You are on page 1of 114

V O L U ME 4 , ISS U E 5 MAY 2 0 2 0

MASS
M ONTHLY A PPL ICATIO N S IN
STRE N G TH SPO R T

E R IC H E LMS | G R E G N UCK O LS | MIC HAEL ZO URDO S | ERIC T REXL E R


The Reviewers
Eric Helms
Eric Helms is a coach, athlete, author, and educator. He is a coach for drug-free strength and physique
competitors at all levels as a part of team 3D Muscle Journey. Eric regularly publishes peer-reviewed
articles in exercise science and nutrition journals on physique and strength sport, in addition to writing for
commercial fitness publications. He’s taught undergraduate- and graduate-level nutrition and exercise
science and speaks internationally at academic and commercial conferences. He has a B.S. in fitness
and wellness, an M.S. in exercise science, a second Master’s in sports nutrition, a Ph.D. in strength
and conditioning, and is a research fellow for the Sports Performance Research Institute New Zealand
at Auckland University of Technology. Eric earned pro status as a natural bodybuilder with the PNBA in 2011 and competes in
unequipped powerlifting, weightlifting, and strongman.

Greg Nuckols
Greg Nuckols has over a decade of experience under the bar and a B.S. in exercise and sports science.
Greg earned his M.A. in exercise and sport science from the University of North Carolina at Chapel
Hill. He’s held three all-time world records in powerlifting in the 220lb and 242lb classes. He’s trained
hundreds of athletes and regular folks, both online and in-person. He’s written for many of the major
magazines and websites in the fitness industry, including Men’s Health, Men’s Fitness, Muscle & Fitness,
Bodybuilding.com, T-Nation, and Schwarzenegger.com. Furthermore, he’s had the opportunity to
work with and learn from numerous record holders, champion athletes, and collegiate and professional
strength and conditioning coaches through his previous job as Chief Content Director for Juggernaut Training Systems and
current full-time work on StrongerByScience.com.

Michael C. Zourdos
Michael (Mike) C. Zourdos, Ph.D., CSCS, has specializations in strength and conditioning and skeletal
muscle physiology.  He earned his Ph.D. in exercise physiology from The Florida State University (FSU)
in 2012 under the guidance of Dr. Jeong-Su Kim. Prior to attending FSU, Mike received his B.S. in
exercise science from Marietta College and M.S. in applied health physiology from Salisbury University.
Mike served as the head powerlifting coach of FSU’s 2011 and 2012 state championship teams. He
also competes as a powerlifter in the USAPL, and among his best competition lifts is a 230kg (507lbs)
raw squat at a body weight of 76kg. Mike owns the company Training Revolution, LLC., where he has
coached more than 100 lifters, including a USAPL open division national champion.

Eric Trexler
Eric Trexler is a pro natural bodybuilder and a sports nutrition researcher. Eric has a PhD in Human
Movement Science from UNC Chapel Hill, and has published dozens of peer-reviewed research
papers on various exercise and nutrition strategies for getting bigger, stronger, and leaner. In addition,
Eric has several years of University-level teaching experience, and has been involved in coaching since
2009. Eric is the Director of Education at Stronger By Science.

2
Table of Contents

6
BY G R EG NUCKOL S

Improving Muscle Growth by Individualizing Training Volume


In a recent study, subjects trained one leg with standardized training volume (the same volume
for everyone), and one leg with individualized volume (a 20% increase above their prior baseline).
Individualized volume led to more quad growth, suggesting that we should be more concerned
about manipulating training volume based on your recent history, rather than aiming to find one
single “optimal” level of training volume for everyone.

16
BY M I CHAEL C. ZOUR DOS

Can You Predict Bench Press 1RM Using Dumbbells?


If you can dumbbell bench 45kg dumbbells for 10 reps, what does that translate to in a barbell
bench press? We’ve all probably been asked as much or thought about what that conversion
is like. This article reviews the first study to examine the ability of the dumbbell bench press to
predict barbell bench 1RM.

26
BY E RI C HEL MS

The First Bodybuilding Refeed Research


Refeeds, carb and calorie cycling, diet breaks, cheat meals and days, zig-zag dieting; in one form
or another, bodybuilders have been implementing nonlinear fat loss tactics for decades. First they
were based on anecdotes, then combined with mechanistic, indirect, and some human data on non-
athletes. However, this study is arguably the first truly relevant trial to date.

39
BY E R I C T R EXL ER

Can Sleeping More Make You Leaner?


We all know that insufficient sleep is unpleasant, but does it hinder our gains in the gym? A new study
found that a sleep education intervention enhanced fat loss throughout a resistance training program,
but with some notable limitations. Read on to find out whether or not your sleep habits might be
getting in the way of your physique goals.

50
BY G R EG N UCKOL S

You Can’t Flex Bone: Predicting Powerlifting Performance Based on


Fat-Free Mass
I think everyone knows that more muscular lifters tend to be stronger lifters, but I don’t think people
recognize just how strong that relationship is among competitive powerlifters. In a recent study on
raw, drug-free powerlifters, fat-free mass was almost perfectly correlated with performance. This
article will explore that relationship, and what we can do with it.

3
69
BY M I CHAEL C. ZOUR DOS

The Scientific and Practical Considerations of Long Inter-Set Rest


Intervals
Longer inter-set rest intervals can lead to greater hypertrophy and strength gains than shorter inter-set
rest intervals. This study demonstrates that eight-minute rest intervals may be needed to maximize
volume performance. But is eight minutes really necessary? This article discusses both the scientific
merit and practical considerations of rest interval duration.

82
BY E R I C TR EXL ER

Has Carnitine Pivoted From a Fat Loss Supplement to a Recovery


Supplement?
Back in the day, carnitine was widely touted as a fat loss supplement. While that hype has mostly died
down, a new meta-analysis suggests that carnitine may reduce post-exercise muscle damage and
soreness. Read on to find out if the results of this meta-analysis are as good as they seem.

97
BY G R EG N UCKOL S

Variety is the Spice of Life: If You Want Well-Rounded Triceps


Growth, You Need Both Compound and Single-Joint Exercises
A recent study examined triceps growth (among other things) in subjects that only did bench press,
only did triceps extensions, or performed both exercises. To maximize growth of all three heads of the
triceps, a combination of both exercises was necessary.

108 BY M I CHAEL C. ZOUR DOS

VIDEO: Bodyweight and Home Training Progressions


In these uncertain times, you may not have gym access. This video first presents scientific data
demonstrating that bodyweight training can help to preserve muscle mass and strength. Second,
this video presents a nine-week at-home training program along with specific examples of household
items that can be used. Further, the video demonstrates various examples of how progressive
overload can be achieved with this approach.

110
BY M I CHAEL C. ZOUR DOS

VIDEO: Nutrition on Lockdown


Obviously it’s less than ideal to be stuck at home, and you or your clients may be struggling with eating
due to boredom, depression, or emotional reasons. However, there are other aspects of being stuck
at home that may be influencing your food behaviors. Likewise, there may be some silver lining to this
change in environment which could be used to establish better nutrition habits that outlast lockdown.

4
Letter From the Reviewers

M ASS is entering its fourth year! If you’ve been here for a long time, we’re grate-
ful you’ve been willing to stick with us, and if you signed up during the anniver-
sary sale and this is your first issue, we’re thrilled to have you aboard.
A MASS anniversary is as opportune a time as any to inform you or remind you about
a couple of the resources in the member site that a lot of you may not be aware of. We
try to avoid jargon as much as is reasonable in our written articles, but we occasionally
need to use some technical terminology; if you’re unsure about the meaning of a word,
or if you’d just like to brush up on some foundational concepts, you should check out the
MASS Glossary. Each issue also includes a “missed the cut” section with cool studies
that we didn’t write about, but which may still interest you. If you’re not confident in
your ability to read and interpret a study for yourself, you should check out our Interpret-
ing Research guide.
Starting with the video department, Eric Helms and Mike both had timely videos for
people who are stuck at home during COVID-19 lockdowns. Eric’s video discusses how
being stuck at home can influence food behavior, and how you can use this time to build
good eating habits that will persist once stay-at-home orders are lifted. Mike’s video dis-
cusses at-home bodyweight training, and actually includes a full program you can follow
to maintain muscle and strength without gym access.
From the nutrition department, Eric Helms’s article this month covers the first study
investigating the effects of refeeds in an athletic population. Eric Trexler’s articles cov-
er the impact of sleep on body composition, and a new meta-analysis investigating the
effects of L-carnitine on muscle damage.
From the training department, Mike reviewed one study investigating the relationship
between dumbbell bench performance and bench press 1RM strength, and another study
on the effects of rest intervals on volume load performance. Greg reviewed studies ex-
amining the relationship between fat free mass and powerlifting performance, the impact
of single- and multi-joint training on muscle growth, and whether individualized training
volume leads to more muscle growth than standardized training volume.
As always, if you have any questions or feedback about any of the content in this issue,
feel free to post in the Facebook group. We’ll do our best to answer any questions you
have about the topics we cover.

5
Study Reviewed: Muscle Hypertrophy Response Is Affected by Previous Resistance
Training Volume in Trained Individuals. Scarpelli et al. (2020)

Improving Muscle Growth by


Individualizing Training Volume
BY G RE G NUC KO LS

In a recent study, subjects trained one leg with standardized training


volume (the same volume for everyone), and one leg with individualized
volume (a 20% increase above their prior baseline). Individualized
volume led to more quad growth, suggesting that we should be more
concerned about manipulating training volume based on your recent
history, rather than aiming to find one single “optimal” level of training
volume for everyone.

6
KEY POINTS
1. Over eight weeks, trained subjects trained one leg with 20% more weekly sets
than they’d been using in their training previously, and trained the other leg with
22 sets per week, regardless of prior training volume.
2. Subjects experienced significantly more quad growth, on average, in the leg that
underwent a 20% volume increase, even though average volume ended up being
similar between conditions.
3. When assigning training volume, it’s better to focus on gradually increasing
volume from one’s current baseline, rather than simply jumping to some level of
volume that’s theoretically ideal for the average person.

I magine a scenario for me: We


fast-forward 20 years in the fu-
ture, and we now have enough
tive to each individual’s baseline train-
ing volumes. “Low volume,” “moder-
ate volume,” and “high volume” don’t
data to accurately and precisely model necessarily mean, say, 0-10, 11-20, and
the “inverted U” relationship between 20+ sets per muscle group per week; I
training volume and muscle growth. see them more like, “less volume than
It’s found that the Optimal(™) training I’m accustomed to,” “about the same
volume is 20 sets per muscle group per amount of volume I’m accustomed to,”
week. A client comes to you, having re- and “considerably more volume than
cently plateaued after previously making I’m accustomed to.”
solid gains on 8 sets per muscle group That’s not the type of thinking that
per week. Do you A) bump their volume generally informs research design, un-
up gradually and see how they respond, fortunately. The presently reviewed
or do you B) move them straight from 8 study (1) is an exception. The research-
sets per week to the Optimal(™) volume ers interviewed experienced lifters and
of 20 sets per week? asked them how many sets of quad
To me, option A seems like a no-brain- training they performed per week. For
er. Population averages may or may each subject, one leg was assigned to an
not apply to an individual, and this cli- individualized condition in which vol-
ent’s baseline ability to adapt to training ume for that leg was increased by 20%
stressors is heavily influenced by their over their pre-study volume. For exam-
recent training. “High volume” or “low ple, if someone had been doing 5 sets
volume” training aren’t fixed points of quad training per week, then that per-
across all lifters, but are fluid constructs son started doing 6 sets; if they’d been
that vary person-to-person, defined rela- doing 40 sets, they started doing 48.

7
The other leg was assigned to a stan- forming knee extensions and leg press.
dardized condition, which consisted of
22 sets for all subjects. After 8 weeks Experimental Design
of training, quadriceps cross-sectional This study utilized a within-subject
area (CSA) increased significantly more unilateral design. That means that for
in the individualized condition than the each subject, one leg was randomized
standardized condition. Thus, when an- into the standardized condition, with
alyzing and assigning training volume, the other leg entering the individual-
it’s more informative and productive to ized condition. Subjects trained twice
look at changes relative to the individu- per week for eight weeks, performing
al’s training history, instead of compar- unilateral leg press followed by unilat-
ing their training volume to some fixed eral knee extensions. Volume for the
reference point. standardized leg was fixed at 22 sets per
week (based on the average weekly vol-
ume used in an allegedly random sample
Purpose and Hypotheses of studies in the literature), which I as-
sume means 5-6 sets apiece, in each ses-
Purpose sion, of unilateral leg press and unilater-
The purpose of this study was to com- al knee extensions per workout. Volume
pare the hypertrophic effects of train- for the individualized leg was based on
ing with a standardized (i.e. 22 sets per each subject’s self-reported quad train-
week) amount of volume versus an in- ing volume in the two weeks prior to
dividualized (+20% of an individual’s the start of the study; the researchers in-
previous volume) amount of volume. creased weekly volume by 20% above
baseline for the individualized leg on a
Hypotheses per-subject basis. In both conditions, all
The authors hypothesized that individ- sets were taken to technical failure, and
ualized volume would lead to more mus- loads were adjusted so that the subjects
cle growth than standardized volume. would reach the point of failure after
8-12 reps per set. Subjects rested two
minutes between sets.
Subjects and Methods Like any longitudinal training study,
testing was performed before and after
Subjects the training program. No strength out-
The subjects were 16 young males comes were tested, so the only testing
with an average of 5.1 years of resis- consisted of measuring quad CSA via
tance training experience, regularly per- ultrasound scans.

8
Figure 1 Muscle cross-sectional area increases from baseline to after 8 weeks resistance training

25 4.5
Standardized

Individualized vs. Standardized (cm2)


Individualized
3.0
20

1.5
15
ÆCSA (%)

0.8 (TE)

CSA
0.0

10 -0.8 (TE)

-1.5

5
-3.0
2.61 (CV)

0 -4.5

A B

A) Dashed line indicates co-efficient of variation: 2.61%


B) Gray circles indicate whether the increases in CSA (cross-sectional area) obtained with the individualized protocol are greater than
(above 0.8 typical error [TE]), smaller than (below -0.8 TE), or similar (between 0.8 and -0.8 TE) to those obtained with the standardized.
Dashed line indicates the measurement typical error: 0.8cm2
* = significantly different from standardized (p < 0.05)

Findings
Both conditions completed similar
Quadriceps CSA increased signifi- volume loads on average, but there were
cantly more in the individualized condi- large differences on an individual basis.
tion than in the standardized condition All subjects increased their volume load
(+9.9% vs. +6.2%; p = 0.042; mean dif- relative to baseline by approximately
ference = 1.08cm2; CI = 0.04-2.11cm2). 20% in the individualized condition. In
Furthermore, of the 16 subjects, 10 had the standardized condition, changes in
superior increases in quadriceps CSA in volume load from pre-training ranged
the individualized condition compared from a reduction of almost 50% to an
to the standardized condition that were increase of 120%. Overall, 4 of the 16
larger than the typical measurement er- subjects decreased their weekly set vol-
ror of the ultrasound CSA assessments. ume in the standardized condition, and
Only 2 out of 16 subjects had superior 8 out of 16 subjects increased their set
increases (in excess of the measurement volume by 30% or more, meaning that
error) in their leg assigned to the stan- only 4 subjects had modest (<30%) in-
dardized condition. Hypertrophy was creases in set volume.
similar between conditions in the other
four subjects.

9
troduces a source of heterogeneity into
Interpretation the results of the study. More important-
This study represents an important ly, we can claim that increasing set vol-
piece of the puzzle that’s been sore- ume should increase muscle growth, but
ly missing from the resistance training training studies aren’t actually designed
literature. When studying the effects of to directly support or refute that idea,
training volume on muscle growth, sub- since training volume assigned to each
jects tend to be assigned to differing vol- participant doesn’t consider the lifter’s
ume conditions without any regard for previous training volume.
prior training history. If a study uses 10 A major strength of this study was
sets per week as a low-volume condition that the volume ended up being similar
and 20 sets per week as a high-volume in both conditions. If volume differed
condition, the 10-set condition may feel significantly between conditions, one
like super high volume and be very dif- could then claim that maybe the indi-
ficult to recover from for someone who vidualized condition just led to more
had previously been training with a vol- growth because volume was higher (or
ume of just 5 sets per week for the target because volume was lower, thus posit-
muscle group; conversely, 20 sets per ing that that other condition led to over-
week could feel like a deload for some- reaching). Since both conditions led
one accustomed to 30 sets per week. Ef- to similar training volume on average,
fects like that should theoretically wash this study supports the idea that optimal
out with random group allocation if the training volume is highly individualized
sample size is sufficient, but it still in- and probably based, in part, on individu-

10
al training histories. I think that’s a more creasing your volume, or if you’re not
productive way to approach the problem making gains and constantly feel worn
than simply trying to discover the single down with 30 sets per week, you could
number that represents the Optimal(™) surmise that you probably don’t need
weekly set volume. If, for example, we to be afraid of knocking your volume
had 10 times as much research and we down. However, I do think that assign-
could pinpoint 15 sets per week as the ing volume based on subjects’ habitual
weekly set volume associated with the training volume is a more informative
largest average gains in the literature, I study design, and it more closely resem-
still don’t think that would be incredibly bles how I approach things as a coach.
helpful. If you’ve been making gains It also makes better sense biologically.
with 5 sets per week, tripling your train- We know that people habituate to pri-
ing volume overnight probably isn’t a or stimuli, that sufficient physiologi-
great idea, and if you’ve been making cal stressors drive adaptation, and that
gains with 30 sets per week, cutting your stressors can be excessive for an organ-
training volume in half also probably ism to adapt positively to (relative to the
isn’t the ticket to improved hypertrophy stressors they’re currently habituated
outcomes. However, if we had base- to). Manipulating training volume rela-
line training data in all of those studies tive to each subject’s established base-
in this theoretical future meta-analysis, line respects what we know about how
and we could determine that a 10% in- people adapt to training stimuli, where-
crease in volume every six weeks was as giving half your subjects X volume
associated with the largest gains, I do per week and half your subjects 2X vol-
think that information would be highly ume per week does not. In other words,
generalizable for most trainees. It may the approach used in the current study
not be helpful to someone who’s already respects what we’ve learned in the past
pushing the boundary of overtraining, century about how organisms respond to
but it would probably be useful for the stress, while the standard approach for
vast majority of lifters. That’s not to studying the impact of volume on mus-
say that the current body of literature is cle growth seemingly pretends like that
without merit; learning more about the body of knowledge doesn’t exist.
levels of volume that produce the best The results of this study raise an in-
outcomes, on average, can still serve as teresting possibility, which we’ve dis-
a useful sanity check. For example, if cussed in MASS before: volume cycling.
we find out that 15 sets per week is Opti- If periodic increases in weekly set vol-
mal(™) and you’re not making progress ume reliably produce increases in mus-
with 5 sets per week, you could surmise cle growth, that’s all well and good, but
that you could probably benefit from in-

11
there must be a limit to that approach. If and 30%, and 20% came out on top, then
you compound 20% increases in set vol- we could claim that, on average, a ~20%
ume over time, even if you only increase increase in volume maximizes hypertro-
set volume every 6 months, you’ll even- phy over a matter of months. There’s a
tually wind up with completely unman- second way to test this idea using the data
ageable training volume. There must be from the present study, though it would
a finite amount of training that people not be able to establish causation. You
can respond positively to. However, we could compare the relationship between
also know that lifters can decrease train- volume changes in the standardized group
ing volume considerably – by at least and hypertrophic results relative to the
two-thirds – and not lose muscle (2). individualized condition. In other words
Thus, we could gradually increase set (sorry, long sentence incoming), if there
volume over time until it becomes un- was a clear inverse parabolic relation-
manageable, reduce training volume to ship, where subjects who decreased their
establish a new baseline, and repeat the volume in the standardized condition
process (4). You could do this with all and subjects who massively increased
muscle groups simultaneously, or (more their volume in the standardized condi-
realistically), with a handful of muscle tion both experienced way less growth in
groups at a time. Currently, the concept their quad assigned to the standardized
of volume cycling is used “in the trench- condition than in their quad assigned to
es,” but it only has indirect theoretical the individualized condition, but subjects
support in the literature. I’d love to see who increased their weekly volume by a
it studied more directly. modest amount in the standardized con-
There’s one more thing I’d like to clear dition experienced similar growth in both
up about the results of the current study: legs, that would then suggest that modest
We know that a 20% increase in training increases in training volume are, in fact,
volume above the subjects’ pre-training optimal for hypertrophy. This method
baseline worked well, but we can’t claim wouldn’t be able to pinpoint 20% as op-
(as the authors of the study do) that a 20% timal, but it could provide additional evi-
increase in training volume “maximize[s] dence that somewhere in the ~20% range
muscle hypertrophic response in trained (give or take 10-20%) is optimal. On the
individuals.” There’s only one way we contrary, if the peak of the inverted pa-
could ascertain the optimal (average) in- rabola was closer to 50%, that may sug-
crease in training volume to maximize gest that the current study was too con-
hypertrophy: We would need a study to servative, and larger increases in volume
compare different volume increases. If a are actually optimal for hypertrophy, in
study tested increases of, say, 10%, 20%, the short-to-medium-term.

