“But Menno, what about Eric Helms’s research proving that you need more protein in a deficit? Doesn’t that show your literature review about the optimal protein intake is wrong?”
– Random person, every day
Let’s review the research Eric Helms did on how much protein you need when cutting. He published a literature review and an experimental study as part of his Master’s dissertation to research if we need more protein when in a deficit. We start with the literature review, because that was the introduction for the actual study in his dissertation.
A systematic review of dietary protein during caloric restriction in resistance trained lean athletes: a case for higher intakes.
by Eric Helms et al.
Introduction: Caloric restriction occurs when athletes attempt to reduce body fat or make weight. There is evidence that protein needs increase when athletes restrict calories or have low body fat. The aims of this review were to evaluate the effects of dietary protein on body composition in energy-restricted resistance-trained athletes and to provide protein recommendations for these athletes.
Methods: Database searches were performed from earliest record to July 2013 using the terms protein, and intake, or diet, and weight, or train, or restrict, or energy, or strength, and athlete. Studies (N = 6) needed to use adult (? 18 yrs), energy-restricted, resistance-trained (> 6 months) humans of lower body fat (males 23% and females 35%) performing resistance training. Protein intake, fat free mass (FFM) and body fat had to be reported.
Results: Body fat percentage decreased (0.5-6.6%) in all study groups (N = 13) and FFM decreased (0.3-2.7kg) in nine of 13. Six groups gained, did not lose, or lost nonsignificant amounts of FFM. Five out of these six groups were among the highest in body fat, lowest in caloric restriction, or underwent novel resistance training stimuli. However, the one group that was not high in body fat that underwent substantial caloric restriction, without novel training stimuli, consumed the highest protein intake out of all the groups in this review (2.5-2.6g/kg).
Conclusion: Protein needs for energy-restricted resistance-trained athletes are likely 2.3-3.1g/kg of FFM scaled upwards with severity of caloric restriction and leanness.
The theory that you need more protein in a deficit
The introduction of Eric’s dissertation – I’m calling the lead author by his first name because we’re friends – makes the case that theoretically, we need more protein when energy intake is below maintenance levels. Eric et al. state:
When in negative energy balance, the anabolic response to protein is enhanced (Saudek & Felig, 1976), which can be erroneously interpreted to mean that less protein is needed during weight loss.
Why would this be an erroneous interpretation? In an energy deficit, the body becomes more conservative with protein. The more efficient the body is with protein, the less it logically needs. This is similar to the finding that the more advanced you are as a strength trainee, the less protein you need, because your body’s protein metabolism becomes more efficient. It’s counter-intuitive, but it’s perfectly logical.
Of course, Eric et al. do offer an explanation:
When supply is limited, efficiency is increased, indicating the body’s increased need for protein in states of negative energy balance (Fielding & Parkington, 2002).
The current recommended intakes of protein for strength is 1.6 to 1.7 g/kg.
The evidence that you need more protein in a deficit
Do you notice any relation between protein intake and weight loss success? I don’t. Of course, visual inspection is not very scientific, so I performed a regression analysis in SPSS. I got a beta coefficient of -1.7, p = 0.36. That means there was no statistically significant relation between these 2 variables.
Technical note: I averaged Umeda’s 2 groups as 1 data point in the graph (not in the analysis) due to the study’s data heterogeneity. Since the higher protein group actually had a worse p-ratio here (11.5, literally off the chart), this only strengthens the lack of a relation.
In fact, the top 5 most successful study groups were by Garthe et al. and Mero et al. Their protein intakes in order of best body composition results were 1.6, 1.6, 1.4, 1.4 and 1.5 g/kg protein. Not exactly very high protein intakes.
The argument from Eric et al. rests mostly on 2 studies. The first is Mettler et al.’s high protein group with a protein intake of 2.3 g/kg. However, going by the data, their lean body mass to total bodyweight loss ratio was 0.2. That was actually just the average in the dataset, so you can’t exactly say they were doing spectacularly well because of their high protein intake. They were just doing average. Which fits all the other data: the extra protein wasn’t harmful, but it wasn’t beneficial either.
The second study is Maestu et al. with a p-ratio of 0.1 and a protein intake of ~2.6 g/kg. So this group performed slightly better than average and had a far higher than average protein intake (the average protein intake was 1.6 g/kg). Was it the protein intake that caused them to be relatively successful in preserving muscle mass during the diet? If you look at Maestu et al.’s study, there’s another factor that just may be relevant: these guys were the only bodybuilder group in the dataset. Not only that, they were national and international level competitors with an average of 8 years of training experience. Training 600 minutes per week, increasing even to ~1000 minutes per week throughout the study. So we have international level bodybuilders training over 16 hours a week and the reason their muscle retention was slightly more successful than average in athletes and regular Joes must have been… their protein intake? Maybe, could it have been their training? Or their genetics? Or name any other factor that makes someone an international bodybuilding champion. Since there was no relevant control group, this study can’t demonstrate a positive effect of a high protein intake.
