Ongoing Study Discussion Thread - Nutrition, Training, Health etc.

[EiFit91]

https://bjsm.bmj.com/content/bjsport.../1365.full.pdf

The fraction of research hypotheses supported in sports and exercise medicine is implausibly high (82.2% of studies surveyed in the above paper reported that the results supported the research hypothesis). In other fields the replication rate of randomized trials is often about 50%, so 82.2% sounds too high to be true.

Would be nice to see a more direct study on this though. It could be that the prior probability of true findings is higher for sports and exercise medicine compared to other fields, and this indirect study cannot really rule out that possibility. But I guess it would be very expensive to run a lot of well-powered replication studies in exercise science to actually test this directly?

[Heisman2]

https://bjsm.bmj.com/content/bjsport.../1365.full.pdf

The fraction of research hypotheses supported in sports and exercise medicine is implausibly high (82.2% of studies surveyed in the above paper reported that the results supported the research hypothesis). In other fields the replication rate of randomized trials is often about 50%, so 82.2% sounds too high to be true.

Would be nice to see a more direct study on this though. It could be that the prior probability of true findings is higher for sports and exercise medicine compared to other fields, and this indirect study cannot really rule out that possibility. But I guess it would be very expensive to run a lot of well-powered replication studies in exercise science to actually test this directly?

I feel this topic, but more broadly the trustworthiness of studies and the provided interpretations of data, deserves its own thread.

[Mrpb]

https://bjsm.bmj.com/content/bjsport.../1365.full.pdf

The fraction of research hypotheses supported in sports and exercise medicine is implausibly high (82.2% of studies surveyed in the above paper reported that the results supported the research hypothesis). In other fields the replication rate of randomized trials is often about 50%, so 82.2% sounds too high to be true.

Interesting paper. Thanks for sharing. Are you aware that you're comparing two very different metrics? 82.2% isn't the replication rate, it's the 'hypothesis confirmation rate'.

In biomedical science the hypothesis confirmation rate is higher, about 90%.

Also good to keep in mind: only 60% of the studies they looked at reported a study hypothesis. 82.2% of those 60% reported that the results supported the research hypothesis. So about 49% of the studies they looked at that supported the research hypothesis.

One possible explanation is that SEM researchers are more likely to test a hypothesis that they've already discovered to be effective in the field. For example drop sets.

[EiFit91]

Interesting paper. Thanks for sharing. Are you aware that you're comparing two very different metrics? 82.2% isn't the replication rate, it's the 'hypothesis confirmation rate'.

In biomedical science the hypothesis confirmation rate is higher, about 90%.

Also good to keep in mind: only 60% of the studies they looked at reported a study hypothesis. 82.2% of those 60% reported that the results supported the research hypothesis. So about 49% of the studies they looked at that supported the research hypothesis.

One possible explanation is that SEM researchers are more likely to test a hypothesis that they've already discovered to be effective in the field. For example drop sets.

I am not sure they are that different, but I am at fault here for being so vague in my above post. I had a specific definition of "replication rate" in mind, I think of a successful replication as a "statistically significant effect in the direction of the original study", so the "replication rate" according to that definition would correspond to the "hypothesis confirmation rate" in the above paper, provided that "hypothesis confirmation" means "statistically significant effect" for the research hypothesis under study. I thought that this would be what "hypothesis confirmation rate" means in the above paper but now I got a bit unsure. It seems to be based on their description at the end of page 1367 but it is not clearly stated in the paper. Note that I would expect that even if the rate of true hypotheses is 82.2%, the hypothesis confirmation rate in hypothetical replication studies would be lower than 82.2% as the statistical power would be lower than 100% (but unlikely to be as low as 50%).

I am unsure why they find that only 60% even reported a study hypothesis. Could it be that the other 40% report a p-value (outcome of a significance test) as their main outcome, but that the paper doesn't explicitly verbally state what the actual research hypothesis is?

I don't think the paper is particularly good, as it doesn't really provide any evidence for presence of "questionable research practices" in the field. To test for that directly, one should collect p-values across a representative sample of SEM studies and look for spikes or discontinuities in that distribution (typically just around the threshold associated with statistical significance).