12
by 6.6 ± 12.4 sets per week. Those dif-
AS WITH ALL METRICS, ferences represented relative increases
of 12%, 33%, and 58%. At first, that
WE SHOULDN’T EXPECT looks like a clear linear relationship,
suggesting that larger increases in vol-
RELATIVE INCREASES IN ume are predictive of more hypertrophy,
and that the 20% increase in the present
TRAINING VOLUME TO BE study was too conservative. However,
we can’t overlook those enormous stan-
PERFECTLY PREDICTIVE dard deviations. For the subjects in the
top tertile (the subjects that experienced
OF HYPERTROPHY. the most hypertrophy), an increase of
6.6 ± 12.4 sets per week suggests that
several of the top-responding subjects
As with all metrics, however, we likely decreased their training volume,
shouldn’t expect relative increases in and several increased their training vol-
training volume to be perfectly predic- ume by more than 100%. In that same
tive of hypertrophy. In support of the study, the subjects’ actual training vol-
notion that volume increases are only ume in sets per week during the study
weakly associated with hypertrophy, a protocol also wasn’t predictive of hy-
recent training study by Aube and col- pertrophy. Changes in leg fat-free mass
leagues (3) performed exploratory anal- and quad muscle thickness were virtual-
yses to see what factors differentiated ly identical between groups performing
high and lower hypertrophic responders. 12, 18, or 24 sets per week. Change in
They split their group into tertiles based training volume was more predictive of
on hypertrophy response (the lowest ter- results than actual training volume was,
tile experienced virtually no growth, and
the highest tertile experienced the most
growth), and they found that their low-
est tertile had increased their training
volume by just 1.8 ± 6.7 sets per muscle
CHANGE IN TRAINING VOLUME
group per week (that was the difference
between their pre-study training volume
WAS MORE PREDICTIVE
and the volume they were assigned for
the study), their middle tertile had in-
OF RESULTS THAN ACTUAL
creased their training volume by 4.3 ± TRAINING VOLUME WAS.
9.3 sets per week, and their highest ter-
tile had increased their training volume

13
APPLICATION AND TAKEAWAYS
Slightly increasing your training volume above its current level will probably help you
build more muscle, as long as you aren’t already straining your ability to recover from
and adapt positively to training. Theoretically, cycling training volume over fairly long
time scales (e.g. multiple months) may lead to more muscle growth than sticking with
a fixed level of training volume over time.

but neither metric was strongly associ-


ated with hypertrophy. So, I still believe
that research should be more focused on
finding ideal changes in training volume
from subjects’ pre-training baselines
rather than in finding a single, magi-
cal, absolute training volume, but we
shouldn’t expect that either metric will
be anywhere close to perfect for predict-
ing muscle growth.

Next Steps
I’d love to see a longitudinal training
study where subjects are assigned dif-
fering levels of volume increases from
baseline. To cover a wide swath of pos-
sibilities, it could include four groups,
with groups increasing weekly set vol-
ume by 0%, 25%, 50%, and 100% to
cover a wide range of possibilities (5).
I would anticipate that either a 25% or
a 50% increase would yield the best
results out of those four options. That
would help us establish a general area
where we should be looking in future
studies.

14
References

1. Scarpelli MC, Nóbrega SR, Santanielo N, Alvarez IF, Otoboni GB, Ugrinowitsch C, Libardi
CA. Muscle Hypertrophy Response Is Affected by Previous Resistance Training Volume in
Trained Individuals. J Strength Cond Res. 2020 Feb 27.
2. Bickel CS, Cross JM, Bamman MM. Exercise dosing to retain resistance training adapta-
tions in young and older adults. Med Sci Sports Exerc. 2011 Jul;43(7):1177-87.
3. Aube D, Wadhi T, Rauch J, Anand A, Barakat C, Pearson J, Bradshaw J, Zazzo S, Ugrinow-
itsch C, De Souza EO. Progressive Resistance Training Volume: Effects on Muscle Thick-
ness, Mass, and Strength Adaptations in Resistance-Trained Individuals. J Strength Cond
Res. 2020 Feb 13.
4. Note that the process of increasing volume doesn’t have to be super rapid. If you’re making
progress at a given level of volume, it may be worthwhile to ride it out until you stop making
progress, rather than being greedy and proactively ramping your volume up prematurely.
That’s a decision you can make for yourself, though.
5. Another option would be to simply increase volume by, say, 1, 3, 5, and 10 sets per week.
Percent increases may be problematic if some of the subjects were already training with real-
ly high volumes. I strongly assume a 50% increase from 6 to 9 sets per week would go swim-
mingly, but a 50% increase from 40 to 60 sets per week would probably be way overboard.

15
Study Reviewed: Relationship of Barbell and Dumbbell Repetitions with One-
Repetition Maximum Bench Press in College Football Players. Heinecke et al. (2020)

Can You Predict Bench Press


1RM Using Dumbbells?
BY MIC HAE L C . ZO URD O S

If you can dumbbell bench 45kg dumbbells for 10 reps, what does that
translate to in a barbell bench press? We’ve all probably been asked as
much or thought about what that conversion is like. This article reviews
the first study to examine the ability of the dumbbell bench press to
predict barbell bench 1RM.

16
KEY POINTS
1. This study examined the accuracy of predicting barbell bench press one-
repetition maximum (1RM) from reps to failure at a moderate load with four types
of benching: 1) barbell bench, 2) bilateral dumbbell (DB) bench, 3) right arm only
DB bench, and 4) left arm only DB bench.
2. Reps performed during all types of benching predicted barbell 1RM with pretty
similar accuracy.
3. While it is difficult to directly apply this data at all times, it is possible that DB
bench progress could be used as a proxy for barbell bench progress in certain
situations; however, more data is needed to confirm this.

I f I can do the 50lb DBs for 10 reps,


then what can I bench?” We have
all heard someone ask a similar
failure (i.e. one arm at a time) with the
same 45.5kg load. Reps performed, on
average, were 13.8 ± 9.2 on the barbell
question to that at some point. To an- bench press and 12.5 ± 9.5 on the bilat-
swer, I usually say I don’t know, then eral DB bench press, suggesting that a
I think of someone I know of similar moderate load was used (i.e. ~65-70%
strength to the person asking and base of 1RM). In brief, the findings showed
my answer off of that person. That’s high correlations between repetitions
not a particularly good response, but it to failure during all types of benching
helps to give something of a reference and barbell 1RM (r = 0.86-0.90). Fur-
point. If you have a new client who does ther, using linear regression equations,
DB bench more than they barbell bench reps to failure with all types of bench-
press, then knowing the relationship be- ing predicted 1RM with pretty similar
tween DB strength and barbell strength accuracy. These predictions were not
may be useful. The reviewed study from perfect, but on average, they predicted
Heinecke et al (1) is the first to examine barbell 1RM with an average error of
the relationship between reps performed 7.3-8.1 kg. This article will discuss the
on the DB bench press and barbell bench finer details of these findings and will
press 1RM. 40 collegiate American present situations when using DB bench
football players performed repetitions strength as a proxy for barbell 1RM may
to failure with a 90.9kg barbell bench be useful.
press load and 45.5kg (in each hand)
DB bench press load on two different
days. On a third day, the athletes per-
formed unilateral DB bench presses to

17
study. In the first session, the athletes
Purpose and Hypotheses completed barbell bench press 1RM
testing. In the second, third, and fourth
Purpose sessions, the athletes completed one set
The purpose of this study was to ex- to failure using a 90.9kg (200lb) barbell
amine the accuracy of predicting barbell bench load, a set to failure with 45.5kg
bench press 1RM from sets to failure (100lb) DBs (in each hand) on the bilat-
with a moderate load during four differ- eral DB bench, and sets to failure on the
ent types of benching: 1) barbell bench unilateral DB bench (one arm at a time),
press, 2) bilateral DB bench press, 3) respectively. The authors noted that the
right arm only DB bench press, and 4) 90.9kg barbell load was chosen because
left arm only DB bench press. they wanted the DB load to be the same,
and from observing normal training, the
Hypotheses athletes had trouble balancing more than
No hypotheses were provided. 45.5kg DBs in each hand.
The authors then examined the rela-
tionship between reps performed during
Subjects and Methods all types of benching and barbell bench
press 1RM. Additionally, the authors ex-
Subjects amined the accuracy of predicting bar-
40 collegiate American football players bell 1RM from each type of benching.
who had been lifting at the collegiate lev-
el for at least one year participated. The
available subject details are in Table 1. Findings
First, the repetitions performed on
Study Protocol each set and the relationship between
This protocol was simple. Four ses- reps performed and barbell bench press
sions, each separated by exactly 48 1RM are presented in Table 2. As you
hours, were needed to complete this can see, there are strong relationships

18
between reps performed and barbell – 7.1kg; DB bench – 7.9kg; right arm
1RM for each type of pressing. Re- DB bench – 8.3kg; left arm DB bench
member, these high R-values are just a – 8.0kg. Figure 1 shows Bland-Altman
correlation; all the correlations tell you plots of each 1RM prediction versus
is that those who performed more reps the actual barbell bench press 1RM.
tended to have a higher 1RM. Observ- A Bland-Altman is an agreement plot,
ing a correlation is completely different which is used because it is much more
than accurately predicting 1RM or say- difficult to get methods to agree as op-
ing that the predicted and actual 1RM posed to just correlate (2). On these
agree with each other. plots, the horizontal line at 0 is the ac-
For the 1RM predictions, none of the tual average 1RM, and each dot is an
predictions were statistically differ- individual data point (one person’s pre-
ent than the actual 1RM (p>0.05); but diction). The dashed horizontal lines
again, that does not mean that the pre- on the top and bottom represent the
dictions “agreed” with the actual 1RM. limits of agreement. For each type of
For example, the average errors (essen- benching, some of predictions were off
tially how many kilograms, on average, by as much as 20kg and many others
each prediction was off from the actu- were off by as much as 10kg; howev-
al 1RM during each type of benching) er, other predictions were almost dead
were as follows: barbell reps to failure on. In general, these plots show good

19
agreement on average, but it is clear- rate. One other point that is noteworthy
ly better to look at these predictions on when looking at the plots is that the re-
an individual level, as the predictions gressions tended to over-predict 1RM
were accurate for some, but way off for for weaker athletes, and under-predict
others. Additionally, the limits of agree- 1RM for stronger athletes.
ment are fairly wide (i.e. almost -20 to
+20kg), which further demonstrates
that some predictions were not accu-

Figure 1 Agreement between actual and predicted 1RM

Barbell reps performed prediction Bilateral DB bench prediction

30
30

20
Predicted - actual 1RM (kg)

Predicted - actual 1RM (kg)

20

10
10

0 0

-10 -10

-20 -20

-30 -30
100 125 150 175 200 100 125 150 175 200

A Actual 1RM bench press (kg) B Actual 1RM bench press (kg)

Right arm only DB bench prediction Left arm only DB bench prediction

30 30
Predicted - actual 1RM (kg)

Predicted - actual 1RM (kg)

20 20

10 10

0 0

-10 -10

-20 -20

-30 -30
100 125 150 175 200 100 125 150 175 200

C Actual 1RM bench press (kg) D Actual 1RM bench press (kg)

20
cise had probably mirrored their barbell
bench press progression, which helped
IF YOU NEVER PERFORM THE the prediction. If you never perform the
DB bench, but barbell bench frequently,
DB BENCH, BUT BARBELL then using a DB bench to predict barbell
1RM would probably be far less accu-
BENCH FREQUENTLY, THEN rate for you. Conversely, if you only DB
bench (possibly a bodybuilder), then
USING A DB BENCH TO DB bench press reps to failure may be
a poor predictor of barbell strength, in
PREDICT BARBELL 1RM that case. If the prediction isn’t terribly
accurate for the bodybuilder, that’s not
WOULD PROBABLY BE FAR too consequential as they are likely per-
forming a rep max test just to get a rough
LESS ACCURATE FOR YOU. estimate of their 1RM. The question is,
when might using DB bench press reps
to failure come in handy for you as a
lifter or coach?
Interpretation For some lifters, a shoulder tweak
To start, let’s return to Figure 1 for a might make high volume benching
moment. When visually comparing the tough at times, but using DBs may not
plots, reps performed on the barbell cause any pain. In that situation, the
bench press is unsurprisingly the best present data could be useful to get a
at predicting barbell 1RM, but not by rough estimate of a barbell 1RM. This
that much. All plots are actually pret- hypothetical individual could even be
ty similar, which is supported by the a powerlifter who can bench from time
similar R-values in Table 2. At first, to time, but they can’t (or don’t) do the
it seems kind of ridiculous to say that majority of their volume training with
reps to failure on the DB bench is just the barbell bench press. The regression
as good at predicting barbell 1RM as equation printed in the paper to predict
reps to failure on the actual bench press. barbell 1RM when using 45.5kg DBs on
However, since all these athletes were the bilateral DB bench was: 1RM (kg) =
on the same football team, that means 2.115 (Reps Performed) + 110.04. This
they probably did the same training pro- is a simple linear regression, or in other
gram, which included both bench press words: y (1RM in kg) = a (intercept) x
and DB bench. Therefore, since these (repetitions performed) + b (slope). The
individuals were used to performing the best way to deem if the prediction de-
DB bench, the progression of that exer-

21
rived from a linear regression equation the linear Epley (1RM = Load X Reps
is practically accurate is to perform this X 0.0333 + (Load)) and Bryzki (Load /
prediction within close proximity (~1 [102.78 – 2.78(reps)] equations to pre-
week) of an actual barbell bench 1RM. dict 1RM (average error of ~3-5kg), but
If the prediction is accurate, then you these predictions were made from using
can continue to use the DB bench pre- higher loads (3RM and 5RM). So, the
diction over time to gauge progress or current study reporting an average error
potentially help with planning attempts of 7.9kg for a linear equation based on
for a 1RM test or a powerlifting meet, bilateral DB bench performance isn’t
but I would stress that you gauge the ac- too bad considering the present study’s
curacy before doing the latter. predictions were based on sets with an
The authors of this study stated that average of 12.5 reps. Another 1RM pre-
the predictions of barbell bench press diction method, submaximal velocity, is
1RM were made with “excellent ac- nice in theory since only working up to a
curacy and only minimal error.” I re- few submaximal singles isn’t very fatigu-
spectfully disagree with that statement. ing. However, the more data that comes
As seen above in Figure 1, these pre- out on submaximal velocity predictions,
dictions were indeed quite accurate for the less convinced I am that these meth-
some individuals; however, the predic- ods are accurate. MASS previously re-
tions were >20kg off for others, and the viewed a study from Perez-Castilla et al
estimation errors were between 7.1-8.3 (5), which reported two- and four-point
kg on average. In my opinion, these submaximal velocity methods to predict
predictions are far from “excellent,” seated cable row 1RM very accurately
and the Bland-Altman plots show that; (average error of ~3kg). Garcia-Ramos
however, the predictions are decent on et al reported a two-point submaximal
the group-level, especially compared to velocity method, which used loads that
1RM predictions from reps performed corresponded to an average concentric
equations and submaximal velocities in velocity of 1.0 and 0.5 m/s on a Smith
other studies. For example, Reynolds et machine bench press to predict 1RM to
al 2006 (3) found that both linear and within -9-3 kg (6). As an aside, here is
nonlinear reps-performed equations a tutorial on how to perform submaxi-
were significantly inaccurate (p<0.05) mal velocity predictions. While the Pe-
when using a 10RM load or higher in rez-Castilla and Garcia-Ramos predic-
the barbell bench press. In compari- tions are quite accurate, those findings
son to Reynolds and the present study, don’t hold up on free-weight compound
Distacio 2014 (4) found more accurate movements. Specifically, Banyard et al
1RM back squat predictions when using 2017 (7) found that even when using a
five-point submaximal velocity method,

22
bench press was capable of predicting
barbell 1RM with similar accuracy to
ESTIMATING YOUR BARBELL bilateral DB bench press, unilateral
training should be included in a pro-
BENCH PRESS 1RM FROM A gram aimed at improving bench press
strength. I don’t buy into that idea on
SET OF DB REPS TO FAILURE this data alone. However, as we have
written about before, unilateral training
WON’T BE PERFECTLY has utility for muscle imbalances and
for increasing strength of the contralat-
ACCURATE, BUT THE eral limb if one limb cannot be trained
due to injury.
ESTIMATE IS COMPARABLE
(OR EVEN BETTER) TO Next Steps
SOME OTHER MORE One idea for follow-up research is to
investigate if other assistance exercis-
CONVENTIONAL METHODS es can predict the 1RM on a main lift.
Right away though, I’d like to see how
OF PREDICTING 1RM. much the accuracy of the DB bench pre-
diction improves with lower reps (~5
reps). Additionally, I’d like to see how
accurate the DB bench is at predicting
which used loads up to 90% of 1RM (20, barbell 1RM for a group of individu-
40, 60, 80, and 90% of 1RM) to predict als who don’t barbell bench very much.
the back squat 1RM, there was an aver- Earlier in this article, it was postulated
age error of 19.8kg, which is far from that this data might be useful for some-
practically accurate. This is all to say one with a shoulder issue who does a lot
that estimating your barbell bench press of DB pressing because they can only
1RM from a set of DB reps to failure bench with low volume and farther from
won’t be perfectly accurate, but the es- failure. If the DB bench is a reason-
timate is comparable (or even better) to able predictor of barbell bench strength
some other more conventional methods in individuals who don’t train the bar-
of predicting 1RM. bell bench particularly heavy, then DB
Finally, the authors of this paper fo- bench progress could be used as a proxy
cused much of their discussion on hy- for barbell 1RM progress.
pothesizing that since unilateral DB

23
APPLICATION AND TAKEAWAYS
1. This study showed that a linear regression equation based on reps performed
on the DB bench press was not only reasonably accurate at predicting barbell
bench press 1RM, but that this prediction was almost as accurate as using a set
of barbell bench press to failure.
2. This data may be useful for someone who cannot currently barbell bench press
heavy due to a shoulder injury but is able to DB bench press with sufficient vol-
ume. In that scenario, the DB bench could be used as a proxy for barbell bench
press.
3. It is important to interpret this data with caution since all of the athletes here were
likely performing both barbell and DB bench presses in their current programs.
If you are not used to performing one or the other, predicting barbell 1RM from a
set of DB bench to failure would probably be less accurate.

24
References

1. Heinecke ML, Mauldin ML, Hunter ML, Mann JB, Mayhew JL. Relationship of Barbell
and Dumbbell Repetitions With One Repetition Maximum Bench Press in College Football
Players. The Journal of Strength & Conditioning Research. 2020 Mar 23.
2. Scott LE, Galpin JS, Glencross DK. Multiple method comparison: statistical model using
percentage similarity. Cytometry Part B: Clinical Cytometry: The Journal of the Internation-
al Society for Analytical Cytology. 2003 Jul;54(1):46-53.
3. Reynolds JM, Gordon TJ, Robergs RA. Prediction of one repetition maximum strength from
multiple repetition maximum testing and anthropometry. The Journal of Strength & Condi-
tioning Research. 2006 Aug 1;20(3):584-92.
4. DiStasio TJ. Validation of the Brzycki and Epley Equations for the 1 Repetition Maximum
Back Squat Test in Division I College Football Players. (Doctoral dissertation, MA thesis,
Southern Illinois University Carbondale).
5. Pérez-Castilla A, Suzovic D, Domanovic A, Fernandes JF, García-Ramos A. Validity of
different velocity-based methods and repetitions-to-failure equations for predicting the
one-repetition maximum during two upper-body pulling exercises. Journal of strength and
conditioning research. 2019.
6. García-Ramos A, Haff GG, Pestaña-Melero FL, Pérez-Castilla A, Rojas FJ, Balsalo-
bre-Fernández C, Jaric S. Feasibility of the 2-point method for determining the 1-repetition
maximum in the bench press exercise. International Journal of Sports Physiology and Per-
formance. 2018 Apr 1;13(4):474-81.
7. Banyard HG, Nosaka K, Haff GG. Reliability and validity of the load–velocity relationship
to predict the 1RM back squat. The Journal of Strength & Conditioning Research. 2017 Jul
1;31(7):1897-904.

25
Study Reviewed: Intermittent Energy Restriction Attenuates the Loss of Fat Free Mass in
Resistance Trained Individuals. A Randomized Controlled Trial. Campbell et al. (2020)

The First Bodybuilding


Refeed Research
BY E RI C HE LMS

Refeeds, carb and calorie cycling, diet breaks, cheat meals and days,
zig-zag dieting; in one form or another, bodybuilders have been
implementing nonlinear fat loss tactics for decades. First they were
based on anecdotes, then combined with mechanistic, indirect, and
some human data on non-athletes. However, this study is arguably the
first truly relevant trial to date.

26
KEY POINTS
1. This study compared body composition and metabolic outcomes in two groups
of trained men and women following the same net weekly energy deficit for seven
weeks; however, one group was in a larger deficit for five days followed by two days
at estimated maintenance, while the other sustained a small deficit on all days.
2. Overall, outcomes were better in the refeed group, which retained significantly
more dry fat-free mass and had a slightly better maintenance of resting metabolic
rate (RMR). The superior maintenance of RMR is likely related to the retention of
fat-free mass, since fat-free mass is a large contributor to RMR.
3. This study isn’t without limitations, but it is at least initial evidence that refeeds
might be beneficial for dieting individuals looking to maintain lean mass. Follow
up research will help us determine if this is a robust finding that stands up to
replication.