As such, the data presented by Eric et al. do not provide any evidence that you need more protein when cutting than when bulking.
In the conclusion, Eric et al. derive their final recommended protein intake from Phillips and Van Loon (2011). Yet those authors state ‘protein intakes in the range of 1.3–1.8 g /kg/day […] will maximize muscle protein synthesis’ with the lower end being for advanced trainees. The abstract does list 1.8 – 2 g/kg may be advantageous when cutting to promote fat loss and the 1.8 – 2.7 g/d figure is given in the text, but those figures are not substantiated by any research or formal analysis.
Basically, Eric came up with the hypothesis that you need more protein when cutting than when bulking. However, the research review itself was just that: it generated an hypothesis. There is no actual evidence in it that shows the hypothesis is true. So Eric then performed an experimental study to test his hypothesis. (Eric is after all a true scientist and I say that with nothing but respect.)
High-protein, low-fat, short-term diet results in less stress and fatigue than moderate-protein, moderate-fat diet during weight loss in male weightlifters: a pilot study
by Eric Helms et al.
PURPOSE: Athletes risk performance and muscle loss when dieting. Strategies to prevent losses are unclear. This study examined the effects of two diets on anthropometrics, strength, and stress in athletes.
METHODS: This double-blind crossover pilot study began with 14 resistance-trained males (20-43 yr) and incurred one dropout. Participants followed carbohydrate-matched, high-protein low-fat (HPLF) or moderate-protein moderate-fat (MPMF) diets of 60% habitual calories for 2 weeks. Protein intakes were 2.8g/kg and 1.6g/kg and mean fat intakes were 15.4% and 36.5% of calories, respectively. Isometric midthigh pull (IMTP) and anthropometrics were measured at baseline and completion. The Daily Analysis of Life Demands of Athletes (DALDA) and Profile of Mood States (POMS) were completed daily. Outcomes were presented statistically as probability of clinical benefit, triviality, or harm with effect sizes (ES) and qualitative assessments.
RESULTS: Differences of effect between diets on IMTP and anthropometrics were likely or almost certainly trivial, respectively. Worse than normal scores on DALDA part A, part B and the part A “diet” item were likely more harmful (ES 0.32, 0.4 and 0.65, respectively) during MPMF than HPLF. The POMS fatigue score was likely more harmful (ES 0.37) and the POMS total mood disturbance score (TMDS) was possibly more harmful (ES 0.29) during MPMF than HPLF.
CONCLUSIONS: For the 2 weeks observed, strength and anthropometric differences were minimal while stress, fatigue, and diet-dissatisfaction were higher during MPMF. A HPLF diet during short-term weight loss may be more effective at mitigating mood disturbance, fatigue, diet dissatisfaction, and stress than a MPMF diet.
Noticed the effect on body composition? Even in a 40% deficit, 2.8 g/kg protein had no beneficial effect on fat loss, muscle retention or strength compared to 1.6 g/kg. Magnitude-based inference showed with 99% certainty that the greater protein intake had no effect on bodyweight or muscle mass, measured in 3 different ways. For fat loss the certainty was 97% and there was actually a 2% chance that the higher protein was harmful compared to a 1% chance that it was beneficial.
What about the beneficial effect of protein on mood? Even if there’s no effect on your body composition, doesn’t this research show extra protein still helps you psychologically? Unfortunately, it doesn’t, or at least this study doesn’t show it. This study was confounded in a major way by using maltodextrin powder as the protein placebo. That meant the medium protein group was consuming 95 grams of pure sugar. When you’re in a 40% deficit and over 25% of your diet is sugar, it’s not surprising you’ll feel poorly. They were likely starving and the major difference in hunger alone could easily explain the difference in mood.
Eric Helms’s research does not support the theory that you need more protein in a deficit. Theoretically, we may even need less, since the body’s protein metabolism becomes more efficient in a deficit. Eric’s narrative review provided a tentative hypothesis based on a cross-sectional comparison of several studies, but not a single experimental study or analysis actually supports this hypothesis.
Don’t worry, Eric. I still love you. In fact, I consider Eric to be one of the absolute top guys in evidence-based fitness. Unfortunately, he, like me, doesn’t publish much, but definitely check out his shared website with 3DMJ.
Interested in more research reviews like this? Check out my PT Course where you’ll learn the full Bayesian Bodybuilding methodology and you can become a certified Bayesian Personal Trainer.
Below is Eric Helms’s full reply. As expected, it is excellent. I will respond to this over the weekend.
Okay folks, first off let me point out that Menno Henselmans was kind enough to message me to ensure the accuracy of this blog post and to get my temperature on it before he posted it http://bayesianbodybuilding.com/eric-helms-protein/. A very respectful, honest thing to do. Much respect.