[EiFit91]

I feel this topic, but more broadly the trustworthiness of studies and the provided interpretations of data, deserves its own thread.

I agree, I am just unsure if the above paper is the right one to spark a discussion on this topic. Perhaps evidence on replication failures from other fields + direct evidence for "questionable research practices" (an excess of reported results are statistically significant) would be better as it could serve as a warning to treat sports/exercise science findings with caution?

I am also unsure if the study conditions that have been pointed out as problematic in other fields (e.g. social psychology) are the same as in SEM/general "sports science" research, though I would suspect that this would be the case. Some of those would be low sample sizes (which would imply low statistical power) and flexibility of analytical decisions available to the researchers (many different statistical tests could potentially be reported in the paper as the main result). You probably know much more about this than me.

[Mrpb]

I am not sure they are that different, but I am at fault here for being so vague in my above post. I had a specific definition of "replication rate" in mind, I think of a successful replication as a "statistically significant effect in the direction of the original study",

That's correct afaik.

Hypothesis confirmation would be like this: I (pretending I'm a researcher) hypothesise that high training volume for the quads results in larger muscle growth than low volume.

In order to test my hypothesis I design a study. Half the the participants do 20 sets leg extensions per week, the other half does 10 sets per week. If after 12 weeks the high volume group has statistically more muscle growth, my hypothesis is confirmed.

Replication would be repeating the same study to see if it confirms the results of the first study.

I am unsure why they find that only 60% even reported a study hypothesis. Could it be that the other 40% report a p-value (outcome of a significance test) as their main outcome, but that the paper doesn't explicitly verbally state what the actual research hypothesis is?

I guess many SEM researchers don't find it that important. They just do a study, for example testing low volume vs. high volume, without forming an formal hypothesis.

I think those study can be useful too if they use good methods. They provide more data for meta analysis and systematic reviews.

[Heisman2]

The way I think about it is that a hypothesis should be generated base on prior work in the field. You read through the totality of the literature pertaining to the topic you wish to study, then in the background section of your paper you summarize all of this and draw a logical hypothesis for your work, which will either be:

1. A replication study where you try to repeat what has been found previously
2. A novel study where you are trying to infer what will happen based on prior results.

As an example, there are some studies showing site-specific hypertrophy, with different areas of a muscle being targeted by either concentric or eccentric portions of reps or by different exercise selection. So now you do a study where you perhaps have one group do several different types of exercises or a muscle group, some of which focus on concentric portions and some of which focus on eccentric, and you compare to a volume/frequency/effort matched group who does only one exercise. You hypothesize that the former group will generate greater total hypertrophy of the muscle than the latter.

On the other hand, using the same 2 groups, you can look at the literature and find there seems to be an upper limit of work that can be done before fatigue sets in and impedes recovery. You know much of the literature shows similarities for hypertrophy between multi-joint and single-joint exercises. One meta-analysis and systemic review found a trend for hypertrophy benefit with eccentric exercise but this was not statistically significant. You hypothesize that the former group that does multiple exercises and focuses on concentric and eccentric exercise will generate less total hypertrophy of the muscle than the latter due to accumulating too much damage/fatigue with the relatively high volume workload that includes eccentric actions while the group doing one exercise that doesn't focus heavily on eccentric actions will be able to recover well enough to see benefit.

This is just an example off the top of my head to illustrate the point that you can take different aspects of the literature and mold it to generate a hypothesis. What's more is you can do the actual study, see the results, and then retroactively choose to include in the background the literature that supports the hypothesis that fits your results. When this starts to happen various research groups can get more entrenched in the line of thought that supports their own work and it can divide the field. People can start citing the papers and authors whose work goes along with their own. My favorite analysis that touches on this is actually one involving dietary sodium intake (https://academic.oup.com/ije/article...485?login=true).

I'm not saying that's happening currently with SEM, but there certainly do seem to be some researchers who put forward similar ideas repeatedly and generate studies that support those ideas. Look at the back and forth in the field regarding the low vs higher volume as one example.

My overall point is that hypotheses themselves need to be taken in context.

Regarding studies without a hypothesis; I have no problem with that. That in a sense is basic science research, where you are exploring something for the sake of exploring, and that can definitely generate new, potentially unbiased knowledge that can lead to hypothesis-generation and application down the line.

