Posted on

Another Look at the Fat-Free Mass Index (FFMI)

For no other reason than because it’s on my mind, I want to look at the topic of the Fat Free Mass Index (FFMI) and the idea that it does or does not indicate a natural limit for bodybuilders/weight trainers.   Actually, I take that back, writing this is due to me finding a post I made several months back in my Facebook group when I looked in some detail at a paper that was referenced at me in an attempt to disprove or dismiss the idea that an FFMI cutoff existed.

The idea, as I’ll detail below had to do with the suggestion that there was a cutoff for FFMI that did or could be used to indicate that someone was or was not using anabolic steroids. Since the original concept was originated back in 1995, there has been a lot of back and forth about the topic.  I suspect it kicked off some of the FAKE NATTY nonsense and I’ve seen various counterarguments against there being a specific FFMI limit that can be used to determine if someone is natural or not.

In addressing the topic again,  I’ll be looking at a lot of different topics in this piece, probably jumping around a lot since I seem to have no organizational skills for these articles.  And this got away from me so it’s gonna be three parts.

First, definitions.

What is the FFMI?

You can think of the FFMI as being conceptually similar to BMI (the body mass index, which is weight divided by height squared) although specifically for fat free mass (FFM).

It is defined as

FFMI = fat free mass / height^2 (that means squared)

Where weight is in kilograms (divide weight in pounds by 2.2) and height is in meters (convert height in feet and inches to inches and then go use an online calculator).  And this represents exactly what it sounds like, the relationship of FFM (everything in the body that is not bodyfat including muscle, organs, bones, water, glycogen, connective tissue, etc.) to height.  I’d note that as discussed in The Women’s Book, there is actual data on FFM per unit height in athletes including in elite bodybuilders that I will address briefly in, I wanna say, Part 2.

The Original Kouri et. al. Paper

The paper that started this mess is titled Fat-free mass index in users and nonusers of anabolic-androgenic steroids by Khouri et. al. (including a researcher named Pope who was really driving this bus) and published in The Clinical Journal of Sports medicine in 1995.  In the paper, the researchers examined 156 men from gyms in the Boston and LA area who went through a bunch of measurements including body composition (by calipers which is far from ideal) but only 134 had full data along with being given questionnaires about their steroid use (i.e. had they used, were they using, etc.)

To this, an an additional 23 who had been recruited for a testosterone cypionate study.  This gave a sample size of 157 individuals of whom 74 had (self-reportedly) never used steroids and 83 had used steroids, with 52 of those having used in the previous year.  Urine testing was administered to the subjects and the researchers felt that:

Briefly, no evidence suggested that any athlete had deliberately misrepresented his steroid use, nor did any urine test contradict an athlete’s verbal report.

This is important because, as we all know, folks lie about their steroid use and someone saying they are clean tends to mean very little.  Hence the urine test.  No, it’s not perfect, someone who used ages ago would show up as clean but this isn’t an issue that can be gotten around completely.

So far as the population used, all of the natural subjects were not just your average gym goer with the researchers stating:

The nonusers included many dedicated bodybuilders. Several had competed successfully in “natural” bodybuilding contests, two held world records in strength events, and many others were recognized by their associates as highly successful weightlifters.  Thus the nonuser group probably included individuals who closely approached the maximum limits of muscularity that could be attained without drugs.

Mind you, more details than that were not provided in terms of what bodybuilding federation or strength sport they were in and this is a huge oversight (an article I will address below brings up this point).  I’ve reproduced the results shown below and I want to make the point that their average body fat percentage was relatively low in both groups (12.8% +- 4.8%) and I will come back to this.

Khouri FFMI Results

Now, as the results above show, even with steroid use, the users were still considerably lighter on average than even the typical IFBB pro with only 175 pounds of lean body mass.   Even in 1995, the pros were significantly bigger than this (in the modern era, some are competing at 300 lbs ripped although they are using far more than just anabolics).   And, again, look at the nonusers. An average lean body mass of 158 pounds which is not huge by any stretch. Of course, that’s not taking into account the error bars. Some of the subjects in each group were larger than this and some smaller.