The only thing I wanted him to change in fact, was that originally he wrote that the research in question was part of my PhD when it was actually a part of my Masters. And of course he made this edit, so really I left the blog post untouched. Now that is not to say I agree with his interpretation. In fact, I disagree with his interpretation in a few important ways, both mechanistically and in terms of practical application. However, I believe his interpretation is a perfectly reasonable one to make even though I disagree. So, I am putting forth the following response to help those reading both of our thoughts come to their own conclusions on protein intake.
-As an important aside, I would please recommend everyone first read Menno’s article on protein intake http://bayesianbodybuilding.com/the-myth-of-1glb-optimal-p…/ and read the full text of my systematic review and double blind crossover trial http://www.researchgate.net/profile/Eric_Helms so they understand the foundation of what is being discussed.
To start, Menno addresses my systematic review of protein intakes in lean dieting resistance trained populations. He starts by putting forth the fact that the efficiency of protein utilization increases as you decrease calories, which is true. This he states is the logical reason why protein needs don’t increase in a deficit, and he states that it is illogical to state that protein needs increase in a deficit, and that in fact based on physiology they would decrease.
In my opinion, this is a simply missing the important point of “why” protein efficiency goes up when you reduce calories. Similar phenomenon occur in many ways in the body. The body increases efficiency in the face of a limited resource. In fact an increase in efficiency signals that there is a limited resource.
If you reduce calories, your energy expenditure decreases, becoming more “efficient” with the utilization of energy to prevent starvation. This doesn’t mean you wouldn’t benefit from eating more food, it means the body is trying to adapt to less food.
If you travel to a very high altitude, you get more efficient with oxygen extraction in an effort to keep you alive. This doesn’t mean you don’t feel the negative effects of reduced oxygen and that you wouldn’t benefit from more oxygen, it means the body is trying to adapt to less oxygen.
An increase in efficiency signals the fact that there is a limited resource. So it is not illogical to posit that increasing protein in the face of a deficit would be a potentially lean mass sparing decision.
Protein breakdown increases in the face of a caloric deficit as more protein is used for energy. This is well established http://www.ncbi.nlm.nih.gov/pubmed/15173435. It is also well established that the leaner you are, the more protein you oxidize for energy since there is less available from body fat http://www.ncbi.nlm.nih.gov/pubmed/10574520. So while you can accurately state that protein efficiency increases in a deficit and be correct, it is also true that you are using more protein for fuel. This is not a one sided equation. If protein efficiency went up only, and you didn’t use more protein, the net change would be an increase in protein balance and you wouldn’t have a huge body of evidence showing lean mass losses while dieting, you’d have a huge body of evidence showing lean mass gains while dieting!
For these reasons, I think it is plainly incorrect to state that protein needs decrease when you diet. It is only accurate to state that protein efficiency increases when you diet.
Honestly, I think Menno would likely agree with that distinction. I think our true disagreement just comes down to how much protein is required to offset the increase in protein breakdown that occurs, and he would argue that 1.8g/kg is enough (and likely more than enough), while I tend to be more conservative and fall on the side of larger safety buffer.
What we also agree on, is the characterization of my literature review as a hypothesis rather than a definitive conclusion. I explained how a caloric deficit increases protein breakdown, how caloric deficits result in lean body mass losses that scale with the severity of energy restriction, and how the leaner you are the more protein is used for fuel. Then, I hypothesized that an effective strategy to offset this would be an increase in protein intake. Yes, a hypothesis.
A final note regarding my literature review, I want to give a huge kudos to Menno for actually doing some statistical analysis on the data that I presented. That really impressed me. I don’t dispute his findings either. With only 6 studies to form my hypothesis it doesn’t surprise me that the correlation is a) very weak between LBM and protein intake and b) non-significant. However, I am curious to see what occurs when Menno reanalyses this information based on grams per kilogram of lean body mass. I actually just provided him that data to run an analysis on. I doubt it will reach significance and show a strong correlation, but I am curious to see if the p value decreases and the inverse correlation increases as it might indicate that something is happening, even if there is not enough data yet to make any definitive, quantitative conclusion.
Now finally, Menno addresses my cross over study comparing 1.6g/kg of protein to 2.8g/kg of protein and acknowledges, like I do in the study, that there was no significant or meaningful differences found in strength, lean body mass or body fat change between the two diets. I think that 1.6g/kg is probably approaching an optimal intake of protein during a deficit, and even if it was a bit short of optimal, I doubt a 2 week period would be an adequate enough time period to show the difference. Additionally, I wasn’t able to secure DEXA, hydrostatic weighing or ultrasound (despite my best efforts) as a method to measure lean mass or muscle mass changes. I only had access to high quality, skilled anthropometric measurements. This means skinfolds, and skinfold derived equations. Skinfolds measurements when taken by a skilled technician themselves (when not put into an equation) are a highly reliable and accurate way of measuring fat mass. But, LBM and body fat percentage derived from equations are not. So based on the time frame and the measurements available, I am truly not surprised that there was no difference in LBM (BTW, of course our lab got an ultrasound…6 months after I finished my masters!).