[Mrpb]

Mine was just a simple example of course. I think that works best for explanation.

[Heisman2]

Mine was just a simple example of course. I think that works best for explanation.

Yeah, simple is better, I was more trying to show how the same body of literature can lead to different hypotheses and why I feel this is important to recognize.

[EiFit91]

That's correct afaik.

Hypothesis confirmation would be like this: I (pretending I'm a researcher) hypothesise that high training volume for the quads results in larger muscle growth than low volume.

In order to test my hypothesis I design a study. Half the the participants do 20 sets leg extensions per week, the other half does 10 sets per week. If after 12 weeks the high volume group has statistically more muscle growth, my hypothesis is confirmed.

Replication would be repeating the same study to see if it confirms the results of the first study.


I guess many SEM researchers don't find it that important. They just do a study, for example testing low volume vs. high volume, without forming an formal hypothesis.

I think those study can be useful too if they use good methods. They provide more data for meta analysis and systematic reviews.

Would "just do a study" imply that the results are just reported as summary statistics (mean and standard deviations for the two experimental groups), or would these studies that don't have a "study hypothesis" also do some formal statistical inference on the data (so report for instance the p-value for a formal hypothesis test comparing muscle growth in the two groups)? I agree that if they just report summary statistics they are still going to be useful for meta-analyses, just trying to understand the difference.

[Mrpb]

Would "just do a study" imply that the results are just reported as summary statistics (mean and standard deviations for the two experimental groups), or would these studies that don't have a "study hypothesis" also do some formal statistical inference on the data (so report for instance the p-value for a formal hypothesis test comparing muscle growth in the two groups)? I agree that if they just report summary statistics they are still going to be useful for meta-analyses, just trying to understand the difference.

I'm not sure. I just looked up a few studies on my harddrive but they all have a hypothesis so far.

But I haven't looked at how stringent their (of the study you linked) criteria were to call something a formal hypothesis or not.

[Heterodox]

I have a question about the following study from a while ago

https://pubmed.ncbi.nlm.nih.gov/23067428/

They compared 10 grams or protein every 1.5 hours; vs 20 grams every 3 hours; or 40 grams every 6 hours.

3 hours was moderately greater than the 6-hour group, and slightly better than the 1.5-hour group

My questions is 3 hours was slightly better than the 1.5 hour group but the 1.5 hour group only had 10g of protein, its possible that 20g every 1.5 hours is even greater than 20g every 3 hours. Are there any other similar studies that have explored this further? I know there was talk about a muscle full effect but i also recall that possibly being debunked?

Does anybody know of any more studies with more frequent protein feedings like the above?

I am tempted to buy a big bag of whey and experiment my self for a few weeks 20g protein every 1.5 hours while i'm here bored at home during lockdown. I have my home gym and nothing else to do.

[SuffolkPunch]

I have a question about the following study from a while ago

https://pubmed.ncbi.nlm.nih.gov/23067428/

They compared 10 grams or protein every 1.5 hours; vs 20 grams every 3 hours; or 40 grams every 6 hours.

3 hours was moderately greater than the 6-hour group, and slightly better than the 1.5-hour group

My questions is 3 hours was slightly better than the 1.5 hour group but the 1.5 hour group only had 10g of protein, its possible that 20g every 1.5 hours is even greater than 20g every 3 hours. Are there any other similar studies that have explored this further? I know there was talk about a muscle full effect but i also recall that possibly being debunked?

Does anybody know of any more studies with more frequent protein feedings like the above?

I am tempted to buy a big bag of whey and experiment my self for a few weeks 20g protein every 1.5 hours while i'm here bored at home during lockdown. I have my home gym and nothing else to do.

The design of that study suggests they are trying to highlight the time course of protein synthesis - which it does seem to do.

If total protein was above adequate levels for all groups then I would not expect extra protein to make any difference. 80g is a little low though for resistance trained males - so you would need to compare 20g every 1.5 hours to 40 g every 3 hours to avoid confounded results.

[Mrpb]

My questions is 3 hours was slightly better than the 1.5 hour group but the 1.5 hour group only had 10g of protein, its possible that 20g every 1.5 hours is even greater than 20g every 3 hours.