The researchers showed this in graphical form and you can see that none of the natural users (white triangles) exceeds a value of 25 while a great many of the steroid users do (black triangles) which led to the basic conclusion that an FFMi of 25 represented a cutoff point for natural athletes, with anyone above that value using anabolics.

It’s worth noting that many of the steroid users were below 25.  Basically, the conclusion was never that steroids would guarantee you exceeded 25 but that you couldn’t exceed 25 without them.


I’d note that the line indicating a cutoff not being a straight line indicates that something else is going on and the researchers speculated that height differences play a role.  The issue is that humans are not two-dimensional but three with body and muscle volume scaling differently.  As they state:

In other words, the tallest athletes are not only taller but also wider and thicker than the shorter athletes of apparently comparable muscularity, thus the tallest athletes scores somewhat higher on the FFMI equation.

This led them to develop a normalized FFMI equation which is

Normalized FFMI = FFMI + 6.1x(1.8-h)

where h is in meters.  Regraphing this, they found that the naturals stopped abruptly at a 25 FFMI while users went above this.

Normalized FFMI Note that many of the steroid users were still below 25.  And many of the non-users didn’t even get close to the abrupt cutoff point at 25.  But none of the non-users got across it.  Not one.

The researchers also examined past Mr. USA results from 1939 to 1959 based on reported heights and weights and assuming 5% body fat (my note: probably not a safe assumption as bodybuilders of the era were not as lean as today) showing that they had a mean FFMI of 25.4 +1.5 with 3 having values above 27.0.

Again I’d note that the assumption of 5% bodyfat is unlikely to be correct.  Guys just didn’t get as lean and I bet if you corrected for that, the FFMI values would fall very close to the supposed 25 cutoff point.  But I’m not doing the math since the weights and heights of the bodybuilders weren’t reported in the paper and I’m just not that driven to look them up.

There is more to the paper, comments about limitations (such as BF% being measured by skinfolds, although calculations show that it wouldn’t have mattered much overall).

One important comment, and one I will be coming back to below is this:

Fourth, our formula may not be satisfactory for fat individuals.  Because a gain in the fat component of the body is consistently accompanied by some gain in the lean component, it is possible taht individuals might be able to exceed substantially an FFMI of 25 without steroids.  Because our sample of nonusers only included six men with bodyfat levels of >20%, it is probably hazardous to generalize the finding of fat individuals.

Read that a couple of times, it will be important in Parts 2 and 3.

And with all of that the researchers report, in a preliminary fashion (admittedly based on a smallish sample) that an FFMI cutoff of 25 kg/m^2 may represent a natural cutoff or screening tool for whether or not someone is using steroids.

Pope would follow this with a book called The Adonis Complex that I honestly never bothered to read.  It seems that there he took the conclusion from this paper to a little bit greater of an extreme and at least some of the pushback against the cutoff is due to the rather absolute statements he appears to have made in that book (i.e. that NOBODY can naturally surpass the limit or some such, again I didn’t read it and am relying on reports of what it said).

And let’s just say that the Internet went fairly batshit over this finding/idea, that there was some specific limit that could not be surpassed with anabolics (or, in reverse, that an FFMi beyond a certain point was PROOF of steroid use).  But people seem to refuse to accept that there are ANY genetic limits to how much muscle one can gain naturally.

Mind you, many seem to think that there are no genetic or natural limits but that’s nonsensical to begin with.  Of course there are limits and if you look at natural bodybuilding, there has been little to no progress in body weight for decades now despite increasing numbers of people entering the sport.  Where that limit may lie might be more debatable.  But to believe that there is NO limit is pie in the sky nonsense.

There are arguably more folks competing, with better nutrition and training, and few are really any heavier than they were way back when.  Clearly there is a limit and folks have hit it.  And note that this is true in other sports where world records have basically stopped improving since 1988.  You’d have to go pretty far up your butt to argue that natural bodybuilders are continuing to move up and up when all other elite athletes (many if not most who are using drugs) have stagnated for nearly 30 years.

