Whey vs Casein. Who will win?

Ok, I lied. I couldn’t find the creatine study in MSSE that we reviewed years ago as part of the Sport Medicine Journal Club. But I did find a more recent study on supplementation timing, which, after reading the article, I noticed was sponsored by AST Sport Science. And while the study on supplement timing was interesting, there was this glowing ad on the AST site that drew my attention. AST Sports Science has a product called VP2 Whey Isolate.

AST claims that their study supports the use of VP2 and that their study showed increased strength in subjects 156% greater than other proteins; and over 615% more lean muscle growth (than what, they don’t say on their web site) while at the same time promoting reduction in body fat. In 10 weeks.

They also tout on their website that this study was selected to be presented in full at the American College of Sports Medicine’s 49th Annual Meeting, 2002 in St. Louis, Missouri and that that ACSM meeting was held in conjunction with the 6th International Olympic Committee’s World Congress on Sport Science. Not to be all jaded, but ACSM has something along the lines of 800 abstracts that get “selected” to be presented. Quite honestly, having presented at ACSM myself, it’s not difficult to “get selected”.

But, with numbers like that, who can resist looking for the study?? Ok, fine. I’m a geek, and I can’t resist.

So here’s Cribb, PJ, Williams, AD, Crey, MF, Hayes, A. The Effect of Whey Isolate and Resistance Training on Strength, Body Composition, and Plasma Glutamine. International Journal of Sport Nutrition and Exercise Metabolism, 16: 494-509, 2006.

Here’s the abstract:

Different dietary proteins affect whole body protein anabolism and accretion and therefore, have the potential to influence results obtained from resistance training. This study examined the effects of supplementation with two proteins, hydrolyzed whey isolate (WI) and casein (C), on strength, body composition, and plasma glutamine levels during a 10 wk, supervised resistance training program. In a double-blind protocol, 13 male, recreational bodybuilders supplemented their normal diet with either WI or C (1.5 gm/kg body wt/d) for the duration of the program. Strength was assessed by 1-RM in three exercises (barbell bench press, squat, and cable pull-down). Body composition was assessed by dual energy X-ray absorptiometry. Plasma glutamine levels were determined by the enzymatic method with spectrophotometric detection. All assessments occurred in the week before and the week following 10 wk of training. Plasma glutamine levels did not change in either supplement group following the intervention. The WI group achieved a significantly greater gain (P Rationale:

The authors state that the purpose of this study was, “…to examine the effects of a hydrolyzed whey isolate in comparison to casein supplementation on strength, body composition and plasma glutamine during a 10 wk intense resistance training program.” According to the researchers, a study comparing whey with casein during a resistance training program in trained individuals hasn’t been done yet.

[Once again, I want you to note the lack of a primary outcome. We know that the researchers are interested in strength, body composition and plasma glutamine. Which is it the most important is not clear. Can you smell a data mining study yet?]

Methods:

The researchers recruited 19 male recreational bodybuilders. To be included in this study, subjects had to have at least 2 years of resistance-training experience and had to agree not to ingest any other nutritional supplement or non-prescription drug that might affect muscle growth or the ability to train intensely during the study. Subjects who a) had current or past history of anabolic steroid use, or b) had ingested any “ergogenic” supplement for 8 weeks prior to the start of the study were excluded from the study.

[Why 19? I have no idea. How did they determine no history of steroid use or previous ingestion of an “ergogenic” (whatever the heck that means) supplement? No idea. I guess they just asked them. What qualifies as a “recreational bodybuilder” anyways? I suppose that’s “recreational” as opposed to “competitive”, but “bodybuilder” vs. what? I mean, you can be a recreational hockey player in that you can play pick-up games of hockey. How can you be a recreational bodybuilder? What defines bodybuilding as an activity or a sport? By participating in spontaneous physique shows? But I digress.]

After baseline testing, subjects were, “…matched for maximal strength in three weight training exercises and then randomly assigned to either supplement.”

[So, what they did was pair up two guys who had similar lift loads in squat, cable pull-down and bench press, and then randomly put one guy in the casein group and the other in the whey group. How they randomly assigned each guy isn’t clear. It could have been a flip of a coin, it might have been as simple as an arbitrary decision (which isn’t actually random), or it could have been not random at all. What they did with the odd guy out (since they recruited 19 guys), isn’t explained either.]

Supplements

The authors claim that immediately after all baseline testing, subjects were given their assigned supplement in a “double blind procedure”. Each subject got identical, unmarked sealed containers. The explanation gets a bit dicey here. The macronutrient content of each supplement was approximately 90g protein, 3g carbohydrate and 1.5g fat per 100 g of supplement (90 + 3 + 1.5 = 94.5…so what’s in the other 5.5g?) For the casein supplement, the 90g of protein was casein. For the whey supplement the protein component was hydrolyzed whey, or simply, VP2.