I've wondered the same thing. I don't think it has been tested in a post workout situation like this paper did. I think it's possible that 20 gram every 1.5 hours does even better.

But keep in mind: they only trained the quads. If you're going to train more muscle groups 20 gram protein won't maximise MPS. I'm expecting 40 grams every 3 hours probably beats 20 grams every 1.5 hours.

I am tempted to buy a big bag of whey and experiment my self for a few weeks 20g protein every 1.5 hours while i'm here bored at home during lockdown. I have my home gym and nothing else to do.

Assuming you're 35 and train more than just your quads I think it makes sense to have more protein per serving.

[Heterodox]

I didn't realise they only trained quads that's very interesting.

I am 35 and train full body, typically i have 20g protein in the morning, 20g before bed and 2 larger meals, lunch and dinner with around 40g, usually around 5 hours between meals. I guess bumping all of them up to 40g would put me pretty much at what's considered optimal in that case.

I am kind of just in one of those moods where I'm feeling really motivated, bored etc due to lockdown and began looking at MPS studies again lol.

20g ever 1.5 hours is pushing beyond the boundaries of what's even practice in reality, 40g every 3 hours is also very interesting, i wish there were more studies. I wouldn't expect there to be a huge difference but its still very interesting and intriguing for us that love to learn about this stuff.

Being at home all the time during lockdown though seems like the perfect opportunity to experiment though.

[Mrpb]

Another way to look at it is that 40 gram protein from whey will probably maximise MPS over 3 hours. So there's probably no point in having more protein again within 3 hours, I suspect.

I am 35 and train full body, typically i have 20g protein in the morning, 20g before bed and 2 larger meals, lunch and dinner with around 40g, usually around 5 hours between meals. I guess bumping all of them up to 40g would put me pretty much at what's considered optimal in that case.

Those 20 gram feedings seem relatively low indeed. Might also be worthwhile to look at the leucine content. 3 gram leucine is probably a good target.

[Mrpb]

Interesting new study by Hall. Keto vs high carb plant based.

"A low-fat, plant-based diet resulted in 550-700 less daily calorie intake (but higher insulin and blood sugar levels) vs. a low-carb, animal-based diet, in new #NIH study"
https://www.nature.com/articles/s41591-020-01209-1

[EiFit91]

Interesting new study by Hall. Keto vs high carb plant based.

"A low-fat, plant-based diet resulted in 550-700 less daily calorie intake (but higher insulin and blood sugar levels) vs. a low-carb, animal-based diet, in new #NIH study"
https://www.nature.com/articles/s41591-020-01209-1

Interesting (I only looked at the preprint version as I cannot access the published one). I don't understand how the standard errors were calculated as there are very few details about this at least in the preprint.

The authors were really careful in drawing any strong implications but can suspect that advocates of a "plant based diet" will use this study to push their agenda. I won't be jumping to a low fat diet any time soon, can sustain it for a short while but then I completely crash and feel awful (way worse than just "hunger") until I up my fat intake.

[AdamWW]

Interesting (I only looked at the preprint version as I cannot access the published one). I don't understand how the standard errors were calculated as there are very few details about this at least in the preprint.

The authors were really careful in drawing any strong implications but can suspect that advocates of a "plant based diet" will use this study to push their agenda. I won't be jumping to a low fat diet any time soon, can sustain it for a short while but then I completely crash and feel awful (way worse than just "hunger") until I up my fat intake.

Yeah I have no desire to get my fats to 10% of total calories...

I usually have between 60g and 80g a day and that’s a pretty good sweet spot for my tastes and energy needs.

[R5GT]

Interesting new study by Hall. Keto vs high carb plant based.

"A low-fat, plant-based diet resulted in 550-700 less daily calorie intake (but higher insulin and blood sugar levels) vs. a low-carb, animal-based diet, in new #NIH study"
https://www.nature.com/articles/s41591-020-01209-1

I actually tried a ultra low fat a couple of years ago i posted a thread back then but didn't get many replies

https://forum.bodybuilding.com/showt...hp?t=176892511


I was eating less than 20g per day just getting efa's from fish oil eating in a 1000kcal surplus, i only did it for a few weeks until i found my self craving fat so much i ended up having a massive fat filled cheat day but before that cheat i was around 4lbs up and i honestly felt like i had got leaner and was super vascular all the time but the cheat day ruined me and i no doubt gained a bunch of fat from the cheat day, i gave up on it after that as i couldn't stomach no eating any fat everything was so bland.