T he concept of a “refeed” has been


around in one form or another
for decades in the bodybuilding
en were split into two groups following a
25% net energy deficit for seven weeks;
one group had a daily 25% deficit, the
world. While there are multiple forms, other group had a 35% deficit for five
the common colloquial definition is an days followed by two days at mainte-
increase in calories, while dieting, that nance. At the post-test, both groups lost
lasts a day or a few days at most. Typi- a significant amount of body fat; how-
cally, this is an increase to maintenance, ever, the continuous diet group also lost
or to a surplus, and often the proportion a significant amount of fat-free mass
of carbohydrate increases to purport- and experienced a significant reduction
edly replenish glycogen and positive- in RMR while reductions in the refeed
ly impact energy expenditure. Ideally, group were non-significant for fat-free
these outcomes would result in higher mass and RMR. Additionally, the refeed
training quality, mitigation of metabolic group retained significantly more “dry
adaptations, a step away from a state of fat-free mass” (defined later) compared
low energy availability, a mental break to the continuous group. In this review,
from dieting, and ultimately better body I’ll go into the history of refeeds and
composition outcomes. While there’s a explain the outcomes of this study, dis-
fair amount of research on intermittent cussing their relevance, limitations, and
dieting, this is the first study (1) where ultimately applications for coaches and
the researchers simulated refeed prac- athletes alike.
tices from physique sport. Specifically,
trained, reasonably lean men and wom-

27
Purpose and Hypotheses Subjects and Methods
Purpose Subjects
The purpose of this study was to com- A mixed-sex group of 27 participants
pare body composition, RMR, and leptin with resistance training experience com-
concentration changes in two groups pleted this trial. You can see the average
of resistance trained men and women baseline characteristics for age, height,
after seven weeks of net energy-defi- body mass, and years of resistance train-
cit-matched dieting, with one group us- ing experience in Table 1. As a side note,
ing a seven-day per week continuous 58 people were originally recruited. 4
deficit approach, and the other group us- never showed up for baseline testing, 3
ing a five-day on (deficit), two-day off never showed up to the study training
(maintenance) intermittent approach. sessions after baseline testing, 21 with-
drew during the study, and data from 3
Hypotheses other participants weren’t included in
The authors hypothesized that both di- the analysis after self-reported nutrition-
ets would result in the successful main- al non-compliance. This ~46% comple-
tenance of lean mass and metabolic rate tion rate is not abnormal for longitudinal
since they both included resistance exer- studies of energy restriction and serves
cise and a relatively high protein intake. as a reminder of why nutrition trials are
They also noted that they were unsure if difficult to complete.
the refeed group would experience bet-
ter body composition or metabolic rate Study Overview
outcomes due to there being no prior re- This was a randomized, parallel-group,
search on refeeds in a trained population. controlled trial. After recruitment, par-
ticipants came to the lab for initial
screening and nutritional tracking in-
struction, and the participants began a
two-week period in which they tracked
their body mass and nutrition to estab-
28
lish their maintenance calories: the in- additional carbohydrates on their two
take estimated to result in no change in refeed days to reach estimated mainte-
body mass over time. Then, participants nance. Participants were assigned an ex-
were randomly assigned to either the perienced member of the research team
refeed or the continuous group, which as a personal nutrition coach to answer
both followed a 25% net weekly energy diet-related questions during the study.
deficit relative to their estimated mainte- To control for post-workout nutrition
nance requirements. The weekly energy and to aid in reaching the daily protein
distributions for the refeed and continu- target, participants were given whey
ous groups are shown in Table 2. protein isolate after each supervised re-
Participants in both groups were in- sistance training session. Self-reported
structed to consume 1.8g/kg of protein baseline and diet-day energy and mac-
per day, with the remaining calories ronutrient intakes are shown in Table 3.
evenly distributed between dietary fat Pre- and post-testing consisted of mea-
and carbohydrate. In the refeed group, suring body mass, RMR via metabolic
participants were instructed to consume cart, total body water via bioimped-

29
ance, and body fat percentage, body fat Supervised resistance training consist-
mass, and fat-free mass via ultrasound. ed of a four-day upper/lower split with
Pre- and post-testing occurred under the two versions of each session type per-
same conditions (time, assessor, equip- formed per week. In week four, a de-
ment) following an overnight fast. Be- load was performed such that only two
fore pre-testing, participants were in- training days occurred: one lower- and
structed to refrain from physical activity one upper-body session. In each session,
for 48 hours, and post-testing occurred two exercises were required, and then
48 hours following the final exercise the participants were free to select from
session. In the refeed group, post-test- the available options for three additional
ing occurred at least two days after the lower body exercises or four additional
last refeed day in an attempt to control upper body exercises, depending on the
for glycogen-mediated water retention. day. Sets progressed from three to four
Additionally, the researchers calculated in most exercises. Repetition ranges var-
“dry fat-free mass” by subtracting the ied slightly by exercise, but appeared to
bioimpedance-derived total body water follow a linear progression such that all
from the ultrasound estimation of fat- exercises were performed in the 6-8 and
free mass as another method to control 8-10 rep range by week seven. Loads
for potential differences in body water were selected so sets were one rep from
between groups (more on this later). failure (1 RIR) while falling in the target

30
repetition range. Tables 4a and 4b dis- however, these decreases in the refeed
play an example of the training with the group were nonsignificant. The only
target rep ranges from week seven. between-group difference was that the
Finally, a subsample of eight partici- refeed group lost significantly less dry
pants had their fasting leptin concen- fat-free mass than the continuous group.
trations assessed; however, this was The body composition and RMR data
not balanced between the sexes (seven are displayed in Table 5 with within- and
males and one female), and the subsa- between-group significance and with-
mple was not large enough to represent in-group effect sizes. Individual partici-
true between-group differences, so I pant changes in fat and fat-free mass are
won’t discuss these outcomes. shown in Figures 1 and 2.

Findings Interpretation
Both groups significantly decreased Since I joined the bodybuilding and
weight, fat mass, and body fat percent- fitness communities in the mid 2000s,
age. Regarding within-group chang- I’ve been aware of the concept of re-
es, fat-free mass and RMR decreased feeds (and other similar concepts). On
significantly in the continuous group; the Muscular Development forums in

31
the mid 2000s, Dave Palumbo had a Infamous underground bodybuilding
well-established contest-prep diet that author Dan Duchaine wrote about carb
was low carb, high protein for 5-6 days, (and, as a consequence, calorie) cycling
followed by a carb cheat meal(s) or in his “Underground Body Opus” pub-
day(s). I remember studying for my first lished in 1996. Finally, my 3DMJ col-
personal trainer certification through league Jeff Alberts – who first competed
the ISSA in 2005 and learning Dr. Fred in 1993 – was aware of bodybuilders’
Hatfield’s “zig-zag diet” – in which you carbohydrate cycling since the 1980s
cut with higher calories on training days (which incidentally lines up with the
and lower calories on other days – in the early research on carbohydrate loading
sixth edition ISSA handbook from 2000. following depletion [2]).

Figure 1 Individual participant changes in fat mass


Figure 2 Individual participant changes in fat-free mass

4
4
3
2
Fat mass (kg)

2
Fat-free mass (kg)

1
-2

0
-4

-6 -1

-8 -2

Refeed group Continuous group -3

-4

Refeed group Continuous group

32
It’s safe to say that something resem- populations studied, whether controlled
bling refeeds has been used in practice or ad libitum periods were included in
for at least three decades. In peer-re- the non-dieting periods, whether in- or
viewed research, intermittent caloric re- out-patient models were used (self-re-
striction has been studied for its effects port, metabolic ward, or somewhere in
on health, obesity, and lifespan, begin- between), and the intended outcomes of
ning with mice in the early 1980s (3). the various interventions.
Today, this area of research has expand- However, bodybuilding-derived ap-
ed. A recent meta-analysis (4) compared proaches entered the peer-reviewed
the effects of continuous restriction on literature only recently. Our own Dr.
health and obesity to various forms of Trexler was the lead author of an open
intermittent caloric restriction including access review on metabolic adaptations
alternate day fasting, the five and two to weight loss in athletes (5). In it are
diet (the opposite of the present study: the first peer-reviewed mentions of re-
two days of aggressive energy restriction feeds as a bodybuilding strategy to mit-
with five days unrestricted), and other igate metabolic adaptation. Last year, I
intermittent approaches. While this me- joined lead author Jackson Peos in an
ta-analysis reported similar outcomes open access review on all forms of in-
in studies of continuous and intermit- termittent restriction viewed through a
tent restriction, there was a great deal lens of theoretical utility for athletes (6).
of heterogeneity in the setup of these Finally, Dr. Brandon Roberts led a paper
intermittent approaches, including the on nutrition recommendations for phy-

33
Metabolic Adaptations to Weight Loss”
in Volume 2, Issue 5 back when he was a
WHILE A LARGE INCREASE IN guest reviewer. If you prefer video con-
tent, I recorded “The Nuts and Bolts of
CALORIES FOR A SHORT TIME Diet Periodization” in Volume 1, Issue 5
PERIOD MIGHT TEMPORARILY back in 2017. I’m happy to say all three
still hold up as representations of best
INCREASE GLYCOGEN, TRAINING practice informed by current data.
INTENSITY, AND RELIEVE DIETARY The present study is the first examin-
ing refeeds as used by modern physique
CRAVINGS, SOME EVIDENCE competitors attempting to counter met-
INDICATES THAT CERTAIN abolic adaptations (as illustrated in this
published interview; 8) in a trained pop-
PHYSIOLOGICAL PROCESSES ulation. In the last couple of years, I’ve
specifically seen the utility of 48-hour
CHANGE IN RESPONSE TO refeeds reported by bodybuilders. I’ve
REFEEDING BASED ON THE personally experienced, and have seen
in our 3DMJ athletes, more consistent
LENGTH OF THE REFEED, NOT weight loss, better visual outcomes, and
JUST THE ENERGETIC CONTENT. more time without menstrual cycle dis-
ruption (okay, I haven’t personally ex-
perienced that one) since we started us-
ing multi- versus single-day refeeds. If
sique athletes (7), which myself and Dr. my anecdotes are accurate observations
Trexler both contributed to. We placed of a causal effect of multi-day refeeds,
refeeds on a spectrum of non-linear fat I’d speculate this could be due to the
loss strategies, as shown in Figure 3. energetic and temporal effect of increas-
There’s a lot of ground to cover in ing calories. While yes, a large increase
this area, which is beyond the scope of in calories for a short time period (say
this review. But, all three of these pub- a meal or a day) might temporarily in-
lications are open access papers target- crease glycogen, training intensity, and
ing athletes and coaches. However, if relieve dietary cravings, some evidence
you’re intimidated by peer-reviewed re- indicates that certain physiological pro-
search, or you simply don’t like formal cesses change in response to refeeding
writing, I discuss some of the overlap- based on the length of the refeed, not just
ping data in a review of a study on diet the energetic content of the refeed. For
breaks in MASS here. Also, Dr. Trexler example, luteinizing hormone pulsatili-
wrote a concept review “Dealing with ty, a marker of menstrual cycle stabili-

34
ty, is dysregulated following substantial physique sport. Intermittent energy re-
energy restriction. In one study (9), a striction may help mitigate that effect.
single day where women consumed far Getting back to our present study, the
in excess of maintenance calories did findings are straightforward, yet prelim-
not reverse this marker of menstrual cy- inary, and they confirm the efficacy, al-
cle disruption; however, another study beit cautiously, of the non-linear dieting
found a 48-hour period of eating rough- practices that science-interested body-
ly at maintenance did restore luteinizing builders and coaches previously pieced
hormone pulsatility (10). Interestingly, together from indirect research and
while the energy distribution pattern anecdotal experience. Simply put, the
during restriction and refeeding were body composition outcomes were better
different, the net energy intake was sim- in the refeed group. At the group level,
ilar in these studies. As I discussed in a they lost nonsignificant amounts of fat-
MASS article, menstrual cycle disrup- free mass while the continuous group
tion is just one of many effects from low lost significant amounts of fat-free mass,
energy availability, which is an almost and the refeed group lost significantly
inevitable consequence of dieting in less dry fat-free mass. Most likely as a
consequence of retaining more fat-free
mass, RMR was slightly better retained
in the refeed group; however, the abso-
lute differences are unlikely to be clini-
SINCE RMR IS PRIMARILY cally meaningful (~40kcal/day).
IMPACTED BY FAT-FREE Regarding the limitations of this study,
I’ve seen a number of them discussed
MASS, AND GIVEN THAT by folks in our community, and I agree
with some of these assessments, but not
THERE WAS SLIGHTLY all. First, one potential limitation is that
the body composition changes could be
BETTER RMR RETENTION artefacts of the refeeds. High carbohy-
drate intakes increase muscle glycogen,
IN THE REFEED GROUP, I associated glycogen-bound water, and
can subsequently mask or magnify body
TEND TO THINK THESE ARE composition changes. While the authors
PROBABLY “REAL” BODY specifically waited at least 48 hours
post-refeed to assess body composition,
COMPOSITION DIFFERENCES. and they calculated dry fat-free mass
(which was the only between-group
change), some have questioned wheth-
35
er 48 hours is long enough to wash out see three individuals who gained very
these effects and pointed out that dry small amounts of fat-free mass, three that
fat-free mass isn’t a validated method of barely lost any fat-free mass, four that lost
controlling for this potential confounder. around ~1.5kg, and then it slopes down-
I don’t disagree with these specific cri- ward along the continuum of participants
tiques, but since RMR is primarily im- who lost from a little over 2kg to a little
pacted by fat-free mass, and given that over 3kg of fat-free mass. Looking at the
there was slightly better RMR retention refeed group in the same figure, to me, it
in the refeed group, I tend to think these basically looks like the continuous group
are probably “real” body composition data set, if all but the two far left partic-
differences. ipants had ~0.5-1kg more fat-free mass.
Another interpretation of the data is that Don’t agree with my visual interpreta-
there were “responders and non-respond- tion? That’s fine! I actually don’t like ei-
ers” to the refeeds, or even that there were ther interpretation, because it’s really not
“outliers.” I think this stems from the vi- appropriate to try to describe “trends” in
sual assessment of Figure 2, which shows such small data sets. In this case, doing so
individual changes in fat-free mass. To be isn’t even figuratively “missing the forest
clear, an outlier has a specific (although for the trees”; rather, it’s akin to being 5
not universal) definition in statistics. miles away from a forest, squinting real-
There are tests and criteria to determine ly hard at just a section of it, and trying
how a data point looks in relation to the to guess what specific type of pine trees
rest of the data, and how many standard make up the forest. Since we only have
deviations from the mean it is; such tests 13-14 data points in each group to assess,
can be used to determine if a data point any small variation due to chance (or
is truly an outlier. Often, especially in a perhaps a real responder/non-responder
small data set, an outlier results in the scenario) would drastically change the
data differing significantly from a normal visual appearance of the figure, and may
distribution. The authors specifically not- or may not be representative of how re-
ed their data were normally distributed feeds operate broadly. Lastly, a potential
and reported no outliers. In a small data limitation Dr. Trexler brought up in peer
set such as this, we would have to see an review is that the refeed group consist-
individual with more extreme changes to ed of seven females and six males, while
get far enough from the mean to qualify the continuous group had six females and
as an outlier. Ok, so were there “respond- eight males. In our published bodybuild-
ers and non-responders”? I personally ing recommendations (7), we noted that
don’t think that’s an accurate way of de- female competitors seem to retain mus-
scribing the data. Viewing the continuous cle mass more consistently than men. The
group in Figure 2 from right to left, we sex-disparity between groups is pretty

36
APPLICATION AND TAKEAWAYS
This is the first study examining bodybuilding-style, two-day carbohydrate refeeds in
mixed-sex groups of lean, resistance trained dieters. After seven weeks with the same
prescribed net energy deficit, better fat-free mass retention was observed in the refeed
group. Prior data support the concept that intermittent approaches to energy restric-
tion, at the very worst, are equivalent to daily energy restriction. This study indicates
that the previously unstudied use of refeeds may aid in fat-free mass retention and
may be worth implementing, although future studies are needed to confirm its efficacy.

small (refeed: 7 female, 6 male; contin- ic measures of muscle mass (ultrasound


uous: 6 female, 8 male), but then again, muscle thickness). Additionally, I’d like
the body composition differences were to see leptin collected in all participants,
also small. I don’t think it’s fair to claim as well as other metabolic and hormonal
definitively that a sex difference contrib- data (thyroid hormone, insulin, testoster-
uted to the outcomes, but it is possible. one, cortisol) and measurements taken
What is fair to say is that like any time midway through the study so we can see
the first study on a specific practice is the time course of changes. Lastly, I’d
published, we should be cautious in in- love to see an assessment of other com-
terpreting the data and fight any urges to ponents of energy expenditure. Some-
get overly excited, drastically change our thing as simple as a step count would be
practice, or leap to any sweeping con- cool, ideally blinded to the participants,
clusions. This is exciting and intriguing, to see if non-exercise activity thermogen-
yet speculative work. We need further esis was impacted. Also, another study
research to confirm if refeeds are truly could be conducted comparing refeeds of
advantageous, and if they are, what the differing macronutrient distributions to
best approach to refeeding is and which assess how critical it is that refeeds are
mechanisms underpin the outcomes. dominated by carbohydrate. Mechanistic
data shows better leptin response to over-
feeding carbs versus fat (11), but does
Next Steps this play out in longitudinal outcomes?
Or could we just increase calories with
I would love to see a similar study us- a more “normal” macronutrient spread?
ing multiple groups with different re-
feed setups (no refeeds, single day, and
48-hour) with performance data report-
ed (1RMs, volume) and also site-specif-

37
References

1. Campbell BI, Aguilar D, Colenso-Semple LM, Hartke K, Fleming AR, Fox CD, Longstrom
JM, Rogers GE, Mathas DB, Wong V, Ford S. Intermittent Energy Restriction Attenuates the
Loss of Fat Free Mass in Resistance Trained Individuals. A Randomized Controlled Trial.
Journal of Functional Morphology and Kinesiology. 2020 Mar;5(1):19.
2. Costill DL, Sherman WM, Fink WJ, Maresh C, Witten M, Miller JM. The role of dietary car-
bohydrates in muscle glycogen resynthesis after strenuous running. The American journal of
clinical nutrition. 1981 Sep 1;34(9):1831-6.
3. Goodrick CL, Ingram DK, Reynolds MA, Freeman JR, Cider NL. Effects of intermittent
feeding upon growth and life span in rats. Gerontology. 1982;28(4):233-41.
4. Harris L, Hamilton S, Azevedo LB, Olajide J, De Brún C, Waller G, Whittaker V, Sharp T,
Lean M, Hankey C, Ells L. Intermittent fasting interventions for treatment of overweight and
obesity in adults: a systematic review and meta-analysis. JBI database of systematic reviews
and implementation reports. 2018 Feb 1;16(2):507-47.
5. Trexler ET, Smith-Ryan AE, Norton LE. Metabolic adaptation to weight loss: implications
for the athlete. Journal of the International Society of Sports Nutrition. 2014 Dec;11(1):7.
6. Peos JJ, Norton LE, Helms ER, Galpin AJ, Fournier P. Intermittent Dieting: Theoretical
Considerations for the Athlete. Sports. 2019 Jan;7(1).
7. Roberts BM, Helms ER, Trexler ET, Fitschen PJ. Nutritional Recommendations for Phy-
sique Athletes. Journal of Human Kinetics. 2020 Jan 31;71(1):79-108.
8. Mitchell L, Hackett D, Gifford J, Estermann F, O’connor H. Do bodybuilders use evi-
dence-based nutrition strategies to manipulate physique? Sports. 2017 Dec;5(4):76.
9. Loucks AB, Verdun M, (With the Technical Assistance of Rebecca Brown and Jean R.
Thuma). Slow restoration of LH pulsatility by refeeding in energetically disrupted women.
American Journal of Physiology-Regulatory, Integrative and Comparative Physiology. 1998
Oct 1;275(4):R1218-26.
10. Olson BR, Cartledge TA, Sebring NA, Defensor RU, Nieman LY. Short-term fasting affects lu-
teinizing hormone secretory dynamics but not reproductive function in normal-weight seden-
tary women. The Journal of Clinical Endocrinology & Metabolism. 1995 Apr 1;80(4):1187-93.
11. Dirlewanger M, Di Vetta V, Guenat E, Battilana P, Seematter G, Schneiter P, Jequier E, Tap-
py L. Effects of short-term carbohydrate or fat overfeeding on energy expenditure and plas-
ma leptin concentrations in healthy female subjects. International journal of obesity. 2000
Nov;24(11):1413-8.

38
Study Reviewed: A Randomized Controlled Pilot Trial of Sleep Health Education on Body
Composition Changes Following 10 Weeks Resistance Exercise. Jabekk et al. (2020)

Can Sleeping More Make


You Leaner?
BY E RI C T RE X LE R

We all know that insufficient sleep is unpleasant, but does it hinder


our gains in the gym? A new study found that a sleep education
intervention enhanced fat loss throughout a resistance training
program, but with some notable limitations. Read on to find out
whether or not your sleep habits might be getting in the way of your
physique goals.

39
KEY POINTS
1. The presently reviewed study (1) evaluated the effects of a sleep education
intervention on body composition changes in response to 10 weeks of resistance
training.
2. Both groups increased lean mass, but the exercise + sleep group lost significantly
more fat mass. However, the study did not attempt to standardize dietary habits,
the details of the training program were unclear, and the sleep questionnaire did not
provide strong evidence that the exercise + sleep group actually got more (or better)
sleep than the exercise only group.
3. Despite the notable limitations, the results of this study generally agree with the
broader body of literature indicating that sleep quality and quantity can be modified
by practicing good sleep hygiene, and that getting adequate sleep yields favorable
effects for performance and body composition.

