2 New Diet Studies

*ugh* journalists

I’m talking to you, Mandy Oaklander!

Regarding the new low carb vs low fat study, she writes: “Popular diets are pretty much the same for weight loss, study finds.

Effects of low-carbohydrate and low-fat diets: a randomized control trial (Bazzano et al., 2014)

Further, “An earlier study in Annals of Internal Medicine did find that low-carb dieters lost slightly more weight than low-fat dieters after one year. The study today reached similar conclusions, but the differences in weight loss were not significant.”

Perhaps Mandy just doesn’t realize there’s a difference between significant, as in “meaningful,” and significant, as in “P<0.05.”  Pro-tip: you can tell them apart relatively easily, because the latter is usually accompanied by a cute little asterisk.  For example, the differences in weight loss were quite statistically significant (P<0.05):

Bazzano BW
She goes on to say “After a year follow-up, some of those pounds crept back for people on both diets…”

To that I say: yeah, but fat mass continued to decline in those on the low carb diet, meaning some of that weight re-gain was muscle:

Bazzano FM

So, between 6 and 12 months, carbs and calories were creeping up in the LC group, yet fat mass was still declining.  Perhaps this way of eating improved their metabolism, or restored the ability to effectively partition nutrients.

***in real-time: at this point, I realize that Mandy was actually talking about the other study, which she was covering accurately.  Sorry, Mandy!***

Bazzano PA

…so maybe the low-carb (LC) diet improved muscle mass because it was also high protein? …perhaps, but 19% vs 24% (71 vs 85 grams) isn’t a very big difference.  Alternatively, since the LC group really just maintained absolute protein intake (86 grams at baseline, 85 at month 12), whereas low-fat (LF) dieters decreased (86 grams at baseline, 71 at 12 months); perhaps this is why LF lost muscle mass..?  Still, those changes in protein intake are small, and I think people can be too quick to chalk up the benefits of LC to “high protein.”

In sum, this is actually one of the more “pro” LC studies.  And it wasn’t even a huge difference in carbs: 198 vs 127 grams/d at month 12 (54% vs 34%).  Big difference in fat mass; and CRP, a marker of inflammation, even declined in the LC group.

Low fat diet advocates have been giving me headaches for years… the low fat diet caused headaches (P<0.05):

Adverse Events 1



Adverse Events 2

The study Mandy was actually talking about: Comparison of weight loss among named diet programs in overweight and obese adults: a meta-analysis (Johnston et al., 2014)

It was a meta-analysis, which is just about the only type of study capable of taking down LC.



…but at least it had this cool chart (modified):

cool chart (modified)

cool chart (modified)


*ugh* scientists


The macro’s in “Low fat” overlap with “Moderate,” implying “Low carb” is “EXTREME”  …the authors’ bias is subtle, I’ll give ‘em that, but I’m getting too old for this.

Dear Obesity Researchers,

If you want to design a study showing a low fat diet is as good as low carb for fat loss, here’s your best bet: recruit young, exercise-tolerant overweight patients who aren’t on any meds.  PROOF (see Ebbeling study).  Or find 10 similar ones and write up a pro-LF meta.

If you want to show low carb is better, recruit patients with obesity.


calories proper

Be Sociable, Share!
  • Thomas Hemming Larsen

    Instead of always focusing on weight loss, researchers should do more on trained or normal weight persons (like the gymnasts) to see what happens on different diets.

    Btw, https://www.linkedin.com/pulse/article/20140902121017-23027997-diet-research-stuck-in-the-stone-age .

    • johnnyv

      Not much money for such a small proportion of the population. Studies on non novices take too long/too many participants to reach significance = $$$$.

      • http://www.caloriesproper.com/ William Lagakos

        ^^^ exactly

      • Thomas Hemming Larsen

        Yeah, but I still think it is relevant.

        • CynicalEng

          Trained and normal weight persons aren’t at health risk or costing healthcare systems money, so nobody is especially interested in their response to different diets.

          • http://www.caloriesproper.com/ William Lagakos

            Some journals, like Strength & Conditioning or Sports Sciences publish a lot of cool studies, but you can always tell very low budget.

          • Thomas Hemming Larsen

            I’m not sure I completely agree with that. I agree that obesity is a huge problem but there are other diseases in the world. I’ve seen people with parkinsons, cancer etc. who were ‘fit’ and lean.