Now I wrote about this paper back in 2009 and in hindsight, my conclusions were a bit stronger at the time than they should have been (and far stronger than they are now).  Because at that time, although I don’t seem to have said it in my paper analysis, I did kind of feel that 25 represented a more or less absolute cutoff for natural athletes.  Even then I suspect or at least hope I acknowledged that:

  • There are exceptions to the 25 cutoff for natural athletes.
  • They are only exceptions.
  • You are unlikely to be one of them because that’s not what an exception is (I will beat this dead horse in this series like you would not believe).

I have certainly made the above points in every asinine argument I’ve had over the topic in recent years.

I do want to mention that there was a minor error in my original piece (or was until I was made aware of it towards the end of LAST year and corrected it).  As I stated, in Pope’s original paper, they examined the FFMI of past Mr. USA winners and I incorrectly stated that only 2-3 surpassed the 25 FFMI limit.  This was incorrect and it was more like 5-6 (3 surpassed 27 which is probably what threw me when I wrote 2-3 mistakenly).

However, this is a totally inconsequential issue as it no way changes the interpretation of the original paper.  Whether the value is 5-6 or 2-3 doesn’t change the fact that, while an FFMI cutoff of 25 might not be an ABSOLUTE cutoff (in the sense of being 100% certain) for whether someone is natural or not, it might as well be.

Because if only a small percentage of the most elite of the elite can break it (i.e. 99.9% of lifters will not even get close to the value to begin with), it might as well be 100%.  Or rather, statistically, you can say with some degree of certainty that anybody going above that is not natural.  Or rather (mark 2), by the definition of what an exception is, they are unlikely to be an exception.  If statistically 99.9% people never get past the cutoff without drugs, by extension anybody who does is MOST LIKELY to be using.

Arguments Against the FFMI Cutoff

A number of different arguments against the proposed FFMI cutoff have been made, many of them aimed at me (or being due to arguments I inserted myself into because so much bullshit is being written to try to dismiss the idea of ANY cutoff).  One is that older data like the Mr. USA data from above isn’t relevant because techniques and methodologies of training and nutrition have improved. Possibly and I won’t deny that.

But as I noted above, this belies the fact that in probably natural athletes or federations, competition body weights haven’t really changed much in the past couple of decades.  The best superheavy weights come into condition in the low 200’s and most guys compete in the lighter classes.

If we really had that much better nutrition and training (and no BFR does not count), why are guys basically the same size?  At most we have more guys in the higher weight classes due to more involvement in the sport.  But bigger athletes in absolute terms?  There aren’t many and the ones that there are can be counted on one hand.  In contrast, pro bodybuilders, who take ALL the drugs continue to get bigger.

One was the claim that old times bodybuilders and strongmen, and here I’m talking about the early 20th century had reported weights and heights that would put their FFMI above 25.  Umm, ok. But here’s the thing: these guys are known to have been NOTORIOUS for lying about this stuff. It was part of the showmanship of the day, you lied about what you could lift, how big you were, how big your arms were, etc.   Do you believe every guy online who benches 600 for reps or who has a friend with a cousin that is 300 lbs ripped but who can’t seem to ever show you video or pictures?  If so, I have oil wells in Texas to sell you.

Without any accurate record of their height and weight (and what is available is assuredly self-reported which means nothing), this is simply unreliable as hell.  Even if it is true and the weights and heights are accurate, if this is the best argument you can make, that a handful of guys 100 years ago exceeded FFMI to say that the cutoff doesn’t exist, well….that’s a reach and I don’t think it proves much of anything.  So 5 guys or whatever in a century got there or whatever. Hardly compelling and if anything it makes the opposite point that 99.9% of people can’t get past it.  The 5 exceptions are meaningless.

Another argument comes from a 2008 article that was sent to me penned by Jan Todd.  Now, Jan and her recently deceased husband Terry are LEGENDS in the strength training world. They kept an amazing library of strength training information here at Austin and Jan was one of if not THE first female powerlifters (as discussed in The History of Women’s Sports Part 2).  I have the utmost of respect for both her and her husband and what they have done.  But I disagree with at least some of what she wrote.