I f you’ve been a MASS reader for a


while, you know that we take our
sleep pretty seriously around here.
odic reminders), while the other group
was blinded to the intervention and re-
ceived no information about sleep. Both
In the past, we’ve covered research groups gained lean mass and lost fat
about naps, glasses that block blue light, mass to some degree, but the exercise
a variety of sleep interventions, and + sleep group lost significantly more
even how macronutrient distribution fat mass and experienced a significantly
affects sleep. Inherently, our interest in larger reduction in body-fat percentage.
these topics is largely predicated on two While these results are both positive
main hypotheses: 1) that sleep actually and intuitive, the study has some pret-
impacts our ability to train, recover, and ty notable limitations that must be con-
adapt to training; and 2) that we can take sidered. Read on to find out what those
some easy, practical steps to improve limitations are, and whether or not your
our sleep habits enough to actually aid sleep habits might be getting in the way
our fitness-related goals. The present- of your physique goals.
ly reviewed study (1) tackles the latter
hypothesis directly, with an extremely
practical and straightforward interven- Purpose and Hypotheses
tion. Two different groups completed a
10-week resistance training program; Purpose
one group received a pretty basic sleep The authors stated that the purpose
health education intervention before the of the current study was to determine
onset of the program (along with peri- if “sleep optimization” beneficially im-
40
pacts changes in fat mass or lean mass thors administered a pre-study ques-
when combined with resistance training. tionnaire to assess pre-existing sleep
Notably, an assumption that is necessar- issues, and subjects were excluded if
ily built into this purpose statement is they reported having sleep problems for
that the sleep health education interven- more than two continuous weeks. Base-
tion they used actually optimized sleep. line subject characteristics are listed in
Table 1. While the study began with
Hypotheses 30 subjects, 4 dropped out due to time
The authors hypothesized that “opti- constraints, and 3 others were excluded
mizing sleep quality and quantity would from the analysis because illness or in-
create additional positive body com- jury led to insufficient adherence (<70%
position effects compared to resistance of exercise sessions completed). The fi-
exercise only.” Again, it seems safe to nal analysis included 22 subjects, with
assume that they believed their sleep 10 in the sleep group (training + sleep
health education intervention would be intervention) and 12 in the control group
effective enough to tangibly improve (training only).
sleep quality and quantity; otherwise,
the study would’ve categorically failed, Methods
which happens sometimes. Subjects were randomly assigned to
one of two groups. The control group
underwent a 10-week resistance training
Subjects and Methods program, while the sleep group complet-
ed the same training program, but also
Subjects received a sleep health education inter-
The researchers recruited 30 male vention at the beginning of the study.
participants for the current study. Par- The sleep education was intended to help
ticipants were untrained (no regular re- participants in the sleep group improve
sistance exercise in the last six months) the quantity and quality of their sleep.
students or university employees with The intervention was based on guide-
no significant sleep problems. The au- lines from the Norwegian Competence
Center for Sleep Disorders and consist-
41
ed of the following pieces of advice: “1) participants picked a load, tried to hit
Keep it dark, quiet and cool in the room about 10 reps, went to failure each set,
you sleep. 2) Go to bed and get up at and started selecting heavier loads when
the same time during the whole week, they began reaching failure at more than
including weekends. 3) Go to bed so 10 reps per set. But that’s essentially a
that you can get 8 hours of sleep. 4) Two guess; there’s simply not enough infor-
hours before bed: a) turn down lights. b) mation to determine exactly what this
Put away pc, pads and mobile phones or training program looked like, or how the
use blue light filter and lowest possible training stimulus was (or was not) pro-
light setting. c) No exercise. d) No food, gressed over time. The paper also states
coffee or black tea, or drinks with en- that training was supervised “during the
ergy. e) Do calm and positive activities first half of the study,” so it sounds like
and do not bring up relational conflicts participants were supervised for the first
before bedtime. 5) When you get up in five weeks, then asked to train inde-
the morning, expose yourself to as much pendently (and keep a training log) for
light as possible right away. 6) Last cup the final five weeks. To assess changes
of coffee for the day six hours before in fat mass and lean mass, body compo-
bedtime. 7) Try to get so much sleep that sition was measured before and after the
you do not need an alarm clock to wake program using DXA. Participants also
up. 8) Do not use electronic devices that completed the Bergen Insomnia Scale
emit light in bed. 9) Learn a relaxation questionnaire and had blood samples
technique, use it to fall asleep at night drawn both before and after the pro-
or if you wake during night.” The sleep gram. The questionnaire was intended
group was regularly reminded about to assess insomnia-related aspects of
these suggestions throughout the study, sleep, and the blood samples were ana-
while the control group was blinded to lyzed for HDL, LDL, triglyceride, fast-
these suggestions. ing glucose, and glycated hemoglobin
The resistance training program con- (HbA1c) levels.
sisted of two full-body workouts per
week. Workouts began with a 10-minute
aerobic warmup, followed by leg press, Findings
chest press, leg extension, narrow grip The exercise + sleep and exercise
cable chest press, lat pulldown, seat- groups did not have significantly different
ed leg curl, and narrow grip cable row. scores on the sleep questionnaire at base-
The paper states that “exercises were line (9.4 versus 13.4 points) or the end of
performed with 2-3 series of 10-repe- the study (7.4 versus 9.1 points). Just to
tition maximum with recommended 90 be clear, lower scores on this question-
seconds inter-set breaks.” I’m assuming naire reflect fewer sleep issues, so a re-

42
Figure 1 Mean group changes in anthropometric variables

Exercise

Exercise + sleep
2.0

1.5

1.0

0.5
Group mean (kg)

0.0

-0.5

-1.0

-1.5

-2.0
Body weight Fat mass Lean mass
change change change

* = Exercise + sleep significantly different from exercise

duced score is a good thing. Similarly, the icant (p = 0.29). Both groups lost fat
change from baseline to post-testing was mass, but the exercise + sleep group
not significantly different in the exercise lost significantly more than the exercise
+ sleep and exercise groups (-2.0 versus group (-1.8 ± 0.8 versus -0.8 ± 1.0kg; p
-4.3 points, respectively). Unfortunately = 0.02). Both groups experienced a re-
I can only report the mean changes here; duction in body-fat percentage, but the
the researchers did not provide standard reduction in the exercise + sleep group
deviations, which isn’t ideal, but we can was significantly larger than the reduc-
only use what we’ve got. tion in the exercise group (-2.07% ver-
Body composition results are presented sus -1.11%, no standard deviations pro-
in Figure 1. Both groups increased lean vided; p = 0.03).
mass throughout the study; the increase Results for the blood biomarkers are
was larger in the exercise + sleep group listed in Table 3. Overall, nothing par-
than the exercise group (1.7 ± 1.1 ver- ticularly interesting happened. There
sus 1.3 ± 0.8kg), but the between-group was a curious difference between groups
difference was not statistically signif- for blood triglyceride levels, but the re-
43
searchers noted that this relationship expensive, and have minimal risk of side
seemed to be largely driven by one par- effects. As you can imagine, I have a pret-
ticular subject with a very atypical in- ty favorable opinion of sleep. It’s hard to
crease. It’d be great to further explore get more practical than something that
exactly how large of an impact that out- everybody does every day. Sleep is free,
lier might have had, but again, standard and extra sleep tends to bring a bunch
deviation values were not provided, so of extra unintended benefits rather than
we really don’t have much to fuel our ex- unfavorable side effects. However, these
ploratory efforts. However, the reported positive aspects alone do not necessarily
outlier’s blood triglycerides increased mean that sleep interventions will neces-
by 1.94 mmol/L; to contextualize that sarily improve our performance or phy-
number, a total triglyceride level of 1.8- sique, and it doesn’t mean we should be
2.2 mmol/L is considered “borderline uncritical of studies that appear to con-
high.” So a change of 1.94 mmol/L is firm our previously held biases.
pretty astronomical and fairly similar to I love the concept of this study and the
seeing your blood LDL level jump up by general approach that the researchers
around 140 mg/dL in response to start- took. It’s hard to get more practical than
ing a basic resistance training program this study’s overarching goal: Teach
and trying to sleep a little better. people about sleep, remind them that it’s
important, have them lift, and track the
gains. Having said that, this study had
Interpretation some pretty notable limitations that can’t
I’m going to be fully transparent about be ignored. The authors drew attention
my biases here: In general, I am very to one of the limitations, which was a
fond of strategies that are practical, in- relatively high dropout rate. Of the 30
44
subjects that enrolled in the study, only if (or how) loads were progressed over
22 completed the study with satisfacto- time, and I’d have no chance of repli-
ry adherence. This presents two poten- cating the training program with confi-
tial issues: 1) with a small sample size, dence. In addition, no efforts were made
we could potentially “miss” important to control, or even monitor, dietary hab-
patterns that would otherwise be sta- its. The paper does not state that partic-
tistically significant in a larger sample, ipants were given guidance to maintain
and 2) if systematic differences exist be- or improve their normal diet, nor were
tween people who dropped out and peo- they asked to log their food intake at any
ple who completed the study, this could time. Even basic food diary information
introduce some bias into the results. For could have provided some very infor-
point #1, there’s not much we can do. mative context to support the observed
Sure, it’s a relatively small sample, but changes in body composition. Finally,
not egregiously low compared to other I view the method for sleep assessment
interventions in our field of study. Plus, as a fairly big limitation. The Bergen In-
there was a large enough sample to de- somnia Scale is a pretty broad tool, with
tect some significant body composition only six total questions. The question-
changes, so that doesn’t seem like a huge naire itself does not specifically track
issue. As for point #2, the authors pre- information like the time of sleep onset,
sented the reasons for dropping out, the the time of waking, or the total duration
exact number of dropouts in each group, of sleep obtained. As a result, the ques-
and also performed something called an tionnaire results did not provide strong
“intent to treat” analysis. Based on the evidence of meaningfully better sleep in
information provided, it doesn’t seem the exercise + sleep group than the ex-
like there’s cause for serious concern ercise only group. Ideally, it would’ve
about bias related to dropouts. been nice to utilize both objective and
Personally, I’m more interested in a subjective indices of sleep duration and
few of the limitations that were not men- quality, and it would’ve been nice to use
tioned in the original paper. First, the a more nuanced questionnaire for the
paper provides extremely minimal in- subjective assessment of sleep.
formation about the details of the train- If this study was the only relevant re-
ing program. Most of the time, there’s a search we had on the topic, it would be
pretty easy trick for determining if the very hard to confidently interpret the
methods were described clearly enough: findings in light of these pretty notable
If you were supposed to start from limitations. Fortunately, the relationship
scratch and replicate the study, could between sleep, diet, and performance is
you do it based on the information pro- a hot research topic right now, and new
vided? In this scenario, I’m still not sure studies are popping up all over the place.

45
Collectively, these studies are painting evaluating sleep restriction typically al-
a clearer picture of how sleep influenc- low for a “sleep opportunity” of rough-
es our performance, dietary behaviors, ly 3-6 hours per night. In such studies,
and body composition. For example, a roughly 7-8 hours is typically considered
more tightly controlled study back in the “sufficient” level to which results
2010 found results pretty comparable are compared. Based on that literature, it
to the presently reviewed study. Over would seem pretty safe to conclude that
a two-week weight loss phase, reduc- you’d want to obtain, at minimum, a sol-
ing the nighttime sleeping opportunity id 7-8 hours of sleep per night. However,
from 8.5 to 5.5 hours per night resulted a fascinating body of literature pertains
in less loss of fat mass and greater loss to actually extending sleep beyond the
of lean mass (2). As the body of liter- minimal “adequate” level, rather than
ature grows, it becomes more apparent simply avoiding sleep restriction. Greg
that insufficient sleep can lead to alter- laid out a nice review of this literature
ations in glycemic control, hunger, ap- back in Volume 2 of MASS, so I’d en-
petite, and ultimately increased energy courage you to check out that article if
intake (3). In addition, sleep restriction you want a deeper dive into the sleep
was recently shown to reduce the rate of extension literature. But, in short, there
myofibrillar protein synthesis (4), which have been multiple studies in athletes
would lead to even more unfavorable showing that increasing sleep time from
body composition outcomes. Of course, ~7 hours to up around ~9 hours can im-
the effects of insufficient sleep do not prove a variety of physical and cognitive
solely impact body composition; a va- performance outcomes (6). In line with
riety of performance-related outcomes these findings, the National Sleep Foun-
are impaired as well. As summarized in dation recommends getting 8-9 hours of
a recent review (5), poor sleep practices sleep per night to support sports perfor-
have the potential to unfavorably affect mance (5), which is just a bit higher than
function of the cardiovascular, respira- the 8 hours that is often recommended
tory, neuroendocrine, neuromuscular, for the general population.
gastro-intestinal, and immune systems. Acknowledging the importance of
Naturally, prolonged sleep restriction sleep is one thing, but actually obtaining
can result in impairment of a wide range more sleep is another thing altogether.
of performance outcomes, in addition to To date, sleep extension appears to be
increasing susceptibility to injury and the most efficacious sleep intervention
illness (5). to actually improve performance (7).
It’s clear that insufficient sleep is bad Generally speaking, obligations of the
news for performance and body compo- modern world dictate that sleep exten-
sition, but how much is enough? Studies sion strategies typically involve going to

46
bed earlier, as relatively few people en- length light close to sleep time, or
joy the ability to simply sleep in longer re-evaluating the macronutrient distri-
than they currently do. Based on what we bution of your diet. However, note that
know about circadian biology, it seems napping strategies can be tricky. While
that an optimal approach to sleep exten- naps are better than nothing in an acute-
sion would ideally involve maintaining ly underslept situation, you’ll recall that
a regular daily sleep time and wake time, regular sleep/wake times and circadian
which would be approximately aligned alignment with light/dark cycles seem
with the natural daily light/dark cycle. If to be pretty important factors for sleep
you’re attempting to implement a pret- quality. So, feel free to utilize naps, but
ty sizable sleep extension, it also seems be mindful of their timing and duration
advisable to ease into it; I would imag- to ensure that they don’t interfere with
ine that abruptly moving your sleep time other aspects of your sleep hygiene. Fi-
two hours earlier would result in a pretty nally, if you’re trying to clean up your
frustrating adjustment and a lot of time sleep habits, make sure your caffeine
spent staring at your ceiling. Evidence intake is in check. The half-life of caf-
suggests that maintaining some simple feine tends to be somewhere around 5-6
sleep hygiene habits, such as the list of hours, and caffeine consumption can
guidelines provided in the presently re- lead to notable sleep disruption several
viewed study, is also a reasonably effi- (6+) hours after ingestion (8), particu-
cacious strategy for sleep improvement larly if total caffeine intake for the day
(7). So, even if you’re unable to work a is relatively high.
couple of extra hours of sleep into your
daily schedule, you can still improve
the quality of your sleep by maintaining Next Steps
some pretty basic habits, which include I love the concept of this study, so I’d
getting plenty of light exposure in the like to see it replicated with some chang-
morning, avoiding light (particularly in es. Specifically, I’d like to see it done
the blue wavelengths) in close proximity with a clearly described, progressive re-
to sleep time, maintaining regular sleep sistance training program, a more thor-
and wake times, and making sure your ough subjective sleep assessment, some
sleeping environment and pre-sleep rou- objective method of sleep assessment,
tine are conducive to restful sleep (6). and longitudinal monitoring of nutrition
If those strategies aren’t cutting it, you habits. Of course basic food logs would
could also explore some of the promis- be informative, but an additional survey
ing strategies that have been covered in or questionnaire to assess things like
previous issues of MASS: taking naps, hunger, stress, and food cravings would
wearing glasses that block blue-wave- be a fantastic addition.

47
APPLICATION AND TAKEAWAYS
The sleep literature continues to suggest that insufficient sleep is disadvantageous for
people who are trying to improve their physique or performance. It seems like 8-10
hours is probably the ideal range for people with such goals, and it seems like some
pretty simple sleep hygiene habits can increase your likelihood of obtaining enough
high-quality sleep to support your goals. The presently reviewed study has some no-
table limitations that prevent us from drawing rock-solid conclusions from this paper
alone. However, the results fit in well with the broader body of literature, and the inter-
vention they used seems like a great starting point for people who are looking for some
simple, straightforward guidelines to improve their sleep habits.

48
References

1. Jåbekk P, Jensen RM, Sandell MB, Haugen E, Katralen LM, Bjorvatn B. A randomized con-
trolled pilot trial of sleep health education on body composition changes following 10 weeks
resistance exercise. J Sports Med Phys Fitness. 2020 Mar 4; ePub ahead of print.
2. Nedeltcheva AV, Kilkus JM, Imperial J, Schoeller DA, Penev PD. Insufficient sleep under-
mines dietary efforts to reduce adiposity. Ann Intern Med. 2010 Oct 5;153(7):435–41.
3. Beccuti G, Pannain S. Sleep and obesity. Curr Opin Clin Nutr Metab Care. 2011 Jul;14(4):402–
12.
4. Saner NJ, Lee MJ-C, Pitchford NW, Kuang J, Roach GD, Garnham A, et al. The effect of
sleep restriction, with or without high-intensity interval exercise, on myofibrillar protein
synthesis in healthy young men. J Physiol. 2020 Mar 11; ePub ahead of print.
5. Chandrasekaran B, Fernandes S, Davis F. Science of sleep and sports performance – a scop-
ing review. Sci Sports. 2020 Feb 1;35(1):3–11.
6. Vitale KC, Owens R, Hopkins SR, Malhotra A. Sleep Hygiene for Optimizing Recovery in
Athletes: Review and Recommendations. Int J Sports Med. 2019 Aug;40(08):535–43.
7. Bonnar D, Bartel K, Kakoschke N, Lang C. Sleep Interventions Designed to Improve Ath-
letic Performance and Recovery: A Systematic Review of Current Approaches. Sports Med.
2018 Mar;48(3):683–703.
8. Drake C, Roehrs T, Shambroom J, Roth T. Caffeine effects on sleep taken 0, 3, or 6 hours
before going to bed. J Clin Sleep Med. 2013 Nov 15;9(11):1195–200.

49
The Relationship Between Body Composition Measured by Dual-Energy X-Ray
Absorptiometry and Maximal Strength in Classic Powerlifting. Ferland et al. (2020)

You Can’t Flex Bone: Predicting


Powerlifting Performance Based
on Fat-Free Mass
BY G RE G NUC KO LS

I think everyone knows that more muscular lifters tend to be stronger


lifters, but I don’t think people recognize just how strong that
relationship is among competitive powerlifters. In a recent study
on raw, drug-free powerlifters, fat-free mass was almost perfectly
correlated with performance. This article will explore that relationship,
and what we can do with it.

50
KEY POINTS
1. In a sample of 15 competitive powerlifters, fat-free mass and bone mineral content
were both strongly associated with weight lifted in the squat, bench press, deadlift,
and total.
2. These relationships held across both sexes, suggesting that the difference in
performance between male and female lifters is primarily (or almost entirely) driven
by differences in fat-free mass.
3. If you know how much fat-free mass you have, you can compare your current lifts
against what you theoretically “should” be able to lift. You can use that information
to guide your medium- to long-term training goals.

I
f you’re even marginally acquaint- This is an article to read for the discus-
ed with the sport of powerlifting, sion. I’m not sure if it completely goes
I’m sure you’re aware that larg- off the rails, or if it’s some of my better
er lifters tend to lift more than small- work. You can be the judge.
er lifters. From that observation, you
can probably intuit that there’s a posi-
tive association between fat-free mass Purpose and Hypotheses
and powerlifting performance. I think
people tend to be surprised at just how Purpose
strong that relationship is, though. The purpose of this study was to ex-
A recent study set out to examine which amine the relationships between pow-
factors were associated with powerlift- erlifting performance and various body
ing performance in a sample of mixed- composition metrics.
sex, raw, drug-free, competitive pow-
erlifters. Unsurprisingly, fat-free mass Hypotheses
and fat-free mass per unit of height were The stated hypothesis was…broad:
strongly associated with performance, “The hypothesis of this study was that
explaining ~86-90% of the variability various powerlifters’ body composition
in the lifters’ totals. Interestingly, bone characteristics would be correlated with
mineral content was almost as strongly maximal strength in the squat, the bench
associated with performance as fat-free press, the deadlift, and the total.”
mass was. I doubt anyone reading this is
surprised that fat-free mass is associated
with powerlifting performance, though.

51
(22 or 23 years old), and the rest com-
Subjects and Methods peted in the open class. They were quite
strong by the standards of published re-
Subjects search but average or just slightly above
The subjects were 15 competitive raw average lifters by competitive standards
powerlifters. Three subjects were juniors (the average female total was 352.3kg at
52
80.7kg [776lbs at 178lbs], and the aver- ables were the subjects’ age, resistance
age male total was 630.6kg at 93.5kg, training experience, and powerlifting
[1390lbs at 206lbs]). You can see more experience (all obtained via interview).
details about the subjects in Table 1.
Findings
Experimental Design The strongest predictors of squat,
The design of this study was incred- bench press, deadlift, and total perfor-
ibly straightforward. Subjects all com- mance tended to be measures of fat-free
peted in the Quebec, Canada provin- mass. Bone mineral content and bone
cial championship on November 10-11, mineral density (for all regions except
2018. The researchers recruited from the head) also tended to be quite pre-
the competitor pool at that meet and got dictive of powerlifting performance;
them into the lab for a DEXA scan prior bone mineral content tended to be more
to January 31, 2019. The lifters’ highest strongly predictive than bone mineral
successful attempts at the powerlifting density. All body composition measures
meet were used as their 1RMs. All of the tended to be more predictive of absolute
body composition measures used in the performance than relative performance
study were derived from the subjects’ (e.g. Wilks score). Below you’ll find
DEXA scans. The only additional vari- tables with the correlation coefficients

53
54
showing the strength of the relationships showing the relationship between fat-
between various body composition mea- free body mass and powerlifting total in
sures and powerlifting performance. this sample (11). Regression equations
The study also supplies a figure show- are provided for each relationship.
ing the relationships between fat-free
mass and performance in the squat,
Interpretation
bench press, and deadlift. For the figures, I know we’ve at least mentioned in
I’ve split those three lifts apart for easier passing in MASS that there’s a very
viewing, and created a fourth scatterplot strong relationship between muscle mass

Figure 1 Relationship between fat free mass and squat 1RM

350
y = 4.365x - 100.1
R2 = 0.798
300

250
One rep max (kg)

200

150

100

50

0
40 45 50 55 60 65 70 75 80 85 90

Fat free mass (kg)

Figure 2 Relationship between fat free mass and bench press 1RM

250
y = 3.3894x - 107.2
R2 = 0.8933

200
One rep max (kg)

150

100

50

0
40 45 50 55 60 65 70 75 80 85 90

Fat free mass (kg)

55
Figure 3 Relationship between fat free mass and deadlift 1RM

350
y = 4.5193x - 83.21
R2 = 0.8317
300

250
One rep max (kg)

200

150

100

50

0
40 45 50 55 60 65 70 75 80 85 90

Fat free mass (kg)

Figure 4 Relationship between fat free mass and powerlifting total

900
y = 12.274x - 290.51
R2 = 0.8773
800

700
One rep max (kg)

600

500

400

300

200

100

0
40 45 50 55 60 65 70 75 80 85 90

Fat free mass (kg)

and strength. If you’ve been following in a sample of competitive powerlifters.


my personal work, I’ve explained the So, just to get one thing out of the way
reasons for that relationship in depth; for the pedants (no judgment; I’m one
I even co-authored some published lit- too): Technically, the study was assess-
erature on the subject (2). However, I ing fat-free mass, not muscle mass. Fat-
believe this is the first study we’ve re- free mass includes muscle, but it also
viewed on that topic in MASS, at least includes bones, organs, and virtually