Subjects were told to consume 1.5g of the supplement per kg of body mass per day, while maintaining their regular diet. The subjects spread out their daily supplement dose into smaller equal servings throughout the day. How many servings per day does not seem to be specified, although the authors provide an example of an 80kg subject who split his daily dose into 4 servings–one with breakfast, one with lunch, one post-workout and one with dinner. They were asked to record the time of day they took the supplement.

Subjects got a new container every week and had to bring their old one back every week to verify compliance with the protocol.

Testing:

Diet was monitored using the standard 3-day recall method. Subjects did one 3-day recall before the start of the study and then in the first and last week of the supplementation period.

Strength testing was done to determine the 1RM for squat, cable pull-down and bench press. This testing was done by an NSCA-certified CSCS who was blinded to group assignment.

Body composition was done by DEXA scan. Plasma glutamine was also measured through blood tests.

Strength training:

Subjects were placed on a 10-week resistance training program (Max-OT from AST Sport Science–I’m sure many of you have already heard of Max-OT, and if not, you can find out more from their site). Subjects were supervised in their workouts either 1 to 1, or 2 to 1 by a CSCS.

Statistics:

Statistical comparisons were done using ANOVAs with one between-group factor (supplement) and one repeated factor (training). Post-hoc analysis was done using t-tests. Changes from baseline within each group were compared using paired t-tests; and changes between groups was tested by unpaired t-tests. Additionally, “…comparisons of the changes made by each group…” were made with unpaired t-tests. The alpha level was 0.05.

[So, for those of you who know what an ANOVA is, it’s confusing as to why “training” is a repeated factor in this analysis, because training doesn’t vary–everyone got the same program. There is a vague reference to “training consistency” as something that was measured, but I’m not aware that it’s anywhere in the Results section, despite the fact that all the subjects kept training logs (also a notoriously pain-in-the-ass way to measure training consistency). Additionally, for those of you who have been following along with the previous reviews I’ve done, we’ve run into the classic “too many tests” problem again. I stopped counting after 22 between-group tests; and didn’t even bother counting the within-group tests.]

Results:

Of the nineteen people who were originally included in the study, only 13 people completed the study. What happened to these 6 people, is left to the imagination of the reader (I personally prefer to think that they were swallowed whole by a gravitational anomaly). But seriously, how this affects the outcome is pretty important, especially with such a small sample size. I would say this is probably one of the biggest weaknesses of this study.

With respect to baseline characteristics (age, height, weight, total strength, lean body mass and fat mass) there were no obvious differences (the researcher did statistically test for them, which is meaningless and does affect the chances you will find a spurious/erroneous significant p-value), but even by eyeballing it, the two groups look comparable–EXCEPT for body mass. The whey group had a total body mass of 84.0 kg (SD 5.0), while the casein group had a total body mass of 79.7 kg (SD 11.2). The means the casein group was not only, on average, lighter, but also had greater variation in their weight, which means in the final analysis, it would be more difficult to see improvements in the casein group. Oddly, the variance in lean body mass and fat mass didn’t look all that different between groups, which is kinda…fishy.

There were no differences detected in energy or protein intake between or within the groups.

The whey group experienced a statistically significant increase in lean body mass and a decrease in fat mass over 10 weeks, when compared to itself. A statistically significant difference was not detected in the casein group when compared to itself. Additionally, the increase in lean body mass and decrease in body fat in the whey group was statistically significantly greater than the casein group.

The whey group went from 67.1 kg (SD 2.7) to 72.1 kg (SD 2.8) in lean body mass, while the casein group went from 62.5 kg (SD 2.4 kg) to 63.3 kg (SD 2.3) in lean body mass.

For fat mass, the whey group went from 13.9 kg (SD 1.5) to 12.5 kg (SD 1.3), and the casein group went from 14.4 kh (SD 1.8) to 14.5 kg (SD 1.8).

In terms of strength gains, the whey group also outperformed the casein group. Statistically significant differences were found in favour of the whey group for all three lifts after adjusting for body mass (i.e. load lifted/kg of body mass). Most impressively, the whey group, on average, almost doubled their squat (mean initial squat 1RM of 80.2 kg (SD 7.3) to a mean final squat 1RM of 155.5 kg (SD 6.0) ) compared to the casein group (mean initial squat 1RM of 71.0 kg (SD 4.5) to mean final squat 1RM of 123.2 kg (SD 8.6) ).

No differences were detected for serum glutamine levels between groups.

Author acknowledged limitations:

This is perhaps the saving grace of this paper, although, it still leaves many questions unanswered. The authors acknowledged the following limitations to their study:

1) There was a high proportion of attrition from the study (a loss of 6 subjects–almost 1/3 of the total subjects). However, they could have reported on the subjects who left the study (in particular, which groups they left from) and could also have done an intention-to-treat anaylsis to avoid the attrition bias.

