Thread by @Dr__Guess

Our carbohydrate dosing study is out as a preprint!

medrxiv.org/cgi/content/sh…

I think this paper is going to cause some... discussion so I wanted to do a thread to hopefully help prevent over-interpretation of the data our study generated.

Design:

Crossover study: n=12.

We tested 5 different "doses" of carbohydrate (10,15,20,25,30%kcal) for 7 days in people with type 2 diabetes.

The menus aimed to keep weight constant.

Protein was kept at 15%kcal

We measured 24-hour glucose concentrations using CGM: the Medtronic iPro2.

Note: we did not measure ketones or insulin. This was a mistake in retrospect but we made this decision to reduce the burden on the participants as much as possible.

RESULTS: We found no difference in 24h, post-prandial, or fasting sensor glucose between the 10% and the 30% dose.

Nor did the data suggest any association between the carb doses and glucose.

Do I think this means that carb restriction has no effect on glucose when weight is maintained and when the carb is replaced with fat and not protein?

No, but I think the issue is far more nuanced than I/the nutrition science community has appreciated...

Here are my interpretations based on our data and those of others:

I think you probably need to get to a very low intake of carb (probably <50g/day?) to see a weight independent-effect on blood glucose concentrations.

We don't have conclusive data on mechanisms but obviously ketones could play a role here.

In our study the lowest dose was 10%kcal but given the kcals needed to maintain weight, this averaged about 65g per day. Some people had 93g at 10%kcal.

If we had measured ketones, this would have helped us with interpretation.

There are two studies (both by Garg) where modest reductions in carb intake (60%kcal vs 35%kcal, and 55% vs 40%kcal ) DID lower glucose substantially but the low carb diets in these two studies had a TON of MUFA, and low SFA.

Our study aimed to keep FA% constant.

And obviously I think it's pretty clear protein is a/the key ingredient for weight neutral reduction in glucose in a non-ketogenic diet.

I am not sure of many things in nutrition, but I am sure protein has important glucose-lowering properties in type 2 diabetes.

Let me also point out some important limitations:

First, we did provide all food, we kept in very close contact with the participants at all times (it was quite a burdensome study for participants) but this was not a ward-based study.

You can never be 100% sure people are sticking to the diets.

Second, I was SO SURE that there would be a difference in glucose concentrations between the 10% and 30% doses, we did not include doses higher than this.

If I could go back in time I would compare 5%, 25% and 45%.

You learn....

The sample size was small (only 12). (It was powered based on seeing 0.9mmol/L difference).

Mostly, my view is another eg 12 people would not have changed this:

The study was already quite a burden for people, so our "control" was pragmatic - we simply placed the CGM a day before they received their study meals to measure the person's glucose during their normal diet for 24-hours before they started each "dose".

Note - I expected there to be a very obvious drop in glucose going from normal to the doses - especially the lowest dose (10%kcal). Nope.

This was VERY surprising to me.

I think the CGM readings (you can see these in the supp data) are telling.

There is striking interday and intermeal variability.

And in many people glucose remains high nearly constantly - night time, fasting and interprandial glucose is high.

Something is needed to shut down the EGP?

Why does protein work? Is it a massive pp insulin surge which lowers pp glucose? But then how does this impact fasting?

Modest reduction in carb not enough? Need ketones to shut down EGP?

I think these are important questions for us to answer not only to help understand clinical management, but also to provide insight into the physiology of type 2 diabetes.

To end:

I am a clinical dietitian as well as a researcher and here is what I am taking from this study in terms of looking after my patients.

⬇️⬇️⬇️⬇️⬇️⬇️⬇️⬇️⬇️

I am no longer excited by the clinical utility of modest carb reductions to manage T2D independent of weight loss.

Though note - modest carb reduction remains a good way of helping people who like this way of eating to lose weight.