Thread by @ethanjweiss

Long thread on some new research. Sorry it is long and sorry it's about nutrition. Also I do some speculating at the end...

New from outstanding team @UBC including @DrJonLittle @kaja_falkenhain @theloctor + @Dy1anLowe: a randomized trial comparing @ww_us or @getkeyto with weight loss at 12 weeks as primary endpoint:
onlinelibrary.wiley.com/doi/abs/10.100…

Background. 1) I'm a co-founder/ advisor to Keyto & hold equity. 2) Keyto was founded in 2018 & is led by 2 amazing co-founders, @lnakamura and Ray Wu. The initial launch focused on developing a breath acetone sensor to enable low carbohydrate/ ketogenic nutrition for weight loss

While the breath sensor was no easy feat and occupied a lot of the early energy, we always aimed to develop a comprehensive tool to help people lose weight and optimize metabolic health. That is, we saw the sensor as necessary but not sufficient

I’ve chronicled by own story many times here:

https://twitter.com/ethanjweiss/status/988115980743200768?s=20

and here:

https://twitter.com/ethanjweiss/status/1086266362052268032?s=20

We believe that low carbohydrate diets offer one thing that no other diet offers in an easy to track biomarker other than weight. One of our key questions was: can the information from the breath sensor enable real behavior change with little or less human intervention.

.@virtahealth and others have shown that low carb nutrition can be very powerful when supported by doctors, health coaches and more. But can this be done with less resources?

We spent/spend a lot of time thinking about behavior. But we also focused a lot of energy on nutrition. As a cardiologist, I had been (and remain) uncomfortable with the make up of conventional ketogenic diets. In particular, I worried about all that saturated fat

In my personal experience, I had adopted a more Mediterranean style diet emphasizing fish and plants and minimizing saturated fats particularly from animal sources. My lipids improved on this diet despite eating a lot of fat (mostly unsaturated)

So we built our program around fish and plant-centered Mediterranean-style low carbohydrate nutrition principles. We wanted to see if you can derive the benefits of low carbohydrate nutrition without the adverse impact on lipids

Anecdotally, I had been having a lot of good experiences with this in patients who came to see me with substantial increases in ApoB (LDL) containing lipoproteins when experimenting with low carb nutrition

Early on, we decided to test our program in as rigorous way as we could. Obviously, that meant a randomized trial, but we went further in comparing not to a control group such as “diet as usual”, but to an active comparator with a long and established history in weight loss

We wanted to know if this thing works and we wanted to be able to compare to something real and something people know

We worked with the UBC to design this study back in 2019. I can’t say enough great things about them or about the support we got from the Canadian Government. The study design was simple and was based somewhat off of our experience with TREAT jamanetwork.com/journals/jamai…

We decided on a primary endpoint of weight loss at 12 weeks, but we pre-specified key secondary endpoints such as weight loss at 6 and 12 months. Full protocol paper is here: ncbi.nlm.nih.gov/pmc/articles/P…

We knew we wanted our trial to be fully virtual (this was the summer of 2019) so we would mail each participant a scale, they would download the app they were randomized to and then we’d mail the breath sensor to those randomized to Keyto

We also had each person go to @QuestDX for blood work at 0, 12, and 48 weeks. We delivered surveys and questionnaires through an app

With that background, what did we find? Important, these are the 12 week and 24 week data. The study will wrap in the fall with the end of the 48-week timepoint

Remember this trial began enrolling in January 2020 and really did not get rolling till March or so of 2020. And yet despite that, people stuck with it and we completed enrollment by the Fall of 2020

It was fortunate that we designed a fully virtual trial before COVID. There is a lot to learn from this experience but one thing is that doing these fully virtual and real world nutrition studies is feasible even during a pandemic

We ended up with 155 people (~70% woman) with an average age of 41, baseline weight ~95 kg and BMI ~ 34. Baseline A1C was 5.4% so this was a group of people with mostly overweight/obesity but not diabetes. Also randomization is beautiful

Here are the primary data: On the left, each dot represents an individual stepping on their wireless scale at home and solid lines showing the daily means for Keyto (orange) and WW (blue). On the right are the waterfall plots showing the percent weight lost for each individual

As mentioned, we pre-specified multiple secondary endpoints and one of the key ones is the weight lost (-8.4 vs. -2.9 kg) at 24 weeks. As you can see, the curves are still separating and participants continue to lose weight

As mentioned, as a cardiologist, I was very worried about the effects of keto on lipids. So we adopted a Mediterranean-style low carb program which was lighter on saturated fat and emphasized foods rich in unsaturated fats such as nuts, olive oil, fatty fish and avocado

We did a comprehensive advanced lipid profile and were reassured that our program did not result in any concerning changes in lipoproteins

We were also impressed with the changes in metabolic markers at 12 weeks. We were particularly impressed by the reduction in A1C & ALT (a marker of liver injury). And given both groups lost weight, we were able to explore whether there might be a weight-independent effect

Now this is important: We surveyed participants on what they ate. Self-reported diet data are notoriously dirty but with a study this size and with a real control group (and active one), it is interesting to look at the changes and the differences

As you’d expect, people said they ate fewer carbs though it was nowhere near the reduction we advised or expected. Who knows if it was this or much lower. But directionally it went down

But what was most interesting was that both groups reported reducing calories but there was no difference between the groups. In fact, numerically, people in the Keyto group said they ate MORE than the WW group despite losing a lot more weight

There are multiple potential explanations for this & I want to be clear that we do not have THE answer but the 2 leading theories are: 1) they ate the same number of calories but low carb offered a metabolic advantage resulting in greater energy expenditure. I seriously doubt it

2) and this is my made up theory just to be clear, the low carb participants under-reported their calories as has been reported by @KevinH_PhD and @GardnerPhD and others pubmed.ncbi.nlm.nih.gov/30672127/

I think this is a much more likely explanation for the paradox that Keyto users lost more weight despite reporting eating more calories. I hypothesize that they perceived that they were eating more than they did because they felt more full

Lots to unpack here and lots to learn. But the bottom line is that these results demonstrate a really exciting option to get all of the metabolic benefits of weight loss and carbohydrate reduction without effects on ApoB containing lipoproteins. And during a pandemic!

Future questions: What predicts success? What happens to body composition- visceral fat? Durability at 1 year? Diabetes prevention? Other metabolic associated diseases (NASH/NAFLD)?

Overall Keyto is a highly scalable intervention that is an option for those looking to lose weight safely

Notably Keyto is also a very small company without lots of resources but had amazing academic partners at UBC. They designed and executed a rigorous and carefully done RCT and against active comparator. Again, people randomized to WW did lose weight, it was just not as much.

Thanks again to UBC team and special mention to @kaja_falkenhain who is just starting her Ph.D. and while she had lots of help from others here, she took the ball and ran with it and has a very bright future in nutrition science.

Look forward to more coming soon…