56
everything else in the human body be- interesting aspects of this relationship.
sides fat. However, skeletal muscle is First, the male and female lifters in the
the single largest tissue constituent of present study both clearly fall along the
fat-free mass, and it’s the tissue compo- same trendline (I believe the females are
nent of fat-free mass that varies the most all of the subjects with less than 55kg of
between individuals. So, fat-free mass fat-free mass, and the males are all of
serves as a good proxy for muscle mass the subjects with more than 60kg of fat-
(especially in a population all compet- free mass on the graphs above, though
ing in the same sport) and, if anything, that’s not made explicitly clear in the
I’d expect the relationships between text). The slope of the trendline for just
strength and actual muscle mass to be the male lifters is incredibly similar to
even stronger than the relationships be- the slope of the trendline for just the
tween strength and fat-free mass seen in female lifters. This suggests that differ-
this study. ences in fat-free mass explain virtually
I doubt anyone’s surprised that fat-free all of the average strength differences
mass is so strongly predictive of power- between male and female lifters. This is
lifting performance. After all, the force what we should expect since we know
you produce to lift weights comes from that male and female muscles can con-
muscle contraction, so more fat-free tract with the same amount of force per
mass should mean more muscle, and unit of mass or cross-sectional area (3).
more muscle should mean greater con- However, I’ve seen people suggest any
tractile force. However, there are some number of reasons for the performance
gap between male and female powerlift-
ers (from males having thicker torsos, to
difference in limb proportions, to differ-
DIFFERENCES IN FAT-FREE ences in pelvis shape, to fiber type dif-
ferences, etc; I’ve suggested this in the
MASS EXPLAIN VIRTUALLY past as well). And who knows: There
very well could be some other factors
ALL OF THE AVERAGE that explain a small proportion of the
strength differences between the sex-
STRENGTH DIFFERENCES es, but it doesn’t seem like there’s very
much left to explain after accounting for
BETWEEN MALE AND differences in fat-free mass.
Another interesting thing to note about
FEMALE LIFTERS. the present study is that bone mineral
content was almost as strongly associ-
ated with performance as fat-free mass
57
was. It’s not surprising that performance this information?”
and bone mineral content were associat- You can use this information to guide
ed – larger people tend to lift more, and training goal selection.
also tend to have greater bone mass –
If we know that fat-free mass and
what was surprising (at least to me) was
strength are strongly associated, then we
that bone mineral content could predict
can predict how much you “should” be
performance almost as well as fat-free
able to lift with the current amount of fat-
mass could. Since the correlation coef-
free mass you have, compared to drug-
ficients for the relationship between fat-
free competitive powerlifters. If you al-
free mass and strength were in the ~0.9
ready lift more than you “should” be able
range, I wouldn’t have batted an eye at
to lift with your current amount of total
correlation coefficients in the ~0.7-0.8
fat-free mass, that suggests that you’re
range for the relationship between bone
already a highly efficient lifter, and your
mineral content and strength, but the
progress is probably bottlenecked by
fact that the bone:performance correla-
muscle growth moreso than your tech-
tion coefficients were also close to 0.9
is somewhat surprising. We know that
muscle and bone are intimately linked
– contractile force from muscle helps
build bone, and osteopenia and sarco- IF WE KNOW THAT FAT-FREE
penia tend to go hand in hand – but an
idea popularized by a researcher named MASS AND STRENGTH ARE
Francis Holway takes things one step
further. Holway proposes that the size of
STRONGLY ASSOCIATED, THEN
your skeleton ultimately constrains how WE CAN PREDICT HOW MUCH
much muscle you can build (4). Thus, in
a cohort of people near their muscular
limits, their muscle mass and bone mass
YOU “SHOULD” BE ABLE TO
should be almost perfectly correlated. If LIFT WITH THE CURRENT
that’s true, then it would be unsurprising
for both bone mineral content and total AMOUNT OF FAT-FREE MASS
fat-free mass to be similarly well-cor-
related with powerlifting performance. YOU HAVE, COMPARED TO
At this point, you may still be think-
ing, “of course fat-free mass and abso-
DRUG-FREE COMPETITIVE
lute strength are strongly correlated. I
already knew that, and this article seems
POWERLIFTERS.
to be a waste of time. What can I do with

58
nique and neuromuscular efficiency. If average to slightly above-average com-
you lift less than you “should” be able to petitive powerlifters, competing raw in
lift with your current amount of total fat- a strictly tested organization. I strongly
free mass, that suggests that you’re a rel- suspect the present sample more closely
atively inefficient lifter, and that you may resembles most of the powerlifters (or
still be able to get considerably stronger people with strength-based goals) in the
just by continuing to hone and develop MASS readership.
technique and movement efficiency. It’s also interesting to compare the pres-
Up until this point, the only study that ent study to the Brechue and Abe study
provided enough data for that sort of to see how the relationship between fat-
calculation was this study by Brechue free mass and strength differs in the two
and Abe (5). It’s a fine study, also us- samples (6). Unsurprisingly, the subjects
ing a cohort of competitive powerlifters, in the Brechue and Abe study tended to
but there are a few major differences have more fat-free mass (~50-117kg,
between Brechue and Abe’s sample and versus ~42-83kg in the present study).
the sample in the current paper. The lift- Where the two samples overlap, the
ers in the prior study were all national- subjects in the Brechue and Abe study
or world-class lifters (so one can assume also tended to lift more. However, the
they were all very genetically blessed), predicted increase in strength per unit of
they competed in equipped powerlifting LBM increase is actually larger in the
(before squat suits and bench shirts ac- present study (the beta coefficient on
tually got good, mind you, but the early the LBM term), whereas the y-intercept
equipment was still doing something), is larger in the Brechue and Abe study.
and they competed untested (in feder- Since the two regression lines have dif-
ations that allowed steroid usage). The ferent slopes, they intersect at ~102kg.
lifters in the present study were basically Functionally, that means that a lifter

59
with 102kg of fat-free mass would have ers with less fat-free mass (<80-90kg or
about the same predicted total in both of so), and a less steep relationship for lift-
these studies, a lifter with 60kg of fat- ers with more fat-free mass. If we plotted
free mass would have a larger predict- the residuals for the linear regression line
ed total using the regression equation in (differences between actual performance
the Brechue and Abe study, and a lifter and regression-predicted performance),
with 110kg of fat-free mass would have you’d see that for the lightest and heavi-
a larger predicted total using the regres- est lifters, the regression equation consis-
sion equation in the present study. tently overestimates performance, while
I’m not totally sure what to make of for the lifters with ~65-90kg of fat-free
these comparisons since the samples mass, the regression equation consistent-
are so different. One weird possibility ly underestimates performance. In other
is that the relationship between fat-free words, the regression equation would
mass (and presumably muscle mass) and probably fit the data better with a break-
strength changes after a person accumu- point somewhere around 85kg of fat-free
lates a very large amount of fat-free mass. mass. Below 85kg, you’d expect a larger
Going back to the Brechue and Abe study increase in performance per kilo of fat-
(5), if we look at the relationships be- free mass gained, and above 85kg, you’d
tween fat-free mass and strength in each expect a smaller increase in performance
lift, we can see that for the squat, the lin- per kilo of fat-free mass gained.
ear relationship between fat-free mass Why would we expect every kilo of
and strength works quite well. However, fat-free mass above a certain threshold
for the bench press and deadlift, it appears to be “worth” less than every kilo of fat-
that there’s a steeper relationship for lift- free mass below a certain threshold? One

60
possibility is that muscle fibers become lifters who are similarly skilled with sim-
too large to maintain efficient function. ilar body composition should have sim-
Larger muscle fibers tend to produce ilar allometrically scaled strength. And
less force per unit of cross-sectional area that’s what we see, up to a point. Up to
than smaller muscle fibers in cross-sec- the 105kg weight class, allometric scal-
tional research (7; in other words, at this ing seems to do a really good job com-
very moment, your larger muscle fibers paring performance between powerlifters
produce less force per unit of CSA than of different body sizes. In higher weight
your smaller fibers). With longitudinal classes, allometrically scaled strength be-
research, we generally see that force gins dropping off. At first, I thought the
per unit of fiber CSA stays the same or only explanatory factor was body com-
increases as fibers grow (8). However, position – heavier lifters carry more fat,
when fibers get too large, that trend may so they have less muscle as a percentage
reverse. Bioenergetic factors may begin of total body mass, so their allometrically
working against the fibers (it’s more dif- scaled strength is lower than it would be
ficult to clear waste products and deliver for lighter lifters with better body compo-
fuel and oxygen when diffusion distanc- sition. However, one “problem” always
es to bordering capillaries increase), or stuck in my mind – Ray Williams. Ray’s
the proportion of each fiber composed an incredible lifter; he’s totaled 1112.5kg
of contractile tissue may begin decreas- at a body mass of 190.4kg. Regardless of
ing. Up to a certain point, as fibers grow, body mass, 1112.5 is an amazing total.
they may change quantitatively (getting Ashton Rouska totaled 881.5kg at a body
larger and stronger), but not change in ef- mass of 92.1kg earlier this year. That’s
ficiency (they still function exactly like also an amazing total. Ashton’s allome-
they did when they were smaller). Past trically scaled total is 43.2. Ray’s is 33.6.
that point, they may change both quanti- At first, I don’t see anything wrong with
tatively (getting larger and stronger), but that. Ashton’s clearly leaner than Ray, so
also decrease in efficiency (with function of course his allometrically scaled total
per unit of size decreasing). should be higher.
It’s worth noting that we have some However, we can work in reverse
degree of anecdotal support for this no- and see how much Ray would have to
tion when looking at powerlifting per- weigh to have the same allometrically
formance. Allometric scaling is predicat- scaled total as Ashton, assuming he still
ed on the idea that strength scales with totaled 1112.5kg. He’d need to weigh
muscle cross-sectional area, while body about 135kg. Maybe I’m a cynic, but I
mass scales with body volume (it’s ex- don’t know that we’ll ever see a lifter
plained in more detail here). Thus, two total 1112.5kg at 135kg under similar

61
conditions (walked-out squats, no dead- elite-level lifter, rather than an outlier
lift bar, two-hour weigh-in, strict drug among outliers. I didn’t even go with his
testing). Among lifters competing under best total, just his most recent. The top al-
similar conditions, Dennis Cornelius is lometrically scaled total in all IPF weight
the only person who’s somewhat come classes below 120kg range from 43.3 to
close (1000 at 132.5). I shouldn’t need 45.7. To match 43.3, Ray would need to
to tell you that 1000 and 1112.5 are very total 1112.5 at 130.2; to match 45.7, he’d
different numbers, though. If we expand need to total 1112.5 at 120.1.
our search to include federations without I recognize that this was a pretty long
drug testing and with different equipment diversion, but my main point is that we
standards, Larry Wheels and Eric Lillieb- also have practical reasons to believe
ridge have both hit 1075 at body weights that strength gains per unit of fat-free
below 140, but they’re still nearly 40kg mass are lower past a certain threshold
away, and the equipment they compete as well; we don’t just have to rely on
on (especially the deadlift bar) is prob- eyeballing the residuals in the regres-
ably massaging their numbers at least a sion analysis from one paper that was
little bit. All of this is to say, even dis- published a long time ago.
regarding differences in body composi-
Last but not least, there’s one final the-
tion between lifters, it seems likely that
oretical issue in play, both in the present
allometric scaling (and thus the mono-
study and in the Brechue and Abe paper:
tonic relationship between muscle mass
Both studies have regression equations
and strength) breaks down when dealing
with pretty large y-intercepts. The y-inter-
with lifters who have enormous amounts
cept in these studies tells you how much
of fat-free mass. Also note, for this ex-
you should expect to lift if you have zero
ample, I chose Ashton as an “average”

62
fat-free mass. Intercepts aren’t always mass and Wilks score, assuming the male
meaningful, but in this case, they are, and female cohorts are comparable. All
since they show how well results com- else being equal, female lifters should
port with theory. In theory, the y-inter- have less fat-free mass than male lifters,
cept should be negative; I’ll explain why but have similar Wilks scores, since Wilks
in a footnote for anyone who cares about was developed to be able to compare per-
the technical reasons (9). In the present formance between the sexes. However,
study, it’s -290.5, which is slightly below the male and female lifters in this study
where it “should” be (see footnote 9). In aren’t directly comparable. The male
the Brechue and Abe study, it was 221, lifters had more training experience and
which is a considerable deviation from higher Wilks scores overall. Within each
theory. In the Brechue and Abe study, the sex, and within the entire cohort, there
breakpoint in the relationship between was also a positive relationship between
fat-free mass and powerlifting total is training experience and Wilks scores (r =
probably the culprit. Smaller increases 0.52-0.78). That’s probably the primary
in strength per unit of LBM for the lift- reason for a y-intercept in the regression
ers with higher LBM decreases the slope equation that’s slightly more negative
of the linear regression line, resulting in than it “should” be. Across all levels of
a large positive intercept. In the present magnification (weaker males vs. stron-
study, the problem may have been differ- ger males, weaker female vs. stronger
ing experience levels. We can see that not females, males vs. females), the weaker
only was there a positive relationship be- group in each comparison had less fat-
tween fat-free mass and absolute weight free mass and likely had less skill and
lifted, there was also a reasonably strong experience. That effect causes the slope
positive relationship between fat-free of the fat-free mass/strength trendline to
mass and Wilks points. At first, that may get steeper, resulting in a more negative
sound intuitive; as people train longer, y-intercept. The lifters with more fat-free
they get better at the lifts and they also mass were lifting about as much as they
build muscle, so more experienced peo- “should” be lifting, while the lifters with
ple will tend to have more fat-free mass less fat-free mass were lifting less than
and also have higher Wilks scores. And they “should” be lifting since they were
within each sex, that’s probably valid less experienced and less skilled.
reasoning; this study also found a posi- So, with all of that out of the way, let’s try
tive relationship between fat-free mass to piece together a set of equations for pre-
and Wilks score within each sex. Howev- dicting performance based on the amount
er, within the entire sample of this study of fat-free mass someone is carrying. As
containing both sexes, there shouldn’t be I’ve discussed in the previous paragraphs,
a positive relationship between fat-free

63
I think the present study underestimates of the Ferland et al equation, the Brechue
strength potential in lifters with lower fat- and Abe equation, and my tweaked ver-
free mass since the lifters with less fat-free sion of the Ferland et al equation.
mass in the present study had less training I think this equation works pretty well
experience, and I think the Brechue and for lifters with up to 86kg of fat-free
Abe study tends to overestimate strength mass. For lifters with more than 86kg
potential in lifters with less fat-free mass of fat-free mass, it’s harder to provide
due to the lifters with high fat-free mass a confident quantitative relationship be-
unduly flattening the regression line. Be- tween fat-free mass and strength. There
tween the two, however, I think the regres- were no lifters in the present study with
sion equation in the present study is much more than 86kg of fat-free mass, and
closer to “correct,” especially for the lift- only five or six lifters in the Brechue and
ers in the sample with more fat-free mass. Abe study had more than 86kg of fat-
So, we can take two fixed points to slightly free mass. However, just eyeballing the
tweak the formula: our theoretical x-inter- graphs in the Brechue and Abe study, it
cept (19.35; discussed in footnote 9), and appears that the slope would decrease by
the top within-sample point for the regres- approximately half, so until more data is
sion equation in the present study (a 765kg published on lifters with extremely high
total with 86kg of fat-free mass). The re- fat-free mass, a slope of 5.74 for each
sulting equation has a slope of 11.48 and a kilo of fat-free mass above 86kg will
y-intercept of -222. It’s a relatively small have to serve as a placeholder.
tweak. In the figures, there’s a comparison

64
So, if someone has less than 86kg of place you in the 98th percentile of lifters
fat-free mass, their predicted total is: in that weight class. That’s … very opti-
Predicted total = fat-free mass × 11.48 mistic. Using the equations straight out
- 222 of the present study, your predicted total
would be 544kg (which should illustrate
And if they have more than 86kg of
my theoretical tinkering only resulted in
fat-free mass, their predicted total is:
a very slight adjustment).
Predicted total = fat-free mass × 5.74
One thing I’ll point out is that, unfor-
+ 271
tunately, this formula doesn’t take height
So, just to give one example, if you into account. Totaling 559kg with 68kg of
weigh 85kg, and you’re approximately fat-free mass may be very doable if you’re
20% body fat, you have 68kg of fat-free 170cm (5’7”), but borderline impossible
mass. Thus, your predicted total would if you’re 200cm (6’7”). For this reason,
be 68 × 11.48 - 222 = 559kg. To sanity fat-free mass per unit of height would be a
check this prediction, that total would better metric. It’s reported in the Brechue
currently rank you around the 61st per- and Abe study, but not the present study.
centile in the world in the 83kg class on I emailed the author of the present study,
the Open Powerlifting rankings: pretty and he indicated that he’s collecting more
good, but not otherworldly. That’s a re- data as part of an ongoing research proj-
alistic number to shoot for. If all we still ect, and that a future paper would report
had was the Brechue and Abe data, the the relationship between fat-free mass per
prediction would be 714kg, which would

65
unit of height in raw, drug-free power- you down with the details, but the way
lifters. When that paper is published, I’ll they went about selecting their factors
update the formula. I think fat-free mass wasn’t ideal, in my opinion. There are
per unit of height would do a better job multiple ways of developing multiple
ensuring the formula works well for fe- regression equations, and they made
male lifters too; even when male and fe- theirs “by the book”: starting with the
male lifters have the same amount of fat- factor that was most strongly associated
free mass, fat-free mass per unit of height with the outcome of interest (powerlift-
is generally greater for the male lifters ing total, in this case) and then including
(though I do still think the formula in the additional factors in descending order of
present study works reasonably well for the strength of their association with the
female lifters too). outcome of interest, as long as they’re
As one final note about this formula: not too strongly associated with some
it’s based on data from competitive pow- other factor already in the model. Unfor-
erlifters. If you don’t train like a power- tunately, the factor most strongly asso-
lifter, odds are pretty good that your ac- ciated with performance in all three lifts
tual total is considerably lower than the was flexed arm circumference, so flexed
total this formula would predict for you. arm circumference was the primary fac-
That doesn’t necessarily mean there’s tor in all of their models. Even if you
anything wrong with the formula, or any- wind up with a high R2 value for your
thing wrong with the way you train. The final model, I’m not wild about using
formula just doesn’t apply to people like flexed arm circumference as the prima-
you, because you don’t train the same ry factor for predicting squat or deadlift
way as the subjects used to derive the for- performance. Calf circumference was
mula. However, if you decide you want also a factor in the bench press model
to give powerlifting a shot, it could help which, again, strikes me as less than ide-
give you some numbers to shoot for. al. An alternate approach to multiple re-
gression is called etiological modeling.
Finally, there was actually another pa-
With an etiological model, you deter-
per published this month on the same
mine the factors that will go into your
topic: predicting performance in com-
model a priori. In other words, if you’re
petitive powerlifters (10). Unfortunate-
studying lung cancer risk, you can say
ly, I’m not too wild about how they
that smoking status and age are going in
developed their regression equations.
your model before you even look at your
They used multiple regression, which
data, whether or age or smoking status
is a good thing if done well, since you
actually turn out to be associated with
can incorporate more predictors (like
lung cancer risk in your particular data-
training age, for example). I won’t bog
set. An etiological multiple regression

66
APPLICATION AND TAKEAWAYS
As you accumulate more fat-free mass, you should be able to lift more. Within samples
of competitive powerlifters, the relationship between fat-free mass and powerlifting
total is consistently strong. We can use that relationship to predict how much you
“should” be able to lift, given your current amount of fat-free mass. If your lifts are
close to the predictions, building more muscle is probably your best way forward. If
your lifts are considerably lower than the predictions, you can probably continue im-
proving your total by improving your technique and lifting efficiency (though building
more muscle certainly never hurts).

model for powerlifting performance may ence in years) - (6.7 × body fat percent-
include fat-free mass per unit of height, age) + (19.4 × flexed upper arm girth in
age, training age, and sex, for example. cm) + (3.9 × thigh girth in cm) - (9.7 ×
An etiological multiple regression mod- calf girth in cm) + (3.9 × hip circumfer-
el for bench press performance almost ence in cm)
certainly wouldn’t include calf circum- If you want to plug some numbers into
ference. I assume it’s clear by now that I any of the equations discussed in this ar-
think an etiological modeling approach ticle, I made a spreadsheet you can play
would have been preferable. However, around with. You can find it here. As a
here were the equations in that study, little bonus, I also included the equations
in case you want to plug some of your from the Brechue and Abe study if you
measurements into them: just restrict the sample to lifters with less
Squat = -145.7 + (4.3 × training expe- than 90kg of fat-free mass. The calcula-
rience in years) - (1.7 × body fat per- tor also includes predictions for each lift,
centage) + (6 × flexed upper arm girth in in case you want to see how one of your
cm) + (1.9 × thigh girth in cm) lifts stacks up against the predictions.
Bench = -82 - (1.2 × body fat percent-
age) + (6.6 × flexed upper arm girth in
cm) + (0.8 × thigh girth in cm) - (2 × calf Next Steps
girth in cm) I’m just waiting for Ferland and col-
Deadlift = -79.8 + (4.9 × training ex- leagues to publish their full dataset ana-
perience in years) - (1.9 × body fat per- lyzing performance as a function of fat-
centage) + (8.1 × flexed upper arm girth free mass per unit of height. I’d love to see
in cm) + (2 × thigh girth in cm) - (3 × an etiological modeling approach used on
calf girth in cm) a larger dataset, with predefined variables
of fat-free mass per unit of height, train-
Total = -391.2 + (10 × training experi-
ing age, biological age, and sex.