However, the study was not completely a waste! Interestingly enough both on the Profile of Mood States and the Daily Analysis of Life Demands for Athletes the participants reported better overall mood state, more diet satisfaction, and less fatigue consuming the higher protein diet.
Menno dismissed this finding saying the following:
“This study was confounded in a major way by using maltodextrin powder as the protein placebo. That meant the medium protein group was consuming 95 grams of pure sugar. When you’re in a 40% deficit and over 25% of your diet is sugar, it’s not surprising you’ll feel poorly. They were likely starving and the major difference in hunger alone could easily explain the difference in mood.”
First off, I think Menno accidentally counted the entire calorie amount of the placebo powder, versus just the maltodextrin portion, as the powder also had 20g of protein in it. The 95g of maltodextrin is 380 calories, and 380 calories is 20.7% of the 1829.3 calories that the moderate protein group consumed not 25%…a minor difference though, so I think his point still stands.
However, I would put forth that his point (like my literature review) is just a hypothesis as to the reason the high protein group felt better. And I would argue, a much weaker hypothesis given that the research on how sugar effects hunger and mood is not as definitive as one might think.
In fact White and colleagues found improvements that only reached significance in a high sucrose diet group in the mental component of a quality of life assessment http://www.ncbi.nlm.nih.gov/pubmed/20095912. This finding is not solitary. Drummond observed subjects consuming a low fat, sugar containing diet who when compared to baseline, reported improvements in perceived quality of life, diet attitude and adherence http://www.ncbi.nlm.nih.gov/pubmed/15369982. Finally, the participants in Raben and colleagues’ highest sucrose group reported higher diet satisfaction and palatability compared to the other groups http://www.ncbi.nlm.nih.gov/pubmed/9347402.
So the contention that the presence of a large amount of sugar in a diet would “easily explain the difference in mood” as Menno puts it is challenged by these controlled trials showing the opposite. In contrast, the hypothesis that a higher protein intake might result in better diet satisfaction and less hunger is one that stands on the top of a great deal of empirical evidence http://www.ncbi.nlm.nih.gov/pubmed/23107521 http://www.ncbi.nlm.nih.gov/pubmed/15867892 http://www.ncbi.nlm.nih.gov/pubmed/18769212. However, as Menno might argue (and I would agree), the majority of this evidence compares low to high protein intakes without comparing a middle of the road intake, like this study http://www.ncbi.nlm.nih.gov/pubmed/16002798 which when analysed on a g/kg basis compares 1g/kg to 2g/kg.
However, recent research by Antonio comparing 1.8g/kg to 4.4g/kg found that a much larger number of participants dropped out of the 4.4g/kg group, and that bodyweight gain was much less than expected based on self-reported energy intake in the 4.4g/kg group http://www.ncbi.nlm.nih.gov/pubmed/24834017. In my opinion this indicates a satiating effect of protein that extends past the 1.8g/kg level and that the participants in the 4.4g/kg group didn’t consume as many calories as they thought they did. A follow up study comparing 2.3g/kg to 3.4g/kg found more weight was gained in the 2.3g/kg group, while the 3.4g/kg group lost more fat, despite a larger self-reported energy intake. Again, I would put forth this likely shows a satiating effect of the higher protein intake and subsequent over reporting of caloric intake which would explain less weight gained, and fat lost in the 3.4g/kg group http://www.ncbi.nlm.nih.gov/pubmed/26500462. To conclude, while it is possible that the differential effects on mood found in the higher protein group in my study are due to the higher sugar intake in the lower protein group, I think it would be premature to conclude this based on the well-established satiating power of protein and the murkier role of sugar.
As a final note I want to emphasize the difference between what Menno and I prescribe with our interpretations of the research. Menno, is happy to use 1.8g/kg as the sole protein intake for all situations (all hail 1.8g/kg!) based on the data as he interprets it. I would agree that intakes over 1.8g/kg are rarely studied. But the absence of evidence on the effect of very high protein intakes (>1.8g/kg) compared to moderately high protein intakes (~1.8g/kg), is not the same as the evidence of absence of an effect of very high protein intakes compared to moderately high protein intakes.
Based on our conversations I know that Menno and I both agree that protein intakes above his recommendations (unless they force other macros too low) aren’t harmful. We share this opinion with other much more well established researchers http://www.ncbi.nlm.nih.gov/pubmed/14971434 and we would also be in agreement with the vast majority of the research on resistance trained individuals showing that there has never been an incidence when a lower protein intake was found to have a statistically significant benefit over a higher intake: https://www.facebook.com/photo.php?fbid=10155242388350441&l=a49477585b
Now the reason I am not comfortable with simply advising 1.8g/kg to everyone (and the reason you will most often find ranges prescribed in the literature) is that there is plenty of individual variability in protein requirements evidenced by this study that found a cyclist requiring 2.8g/kg of protein to reach nitrogen balance compared to the mean requirement of 1.63g/kg http://www.ncbi.nlm.nih.gov/pubmed/24476478.