Titled Size Matters and published in Iron Game History in August of 2008 where she obviously disagrees with some of the strong statements that were made surrounding Pope’s data.  Most of her disagreement has to do with the idea that people will take the FFMI cutoff as PROOF that anybody above it is using steroids which is inherently problematic.  As well she contends that the idea might stimulate some men to start using anabolics (presumably based on the idea of a genetic limit) which I consider a bit of a reach. But no matter.

She does make it VERY clear that she is not arguing against their being a genetic limit, or that athletes in bodybuilding and other sports aren’t using.  Rather it’s more of an argument against profiling athletes and decided they are drug users due to exceeding the FFMI cutoff.  In that sense, I could probably take this bit out of the article and not change much as any disagreement I have is fairly minor and irrelevant and more a matter of semantic nitpicking than frank disagreement.   But it’s already written so nyah.

One thing she rightly takes issue with are pictures in Pope’s popular book showing what they considered a “Bodybuilder” and certainly this is not consistent with what most would think as the elite of the sport. To make this point she brings up Dave Goodin, an Austin bodybuilder who I happen to know.  Dave has been an amazing bodybuilder for decades, owned the old school Hyde Park Gym and put on a natural bodybuilding contest for years (one year I got tricked into judging).

Knowing him, I have no reason to doubt that he’s natural.   He also has amazing genetics, I’ve seen him squat in the 400-500 lb range and he’s clearly not representative of the majority of lifters.  He is just another exception in a small pool of exceptions, representing the top 1% of the 1% who ever touched a weight.  But her argument comes down to the idea that the original sample size was not representative of elite bodybuilders.

She also disagrees with Pope’s analysis of the Mr. USA competitors with her and Terry’s own similar analysis putting the mean FFMI at 26.4 and 15 above 26 and compare this to Pope’s mean value of 25.4 and 8 clearing 26.  And this is just nitpicking so far as I’m concerned.

The first value is not enormously different than the mean found in Pope’s original study although many more exceed that 26 value (15 vs. 8).  At best this still simply points out that the top 1% of elite bodybuilders might surpass the 25 value.  But it’s still at most 15-20 people.  Grimek, a beast of a man was estimated with an FFMI of 31.  He’s also not representative of the average trainee.

She also brings up the issue of early 20th century strongmen with pictures of 9 of them but, as above, I don’t think that means much.  Without accurately recorded measures of literally anything they did including their height or weights, it’s all speculation.  Certainly they were large and muscular, nobody is doubting that.  But any assertions about FFMI are only that.

Even if you accept that the 9 men she shows in her article exceeded the FFMI cutoff, well, that’s 9 added to 15 in the Mr. USA sample and Grimek.  So that’s 25 men up to that point in the weight game.  This shows that there can be exceptions to be sure.  But among perhaps a small percentage of the top 1% of folks.  Yes, fine, perhaps there were more than this during the time period. We have no indication of whether this is or is no the case.  I’m simply making the point that so far the countarguments come down to “Here are a few dozen genetics freaks who surpassed it, therefore the cutoff is wrong.”

Once again, her argument is less with the FFMI per se so much as it is with the idea that the idea of an ABSOLUTE cutoff exists (or that the value is 25 per se) or that it can be used to decided whether or not someone is using steroids or not.  I don’t disagree with this.  As I stated above, clearly there can be exceptions.  And it did get a bit absurd when the idea that ANYBODY WHO CROSSED THE 25 FFMI VALUE WAS USING STEROIDS.  Clearly this isn’t true and I’m really hoping that I never said something that absolute.  If I did years ago, I was wrong.

But if the best you can do is come up with a couple of dozen exceptions, that doesn’t make a strong case for their being MANY exceptions.  As above, and I will repeat this a dozen times, a 25 FFMI cutoff may not be absolute in a literal sense.  But it sure as hell is in a practical sense.  If less than 1% of everybody who has ever touched a weight has a chance to get to it much less past it naturally, with 99%+ certainty I can assume that anybody who does surpass it isn’t natural.

And that’s where I’ll cut it today.

Read Part 2.

Similar Posts:

Facebook Comments