67
References
1. Ferland PM, St-Jean Miron F, Laurier A, Comtois AS. The relationship between body com-
position measured by dual-energy X-ray absorptiometry and maximal strength in classic
powerlifting. J Sports Med Phys Fitness. 2020 Mar;60(3):407-416.
2. Taber CB, Vigotsky A, Nuckols G, Haun CT. Exercise-Induced Myofibrillar Hypertrophy
is a Contributory Cause of Gains in Muscle Strength. Sports Med. 2019 Jul;49(7):993-997.
3. Bishop P, Cureton K, Collins M. Sex difference in muscular strength in equally-trained men
and women. Ergonomics. 1987 Apr;30(4):675-87.
4. The most expansive explanation of this hypothesis I’ve seen was in the book “The Sports
Gene” by David Epstein. It’s really good. You should pick up a copy
5. Brechue WF, Abe T. The role of FFM accumulation and skeletal muscle architecture in pow-
erlifting performance. Eur J Appl Physiol. 2002 Feb;86(4):327-36.
6. It’s actually fat-free mass in the Brechue and Abe study, but LBM and FFM are similar
enough that using them interchangeably shouldn’t have a meaningful effect in this context
7. Krivickas LS, Dorer DJ, Ochala J, Frontera WR. Relationship between force and size in hu-
man single muscle fibres. Exp Physiol. 2011 May;96(5):539-47.
8. Dankel SJ, Kang M, Abe T, Loenneke JP. Resistance training induced changes in strength
and specific force at the fiber and whole muscle level: a meta-analysis. Eur J Appl Physiol.
2019 Jan;119(1):265-278.
9. At first, it would be intuitive to assume that the y-intercept should be zero – if you have zero fat-
free mass, you should be able to lift zero weight. However, you need to keep in mind that fat-free
mass doesn’t produce force. Skeletal muscle mass does. Thus, if a theoretical human had all of their
normal fat-free mass other than muscle mass, you would still expect them to produce zero force.
So, we need to ballpark a plausible lower LBM-sans-muscle-mass figure for adult powerlifters. The
lightest female weight class is the 43kg class. Muscle mass accounts for ~33% of womens’ body
mass, on average. Body fat percentage is more of a guess, but lightweight female lifters tend to be
quite lean, so we’ll ballpark the body fat percentage at 22%. Thus, non-muscle fat-free mass would
account for 45% of this theoretical person’s body mass, or 19.35kg. That’s our x-intercept. We can
use that information to calculate the y-intercept. Skipping ahead a few steps, it’s -237.2.
10. Ferrari L, Colosio AL, Teso M, Pogliaghi S. Performance and Anthropometrics of Classic
Powerlifters: Which Characteristics Matter? J Strength Cond Res. 2020 Mar 12.
11. The figures in the text actually report the correlations between lean soft tissue mass and power-
lifting performance. Lean soft tissue mass is fat-free mass minus bone mineral content. Fat-free
mass is the metric most MASS readers will be more familiar with, and enough data was reported
in figures for me to add up each subject’s total fat-free mass, so that’s what I showed in the figures
here. If you pull up the full text of the study and have a very keen eye to notice that all of the x
coordinates are 2-4kg higher in my figures than the figure in the text of the study, that’s why.

68
Study Reviewed: Effect of Rest Interval Duration on the Volume Completed During a
High-Intensity Bench Press Exercise. Hernandez et al. (2020)

The Scientific and Practical


Considerations of Long
Inter-Set Rest Intervals
BY MIC HAE L C . ZO URD O S
Longer inter-set rest intervals can lead to greater hypertrophy
and strength gains than shorter inter-set rest intervals. This study
demonstrates that eight-minute rest intervals may be needed
to maximize volume performance. But is eight minutes really
necessary? This article discusses both the scientific merit and practical
considerations of rest interval duration.

69
KEY POINTS
1. This study compared the difference in reps performed during four sets to failure
at 85% of one-repetition maximum (1RM) when trained men took two minutes,
five minutes, and eight minutes of inter-set rest.
2. Unsurprisingly, more reps were performed over the four sets with eight minutes
rest compared to both five and two minutes. Further, more reps were performed
with five minutes of inter-set rest versus two minutes of rest.
3. This study clearly shows that greater inter-set rest times are associated with
better maintenance of reps performed from set-to-set when training to failure.
Importantly, longitudinal studies do support longer rest times for hypertrophy;
however, eight-minute rest periods may not be necessary or practical. Therefore,
this article provides a nuanced discussion of both the cross-sectional and
longitudinal inter-set rest interval data, and provides specific recommendations
related to rest intervals.

N
ot too long ago, it was standard short rest intervals, then what about re-
to recommend 30-90 second in- ally long rest intervals? This crossover
ter-set rest intervals to maximize design study from Hernandez et al (1)
hypertrophy. A few longitudinal studies had 15 trained men perform four sets
now exist which show either no differ- to failure at 85% of 1RM in three dif-
ence (2) or a potential benefit of long ferent conditions with variable inter-set
rest intervals (3, 4) for hypertrophy. rest periods. The conditions included:
Similarly, longitudinal studies focused 1) Two minutes inter-set rest, 2) Five
on strength show no difference between minutes inter-set rest, and 3) Eight min-
short and long rest intervals (2, 3) or utes inter-set rest. Not surprisingly, the
enhanced strength with long rest inter- greatest number of total reps performed
vals (4, 5). Importantly, various cross- across all four sets was in the eight-min-
over design studies have shown that ute condition, followed by the five-, and
when the number of sets are equated then the two-minute condition. These
between short and long rest interval pro- results translated to 22.6% more volume
tocols, the short rest intervals can lim- in the eight-minute inter-set rest condi-
it total volume by reducing the number tion versus the five-minute inter-set rest
of reps completed per set (6); however, condition and 19.2% more total volume
the longest rest interval investigated to in the five-minute inter-set rest condi-
date for this purpose is five minutes (7). tion versus the two-minute inter-set rest
So, if long rest intervals are better than condition. The straightforward implica-

70
tion of these results is that taking longer from the introduction, it seems clear that
inter-set rests will lead to more volume, the authors thought total reps performed
which would potentially promote great- would scale with greater inter-set rest.
er hypertrophy and strength adapta-
tions compared to shorter inter-set rests.
While that interpretation is logical, tak- Subjects and Methods
ing eight-minute rest intervals is not
always practical. Therefore, this article Subjects
will review the current state of the liter- 15 men with an age range of 20-34
ature related to both cross-sectional and years old participated. All subjects had
longitudinal rest interval studies; then, trained at least three days per week for
it will provide a nuanced discussion re- the past year leading up to the study. The
garding the practicality of short and long remainder of available subject details
rest intervals. Specific examples of how are in Table 1.
and when to implement both short and
long rest intervals will be provided.

Purpose and Hypotheses Figure 1 Total volume performed

Purpose
3000

The purpose of this study was to com- 2 min

5 min

pare the number of reps performed


Total volume (kg)

8 min
2000

during four sets to failure at 85% of


1RM in the bench press in three differ- 1000

ent conditions consisting of two-, five-,


and eight-minute inter-set rest intervals. 0
2 min5 min 8 min
Condition

Hypotheses From Hernandez et al. (1).

No hypotheses were given. However,


* = significantly less volume than the 8-minute interest condition
# = significantly less volume than the 5-minute interest rest condition

71
Study Overview length of time of the inter-set rest inter-
Four laboratory visits, each separated val. The researchers then compared to-
by at least 48 hours, were needed for tal volume (sets × reps × weight lifted)
this crossover design study. The first and repetitions completed between con-
visit tested bench press 1RM, then one ditions, and they examined the degree
set to failure at 85% of 1RM was com- to which the number of repetitions per-
pleted. The next three visits were the formed decreased from set to set within
testing sessions, which were completed each condition.
in a randomized order.

Testing Sessions Findings


The three testing sessions were identi- The findings of this study were
cal except for the inter-set rest interval, straightforward. More total volume was
which was two minutes in one condition, performed and better maintenance of
five in another, and eight minutes in the reps performed from set to set occurred
final condition. For each condition, sub- when longer inter-set rest intervals were
jects performed four sets of bench press taken. This data is shown in Figure 1
at 85% of 1RM. The first set was per- and Table 2. Below is a more detailed
formed for the exact number of repeti- description of those findings.
tions completed during the 85% set to
failure on the 1RM day. Subjects then Total Volume and Reps Performed
performed three sets to failure at 85% Total volume performed was sig-
following that first set. Again, the only nificantly greater (p < 0.01) in the
difference between conditions was the eight-minute rest interval condition

72
(2,207 ± 372 kg) than in the five-min- interval than the five-minute rest inter-
ute condition (1,793 ± 315 kg). Further, val. Specifically, the five-minute rest
greater total volume was performed in condition performed three fewer reps
the five-minute rest condition compared on average in set four versus set one.
to the two-minute condition (1,448 ± In the eight-minute rest condition, sub-
215 kg) (Figure 1). jects performed only one fewer rep on
For reps performed from set to set, there set four versus set one. In short, these
was a decrease in reps from set one to set findings suggest that lifters training for
two in both the two-minute and five-min- hypertrophy and strength should take
ute rest conditions (p < 0.05). Further, longer rest intervals, which will lead to
both the two- and five-minute conditions more volume and potentially more gains
performed fewer reps than the eight-min- over the long-term. Fortunately, there
ute rest condition in sets two, three, and are a handful of similar cross-sectional
four (p < 0.05). Lifters also performed studies that have also looked at rest in-
fewer reps on sets two, three, and four in tervals and reps performed, along with a
the two-minute rest condition compared few longitudinal studies comparing rest
to the five-minute rest condition (p < intervals for strength and hypertrophy.
0.05). In the eight-minute rest condition, Let’s now take a look at those studies
reps performed did not statistically de- before interpreting how the collective
crease from set one to set four (p = 0.06); findings translate into practice.
however, there was a decrease of -15.63%
(effect size = 0.85), which equated to an Other Acute Studies Looking at Rest
average decrease of one rep. Intervals and Reps Performed
As stated above, a handful of oth-
er studies have looked at the effects of
Interpretation “long” versus “short” rest intervals on
The findings of this study by Hernan- reps performed during sets to failure,
dez et al (1) should not be too surpris- and for the most part, these studies agree
ing. Certainly, most of us would have that longer rest intervals lead to more
hypothesized that taking longer rest reps performed in a single training ses-
intervals between bench press sets to sion. I put “long” and “short” in quotes
failure would result in better mainte- above because those are really relative
nance of repetitions per set. The only terms, and they’re mostly used in a bi-
thing that may be slightly surprising is nary fashion when comparing one rest
that subjects maintained reps performed interval to another, even if both rest in-
so much better in the eight-minute rest tervals are relatively short. For example,
Senna et al observed that ~36% more

73
bench press reps were performed with versus both 30-second and 1-minute
three minutes of inter-set rest versus one rest intervals. Lastly, Scudese et al 2014
minute during five sets to failure at a (10) had trained men perform five sets
10RM load (6). Similar to Senna, Tiba- to failure with a 3RM load in four dif-
na et al (8) found that trained men per- ferent rest interval conditions: one-min-
formed ~26% more reps with three-min- ute, two minutes, three minutes, and five
ute rests versus one-minute rests over minutes. Interestingly, Scudese reported
five sets to failure on the Smith machine that the fewest reps were performed in
bench press. Willardson and Burkett the one-minute rest condition, but there
(7) unsurprisingly found that five-min- was no difference in reps performed be-
ute inter-set rest intervals allowed for tween the other conditions. The differ-
more reps over five sets to failure on ence in the Scudese studies compared
both the squat and bench press with an to the Senna and Tibana studies is likely
8RM load versus both two minutes and explained by the high load used by Scu-
one minute of inter-set rest. Another dese. These studies, along with the pres-
study from Willardson and Burkett et al ent study, are summarized in Table 3.
(9) found that two-minute rest intervals So, the novelty in the present study lies
allowed for better sustained repetition in the length of rest used, as the longest
performance on five sets to failure with rest interval in these previous studies was
a 15RM on the squat and bench press

74
five minutes. Therefore, it is interesting utes rest. So, it may make sense to take
that the presently reviewed study (1) used eight-minute rest intervals if training to
a load (i.e. 85% of 1RM) not too dissimi- failure on the compound movements, but
lar from the Scudese study (3RM or ~90% this seems unnecessary in many other
of 1RM), and found eight-minute rest situations. The studies mentioned in this
intervals to yield so much more volume section only contain acute data, so let’s
than five-minute rest intervals. So, do now take a look at the longitudinal stud-
you need to rest eight minutes? Probably ies comparing different inter-set rest in-
not from a purely scientific perspective. tervals to see how this actually plays out.
From a practical perspective, eight min-
utes is generally unsustainable, but we’ll Longitudinal Studies Comparing Dif-
return to the practical discussion later. I ferent inter-set Rest Intervals
suggest that you probably don’t need to After examining the acute data above,
rest eight minute between sets, because it seems logical to deduce that taking
all of the aforementioned studies used longer inter-set rest intervals to perform
failure training on every single set. How- more total volume is advantageous for
ever, it is unlikely that you are training to hypertrophy. However, up until pretty
failure on every set. So, if the goal is in- recently, this data was largely ignored
deed to perform as many reps as possible in textbooks and mainstream exercise
on every single set, then eight minutes physiology academia. Rather, right after
– or potentially even more – of inter-set the old guard would explain how they
rest may be necessary. However, if you walked uphill to school both ways in the
are performing more traditional volume snow, they would then repeat the narra-
(i.e. 5×8 at 5-8RPE), then each set won’t tive that acute testosterone and growth
be as fatiguing from both a neuromuscu- hormone responses to resistance train-
lar and metabolic perspective, meaning ing were causative factors in skeletal
you can almost certainly get away with muscle hypertrophy. Further, the largest
a shorter rest interval than eight minutes. post-exercise transient response of an-
Further, the Scudese study suggests that abolic hormones tends to happen when
with higher intensities, shorter rest inter- rest intervals are really short (i.e. 30
vals can maintain repetition performance. seconds to 1 minute) (11); thus, the old
Additionally, different rest intervals are guard recommended short rest intervals
likely needed on different exercises. You to maximize hypertrophy, and this theo-
may need five minutes of rest to success- ry became widely accepted. Sure, early
fully maintain the same load for squats data did show a relationship between the
during 5×8 at 5-8RPE mentioned above, acute anabolic hormone response and hy-
but could complete such a prescription pertrophy (12), but it simply tended to be
on rows or curls with only 2-3 min-

75
that hypertrophy-type protocols that used lieve it is, then the presently reviewed
short rest intervals also had more volume study and the other acute studies men-
programmed. More recent literature has tioned above may indeed be indicative
clearly demonstrated that the acute hor- of what long-term rest intervals should
monal response is not a causative factor be. On the other hand, Ahtiainen et al
in hypertrophy (13). So, with the hor- compared five-minute rest intervals ver-
mone hypothesis discounted, what is an sus two-minute rest intervals for quad-
appropriate inter-set rest time? riceps hypertrophy and found no differ-
Both Buresh et al 2009 (3) and, more ence between groups for muscle growth
recently, Schoenfeld et al 2017 (4) com- (2). However, since both groups trained
pared the effects of different inter-set to failure, Ahtiainen had the short rest
rest interval lengths on long-term hyper- group add sets to equate for volume
trophy. Buresh et al (3), had two groups since fewer repetitions per set were per-
train twice per week for 10 weeks with formed in the short rest group; whereas,
the number of sets equated and the only Buresh and Schoenfeld equated for sets
difference between groups being rest- between groups (more on this later).
ing 2.5 minutes or 1 minute between The Buresh (3) and Schoenfeld (4)
sets. Buresh found no statistically sig- studies also examined strength gains.
nificant differences between groups for Surprisingly, Buresh did not find a dif-
arm cross-sectional area; however, there ference between groups for gains in
was an effect size of 0.32 in favor of the 5RM squat and bench press strength.
longer rest interval group (2.5 minutes). However, Schoenfeld et al reported
Schoenfeld et al (4), also showed a po- greater increases in 1RM squat and
tential benefit for long rest over short bench with long rest versus short rest.
rest for hypertrophy. In the Schoenfeld Similar to Schoenfeld, De Salles et al
study, trained men lifted three times (5) found that five and three minutes of
per week for 10 weeks with either three inter-set rest resulted in greater leg press
minutes or one minute of inter-set rest. strength over 16 weeks versus one min-
Hypertrophy over the 10 weeks was ute of inter-set rest. Further, De Salles
similar in the biceps, triceps, and lat- et al found that both the three- and
eral quadriceps; however, hypertrophy five-minute groups performed more to-
was significantly greater in the anterior tal volume than the one-minute group.
quadriceps. Schoenfeld and colleagues In contrast to De Salles and Schoenfeld,
attribute the greater thigh hypertrophy Willardson and Burkett (13) found no
to more volume completed in the long difference in changes in squat 1RM over
rest group. If that assertion is the reason 13 weeks when comparing trained men
for the hypertrophy findings, and I be- who trained twice per week and took ei-

76
ther two or four minutes between sets. you may need it in some cases such as sets
However, Willardson and Burkett did of squats with 15-20 reps close to failure,
find that lifters in the four-minute in- but this probably becomes impractical
ter-set rest group performed more total from a time perspective. Honestly, taking
volume than the two-minute rest group. even five minutes between every set could
While it might be a stretch to say that become impractical. But of course, rest
there is an overwhelming amount of data intervals can change with different exer-
to support longer rest intervals, I do think cises, so if you have three or four sets of
the data is clearly leaning in this direction squat or bench at a submaximal RPE, then
for both strength and hypertrophy. Impor- taking about five minutes rest is probably
tantly, there is no study in young individ- doable. Then, when you get to assistance
uals that has shown better hypertrophy or movements, shorter rests of 1-3 minutes
strength gains with short rest; thus, the should be sufficient and won’t keep you in
findings are either null or in favor of long the gym for an exorbitant amount of time.
inter-set rest. There are, however, a few The main caveat for the studies that
caveats and practical considerations to most people point to (usually Buresh and
this, which we’ll now discuss. Schoenfeld) in support of long rest for
hypertrophy is that the number of sets
Practical Considerations of Long In- is equated between groups in these stud-
ter-Set Rest ies. When sets were added to the short
First, let’s return to the presently reviewed rest group in the Ahtiainen study (2),
study, which showed that eight-minute there was no difference in lower body
rest intervals yielded more reps per set and hypertrophy. I don’t think adding sets
total volume than five- or two-minute rest is a better design; it’s just answering a
intervals. Despite this, I don’t think any- slightly different question. Specifically,
one should interpret this as an indication adding sets in the Ahtiainen study just
that eight-minute rest intervals are nec- suggests that if volume load is equated,
essary all the time. If you train to most- then rest intervals may not be all that
ly submaximal RPEs for volume, the five important for hypertrophy. Therefore,
minutes is probably sufficient on the main while longer rest intervals do allow for
lifts. In fact, a previous study (15) that we more volume and better average inten-
reviewed found that when lifters rested sity when training to failure, you could
intuitively (i.e. not timed), they tended to simply equate for volume by adding an
rest about 4-5 minutes and were able to extra set and taking short rest, which
successfully complete a 5×5 training ses- may actually take less time. For exam-
sion at a 5RM load. Now, you can certain- ple, if you performed 5×8 on squats with
ly take eight minutes of inter-set rest, and 100kg and took five-minute rests, that is

77
20 minutes of total inter-set rest, but per- intervals. Theoretically, in this scenario
forming 6×8 with 90kg and two minutes you could even add a couple more sets
of inter-set rest only requires 10 minutes with the short rest protocol and still fin-
of rest intervals, and certainly less total ish in less time. I’m not necessarily ad-
training time. The 6×8 workout is also vocating for adding more sets, and there
slightly more total volume. If you are is merit to limiting the amount of sets
worried about proximity to failure in this per session on one exercise, but the op-
scenario, I don’t think that’s an issue, as tion is there if you are someone that can
the short rest will probably cause RPE handle really high volumes. However,
to go up each set, and if the 100kg sets if you typically take about five-minutes
are nearly to failure, the 90kg sets won’t rests on the main lifts and are short on
be that far from failure with short rest time one day, then consider making the
above adjustment to get in your neces-
sary volume for the day. If you are short
on time and strength is your main goal,
IF YOU ARE SHORT ON TIME then I would keep your first set at your
programmed training load so that peak
AND STRENGTH IS YOUR intensity is not affected, which should
maintain strength, and then simply take
MAIN GOAL, THEN I WOULD a shorter rest and lighten the load on the
KEEP YOUR FIRST SET remaining sets. In general, more insight
into making training adjustments for
AT YOUR PROGRAMMED time efficiency was written about here,
and specific examples of how to manip-
TRAINING LOAD SO THAT PEAK ulate exercises when short on time are
provided in the program troubleshoot-
INTENSITY IS NOT AFFECTED, ing video. For assistance movements,
you could add in variations of rest-pause
WHICH SHOULD MAINTAIN and agonist-antagonist supersets when
STRENGTH, AND THEN SIMPLY you are short on time. Additionally, you
can probably get away with shorter rest
TAKE A SHORTER REST intervals on assistance movements (in
general) than you can on the main lifts.
AND LIGHTEN THE LOAD ON That being said, I still think that taking
sufficient rest is a good idea, rather than
THE REMAINING SETS. just packing in more volume with shorter
rest intervals. In other words, if you can

78
APPLICATION AND TAKEAWAYS
1. This study demonstrates that taking longer inter-set rest intervals is associated
with greater repetitions performed during sets to failure at 85% of 1RM on the
bench press.
2. Indeed, there are longitudinal studies that support the idea that longer rest intervals
can be superior for both hypertrophy and strength. However, the evidence is not
overwhelming in this area; rather, it is merely leaning in that direction.
3. While I do think it is generally a good idea to rest long enough to maximize perfor-
mance, this may not always be practical. Therefore, adding sets and reducing rest
times is a viable strategy when time is constrained.

manipulate performance to be better, I and does not have lifters train to failure.
think that’s a good thing. Lastly, I think Perhaps one-minute, three-minute, and
taking longer rest may reduce the num- five-minute inter-set rest interval groups
ber of total sets you need to maximize would be the way to go. In this setup,
hypertrophy. For example, we previously the group with one-minute rest inter-
reviewed a study (16) that suggests that vals would have the most sets, followed
30+ sets per week may be optimal for by the three-minute rest group, and the
upper and lower body hypertrophy and five-minute inter-set rest group would
squat and bench strength; however, this have the fewest sets.
study used only 60-second rest intervals.
In that review, I postulated that if longer
rest intervals were used, perhaps the lift-
ers may have been able to get in more
quality sets, which could mean that hy-
pertrophy might be optimized with fewer
sets than 32 on a group level. As always,
understanding this topic conceptually
can help you to apply the appropriate rest
interval duration to your training.