            Besides, I’m obviously biased because I’m not overweight :)

        • http://www.caloriesproper.com/ William Lagakos

          If you didn’t know how a healthy liver worked, you wouldn’t know a sick one if you saw it. We need to understand how biology operates on all levels, not just obese IR.

          • johnnyv

            Absolutely, however the most funding for studies comes from the eat less, move more and eat your healthy cereals school of thought government. Body composition and health due to “willpower” rather than biochemistry. I am sure there will be no problems in obtaining funding for this kind of study…..

          • http://www.caloriesproper.com/ William Lagakos

            This. “Body composition and health due to ‘willpower,’ rather than biochemistry.”

  • Gerard Pinzone

    One of the criticisms of this study was that the low carb and low fat weren’t extreme enough. You’d assume this might have been levied by low carb advocates considering saturated fat was limited. But no! It was suggested that if the low fat diet was even stricter, then the results could have miraculously flip-flopped: http://www.huffingtonpost.com/neal-barnard-md/diet-and-nutrition_b_5761450.html

    Fortunately, the A TO Z study from Stanford University pitted Ornish vs. Atkins (and two other “moderate” diets) and the results were in line with this study: http://nutrition.stanford.edu/documents/AZ_press.pdf

    • http://www.caloriesproper.com/ William Lagakos

      “One of the criticisms of this study was that…” low carb won.

      • alan2102

        That’s right, it was not asking the right question. The “low-fat” diet was not low fat. It was 29% fat — only a scant bit below SAD. Bad study, dismissable out of hand on this basis, unless I’ve got the numbers wrong.

        Also, the low-carbers in the study ate fewer calories — enough fewer, mechanistically, to account for the better weight loss result. Again, lousy study, unless I’ve got the numbers wrong.

        (if someone wants to step in and prove that I’ve got the numbers wrong, that’s fine. I welcome that.)

        Also, your claim about “loss of muscle mass” on the low fat diet is not convincing, based on the numbers you presented. It looks from your graph like the LF folks lost a whopping 1/4th kg over a year’s time — slight enough to be statistical noise, analytic technique error, or whatever, and in practice non-appreciable even if certifiably precisely correct. Likewise the improved lean mass on the LC diet was modest and could indeed be accounted for by the higher protein intake. 14 grams is not much on a daily basis, but multiplied by 365 that’s 5 kg of pure dry protein, (which could represent 25 kg of hydrated lean tissue, if that tissue were 20% protein), possibly enough to account for the difference. That’s especially true when we’re dealing with low protein intakes; i.e. the marginal impact of an extra 10 or 20 grams in a low-protein context could be quite high, over time, versus the marginal impact of same in a high-protein context.

        I’m not arguing for LF or against LC necessarily. I’m just talking about this one study, and the need for fairness in its assessment. It was a crappy study which told us nothing about LF, and (perhaps) a little about LC, though nothing we did not already know.

        • http://www.caloriesproper.com/ Bill Lagakos

          Another way of looking at it: low fat dieters were in a significant energy deficit, down about 500 kcal from baseline (Table 2) — yet fat mass was steadily increasing (Figure 2).

          • alan2102

            1. Again: “low-fat dieters” were not low-fat dieters. They ate a junky, refined-carb-based (as reflected by the very low fiber) slightly-fat-reduced (circa 20% less than the typical 35-7%) SAD diet. Slightly tweaked SAD.

            2. Fat mass was not “steadily increasing”. 250 GRAMS difference is trivial. Realistically, you could say that their fat mass was roughly stable, while it declined modestly on LC, suggesting a modest LC advantage.

            3. An interesting sidelight, not discussed anywhere that I can see, is that LF subjects lost quite a bit of waist circumference — a good deal more, proportionally for the weight lost, than in the LC group. Put another way: the LC people should, given their larger weight loss, have lost a good deal more waist circumference than they did. The results suggest a modest LF advantage in that particular area.

            4. What this study really demonstrates is that both diets, LF and LC, performed very poorly. Essentially, what happened is that subjects reduced calories from about 2000 to about 1500, for a YEAR, and realized trivial weight loss. A 500 calorie daily deficit should (assuming, for a moment, CICO) result in about one pound loss per week, or FIFTY pounds in a year. What actually happened is that subjects lost around 4-12 pounds, or about 10-20% of what would be predicted from the calories — a really miserable showing. My point is not to defend CICO, but to underscore just how tepid and mediocre the results were, relative to the (large) sacrifice these people were making. These were porky people, mean BMI of 35; they needed to lose 40 pounds at least, preferably 50+, (and that would be just to get well into the “overweight” range!), but they only lost 10-25% that much. That means that it would take them 4-10 years just to get down clearly to “overweight”, let alone normal! That’s TERRIBLE. A pox on both diets.

