The New Paleo Diet Study: The Weight Loss

I've written previously about the diet used in the latest paleo diet study. In this post I want to address the results.  Here's the study:  A Palaeolithic-type diet causes strong tissue-specific effects on ectopic fat deposition in obese postmenopausal women.   In this post I am going to address the weight loss (and change in intake) only.

On the "con"side we have the following:

• Small, n=10
• No control group
• Short - 5 weeks

On the "pro" side, there appear to be (see small type below) significant controls in place to ensure compliance, accurate reporting of intake, etc.

Summary:  This group spontaneously decreased caloric intake by an average of 25% with a mean weight loss of 4.5 kg in 5 weeks.  That works out to 2 lbs per week weight loss ... a decent clip.  Two things I would like to highlight:

► Protein rules!    

Very early in the days of the Asylum, I blogged on a study where trading out fat for protein and allowing ad libitum feeding spontaneously led to a 450 cal/day reduction in obese women:   High Protein Diet Induces Sustained Reduction in Ad Libitum Intake Despite Diurnal Leptin Compensation.  It is a study I've discussed many times over the years, initially because it pretty much put Taubes' theories to the trash heap, but as it turns out, so much more.  In that study, carbs were held constant at 50% of energy intake while protein was doubled from 15 to 30%.  In the current study, protein increased from 16 to 28% but carbs were cut significantly, and energy intake was reduced an average of 25% -- median change 520 cal.  

There are obvious limits to protein's satiety effect, and it has been demonstrated, most recently by Bray, that overeating protein is no "freebie" as overeating a high protein diet results in more weight gain, albeit lean mass along with comparable fat.  Still, I believe part of the problem with the so-called misguided "eat less" guidelines has been the inattention to absolute protein intake -- especially for women.  Pegging protein at 15% and reducing calories to 1200-1500 cuts absolute protein intake in half for most women (particularly the obese who are consuming closer to 3000 cal/day).   Low carb diets generally do not increase protein intake much in absolute amounts, but the higher percentage triggers the satiety at a lower total caloric intake.  Magic I tell ya!

► Variety, "Food Vehicles", Portioning/Measuring & Logging may influence intake:

While small in number, there was some interesting info to be found by the manner in which the authors chose to present their data.  Rather than mean & standard deviation, they reported median and interquartile range (Q1 and Q3, the 25th and 75th percentile cut-offs).  For a group of 10 subjects, as shown on right, this information gives us actual data values for two subjects and a better feel for the distribution of values.    Data presented as mean ± s.d. can be (unintentionally) misleading as the ± implies symmetric distribution about the mean, when (a) it may not be symmetrical, and (b) the magnitude of the s.d. can be highly impacted by an outlier (especially in small sample sets).`

I was reminded of the CAF Rat study and how the rat weights varied.  In that study, the weights were all bunched together for the standard chow, there were a couple of high outliers but generally bunched weights for the high sugar and high fat diets, and spread out weights for the "cafeteria diet" fed rats that took in more than one-third more calories than their counterparts when given a rotating selection of human "junk foods".   The CAF rats were exposed to a variety of foods while the other rats were all fed a homogeneous chow.  They ate more, and got fatter, but not all in similar fashion.

Humans are not laboratory rats.  We self select foods and "procure" them in some fashion.  There are a whole lot more factors involved in overeating, and something I haven't mentioned as much of late, "passive overeating" plays a role.  When you eliminate grains, dairy and all added fats except for a couple of cooking oils (olive, canola), you not only eliminate those foods, but entire swaths of a "normal" diet (I put that in quotes so folks can take that as they see fit - grin).  No paleo meatzas or coconut cream puffs either.  Further, this was Palisystem Watchers non-Anonymous System:

Participants were given prepared meal portions that were intended to provide an average intake of about 30 energy percentage (E%) protein, 40 E% fat (mostly unsaturated) and 30 E% carbohydrates for breakfast, lunch and dinner, together with 40 g nuts (walnuts and sweet almonds) on a daily basis for 5 weeks. All meals were prepared by the food service at Umeå University Hospital and were weighed and frozen after preparation. The diet included lean meat, fish, fruit, vegetables (including root vegetables), eggs and nuts. Dairy products, cereals, beans, refined fats and sugar, added salt, bakery products and soft drinks were excluded.  Participants were instructed to complement the provided food with other included food items from the list, ad libitum. To enable preparation of additional complete meals at home, the women received 14-day menus together with recipes and instructions regarding portion sizes. They were also advised to use only rapeseed or olive oil in food preparation. The recommended alcohol intake was less than two glasses of red wine per week. Once a week, participants had one-to-one meetings with the study dietician to check their weight, collect prepared meals for the next 7 days, and review and complete their food records. 

Food intake was assessed by individual diet history interviews at baseline (i.e. before the start of the study) and at the end of the study. During the interviews, food intake during the previous month including meal pattern, meal content, beverage intake, between-meal snacks, frequency of intake of individual food types and portion sizes were assessed. In addition, participants kept daily food records during the intervention of intake of provided food as well as food and beverages that they chose. Women were instructed to record the weight of any leftovers from the provided meals using a household scale. To estimate the amounts of food eaten, participants were instructed to use household utensils (e.g. cup, spoon) and a book of photographs of food portions and standard weights of food items.

So ... having said all of that, I went to plot the intake data (initially for a different reason), but when plotting it something really jumped out at me.  Here are the Q1, Med & Q3 intake data for before and after the paleo diet.

Now we have no way of knowing if blue square = subject 3 in ascending rank of intake is the same subject before/after, but the group as a whole reduced intake considerably.  The gray ovals are *my guess* as to what the distribution of the two highest and lowest points might be.  This is really just a guess based on the larger 3rd vs. 2nd quadrant spread of values.

But as a group as they reduced intake significantly, the variation in intake diminished dramatically as well.  The spread from Q1-to-Q3, called the interquartile range, representing the middle half of the data, was 650 calories "before" but only 174 calories "after".  There appears to be much less range of intake due to one or more or all of the factors described above.  There were lots of checks and such, but in the end, this was ad libitum intake.  I'm sure some would attribute the previous overeating to carbs, wheat, dairy or some combo of it all.  My bets are on limited food choices, simply prepared, avoiding all those typical "problem foods" for many, accountability -- both to a 2nd party and being mindful of intake, and the relatively short duration of 5 weeks all playing a part in "normalizing" intake levels.  

► Some final thoughts on the weight loss

Here is where the lack of a control group really comes in.  There are a veritable crapload of factors I lumped in after the protein content up there ^ ^ .  I imagine just logging food and reporting in (complete with weigh ins) might alter the "spontaneous" eating behavior of most post menopausal women.  These women averaged cusp-of-obese (30).   For many, just becoming aware of how much they are consuming may be enough to enact change irrespective of the diet.  Having most foods prepared and menus to work off of, for 5 weeks?  Pretty much a no brainer that the motivated person can adhere.

Many may say "see, sugar is obesogenic" because when you cut it out, folks lose weight.  You can fill in the blank for wheat, etc.  But the factors these influence are myriad.  Once you get the paleo bread (yeah, I know Mr. Nutty disKetordered Eating laughs at that now) and the various and sundry substitutes for everything (cauliflower can be anything folks!!) much of that may well disappear as innumerable anecdotal reports certainly support.

In the end, they ate less and lost weight.  I doubt "paleo" per se had much to do with that as restrictive diets of all nature tend to produce similar results in comparable time frames.  Imagine that.  

---

Final followup post:  The New Paleo Diet Study: Just the Fats Ma'am