Next Steps
We need a longitudinal study that is
not equated for sets but for total volume

79
References

1. Hernandez DJ, Healy S, Giacomini ML, Kwon YS. Effect of Rest Interval Duration on the
Volume Completed During a High-Intensity Bench Press Exercise. The Journal of Strength
and Conditioning Research. 2020 April.
2. Ahtiainen JP, Pakarinen A, Alen M, Kraemer WJ, Häkkinen K. Short vs. long rest period
between the sets in hypertrophic resistance training: influence on muscle strength, size, and
hormonal adaptations in trained men. The Journal of Strength and conditioning Research.
2005 Aug 1;19(3):572.
3. Buresh R, Berg K, French J. The effect of resistive exercise rest interval on hormonal re-
sponse, strength, and hypertrophy with training. The Journal of Strength & Conditioning
Research. 2009 Jan 1;23(1):62-71.
4. Schoenfeld BJ, Pope ZK, Benik FM, Hester GM, Sellers J, Nooner JL, Schnaiter JA,
Bond-Williams KE, Carter AS, Ross CL, Just BL. Longer inter-set rest periods enhance
muscle strength and hypertrophy in resistance-trained men. Journal of strength and condi-
tioning research. 2016 Jul 1;30(7):1805-12.
5. de Salles BF, Simão R, Miranda H, Bottaro M, Fontana F, Willardson JM. Strength increases
in upper and lower body are larger with longer inter-set rest intervals in trained men. Journal
of Science and Medicine in Sport. 2010 Jul 1;13(4):429-33.
6. Senna GW, Figueiredo T, Scudese E, Baffi M, Carneiro F, Moraes E, Miranda H, Simão R.
Influence of Different Rest Interval Lengths in Multi-Joint and Single-Joint Exercises on
Repetition Performance, Perceived Exertion, and Blood Lactate. Journal of Exercise Physi-
ology Online. 2012 Oct 1;15(5).
7. Willardson JM, Burkett LN. A comparison of 3 different rest intervals on the exercise vol-
ume completed during a workout. The Journal of Strength & Conditioning Research. 2005
Feb 1;19(1):23-6.
8. Tibana RA, Vieira DC, Tajra V, Bottaro M, Willardson JM, de Salles BF, Prestes J. Effects
of rest interval length on Smith machine bench press performance and perceived exertion in
trained men. Perceptual and motor skills. 2013 Dec;117(3):682-95.
9. Willardson JM, Burkett LN. The effect of rest interval length on the sustainability of squat
and bench press repetitions. Journal of Strength and Conditioning Research. 2006 May
1;20(2):400.
10. Scudese E, Willardson JM, Simão R, Senna G, de Salles BF, Miranda H. The effect of rest
interval length on repetition consistency and perceived exertion during near maximal loaded
bench press sets. The Journal of Strength & Conditioning Research. 2015 Nov 1;29(11):3079-
83.
11. Kraemer WJ, Ratamess NA. Hormonal responses and adaptations to resistance exercise and
training. Sports medicine. 2005 Apr 1;35(4):339-61.

80
12. McCall GE, Byrnes WC, Fleck SJ, Dickinson A, Kraemer WJ. Acute and chronic hormonal
responses to resistance training designed to promote muscle hypertrophy. Canadian Journal
of applied physiology. 1999 Feb 1;24(1):96-107.
13. Morton RW, Oikawa SY, Wavell CG, Mazara N, McGlory C, Quadrilatero J, Baechler BL,
Baker SK, Phillips SM. Neither load nor systemic hormones determine resistance train-
ing-mediated hypertrophy or strength gains in resistance-trained young men. Journal of ap-
plied physiology. 2016 Jul 1;121(1):129-38.
14. Willardson JM, Burkett LN. The effect of different rest intervals between sets on volume
components and strength gains. The Journal of Strength & Conditioning Research. 2008 Jan
1;22(1):146-52.
15. Ibbott P, Ball N, Welvaert M, Thompson KG. Variability and Impact of Self-Selected in-
ter-set Rest Periods During Experienced Strength Training. Perceptual and motor skills.
2019 Jun;126(3):546-58.
16. Brigatto FA, Lima LE, Germano MD, Aoki MS, Braz TV, Lopes CR. High Resistance-Train-
ing Volume Enhances Muscle Thickness in Resistance-Trained Men. Journal of Strength and
Conditioning Research. 2019 Dec 20.

81
Study Reviewed: The Effect of L-Carnitine Supplementation on Exercise-Induced
Muscle Damage: A Systematic Review and Meta-Analysis of Randomized Clinical
Trials. Yarizadh et al. (2020)

Has Carnitine Pivoted From a Fat


Loss Supplement to a Recovery
Supplement?
BY E RI C T RE X LE R

Back in the day, carnitine was widely touted as a fat loss supplement.
While that hype has mostly died down, a new meta-analysis suggests
that carnitine may reduce post-exercise muscle damage and soreness.
Read on to find out if the results of this meta-analysis are as good as
they seem.

82
KEY POINTS
1. The presently reviewed meta-analysis (1) evaluated the effects of L-carnitine and
L-carnitine-L-tartrate (which the authors collectively referred to as “L-carnitine”)
on a variety of outcomes related to exercise-induced muscle damage and
recovery.
2. The authors reported that muscle soreness was significantly reduced by L-carnitine
supplementation at all time points (0, 24, 48, 72, and 96 hours), and significant
effects on creatine kinase, myoglobin, and lactate dehydrogenase were observed
at the 24-hour mark. After re-running the data differently, my results suggest that
only soreness and creatine kinase were significantly impacted, and only for about
48 hours post-exercise.
3. I’d like to see a larger body of more consistent literature before I can confidently
recommend carnitine supplementation as a performance or recovery aid,
especially for young, healthy, well-trained lifters and athletes.

C
arnitine has been popular in the looked at creatine kinase, myoglobin,
supplement world for quite some lactate dehydrogenase, and muscle sore-
time. Back in the mid-2000s, ness values over a range of time points
it seemed like everybody was taking within 96 hours after exercise. After
L-carnitine to (purportedly) either boost pooling the results from the seven stud-
fat loss or facilitate muscle gain. It’s ies that met the inclusion criteria, the
also possible that you’ve heard of some authors reported that L-carnitine sup-
related forms of carnitine; for exam- plementation reduced creatine kinase,
ple, some studies evaluating effects on myoglobin, and lactate dehydrogenase
performance or blood flow have used values at 24 hours post-exercise, while it
L-carnitine-L-tartrate or glycine propi- reduced soreness values at 0, 24, 48, 72,
onyl-L-carnitine, and acetyl-L-carnitine and 96 hours post-exercise. They also
has become a popular supplement for did some additional subgroup analyses
cognition. The presently reviewed me- suggesting that results were generally
ta-analysis (1) evaluated the effects of more notable in untrained subjects than
L-carnitine and L-carnitine-L-tartrate trained subjects, and the creatine kinase
supplementation, which the researchers effect was more notable for resistance
collectively referred to as “L-carnitine,” exercise than aerobic exercise. Howev-
on a variety of outcomes related to ex- er, as MASS readers might recall from
ercise-induced muscle damage and re- some of my previous articles on creatine
covery. Specifically, the meta-analysis and vitamin D, I have a nasty habit of

83
re-doing meta-analyses. So, read on to and 9 women; 34 were untrained prior
find out how and why my calculations to participating in one of the studies, and
differ from the reported results, and 45 were trained. The mean age of each
what this means for your decisions about sample ranged from 22.0 to 51.9 years.
L-carnitine supplementation. Notably, a widely recognized “scoring
system” called the Jadad scale indicated
that three of the studies were “low quali-
Purpose and Hypotheses ty” and four of them were “high quality”
based on their methods and reporting.
Purpose
The purpose of the current meta-anal- Methods
ysis was to summarize the literature The researchers did a systematic
evaluating the effects of L-carnitine search of the literature, which yielded
supplementation on delayed-onset mus- 604 unique results. Then, they eliminat-
cle soreness and biomarkers of muscle ed studies based on their inclusion and
damage. exclusion criteria. They were looking
for studies that were 1) randomized con-
Hypotheses trolled trials, 2) evaluated at least one
The authors did not directly state a hy- of the outcomes of interest, 3) included
pothesis. In the introduction, they indi- measurements at baseline and at some
cated that multiple studies have shown point 24-72 hours after exercise, and
L-carnitine to improve various outcomes 4) recruited generally healthy subjects.
related to soreness and muscle damage, Studies were excluded if they included
but acknowledged that such findings are other interventions alongside carnitine
not entirely consistent. supplementation (such as another sup-
plement or a dietary modification), or
lacked a control group or condition.
Subjects and Methods If you’ve seen previous meta-analyses,
you might be used to them reporting stan-
Subjects dardized effect sizes, such as Cohen’s D.
The presently reviewed study was a In this case, the studies all reported the
meta-analysis, so they did not recruit same outcomes, so the researchers were
subjects of their own. Rather, they com- able to leave the variables in their raw
bined the results of previously conduct- units (which is preferable). Then, the re-
ed studies. They included seven studies searchers jammed these values into one
in their final analysis, containing a to- big model for each outcome (for exam-
tal of 79 subjects. This included 70 men ple, a single “soreness” analysis, with

84
several subgroups representing each should allow me to concisely summarize
time point). In the methods, they stated the findings without writing a novel.
that they used random effects models if In a nutshell, the authors reported that
there was “evidence of between-stud- muscle soreness was significantly reduced
ies heterogeneity,” but used fixed effect by L-carnitine supplementation at all
models if not. However, I’m not sure if time points (0, 24, 48, 72, and 96 hours).
they adhered to that with the analysis for L-carnitine significantly reduced creatine
muscle soreness; I could be misunder- kinase values at 24 hours, but not at 0,
standing their statement, but the muscle 48, or 96 hours. L-carnitine significant-
soreness forest plot (see Figure 1) re- ly reduced myoglobin levels at the “<24
ported significant overall heterogeneity, hours” time point, but none of the others
between-group heterogeneity, and with- (0, 24, 48, and 96 hours). L-carnitine sig-
in-group heterogeneity, but used a fixed nificantly reduced lactate dehydrogenase
effect model. Anyway, they made four levels at 24 hours, but not at the other time
forest plots, with subgroups for each time point evaluated (0 hours). Notably, signif-
point. They also made some funnel plots icant heterogeneity and significant funnel
and ran some tests to make sure the fun- plot asymmetry were observed for three
nel plots were adequately symmetrical; of the four variables. The lactate dehydro-
asymmetrical funnel plots can be prob- genase funnel plot was not significantly
lematic, since they may suggest the pres- asymmetrical, and its between-group and
ence of publication bias or some other overall heterogeneity tests were not signif-
latent effect that would impact how you icant, but don’t get too excited – the anal-
interpret the data. Finally, the researchers yses only included four data points from
did some follow-up tests to determine two total studies, so there really wasn’t
if effects were larger for untrained ver- much data to test.
sus trained subjects, L-carnitine versus
Finally, the researchers conducted
L-carnitine-L-tartrate studies, or aerobic
some subgroup analyses to assess the
versus resistance exercise bouts.
impact of some key variables, including
training status, form of supplement, and
form of exercise. They found that results
Findings were significantly larger in untrained
You could theoretically write a very, subjects compared to trained subjects,
very long section to summarize these and that results were larger following
findings. We’ve got four outcomes, sev- resistance exercise compared to aero-
eral time points for each, and plenty of bic exercise. The form of supplement
other details to discuss. So, I’ve present- (L-carnitine versus L-carnitine-L-tar-
ed all of the forest plots in Figure 1; this trate) didn’t seem to matter much.

85
time. I think a lot of people associate
Interpretation carnitine with fat loss or neurological
Before I begin, I’d like to commend applications, but generally aren’t aware
these authors for carrying out an inter- that multiple papers dating back to the
esting project. This body of literature early-1990s have been investigating
has flown under the radar for quite some the effect of L-carnitine supplementa-

86
tion on muscle damage and soreness. people?
So, I think this review is a great way to The answer to questions #1-3 is yes,
bring attention to that literature, which and the answer to question #4 is no. I
should (hopefully) prompt some more would argue that the analyses conduct-
follow-up work in this area. ed in this paper assume the incorrect an-
We tend to place meta-analyses at the swer for all four of these questions. As
very top of the “Hierarchy of Evidence,” a result, the confidence intervals are too
suggesting that their results are the most narrow, the p-values are too low, and
reliable and robust form of scientific evi- some of the pooled effect sizes are a bit
dence. While this makes sense in theory, off or susceptible to misinterpretation.
I fear that it causes us to sometimes read One of my first issues is that the effect
meta-analyses in a relatively uncritical of “time” definitely mattered, and there
manner. We don’t stop to consider exact- was a ton of funnel plot asymmetry, but
ly what the result means, or what assump- nobody really did anything about it. They
tions we have implicitly agreed to. When did the work to identify the issues, which
it comes to the presently reviewed study, is definitely a good thing. However, there
I want you to consider four questions: were no follow-up tests or adjustments to
1. Do soreness and muscle damage assess the impact of the funnel plot asym-
change over time following an ex- metry, and the effect of time wasn’t di-
ercise bout? rectly incorporated into the model. There
2. Is there more variability in soreness are a couple of other considerations to
or muscle damage levels a day or keep in mind. Whenever we have cor-
two after an exercise bout com- relations built into a data set, we have to
pared to several days after, when account for them whenever possible; this
most people are essentially back to is a foundational principle of statistics,
normal? and the exact reason why we have special
analyses for repeated measures data (for
3. If you measured some blood bio-
example, independent samples t-tests ver-
marker of muscle damage on four
sus paired samples t-tests). This study’s
separate occasions for a huge group
analyses do not account for the fact that
of people, would a person’s four
multiple values from the same study are
values be more related to each other
correlated with each other. Along the
than to a random value from some-
same lines, it doesn’t seem like the au-
one else in the group?
thors adjusted the analysis to reflect the
4. If you measured two people on four fact that the same people were measured
separate occasions each, did you several times. For example, consider two
actually measure eight different people getting measured four different

87
times; the sample size was two, but this a heteroscedastic compound symmetric
analysis seems to assume that the sample covariance structure, while using the
size was eight. restricted maximum-likelihood method
Based on these considerations, I decid- for estimating heterogeneity variance.
ed to reconstruct the analyses a bit dif- Let’s move past the most jargon-heavy
ferently. These specific analyses are just paragraph in the history of MASS.
a tad outside of my comfort zone, but You might be wondering why this
I took a crack at it. If you’re not super stuff matters, which is fair. The present-
interested in stats, just blow right by this ly reviewed study included some funnel
very small paragraph. I assumed that the plots; ideally, all of the data points fall
authors correctly reported the results within a nice funnel shape. Whenever
of each individual study, then re-did you see asymmetry (way more values on
the analysis differently. I took a gener- one side of the middle line than the other
al linear mixed model approach, used a side) or values way outside of the funnel
compound symmetric covariance struc- “boundaries,” this generally gives you
ture to model within-study correlations reasons to wonder exactly how accu-
with an assumed value of R = 0.6, and rate the pooled effect estimate is. In this
modeled time as a random effect using particular study, significant funnel plot

88
asymmetry was found for three of the significant at either time point available
four outcomes, and the fourth outcome (0hrs = -43.2 [-106.6, 20.1] U/L; 24hrs
only had four data points. If a ton of val- = 15.8 [-246.9, 278.4] U/L). With only
ues are to the left of the centerline, you four data points from two studies, there
might suspect that the analysis is actu- was never much we could’ve done with
ally exaggerating the effect of carnitine. the LDH data, even with the most ele-
For example, the current study reported gant and sophisticated modeling imag-
an overall pooled effect (across all time inable. For creatine kinase, results were
points) of -18.2 units for creatine kinase, statistically significant at 24 hours (-28.4
with a p-value of <0.001. If you apply [-48.0, -8.8] U/L), and pretty notable but
a very basic method of “correcting” for non-significant at 48 hours (-23.2 [-50.1,
this funnel plot asymmetry (the trim and 3.7] U/L), while differences at 0 hours
fill method), the effect gets cut in half and 96 hours were essentially negligible
(-9.9 units), and the p-value jumps to p (no differences larger than 5 U/L, and
= 0.085. Just like that, an effect size gets both p-values above 0.29). For muscle
cut in half and goes from clearly signif- soreness, differences were significant
icant to non-significant. But, we can do at 0, 24, and 48 hours (0hrs = -1.19
better than just using the trim and fill [-1.61, -0.78] cm; 24hrs = -1.69 [-2.20,
method. If we instead model the data as -1.18] cm; 48hrs = -1.31 [-2.04, -0.58]
I described above, we get more robust cm). Differences were not significant at
results. For a concrete example, let’s 72 hours (-0.60 [-1.91, 0.71] cm) or 96
look at the muscle soreness funnel plots hours (-0.40 [-0.83, 0.04] cm). For sore-
(Figure 2). While the caption describes ness values, I believe the researchers
some nuanced details to facilitate inter- reported centimeter values as millime-
pretation, the impact of the two differ- ters, which is important to note. Finally,
ent approaches (the version in the paper for myoglobin, differences at each time
versus my analysis) on the funnel plots point were basically negligible, with no
suggests that we’ve made some signifi- differences larger than 15 μg/L, and no
cant improvements. p-values under 0.29.
When it comes to the “new” analyses, So, the current data do suggest that
the results look a little bit different. Note carnitine supplementation might reduce
that I’m reporting the estimated differ- muscle soreness and creatine kinase lev-
ence between the carnitine treatment and els following exercise. However, the re-
placebo treatment at each time point, sults reported in the presently reviewed
along with the 95% confidence inter- meta-analysis probably overstate the con-
val [lower boundary, upper boundary]. fidence of the findings and the number of
For LDH, results were not statistically time points at which statistically signif-

89
onset muscle soreness suggests that in
addition to the initial mechanical damage
WHEN WE BOIL THE RESULTS from exercise, contributing factors may
include localized hypoxia and the sub-
DOWN TO PRACTICAL sequent responses involving oxidative
stress and inflammation. As described by
TERMS, IT LOOKS LIKE Volek et al (2), hypoxic conditions cause
carnitine release from endothelial cells,
CARNITINE REDUCED increased oxidative stress, and impaired
MUSCLE SORENESS FOR THE blood flow regulation. There is some evi-
dence to suggest that L-carnitine may en-
FIRST 48 HOURS AFTER hance nitric oxide production and blood
flow, in addition to having an antioxidant
EXERCISE. THE MAGNITUDE effect. Along those lines, one study found
that L-carnitine-L-tartrate supplementa-
OF EFFECT WAS ABOUT 1-1.5 tion enhanced flow-mediated vasodila-
tion (2), and others have found increases
CM ON A 10CM SCALE. in nitric oxide metabolites (3) and total
antioxidant status (4) following L-car-
nitine supplementation. So, the idea is
that carnitine supplementation can main-
icant results were observed. When we tain adequate carnitine availability for
boil the results down to practical terms, it endothelial cells during exercise, which
looks like carnitine reduced muscle sore- prevents hypoxic stress during exercise,
ness for the first 48 hours after exercise. thereby reducing the degree of muscle
The magnitude of effect was about 1-1.5 damage and subsequent muscle soreness.
cm on a 10cm scale. So, if you were rat-
That seems pretty cool, but it doesn’t
ing your soreness from 0 to 10, this would
necessarily mean that L-carnitine sup-
knock your rating down by 1-1.5 points
plementation will necessarily improve
(from an 8 to a 6.5 or 7, for example),
performance. When it comes to perfor-
but by 72 hours after exercise, the effect
mance on a single bout of exercise, the
would be pretty negligible. The creatine
carnitine literature is fairly mixed, but
kinase results generally lend support for
the general consensus is that supplemen-
this finding, while the LDH and myoglo-
tation does not yield meaningful per-
bin results were unremarkable. There’s
formance benefits (5). It’s important to
also a plausible mechanistic basis for
note, however, that single-bout exercise
the observed effects. Our current under-
performance would not be impacted by
standing of muscle damage and delayed

90
any purported benefits related to muscle pretty straightforward outcomes like
damage or soreness, and would have to lactate accumulation.
be induced by a different mechanism As the authors noted in the presently
altogether. Theoretically, you could sur- reviewed meta-analysis, their subgroup
mise that carnitine’s effects on muscle analyses added some nuance to the find-
damage and soreness might more read- ings. The findings appeared to be most
ily translate to a benefit for longitudi- notable in studies using resistance train-
nal training adaptations, which involve ing as the exercise modality, and in stud-
protocols with repeated exposures to ies that tested untrained participants.
exercise stress and recovery. Unfortu- These intuitively make a lot of sense.
nately, such research is scant and pret- Naturally, untrained people are going
ty inconclusive. For example, one study to be more susceptible to acute muscle
found that supplementing with glycine damage, and while you could theoretical-
propionyl-L-carnitine throughout an ly create a pretty brutal, damage-induc-
eight-week endurance training program ing aerobic protocol with some intense
failed to improve aerobic or anaerobic downhill running, I think the resistance
performance outcomes more than a pla- exercise protocols were (in aggregate)
cebo (6). In another study (7), partici- more consistently damage-inducing
pants received either 1.5, 3.0, or 4.5g/ than the aerobic protocols in this body
day of glycine propionyl-L-carnitine for of literature. However, there’s another
28 days, while maintaining their typical source of variation to keep in mind here,
training habits. If anything, the larger particularly with the resistance training
doses in that particular study were ac- observation. The entire meta-analysis
tually associated with performance dec- included nine samples from seven sep-
rements on a repeated sprint test. More arate studies. Six of the samples utilized
recently, a study evaluated the effects resistance exercise, but five of those six
of 2g/day of L-carnitine supplementa- were all from the exact same lab. The
tion throughout a nine-week resistance carnitine literature is very inconsistent,
training program (8). The study reported but it seems like a big percentage of the
that carnitine improved leg press, bench positive findings are coming from one
press, and sprint performance, but fat- specific lab, and another large percent-
free mass was not significantly affect- age are coming from a specific country.
ed, and the statistical analyses for the Please don’t mistake that for a conspir-
study were pretty lenient. Overall, there acy; different labs and different regions
isn’t enough evidence to conclude that often have different preferences in terms
L-carnitine will acutely or chronically of the methods they implement and the
enhance performance, and the evidence population they are sampling. It does,
that exists is extremely mixed, even for