Additionally, I would rather err on the side of over prescribing protein versus under prescribing given the info graphic I linked above. Specifically, I would turn your attention to the 7 studies at the bottom of this infographic that found a benefit of a higher protein intake versus a lower one. 5 of these 7 compared protein intakes in the range of 1.4-2.2g/kg to intakes in the range of 2.4-3.2g/kg. Sure, you can find flaws in each study. But if you have to explain away five studies to maintain your position, perhaps you need to rethink how strongly you hold your position. We can’t know for sure that the confounding variables in each of these studies fully explains their findings and that’s enough to make me more comfortable prescribing 1.8g/kg as a minimum intake versus a solitary intake to ensure I cover the possibility of it being too little in some cases. For the record, I normally prescribe either 1.8-2.8g/kg of total body mass or 2-3g/kg of lean body mass for protein intake for recreational and competitive strength and physique athletes. Using the lower half for a caloric surplus/maintenance, and the upper half for a deficit.
As a closing note, I want to look to the future because in my opinion we need more research on this to truly make a definitive conclusion. Thus far, the only true controlled comparisons we have of one protein vs another in lifting populations using progressive resistance training while dieting are:
Walberg showing 1.6g/kg beating out 0.8g/kg in lean mass retention in bodybuilders over just a single week http://www.ncbi.nlm.nih.gov/pubmed/3182156
Mettler showing 2.3g/kg beating out 1.0g/kg in lean mass retention in trained lifters over a 2 week period http://www.ncbi.nlm.nih.gov/pubmed/19927027
Helms (me) showing 2.8g/kg beating out 1.6g/kg in mood state in trained lifters over a 2 week period with no difference in strength of body comp http://www.ncbi.nlm.nih.gov/pubmed/25028958
And a thesis you probably haven’t seen by Longman showing 2.4g/kg beating out 1.2g/kg in lean mass retention in trained lifters over a 4 week period https://macsphere.mcmaster.ca/handle/11375/15355
We don’t have enough data. We disagree because we are filling in gaps with our interpretations of what is available. Hopefully a meta-analysis or a better designed, longer RCT will come out that can give more evidence. I know Menno well enough that he would revise his thoughts on protein if convincing data was to emerge (in fact that is the essence of the Bayesian approach). I commit to do the same. Protein is expensive and won’t be getting cheaper as our population increases, so as soon as I feel that I can responsibly tell lifters to eat less of it, I will. But as it stands, I know there is no harm erring on the higher side (if in fact that’s what I’m doing), and without stronger evidence to show a lack of potential harm of lower intakes, I can’t commit to a lower prescription.
What does the other literature say about the optimal protein intake in a deficit?
Thus far, the only true controlled comparisons we have of one protein vs another in lifting populations using progressive resistance training while dieting are:[4 studies]
- Jose Antonio et al. recently published a study in strength trained subjects in a deficit. There was no effect of 3.4 vs. 2.3 g/kg protein on changes in lean body mass or strength. (I’ll have more to say on the supposed fat loss benefit in a different article, but in short, I think it’s clear that was due to compliance issues, not some magical effect of protein on fat loss that no other study has found, especially the next one.)
- Prior research from Jose Antonio in strength trained individuals had found the same thing: increasing protein from 1.9 to 4.4 g/kg did not affect body composition (including fat loss) or strength. Now, technically these subjects weren’t supposed to be in a deficit, but body fat percentage went down in both groups. This was non-significant, but since it’s yet another study finding no benefits of very high protein intakes, it’s worth taking into account.
- Pasiakos et al. studied physically active individuals (3-4 x per week) in a 40% deficit. Many subjects were military personnel. There was no difference in fat or muscle losses between the group consuming 1.6 and the group consumign 2.4 g/kg protein. A major strength of this study is that they also looked at muscle protein synthesis, anabolic signaling, gene expression, nitrogen balance ánd resting metabolic rate. None of the measures were different between the 2 protein intakes. Note: Eric pointed out the higher protein group did have higher postprandial protein synthesis levels, but given the non-significantly greater lean body mass loss in the moderate protein group, the relevance of this is highly questionable. A limitation was that the strength training was pretty light: full-body training 3x per week with 3 sets of 15 reps per muscle group. For beginners that’s not terrible though. They also did a decent amount of cardio, which would further increase protein needs. These subjects were not strength trained before the study, but that only makes the results more convincing, because we know that strength training experience decreases protein needs.
- Verdijk et al. (2009) found no difference in fat loss, muscle growth, nitrogen balance or strength development between a group consuming 1.1 g/kg protein and a group adding 20 g casein to the diet. The subjects were elderly men who performed strength training 3x per week. They only trained their lower body though, which probably explains the very low protein intake. In case you think their age is a confounder, it is, but most research actually suggests the elderly have a higher protein requirement, if anything, due to anabolic resistance, i.e. less efficient protein metabolism.
