” without actually setting any strict rules.
- My cliché example I always use here is that if I put people on an “all red” diet, I bet they’d lose tons of weight in spite of stuffing their faces with tomatoes, raw meat, skittles, catsup, and kool-aid. You’d get bored really quick, and return to eating for hunger instead of taste.
- I’m biased based on a couple months in South America, where I thought the mainstay dishes of beans and rice were delicious, so I’d have an extra helping (we had them at pretty much every meal). The locals thought I was crazy and frequently wouldn’t even finish their first helping (bored, lack of variety). They were thin, I gained weight.
- Off topic, but our ancestral diet shows some astounding variability. In spite of the “paleo” fanatics that think we all have a lizard deficiency and should eat all-whale-blubber-all-the-time, some hunger-gatherer societies ate incredible amounts of honey, and most estimates I’ve seen suggest that a moderate carb intake (40%+) was pretty routine.
- Additionally, it really matters how strict you are with your carbs. I did 16 weeks at <30g / day, and I’ve done a month or so at <10g / day. Even cashews and peanuts were off limits at this level, as were most vegetables, all fruit. Heck, I was counting the carbs in my eggs and beef (glycogen). For me, one of the most effective and psychologically challenging features of such a diet was limitation of variety. Perhaps not if I had a chef and lots of money, but there is interplay between time, money, and dietary variety, and a change in one will often affect the others. </ol> Well, I truly intended for this to be a brief response (I only had 2 points to rebut — and even numbered them), but again I fail. I’m 100% serious when I say that I decided to go to MS because I was spending hours and hours per day on this stuff anyways, and it seemed like a waste not to put it to use. Here I am going into emergency! haha. Well, I’m thinking of taking these two emails I’ve sent you and compiling them into a hasty blog post. Hope that’s okay with you. -Nate</blockquote>