2) Although no statistical difference was detected between groups with respect to caloric consumption adjusted for body mass, it was found that the whey group was consuming 250 kcal per day more than the casein group consistently throughout the study.

3) The whey group started at a higher weight, with a narrower variance. This may have made it easier for the whey group to appear better than they really were, when compared to the casein group. One way of getting around this would have been to compare all differences adjusted for variance, or, alternatively, by body weight–though the limitation to this “correction” is that it assumes that growth is linear (that is, no matter what your starting weight is, your weight gain increases at the same rate). This may or may not actually be the case, since it may be the case that once you hit a certain weight, it maybe be easier to put on more weight (because your strength gains aren’t linear) and is also very likely that you will lose fat more easily because of an increased lean body mass (since the number of calories consumed weren’t very different pre- and post- study in either group).

Disclosure:

MSSE requires conflicts of interest be published in their studies. Paul Cribb is a consultant for AST, and AST supplied all the supplements for the study.

What makes this study good:

This study isn’t totally horrible. What impressed me the most with this study was the level of concealment used to blind the subjects from one supplement to the other. I suppose it’s possible the consistency or taste of the drinks might have been different, but whether you’d be able to tell you got whey vs. casein purely from these qualities, is not something I know.

I also liked the fact that they tried to find trained individuals and put them on a published standard program for the duration of the trial. It’s not a huge strong point, but it does help in isolating the effects of the supplement vs. variation in outcomes due to variation in training (though, in a randomized controlled trial, the variation in training program would theoretically be equal between groups, and not affect their compariability).

What makes this study bad:

I think I’ve mentioned most of the bad things already, and most of them are quite bad, but I was very disappointed that initial body mass was not controlled for in the final analysis. This could have been accomplishing using a regression model (if they could decide on the MOST important outcome–which, in my opinion would have been change in lean body mass) to determine if changes in lean body mass were dependent on the supplement used, the initial body mass or both, and also whether there were interactions between the two (e.g. the situation in which change in lean body mass was dependent on initial body mass for the whey group, but not dependent for the casein group).

I’m also quite sick and tired of seeing these tiny trials with such miniscule sample sizes that aren’t justified. I can understand that you might not need 100 guys to be in a study, and a larger sample size doesn’t mean your study is necessarily better; but if you’re going to recruit just 20 people, one third of which are going to drop out, then you should have a damn good reason for it. At the very least, if you knew 1/3 of the subjects were going to drop out, you should have recruited 30 people to hit your initial randomly-chosen 20 subject sample size.

The last comment is just my jaded sense of…well, jadedness, in that this was an industry sponsored trial in which the comparison product was supplied by the manufacturer of the product of interest. I have no reason to believe that AST would have given the casein group some horrible quality casein (I don’t even know whether casein comes in different qualities, to be honest), but because it wasn’t a placebo trial, you do have to ask the question as to whether there was pre-trial bias to begin with, with one group getting the BEST possible product and the other group getting what was _technically_ a casein product, but perhaps not the best possible casein product.

The bottom line:

Despite all the limitations to this study, hydrolysed whey does come out on top of casein with respect to strength gains. The lean body mass gain result is dubious at best because initial body mass wasn’t controlled for in the analysis, so we have no way of knowing which way that would have gone. Does this mean you should stop using casein? Well, on the basis of this particular study, if we assume that casein was given a fair shot, I would say that I would prefer to use whey, but that I would still consider a casein-containing product for nighttime, due to the slower absorption profile. What really grinds my gears is that we don’t know anything about the missing one-third of the subjects that left the study, so even though I say that whey does come out on top, I also say that with some trepidation, because there are still a handful of unanswered questions. If I had peer-reviewed this manuscript for publication, I would have required an intention-to-treat analysis, or at the very least, publication of the data available for the missing subjects. However, note that my final interpretation of this study is NOT based on any of the statistical tests they performed, but rather on the relevance of the published numbers.

I was going to calculate the 95% confidence interval around the difference in strength gain between the two groups, but unfortunately, the variance in the squat numbers for the casein group went really wide (from 4.5 to 8.6), which makes it more complicated than I was prepared to invest time in. It looks like there were possibly subjects who did extremely well (possibly better than the whey guys) on casein, and some who didn’t, so it’s possible that casein has one of those binary response distributions–you either do REALLY well on casein or you do quite bit less well, with not much of an in-between response. It’s just not possible to tell without seeing some of the raw data.

So…to make a committed statement: This study is weak-to-moderate evidence in favour of using hydrolysed whey protein as the main source of protein supplementation over casein.