- Campbell et al. recently found no difference in fat loss, strength development or muscle growth in subjects consuming 1 – 1.2 g/kg protein compared to more than 1.2 g/kg. The subjects were overweight men and women that performed full-body strength training twice a week along with 1 weekly cardio session of 60 minutes. A notable strength of this study was that it had 117 subjects and it lasted 9 months.
- Maltais et al. recently found no difference in lean body mass gains in strength training elderly subjects consuming 1.1, 1.3 or 2.1 g/kg protein. They did 3 full-body strength training sessions per week for 4 months. The 2.1 g/kg protein group did lose fat though while the 1.3 g/kg group didn’t. This was probably in part because this group was consuming soy-based protein while the 2.1 g/kg group was consuming dairy. The 1.1 g/kg group actually lost only 200 grams less fat than the 2.1 g/kg group (0.9 kg vs. 1.1 kg fat), but their fat loss did not reach statistical significance.
“there has never been an incidence when a lower protein intake was found to have a statistically significant benefit over a higher intake”
More evidence against the idea we need more protein in a deficit
The results of the above studies in cutting strength trainees are supported by nitrogen balance studies and research in untrained individuals.
- We have the aforementioned Walberg et al. study. Here nitrogen balance was maintained in experienced weightlifters in a severe deficit consuming 1.6 g/kg protein.
- In my article on the optimal protein intake I also mentioned Pikosky et al. where nitrogen balance, protein synthesis and total body protein turn-over were maintained in endurance trainees in a thousand calorie deficit on 1.8 g/kg protein.
In both these studies there was still lean body mass loss, but note that it is circular reasoning to say that lean body mass loss implies insufficient protein intake. You can only deduce this if you first assume that increased protein intakes are required to protect you from muscle loss compared to being at maintenance or in a bulk. And this whole review shows this assumption is highly questionable.
In fact, we have good research in untrained individuals that shows increasing your protein intake in a deficit does not protect your muscles , even when doubling your protein intake at just 440 calories a day .One particularly nice study is the recent Hill et al. study that looked at 3 different protein intakes at weight maintenance and during weight loss. Protein content did not affect body composition and this effect was not different between weight maintenance and weight loss phases. How much muscle you lose is instead determined by things like the rate of weight loss, your training and your genetics.
As for these subjects being untrained, we know that strength training increases protein requirements. This is a direct effect with a clear mechanism: increased protein synthesis and turn-over. But to invalidate the finding that protein needs are not increased in a deficit in untrained individuals, there would have to be an interaction effect between strength training and the relation between energy balance and protein intake. That’s quite far fetched, since these kind of interaction effects are exceedingly rare in nature (remember Occam’s razor/law of parsimony, for those educated in scientific methodology).
We also know that increasing your protein intake has little effect on muscle atrophy during disuse, like during bedrest. If extra protein can’t ward off that kind of catabolism, why would it help during a cut?
Back to Eric’s review
Other than a lack of research, another main argument from Eric was that he saw a subjective trend towards benefits of higher protein intakes in the dataset from his research review. As I analyzed in the first part of this now-very-long post, objectively there was no significant relation between protein intake and weight loss success. However, Eric rightfully pointed out that I used protein relative to total bodyweight in my analysis, i.e. protein per kilogram bodyweight per day. So he provided me with the data on protein intake relative to fat free mass and I redid my analysis.
*drum rolls please*
It didn’t change the result. The p-value increased from 0.36 to 0.43 and explanatory variance (R^2) went down from 8% to 6%. In other words, calculating protein relative to lean body weight instead of total body weight did not strengthen the relation between protein intake and weight loss success: there was still no significant trend towards higher protein intakes being beneficial. Technically, the relation was actually slightly worse, but I don’t think this means anything other than that we don’t have to calculate protein relative to lean body mass and we can just use total bodyweight instead.
Note: So everyone can check my analysis, I have uploaded the outputs of my analyses with protein based on bodyweight vs. fat-free mass here.
Why error margins inflate protein requirements in research
This is a rather technical argument, so I’ve saved it for last. If you’re not statistically inclined, feel free to skip to the conclusion below.
The argument is that dietary protein error margins give the impression in research that there is a trend towards benefits of higher intakes when in fact there isn’t (or at least needn’t be). For example, in Hoffman et al. the protein intake was 1.74 ± 0.13 g/kg. That means the lower end of protein intake’s 95% confidence interval, assuming a normal distribution (which we can, otherwise it would have impacted the study statistics), was only 1.48 g/kg. And that’s assuming the subjects’ self-reported intakes were correct, which is implausible, so the standard deviation was likely larger in actuality and some subjects were likely dipping below 1.48 g/kg on average, not to mention the occasional day where they forgot their protein supplements altogether, just to give an example. These subjects with a low protein intake could easily explain the trend in study towards better results in the higher protein group. The beauty of statistical significance testing and why we don’t just look at averages, is that statistical testing takes this data variance into account.