            5. As you point out, the “LC” was not really very LC by the end of 12 months (127g carb/day), and even at the beginning was near 100 grams — not truly low, though it could be called lowER (than SAD). This is significant.

            6. Hence: the “LF” was not really LF, and the “LC” was not really LC. IOW, we have a duel of two mediocre, faux-therapeutic diets, neither coming close to their billing (as “low” this or that), both inadequate to produce striking results, as should have been clear from the getgo, and both actually producing, in the end, mediocre or even depressing results. So… why bother? What was the freaking point of doing this study?

            7. Title of paper should have been: “Both Slightly-Tweaked SAD and Weak-Ass Almost-Low-Carb Diet FAIL MISERABLY To Produce Substantial Weight Loss After One Year”.

            8. For the record: I’ve been an advocate of LC dieting for decades. There is a ton of good reason, both theoretical and empirical, to support the LC diet. (I mean, REAL LC, not this bazzano lame-ass crap.) There is no profit in pointing to anemic studies like this one as evidence that LC is great and has won over LF. That just makes the claimant look foolish to anyone who actually reads these studies and thinks about them. (Which is, granted, only 0.000014% of the population, but still…)

            9. In case anyone is interested, the bazzano full text as pdf is here: http://www.normanmarcuspaininstitute.com/wp-content/uploads/2014/09/Effects-of-Low-Carbohydrate-and-Low-Fat-Diets.pdf

    • johnnyv

      Don’t bring data to a religious/ideological argument.
      HCLF is pining for the fjords.

      • This Old Housewife

        And nailed to its perch!

    • alan2102

      Barnard is right. If you can ignore his hysterical jibes at LC, the basic points he makes about this study are correct. The “LF” diet (29% fat) not only needed to be “stricter”, it needed to EXIST. It was not a low-fat diet by any meaningful measure. A fairly-designed study would have cut fat to under 15%, or 10%. That’s the level at which the magic starts happening — for at least some subjects. (NOT ALL; perhaps not the majority.) In an analogous way, it is only when carbs are cut heroically — like down to under 80 grams a day — that the magic starts happening on an LC diet. Bullshit “LC” diets that still have 150 grams of carb are just that — bullshit. (Note that this was a common error of researchers in the days before the real LC concept became well known.) If the Bazanno study had allowed the LC group to eat 150 grams of carb per day, you could rightly complain that it did NOT test LC.

      Another good point that Barnard made was that the fiber intake in this study was very low — perhaps 1/4th or less of what would be expected on a true low-fat whole foods-based diet. If you’ve followed the fiber literature over the years, you know that this could be significant.

      Also, the A TO Z study found that the Ornish group was consuming nearly 30% calories as fat at the end of the study; i.e. they did not stay on the diet. Therapeutic diets only have an impact if they are adhered-to. Now, you could say that that very fact — poor compliance — is intrinsically the fault of the diet, and maybe you would be right. I don’t know. Some people have an easy time adhering to LF; others do not.

      • Gerard Pinzone

        The ATOZ study didn’t truly measure the success of the diets under tight controls. In fact, it was a complaint of Dean Ornish and he’s got a point. The subjects weren’t in a metabolic ward and forced to comply. What the ATOZ study show is that “real world” compliance is tough for any diet over time, but the lowest carbohydrate goal wins across the board. Of course, this is general advice based on the sum total of the results. Individual compliance and success are another matter. There indeed were folks who lost great amounts of weight and kept it off by being in full compliance with an Ornish program. Nevertheless, it’s tough to untangle the reasons why. Heck, the fact that an individual complied with ANY diet skews the results! Compliance is a bias in of itself. http://www.nytimes.com/2007/09/16/magazine/16epidemiology-t.html?pagewanted=all

      • http://www.caloriesproper.com/ Bill Lagakos

        Regarding the “magic” of a very low fat diet:

        My response to this is the same as my response to Ma-Pi 2. Has it ever been tested against low carb in obese or T2DM?

        • alan2102

          Sure it would be nice to have a head-to-head, and sooner or later someone will get around to doing one.