[Strawng]

Yeah I have no desire to get my fats to 10% of total calories...

I usually have between 60g and 80g a day and that’s a pretty good sweet spot for my tastes and energy needs.

Me too. I LOVE my carbs but I typically like to have at least 50 g of fat per day. I never really go much over 80-90. My problem with the study is the duration. I've eaten very low fat in the past and I felt great a couple weeks in. Cut to a month or two later and I was killing jars of pb with 0 interest in ever eating that way again. I know keto's not the most "sustainable" diet out there by any means, but I think this one might be even harder to maintain long term for most people.

[AdamWW]

Me too. I LOVE my carbs but I typically like to have at least 50 g of fat per day. I never really go much over 80-90. My problem with the study is the duration. I've eaten very low fat in the past and I felt great a couple weeks in. Cut to a month or two later and I was killing jars of pb with 0 interest in ever eating that way again. I know keto's not the most "sustainable" diet out there by any means, but I think this one might be even harder to maintain long term for most people.

Yeah for me anything below 50g would mean purposefully avoiding nut butters, avocado, and protein bars....

I can’t do that

[Mrpb]

I haven't read the Hall study yet. At first glance I found it interesting that the keto diet wasn't suppressing energy intake like usually expected.

For anyone interested in protein, there's an interesting discussing going on in the supp section: https://forum.bodybuilding.com/showt...9549103&page=1

Topics discussed: Are EAAs a good supplement for muscle building? Is optimising MPS important for hypertrophy?

I've even got Stu Phillips involved to clear up some misconceptions.

[EiFit91]

I haven't read the Hall study yet. At first glance I found it interesting that the keto diet wasn't suppressing energy intake like usually expected.

For anyone interested in protein, there's an interesting discussing going on in the supp section: https://forum.bodybuilding.com/showt...9549103&page=1

Topics discussed: Are EAAs a good supplement for muscle building? Is optimising MPS important for hypertrophy?

I've even got Stu Phillips involved to clear up some misconceptions.

It's very interesting! But the media spin on this study is probably not going to be "wow, the carbohydrate-insulin model is inconsistent with experimental data", it's going to be an army of millitant Twitter vegans using it to command everyone to stop eating meat...

[Heisman2]

I haven't read the study yet but they appear to just do 2 weeks in each group which may not be long enough to see the appetite-suppressing effects.

Regarding the "carbohydrate-insulin model", there is a big back-and-forth in the literature between Kevin Hall and David Ludwig (plus their respective associates). I can pretty much guarantee there will be a detailed critical response of this study from Ludwig's group.

[EiFit91]

https://academic.oup.com/ajcn/articl...5/1212/5897225

Interesting recent study on metabolic adaptation. Main findings:

- Metabolic adaptation does not predict weight regain after 1 year follow up
- Metabolic adaptation is reduced/eliminated after a period of weight stabilization

[Mrpb]

Thanks for the share EiFit91 ^.

I found this an interesting abstract. TL/DR: Google sucks.

Using the Google™ Search Engine for Health Information: Is there a Problem?

There was no correlation between high quality scores and early appearance in the sequence of search results, where results are presumably more visible. Also, 496 advertisements, over twice the number of search results, appeared. We concluded the Google™ search engine may have shortcomings when used to obtain information on dietary supplements and cancer.

https://academic.oup.com/cdn/advance...zab002/6127311

Telling people to use Google may be problematic if they don't know how to differentiate between good and bad info.

[Strawng]

https://academic.oup.com/ajcn/articl...5/1212/5897225

Interesting recent study on metabolic adaptation. Main findings:

- Metabolic adaptation does not predict weight regain after 1 year follow up
- Metabolic adaptation is reduced/eliminated after a period of weight stabilization

Fascinating! This really flies in the face of all that HAES nonsense & that “Biggest Loser” study they always cite to claim that long-term weight loss permanently suppresses the metabolism. It also falls more in line with my personal experience.