Alright folks, the protein debate continues between Menno? and myself. For those not following along, Menno respectfully critiqued my research and I responded in kind, and most recently he responded to my rebuttal http://bayesianbodybuilding.com/eric-helms-protein/.
In what I highly doubt (lol) will be the last round of this debate, I submit the following in response to Menno’s response to my rebuttal (if you get the chance please add this to your site as well Menno Henselmans):
Okay, I’m not going to bring any new literature to the table in this response, so please refer to the citations that Menno and I provided in our previous blog posts.
Now to respond to Menno’s response to my response, let’s start off with the initial portion where Menno says
“As for Eric’s own study…the complete lack of any trend towards benefits even in a 40% deficit in 1.6 vs. 2.8 g/kg protein is valuable evidence.”
I agree it is valuable evidence, but I would like to reiterate that there was a difference in mood state favoring the 2.8g/kg group, and again, the explanation of sugar intake in the 1.6g/kg being the cause is shaky based on the available research.
Menno then digs into “evidence” I didn’t cover, stating;
“However, more importantly, Eric is incorrect in saying this is all the literature we have in strength training individuals in a deficit. (Honestly, I’m sure he knows about most or even all of these and he just forgot.)”
Actually, I didn’t forget about these studies. In fact the first two he cites, I included in my original rebuttal to his blog. But, nonetheless let’s dig in further:
The first study Menno brings up is Jose Antonio’s recent study (which I already brought up) comparing 3.4 to 2.3g/kg of protein. This 8 week study is not a dieting study, the 2.3g/kg group gained 0.8kg and the 3.4g/kg group gained 1.7kg of bodyweight on average by the end of the study. And yes, Menno is correct the higher protein group lost more fat.
The second study by Antonio (which I also already brought up) was comparison of 4.4g/kg to 1.9g/kg. Once again, this is not a dieting study. Only the 4.4g/kg protein group could be said to be in a deficit (barely). The mean change in bodyweight was +1.3kg (yes weight gain) in 8 weeks in the 1.9g/kg group, and the HP group lost only .1kg in 8 weeks. Yes that’s 100g of weight loss in 2 months. Not a dieting study.
Now we get to Pasiakos 2013, that actually was a dieting study (I analyse this in my lit review by the way). In this study, Menno is correct, there wasn’t a statistically significant difference in body composition between the 2x RDA (1.6g/kg) and the 3x RDA (2.4g/kg) group. But, he is incorrect about one thing, the researchers actually did find one statistically significant difference between the 2x and 3x RDA groups. The 3x RDA group was the only group that had a statistically significant rise in post prandial muscle protein synthesis during energy deficit. But I would agree given the lack of significant difference in body composition this likely doesn’t have practical significance. Additionally, I would challenge the notion that this was a true study on people performing resistance training. Here is the exact quote from the study describing the training: “to minimize the potential of an unaccustomed, anabolic stimulus influencing study outcome measures, the intensity and volume of the resistive-type exercise was low. Specifically, volunteers performed one single-joint movement per major muscle group (3 sets of 15 repetitions) using workloads determined during the prestudy period. Frequency, intensity, mode, and volume of resistive-type activities did not change during the 31-d study.” This is not progressive resistance training, which is why it wasn’t included in the primary analysis of my lit review and was only in the discussion.
Unfortunately, then we get back to non-dieting studies. Verdjik (2009) is also not a dieting study, a deficit was not imposed, bodyweight was not statistically significantly different from pre to post. In fact, in the placebo group after 12 weeks bodyweight was 80.2 vs 80.1kg and in the higher protein group it was 79.2 vs 78.9kg. Over 12 weeks, 100-300g of bodyweight loss. This is not a dieting study.
As we move onto Campbell et al 2015, again we are confronted with a study that is not on people dieting. I will quote the authors: “Over time, BM increased in the group that consumed <1.0 g · kg?1 · d?1 and was unchanged in the groups that consumed ?1.0 to <1.2 and ?1.2 g · kg?1 · d?1. ” So either no change or a gain in bodyweight, yes, again weight gain. Menno is correct to say there were not significant differences in body composition, but it is worth pointing out that the authors stated the following: “Compared with baseline, daily hunger and desire to eat were lower in the highest TPro group (?1.2 g · kg?1 · d?1) but unchanged in the other 2 groups (P < 0.05)”.