          For now we’ll just have to be satisfied with good-to-excellent results:

          Diabetes Metab Res Rev. 2014 Mar;30 Suppl 1:55-66. doi: 10.1002/dmrr.2519.
          Ma-Pi 2 macrobiotic diet and type 2 diabetes mellitus: pooled analysis of short-term intervention studies.

          It is kind of amazing that they’re getting such good results with GRAIN-based, very high-carb fare. I would suggest less carb, less grain as source of carb, and more protein, in the form of beans and low-fat dairy, but it is hard to argue with results.

          I note that the Ma-Pi 2 is used largely in developing countries. This is significant. In resource-constrained environments — which is to say most of the world, still — our high-animal-product low-carb diets are inappropriate/impossible. People who live on $US 3000/year cannot do Atkins, and indeed the earth itself cannot possibly supply Atkins-style diets for 7 billion people. As such, stuff like Ma-PI2 has much greater public health significance — globally, and for the long term — than low-carb.

          Back to biology: I note the improvement in uremia on Ma-PI 2. This, too, is significant. Reminds me of the Kempner rice diet results, recently discussed in the blogzone: substantial and lasting improvement of kidney function.

  • jasmine johend

    How much truth is there in the notion that eating a high protein diet causes muscle gain, especially in females? I know how hard it is for men to put on muscle despite going to the gym, lifting heavy weights. I am an omnivore, eat shedload of meat, exercise by walking/cycling everyday if I can, and once weekly half hearted weights (don’t enjoy this just do it for health benefits) and find my weight creeping up lately. Mind you I’ve started some new meds that could be causing weight, but it does mess with my mind. Is it really that easy to put on muscle gain from diet alone?

    • CynicalEng

      No, especially not in a calorie deficit.

    • http://www.caloriesproper.com/ William Lagakos

      Dietary protein is better at ‘maintaining’ than ‘building’ muscle. If you were low/deficient to begin with, or unwittingly started exercising more, then it could be added muscle mass.

      But this whole thins is very context dependent, ymmv, because some studies show huge increases in protein intake have no impact on body comp (eg, http://bit.ly/1olqBkS), whereas others show muscle gain and fat loss (eg, http://caloriesproper.com/nutrient-partitioning-a-very-high-protein-diet/).

    • This Old Housewife

      And what if some of this “shedload” of meat turns into glucose, such as what happens to my husband? How does THAT build muscle?

      • http://www.caloriesproper.com/ William Lagakos

        Hubby is a rarity …here’s what happens when *most* humans overeat protein: http://bit.ly/1onCgzs = no appreciable impact on blood glucose levels.

        • Jack Kruse

          BOOM Why? Because all that protein is turned into heat and we use it to build our hydrogen bonding network in network in water. This is where Peter’s proton series dovetails with my own. Good stuff Bill.

          • Jack Kruse

            The length of hydrogen bonds depends on bond strength, temperature, and pressure in the vessel that water is contained. The bond strength itself is dependent on temperature, pressure, bond angle, and environment. This is usually characterized by local dielectric constant. Water has a very high dielectric constant but it is quickly adjustable by the environment. This is why life is metastable. It changes because the hydrogen bonding network changes as soon as the environment changes. When water is in or part of a topologic insulator its dielectric constant varies over a large range and seems to be controlled by ROS/RNS and light polarization within the cell. The hydrogen bonding around proteins is very complex and varied because of the environmental changes a cell faces. What do the hydrogen bonding do to proteins? Several studies have shown that hydrogen bonds play an important role for the stability between subunits in multimeric proteins. Light transforms our protein polymers to morph. This is the essence of complaint design in all life forms. The protons from the OH group of water fuels those alterations. (Gilbert Ling, Gerald Pollack, Emilio Del Giudice 101. http://www.mdpi.com/1099-4300/16/9/4874)

  • libfree

    From my own experience, I won’t lose much weight with 40% of my diet as carbohydrates but I will remain relatively stable in weight at that amount. My heavy weight loss really occurred at about 10% carbohydrate 25% protein and 65% fat. For what its worth, I gained strength while weight lifting at those amounts.

    • http://www.caloriesproper.com/ William Lagakos

      Nice! The weight lifting and relatively high protein intake probably also helped maintain muscle mass.

      • libfree

        Is that a relatively high protein intake? It didn’t seem like it to me.

        • http://www.caloriesproper.com/ William Lagakos

          good point. It’s slightly above average, but I guess if you’re in an energy deficit, 25% is on the low side.

  • Snackscriber

    As an (ugh) journalist, I am humbled:)