Have you ever asked your kid to brush his teeth?  And then an hour later, your kid has written 3 novels, drawn 7 pictures, rearranged his bed, transferred all his Pokemon cards to a new storage compartment, and made a new Christmas Wish List (even though it’s July).  A bunch of stuff happened!  But the objective was “brushing teeth,” and his rancid breath and scummy incisors are clear indicators that pretty much everything else EXCEPT the goal was achieved!

If there had been a competition for “Article Most Likely to Leave You With Residual Tartar and Bad Breath,” THIS ONE would almost certainly take the Low-Glycemic Index cake.  Some anti-protein propagandists researchers sought to compare the Lipid Profiles of kids born to mothers adhering to either a “High Protein Low Glycemic Index” (HPLGI) diet or to a “Moderate Protein Moderate Glycemic Index” (MPMGI) diet during their pregnancies.  The children were followed from birth to 9 years, and although a bunch of stuff happened, the likelihood of the results meaningfully contributing to clinical applicability is about as likely as your kid actually brushing his teeth.

The article starts out promisingly enough by acknowledging the increased pregnancy risks to both mother and child with maternal obesity.  And all the women in this trial had a pre-pregnancy BMI of 28-45 kg/m2, meaning the majority of these moms would be classified as obese by BMI criteria.  And (this is critical), the study was powered, which means it was designed to detect a true effect, between various diets on maternal weight gain.

Just like the question asked in the original vignette was, “Did you brush your teeth?” the question asked here was, “Did different dietary patterns result in different amounts of maternal weight gain?”  And the HPLGI group gained about 3 pounds less than the MPMGI group.  So all the other stuff we are going to discuss here is ultimately sort of like Christmas Wishes and Pokemon Cards.

So before we dive into this cesspool, let’s clarify a few things about Glycemic Index and the actual amount of protein consumed by each group.  Fun!

A food with a High Glycemic Index would be something that leads to a more profound spike in blood sugar, which, assuming adequate pancreatic beta cell mass and preserved function, would be accompanied by a concurrent spike in insulin.  Conversely, a food with a Low Glycemic Index would lead to a less profound spike in blood sugar and insulin levels.

The cutoff for the HPLGI diet was a Glycemic Index of <55.  This isn’t particularly strict, as foods such as bananas and popcorn are right around 55, and other fruits such as kiwis and grapes are actually below this number.  Not exactly a Keto Kitchen.  Additionally, these gals were encouraged to eat 25% of their daily calories from protein.

The MPMGI diet group actually ended up having a Glycemic Index of right around 55 despite being encouraged to “not worry about Glycemic Index” and keep their protein right around 18% of total energy intake (and they ended up at a reported 18.9%).

Overall, if these gals were eating a 2,000 calorie diet, the HPLGI diet folks would have consumed 125g/protein per day, whereas those on the MPMGI diet would have consumed 94.5 g/protein per day.  Basically a daily chicken breast difference between groups.  Cluck cluck!

So that wasn’t too bad, but here’s where it starts to go downhill:

The dropout rate in this study was 45.5%.  Of the initial 209 babies, only 114 were analyzed over the 9 year follow-up period.

The babies born to the HPLGI moms actually had lower LDL-C at birth, although it wasn’t significantly different than the MPMGI group.  And then at age 9 years, the HPLGI kids had an LDL-C of right around 92 mg/dL, whereas the MPMGI kiddos had an LDL-C of 83.  Uh-oh!  HDL-C was identical, but the authors threw in an LDL-C to HDL-C ratio into the analysis as well, which, of course was higher in the HPLGI group.  Ratios are pretty lame in general, and comparing ratios with equal denominators is particularly useless. Thanks guys!  Insulin levels and triglyceride levels were also higher in the kids who were born to the HPLGI moms, although this did not reach statistical significance.

And then, the Discussion section of this paper takes us on a bizarre Odyssey of largely unrelated, mostly cherry-picked, and ultimately confusing talking points.  A discussion of insulin receptor activity in pregnant rats was interwoven with theoretical epigenetics and sprinkled with a dusting of statin treatment in adults.  Uh ok…but did you brush your teeth?

And then, almost as an afterthought, these two sentences near the paper’s conclusion captures the essence of the study:

“In addition, there was a considerable proportion of dietary data for the children at the 9-year follow-up that was missing, which may reduce the precision of the dietary estimates and limit the ability to fully explore the difference in dietary intake. Moreover, the RCT study was powered to detect a difference in maternal gestational weight gain, not offspring outcomes.”

I think the diets of these 9 year-olds probably had a little more to do with the subtle differences between groups in insulin and lipid levels…I’d tend to blame a current diet with a few too many goldfish crackers and a little too much soda over epigenetic changes from Mom’s extra pregnancy piece of chicken if I had to guess…

Yep, a bunch of stuff happened in this study.  But if there’s still scum on your teeth, I’m not sure that we really accomplished what we were asked to do.  Although if that Christmas Wish List you made while you were supposed to be brushing your teeth included, “Publish Better Studies,” then I won’t be too upset😊