[EiFit91]

Fascinating! This really flies in the face of all that HAES nonsense & that “Biggest Loser” study they always cite to claim that long-term weight loss permanently suppresses the metabolism. It also falls more in line with my personal experience.

I agree! Somewhat related to this topic, this just came out:

https://www.nejm.org/doi/full/10.105...=featured_home

Extremely promising drug had an enormous effect on weight loss in obese individuals. Almost too good to be true, but seems like a really solid study.

[Heisman2]

https://academic.oup.com/ajcn/articl...5/1212/5897225

Interesting recent study on metabolic adaptation. Main findings:

- Metabolic adaptation does not predict weight regain after 1 year follow up
- Metabolic adaptation is reduced/eliminated after a period of weight stabilization

I saw this when it first came out (I browse AJCN) and even included it in the nutrition/weight management course on my website. I wish the general public would get this. The Biggest Loser study got way too much popularity compared to the several other studies suggesting metabolic adaptation is generally not more than 50 kcal/d. The Biggest Loser study also obviously took things to the extreme and should not be considered representative of the general population anyway in my opinion.

Separately though, this also casts more doubt on Hall's findings in the metabolic chamber studies where the subjects do not maintain perfect energy balance throughout (this was one of the criticisms of his work). Unfortunately metabolic chambers are very expensive to run so those studies always have small numbers of subjects and it is hard on a short time scale to determine if one is in perfect energy balance or not.

Thanks for the share EiFit91 ^.

I found this an interesting abstract. TL/DR: Google sucks.



https://academic.oup.com/cdn/advance...zab002/6127311

Telling people to use Google may be problematic if they don't know how to differentiate between good and bad info.

There's a reason the medical community collectively rolls its eyes at people trying to get health information from google. Personally I don't mind when people google and then check the information with a doctor but when people draw conclusions from google it can certainly be problematic. My suspicion though is that this extends far beyond google and misinformation is so rampant on the internet.

I agree! Somewhat related to this topic, this just came out:

https://www.nejm.org/doi/full/10.105...=featured_home

Extremely promising drug (***) had an enormous effect on weight loss in obese individuals. Almost too good to be true, but seems like a really solid study.

I saw this as well. Very exciting. We are not supposed to mention specific medications so I edited it out of the quote; you may want to edit your post and do the same. A cousin of this med (that starts with "lira") is already used for type 2 diabetes and obesity. The downsides are that the one already used is a daily injection, not an oral med, and it is quite expensive.

Important aspects of this study:
- they got weekly (not daily) injections starting at 0.25mg, and every 4wk increased the dose until getting to the max dose of 2.4mg at week 16
- they had counseling sessions every 4 weeks to help adhere to a 500kcal deficit per day and at least 150 minutes exercise weekly
- ~75% of participants were female, ~75% were white, average bodyweight was 105kg, average BMI was 37.9 at the start, 43.7% had prediabetes
- weight loss was seen at week 4 and a nadir was seen at week 60, by week 68 weight loss on average was 14.9% (15.3kg)
- the placebo group lost 2.4% (2.6kg) by week 68
- 7.0% in the med group and 3.1% in the placebo group discontinued treatment due to adverse events

Looking at the graphs the weight loss really began to taper off around week 52. This med works by decreasing appetite. Thus, when doing a lifestyle intervention and trying to stick to a deficit this drug will help with that. We know that when the weight loss tapers off that means the participants have already been increasing their caloric intake for a long time. It's possible had this trial been extended to 2 years these subjects would have started to gain back some of the weight. Also, the placebo group had sustained weight loss as well, which speaks to the quality of the lifestyle intervention as most people in dietary studies reach a peak weight loss at 6 months and then gain it back but the placebo group here did not gain back the weight they lost. The point is that I'm skeptical this drug will be a game changer for obesity at large because most of the general public will not be part of a concomitant lifestyle intervention (also I imagine the cost will be an issue). Additionally, while there is so much obesity and only a handful of medications that are designed to treat obesity, there are actually relatively few healthcare providers who prescribe these meds, in part because people are supposed to be on point with their nutrition/exercise regimen before starting them.

That said, this could be a game changer for a subset of people, and more treatment modalities are always welcome.