Finally we get to the last study Menno cites Maltais 2015, which unbelievably, is also not a dieting study! In this study (on sarcopenic 65 year olds by the way) all three groups increased their bodyweight from start to finish (yes weight gain). However, at least in this case despite the study being misrepresented as a diet study, at least Menno represents the results accurately. He correctly stated that no differences were found between the 1.1, 1.3, and 2.1g/kg of protein groups, except that the 2.1g/kg group was the only one to lose a statistically significant amount of fat. One thing that wasn’t mentioned by Menno was that most likely due to this fat loss, there was an improvement in lean mass to fat mass ratio only in the 2.1g/kg group. A stated by the authors: “In addition, we observed a significant increase in muscle mass to fat mass ratio in the dairy shake group only (0.6 kgLM/kgFM, p?0.05)”
Now that we’ve addressed a bunch of studies that honestly don’t speak to the point of whether or not higher protein intakes benefit resistance trained athletes performing progressive training in a deficit, we move on to a statement I made in my rebuttal: “no study has ever shown a benefit of a lower protein intake to a higher protein intake”. Menno appropriately brings up the 2012 Campbell and Meckling study to prove this statement false. Well done, I concede that I forgot about this study and I will admit that yes, there is 1 study in existence that shows a statistically significant benefit to a lower protein intake vs a higher one.
However, given this concession on my part I want to bring up that Menno’s statement “The fact that not a single study has found statistically significant benefits of protein intakes above 1.6 g/kg…” is equally false.
In the infographic I posted in my first rebuttal, there are indeed studies showing benefits of a higher intake than 1.6g/kg, in fact there are 3 that directly compare ~1.6g/kg to a higher intake https://www.facebook.com/photo.php?fbid=10155242388350441&set=p.10155242388350441&type=3 Candow 2006, Cribb 2007, and Hoffman 2009. Additionally, there is a 2006 Hoffman study that might have found a benefit of 2.4 vs 1.7g/kg on the bench press http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2129168/table/T3/ but I say “might” because the text states there was not a difference despite this table showing there was. Even so, based on the change scores, it does appear there are likely some trends towards the 2.4g/kg intake being superior (which Menno brings up at the end of his response). But it is difficult to evaluate as they did not report exact P values.
Finally, to bring it up yet again my study did find a mood state benefit of comparing 2.8 to 1.6 and I already pointed out why sugar was likely not the cause for this, So that means we have 4 maybe 5 studies that I can think of where there is a benefit shown of protein intakes higher than ~1.6g/kg directly compared to 1.6g/kg.
As I stated in my initial response to Menno, these studies are not perfect, you can find flaws in each, and dismiss the findings if you choose…but when you have to dismiss 7 studies (the 5 at the bottom of the infographic, my study, and the Hoffman study that I linked the table from), it should give one pause as to the strength of their claims. That’s a lot of dismissing to do.
So now that we have more accurately represented the field of research, I want to bring up the argument Menno uses of citing using Pikosky 2008 and Walberg 1988 as evidence that you don’t need more protein in a deficit.
Menno, I’m confused as to why these studies showing nitrogen balance being maintained at 1.6 and 1.8g/kg of protein are worthy of being used as evidence to support your case, but in your original critique of my research you dismissed my usage of Butterfield and Celejowa to support my position? To clarify, your position seems to be that Butterfield and Celejowa showing nitrogen balance not being maintained during slight deficits while consuming 2.0g/kg protein is not fair game and cannot be used…but Walberg and Pikosky showing positive nitrogen balance at 1.6 and 1.8g/kg is good to go? That is not logically consistent and comes across as biased. I don’t think it is intentional, but it’s good to point this out so we can remain objective.
Next Menno kicks some statistical ass and does me the solid of analyzing my results from my lit review at my request, but by g/kg of FFM. Essentially the results are the same, P ratio is not significantly related to protein intake by g/kg or g/kg of FFM. Thank you Menno, and I still completely agree that based on the hard data, my lit review should be viewed as a hypothesis rather than fact. Kudos!
Finally, I want to applaud Menno for pointing out that statistically, due to compliance issues with consuming supplied protein supplements, some studies might show a trend towards a benefit of a higher protein intake that is actually just due to both groups consuming less than is reported.
This is true, but it’s also an argument for why it’s a good idea to recommend intakes in a practical setting that are slightly higher than what should maximize outcomes. If telling someone to consume 1.7g/kg results in 1.4g/kg at times, and this can be detrimental (such as a trend might indicate in research) then perhaps giving a higher recommendation, so that they don’t fall so low would be prudent.
Let’s assume that yes, clients or study participants often accidentally consume less than their prescribed intakes and like Menno pointed out, a .3g/kg drop in protein intake might occur sometimes. Then let’s also assume that 1.8g/kg is where the maximum benefit from protein intake occurs.
Well, let me remind everyone that we’ve both agreed that a protein intake higher than 1.8g/kg (if not taken to an extreme) is not harmful. Okay…hmmmm…then what is a sensible general recommendation (if we assume 1.8g/kg is the highest protein to provide benefit)? Oh, maybe somewhere around 2.1g/kg, or almost exactly 1g/lbs…
In the infographic I posted in my first rebuttal, there are indeed studies showing benefits of a higher intake than 1.6g/kg, in fact there are 3 that directly compare ~1.6g/kg to a higher intake https://www.facebook.com/photo.php?fbid=10155242388350441&set=p.10155242388350441&type=3 Candow 2006, Cribb 2007, and Hoffman 2009.