91
however, give us reason to believe that far the biggest source of dietary L-car-
in order to see a measurable performance nitine, providing up to 100-150mg per
or recovery benefit from carnitine sup- serving; other meat, fish, and dairy prod-
plementation, the conditions have to be ucts are typically down in the 5-30mg
fairly specific, which limits the ability range, and plant sources are quite low.
to broadly recommend it as a reliable Despite the fact that vegetarians have
supplementation strategy. I’ll also note very low dietary carnitine intakes, their
that the carnitine literature seems to plasma and muscle levels of L-carni-
have major, major geographic discrep- tine are only slightly lower than those
ancies in the reported findings; a recent of omnivores, and the difference doesn’t
meta-analysis looking at carnitine’s ef- seem to be functionally meaningful. It
fects on blood glucose and blood insulin seems that vegetarians are able to large-
levels found that results from “eastern ly compensate for low intake by re-
countries” were about 5-7x larger than ducing urinary carnitine excretion, and
those from “western countries,” and this potentially by increasing dietary absorp-
was a pattern observed over a decently tion of the carnitine that is consumed.
large body of evidence (for example, the On that note, it’s important to keep in
comparison for glucose included at least mind that L-carnitine bioavailability is
20 studies from each region). higher from food sources (54-86%) than
While we’re on the subject of carni- from a 2g supplement dose (9-25%) (9).
tine, it’s worth discussing a couple other Anyway, the point is that people gen-
common applications of carnitine sup- erally have plenty of carnitine around
plementation. People have been tout- for their fat-burning needs; as such, a
ing it as a fat loss supplement for ages, recent meta-analysis found that L-car-
primarily because carnitine is involved nitine supplementation did not signifi-
in one of the major metabolic steps in cantly enhance fat loss in normal-weight
the process of burning fat for energy. individuals, and had a very modest ef-
However, carnitine deficiency is gener- fect on weight loss in individuals with
ally caused by a rare genetic condition, overweight or obese BMI classifications
and in most other contexts, our bodies (10). There are some specific instances
are pretty good at making sure we’ve in which carnitine supplementation has
got enough carnitine to get the job done. been found to favorably impact fat loss,
Interestingly, this holds true across a muscle gain or retention, and physical
wide range of dietary carnitine intakes. performance, but these findings are typi-
We tend to make about 25% of our dai- cally observed in older samples (70s and
ly carnitine needs, with the other 75% beyond), individuals with sarcopenia,
coming from food (9). Red meat is by or individuals with unique medical cir-
cumstances, such as genetic conditions

92
affecting energy metabolism or ongoing reabsorption of L-carnitine, and prefer-
hemodialysis treatment (9). ential excretion of acylcarnitine esters
Finally, while it’s a bit beyond the scope (such as acetyl-L-carnitine) in the kid-
of this article, I should at least acknowl- ney (12). In any case, the simplest ap-
edge carnitine’s popularity for a variety proach is to just use the carnitine form
of brain-related outcomes. There’s a that is most commonly used for the out-
growing body of research indicating that come you’re seeking; much of the litera-
carnitine can have positive effects on ture assessing vascular effects uses gly-
things like depressive symptoms, men- cine propionyl-L-carnitine, much of the
tal fatigue, and prevention of cognitive literature assessing performance or body
decline, and there are plenty of plausi- composition outcomes uses L-carnitine
ble mechanisms to explain such effects. or L-carnitine-L-tartrate, and much of
However, this literature is largely based the literature assessing neurological out-
on animal studies or human trials in pop- comes uses acetyl-L-carnitine.
ulations with some form of cognitive In conclusion, the data available indi-
decline or neurological condition; gen- cate that supplementation with L-carni-
eralizing these results to young, healthy tine or L-carnitine-L-tartrate may reduce
individuals with no clinical conditions muscle soreness and creatine within 48
is an uncertain extrapolation. Studies hours of exercise-induced muscle dam-
looking at more neurologically focused age. The effect on soreness was enough
outcomes tend to use acetyl-L-carnitine to lower subjective sensation by about
rather than regular L-carnitine, like- 1.5 “points” on a scale from 0-10. This
ly because evidence suggests that ace- effect seemed to be more noticeable in
tyl-L-carnitine more readily crosses the studies involving untrained subjects and
blood-brain barrier. It’s interesting to resistance exercise. However, the carni-
note, however, that a great deal of oral- tine literature in general seems to be char-
ly ingested acetyl-L-carnitine appears to acterized by inconsistency. As it current-
be de-acetylated (to L-carnitine) during ly stands, I think it’s plausible to believe
or immediately after uptake from intes- that carnitine can facilitate recovery from
tinal cells (11), there seems to be a great damaging exercise. However, there are
deal of inter-conversion between L-car- plenty of other ingredients that also facil-
nitine and acetyl-L-carnitine after it en- itate recovery from exercise, while pro-
ters the bloodstream, and the kidney also viding a variety of other potentially bene-
does a ton of interconversion between ficial nutrients and bioactive compounds.
the two forms in the process of filtering As I stated in my recent article about tart
and excreting excess carnitine (12). In cherry juice, my go-to strategies for sup-
addition, there appears to be preferential porting recovery involve 1) making sure

93
The main obstacle with the carnitine
literature is inconsistency. However, as
BASED ON THE EVIDENCE more and more research becomes avail-
able, we tend to see that perplexing in-
AVAILABLE, I’D LIKE TO consistencies become explainable pat-
terns. So, I’m hoping that we’ll see more
SEE A LARGER BODY data not only evaluating carnitine’s ef-
fects on acute muscle damage and sore-
OF MORE CONSISTENT ness, but also on longitudinal resistance
training programs that should theoreti-
LITERATURE BEFORE cally benefit from more rapid recovery.
Ideally, future carnitine research would
I GIVE CARNITINE specifically utilize resistance training
as the exercise modality, and recruit
SUPPLEMENTATION healthy, well-trained participants. I’m
also interested in seeing how the neuro-
A SHOT. logical side of this research line evolves
in the years to come; I’m hopeful that
some of the promising observations
from rodent models and clinical popu-
my training program is appropriate and 2) lations will reliably translate to younger,
increasing my intake of fruits, vegetables, healthy individuals, but time will tell.
and their juices containing high amounts
of polyphenols and other phytonutrients.
When it comes to carnitine’s effects on
body composition and performance, out-
comes are pretty inconsistent, and seem
to be least promising in young, healthy
individuals. The neurological effects of
acetyl-L-carnitine are quite fascinating,
but more research is needed to determine
which form(s) of carnitine supplementa-
tion may confer meaningful cognitive or
neurological benefits for young, healthy
individuals with no underlying neurolog-
ical or psychological medical diagnoses.

Next Steps

94
APPLICATION AND TAKEAWAYS
The body of literature looking at carnitine’s effects on soreness and muscle damage
is small, but suggests that beneficial effects are possible. For now, I can’t really judge
you for trying out carnitine supplementation as a strategy for improving recovery from
resistance exercise. If you go that route, it would seem as if just about any form of
L-carnitine would be equivalent to the others. You might look at the carnitine literature
as a whole and be lured in by the possibility of improvements for such a broad range
of outcomes, but it’s important to keep three things in mind: (A) a supplement that im-
proves dysfunction will not necessarily super-charge normal function, (B) inconsistent
findings for many different outcomes don’t sum together to create a large “pooled”
confidence level, and (C) several different small (but practically unnoticeable) effects
don’t sum together to create a worthwhile effect; if you take a supplement that causes
six different virtually imperceptible improvements, that’s going to feel like it did nothing
at all, and none of the effects (alone or in combination) actually justify the use of that
supplement. Personally, based on the evidence available, I’d like to see a larger body
of more consistent literature before I give carnitine supplementation a shot.

95
References
1. Yarizadh H, Shab-Bidar S, Zamani B, Vanani AN, Baharlooi H, Djafarian K. The Effect of
L-Carnitine Supplementation on Exercise-Induced Muscle Damage: A Systematic Review
and Meta-Analysis of Randomized Clinical Trials. J Am Coll Nutr. 2020 Mar 10;1–12.
2. Volek JS, Judelson DA, Silvestre R, Yamamoto LM, Spiering BA, Hatfield DL, et al. Ef-
fects of Carnitine Supplementation on Flow-Mediated Dilation and Vascular Inflamma-
tory Responses to a High-Fat Meal in Healthy Young Adults. Am J Cardiol. 2008 Nov
15;102(10):1413–7.
3. Atalay Guzel N, Erikoglu Orer G, Sezen Bircan F, Coskun Cevher S. Effects of acute L-car-
nitine supplementation on nitric oxide production and oxidative stress after exhaustive exer-
cise in young soccer players. J Sports Med Phys Fitness. 2015 Feb;55(1–2):9–15.
4. Parandak K, Arazi H, Khoshkhahesh F, Nakhostin-Roohi B. The Effect of Two-Week L-Car-
nitine Supplementation on Exercise -Induced Oxidative Stress and Muscle Damage. Asian J
Sports Med. 2014 Jun;5(2):123–8.
5. Kerksick CM, Wilborn CD, Roberts MD, Smith-Ryan A, Kleiner SM, Jäger R, et al. ISSN
exercise & sports nutrition review update: research & recommendations. J Int Soc Sports
Nutr. 2018 Aug 1;15(1):38.
6. Smith WA, Fry AC, Tschume LC, Bloomer RJ. Effect of glycine propionyl-L-carnitine on aero-
bic and anaerobic exercise performance. Int J Sport Nutr Exerc Metab. 2008 Feb;18(1):19–36.
7. Jacobs PL, Goldstein ER. Long-term glycine propionyl-l-carnitine supplemention and paradoxi-
cal effects on repeated anaerobic sprint performance. J Int Soc Sports Nutr. 2010 Oct 28;7(1):35.
8. Koozehchian MS, Daneshfar A, Fallah E, Agha-Alinejad H, Samadi M, Kaviani M, et al. Ef-
fects of nine weeks L-Carnitine supplementation on exercise performance, anaerobic power,
and exercise-induced oxidative stress in resistance-trained males. J Exerc Nutr Biochem.
2018 Dec 31;22(4):7–19.
9. Fielding R, Riede L, Lugo JP, Bellamine A. l-Carnitine Supplementation in Recovery after
Exercise. Nutrients. 2018 Mar 13;10(3):349.
10. Askarpour M, Hadi A, Miraghajani M, Symonds ME, Sheikhi A, Ghaedi E. Beneficial ef-
fects of l-carnitine supplementation for weight management in overweight and obese adults:
An updated systematic review and dose-response meta-analysis of randomized controlled
trials. Pharmacol Res. 2020;151:104554.
11. Gross CJ, Henderson LM, Savaiano DA. Uptake of L-carnitine, D-carnitine and acetyl-L-car-
nitine by isolated guinea-pig enterocytes. Biochim Biophys Acta. 1986 May 29;886(3):425–33.
12. Reuter SE, Evans AM. Carnitine and acylcarnitines: pharmacokinetic, pharmacological and
clinical aspects. Clin Pharmacokinet. 2012 Sep 1;51(9):553–72.

96
Study Reviewed: Varying the Order of Combinations of Single- and Multi-Joint
Exercises Differentially Affects Resistance Training Adaptations. Brandão et al. (2020)

Variety is the Spice of Life: If You Want


Well-Rounded Triceps Growth, You Need
Both Compound and Single-Joint Exercises
BY G RE G NUC KO LS

A recent study examined triceps growth (among other things) in


subjects that only did bench press, only did triceps extensions, or
performed both exercises. To maximize growth of all three heads of the
triceps, a combination of both exercises was necessary.

97
KEY POINTS
1. For 10 weeks, groups of untrained men either trained bench press alone, triceps
extensions alone, bench press followed by triceps extensions, or triceps extensions
followed by bench press. Subjects in the groups performing both bench press
and triceps extensions did the same number of sets of each exercise as the two
groups only performing one exercise, which means they were performing twice
as many sets overall.
2. Bench press did a subpar job of building the middle and long heads of the triceps,
while triceps extensions did a subpar job of building the lateral head of the triceps.
The combination of bench press and triceps extensions caused robust growth in
all heads of the triceps. All three groups performing the bench press experienced
fairly similar pec growth, though growth may have been slightly lower in the group
performing triceps extensions before bench press.
3. Triceps extensions strength gains were fairly similar in all groups, but the group
that only performed triceps extensions tended to have the smallest increase in
bench press strength (unsurprisingly).
4. If you want to maximize the growth of muscles with multiple heads (like the
triceps and quads), it’s probably wise to include both multi-joint and single-joint
exercises in your training program.

I
f you come across a freakishly more single-joint triceps training than
strong bench presser, there’s a good powerlifters do.
chance they’ll have freakishly big The present study compared four
triceps. There are exceptions, of groups of novice lifters who trained for
course, but most world-class benchers 10 weeks. One group only performed the
look like they have a beef tenderloin bench press, one group only performed
hanging off the back of their arm. Nota- lying triceps extensions, one group per-
bly, however, you’ll find very few who formed bench press followed by triceps
have pronounced horseshoe-shaped tri- extensions, and one group performed
ceps, whereas if you go to a high-level triceps extensions followed by bench
bodybuilding show, there will be horse- press. There were a lot of outcomes,
shoe-shaped triceps as far as the eye can and most of them weren’t terribly excit-
see. Why is that? It could partially be ing (they were mostly in line with the
a matter of people self-sorting into the principle of specificity). The most inter-
sports they’re best-suited for, but there esting finding, however, was that bench
could be training considerations as well. press and triceps extensions caused dif-
Namely, bodybuilders tend to do way ferent amounts of growth for each head
98
of the triceps. Bench press caused a lot before bench press, that exercise order
of growth of the lateral head, but did in the two bench press plus triceps ex-
a poor job developing the middle and tensions groups wouldn’t affect triceps
long heads of the triceps, whereas tri- growth, and that the combination of
ceps extensions developed the long and bench press and triceps extensions would
middle heads well, but didn’t do a great lead to more growth of all heads of the
job of developing the lateral head. Thus, triceps than only performing bench press
it appears that a combination of bench or triceps extensions in isolation.
press and triceps extensions causes
more well-rounded triceps growth than
performing either exercise in isolation. Subjects and Methods
This finding is probably generalizable to
other muscles with multiple heads. Subjects
The subjects were 43 healthy young
males who had not performed regular
Purpose and Hypotheses resistance training for at least six months
prior to the study.
Purpose
The stated purposes of this study were Experimental Design
to test the effects of single-joint versus This was a pretty straightforward lon-
multi-joint training and to test the ef- gitudinal training study. Before and af-
fects of exercise order on strength gains ter 10 weeks of training, the researchers
and hypertrophy. More concretely, how- tested the subjects’ 1RM bench press and
ever, the authors were comparing the triceps extension, and performed MRI
effects of training consisting of bench scans of the subjects’ pecs and triceps
press alone, triceps extensions alone, to measure muscle cross-sectional area
bench press followed by triceps exten- (CSA). Subjects were familiarized with
sions, and triceps extensions followed the tests prior to recording their pre-train-
by bench press. ing 1RMs by performing repeated 1RM
tests on different days until their day-to-
Hypotheses day strength variation was ≤5%.
The authors hypothesized that strength The subjects were randomized into
gains would follow the principle of four groups. One group only performed
specificity (that subjects would improve the bench press during each training
strength more on the exercises they actu- session; one group only performed the
ally trained), that less pec growth would triceps extension; one group performed
occur in the group that only trained tri- bench press followed by triceps ex-
ceps and the group that trained triceps
99
throughout the program, since the BP-
TE and TE-BP groups were performing
both exercises. Subjects started with a
load equal to 80% of their pre-training
1RMs and performed all sets to concen-
tric failure. Subjects re-tested their max-
es following the fifth week of training to
adjust their training loads.
Bench press and triceps extension
technique were standardized. All sub-
jects took a bench grip of 200% of bi-
acromial breadth (which is a reasonably
wide grip). For the triceps extensions,
the shoulder angle was held at 90 de-
grees of flexion (arms straight up when
you’re lying on your back), and the sub-
jects went through 90 degrees of elbow
flexion; a metallic bar would stop the
triceps extension bar to inhibit further
elbow flexion during each rep.

Findings
There weren’t statistically significant
tensions (BP-TE); and one group per- differences in strength gains between
formed triceps extensions followed by groups for any of the exercises. Howev-
bench press (TE-BP). All groups trained er, the average increase in bench press
twice per week, and rested for three min- 1RM did tend to be lower in the triceps
utes between sets. During the first four extension-only group than the other
weeks, subjects performed three sets of three groups (+9.9% vs. +22.3-27.1%).
their assigned exercise(s), followed by
The triceps extension-only group had
four sets for the next three weeks, and
a nominal (non-significant) decrease in
five sets for the final three weeks. Thus,
pec CSA, whereas all other groups had
total set volume increased for all groups
a significant increase in pec CSA; the
during the program, but total set volume
changes were significantly different be-
for the triceps was higher for the BP-TE
tween the triceps extension-only group
and TE-BP groups than the bench-on-
and all other groups. The bench-only
ly and triceps extension-only groups
100
bench-only group also tended to have
smaller increases in CSA of the middle
head of the triceps as well (+7.3% vs.
+14.0-16.2%), but the difference wasn’t
statistically significant.
Bench press volume load was signifi-
cantly greater in the bench-only group
than the SJ+MJ group and triceps ex-
tension-only group (obviously). Bench
press volume load also tended to be
greater in the BP-TE group than the TE-
BP group, but the difference wasn’t sta-
tistically significant. Triceps extension
volume load was significantly greater in
the triceps extension-only group and the
TE-BP group than the bench-only group
(obviously) and the BP-TE group.

Interpretation
I wanted to review this study for a cou-
ple reasons:
1. Much of the research we have com-
paring multi-joint and single-joint
group had a nominally smaller increase training uses fairly simplistic mea-
in triceps whole-muscle CSA than all surement techniques, and I suspect-
other groups (+4.8% vs. +9.5-11.5%), ed we were missing some nuance in
but the differences between groups the process.
weren’t statistically significant. 2. This study emphasizes a point about
When splitting out the different heads biarticular muscles generally, and
of the triceps, the triceps extension-on- the triceps specifically, that I’ve been
ly group had a significantly smaller in- making for years, but didn’t yet have
crease in CSA of the lateral head of the enough data to fully support.
triceps than all other groups. For the We’ve reviewed several studies in
long head of the triceps, the increase was MASS comparing multi-joint and sin-
significantly smaller in the bench-on- gle-joint exercises, or the effects of add-
ly group than all other groups. The ing single-joint exercises to a program
101
DEXA (not great), and several others (4,
5) have just measured limb circumfer-
ences (far from ideal). Though the re-
searchers in the present study only ex-
amined one site along the length of each
muscle, they did their due diligence by
splitting the heads of the triceps apart
(fantastic), and they used MRI, which is
the gold standard for assessing changes
in whole-muscle cross-sectional area.
Thus, on measurement quality alone,
this study is a big step up from what we
usually see in this body of research. Ku-
dos to the researchers.
So, how does this research look from
a bird’s eye view? De França et al (6)
found that adding curls and triceps ex-
tensions to a program consisting of vari-
ous pressing and upper body pulling ex-
ercises didn’t significantly increase arm
muscle circumference (estimated based
on changes in arm circumference and
skinfold thicknesses) relative to just per-
forming the multi-joint exercises, though
non-significant differences leaned in fa-
vor of the program that also included
the single-joint exercises (+1.33% vs.
+3.17%). Gentil et al (7) used a similar
experimental design, but didn’t find a
significant or meaningful difference in
composed solely of multi-joint training elbow flexor thickness (+6.5% for only
(one, two, three, four). In most of the multi-joint, versus +7.04% for multi-
studies in that body of literature, the joint plus single-joint). Other studies
measures taken are pretty rough; one have largely had similar results: nonsig-
study (2) used ultrasound at three points nificant-differences leaning in favor of
along the length of the biceps (good), adding single-joint training, but few sta-
but other studies (3) have just exam- tistically significant differences (see Ta-
ined changes in lean mass assessed via ble 3 in this article). In the present study

102
THE CURRENT BODY
OF LITERATURE LEANS
SLIGHTLY IN FAVOR OF
ADDITIONAL SINGLE-
JOINT TRAINING
CAUSING MORE GROWTH
THAN MULTI-JOINT
TRAINING ALONE.

is that single-joint training was simply


added to multi-joint training instead of
replacing multi-joint training. In oth-
er words, they compare something like
four sets of pull-downs to four sets of
(1), there also wasn’t a significant dif- pull-downs plus four sets of curls, rather
ference in whole-muscle triceps growth than two sets of pull-downs plus two sets
between the bench-only group and the of curls. Thus, the results may be driven
BP-TE or TE-BP groups, though again, by the fact that the groups performing
the non-significant differences leaned in both multi-joint and single-joint training
favor of the groups also performing tri- are also training with higher overall vol-
ceps extensions. umes for the target muscle group(s).
Thus, overall, I’d say the current body Finally, I’m glad the researchers re-
of literature leans slightly in favor of ported the growth in each head of the
additional single-joint training causing triceps. If the researchers only reported
more growth than multi-joint training overall triceps growth, we’d come away
alone. One drawback to these studies, with the impression that bench press
however (including the present study),
103
alone may be suboptimal for triceps takes both exercises to maximize growth
growth, and may even be redundant for of all heads of the triceps. This confirms
triceps growth if you’re also perform- the findings of a prior study (8) looking
ing triceps extensions in your training at quad growth. When equating for set
(since the group performing only triceps volume, squats alone caused a similar
extensions had similar triceps growth as amount of growth when compared to a
the two groups performing both triceps variety of lower body exercises (squats,
extensions and bench press). However, deadlifts, leg press, and lunges), but the
results for each head of the triceps tell a groups performing a wider variety of
very different story. Bench press seems exercises had more even growth across
to be awesome for growing the lateral all four heads of the quads. For muscle
head of the triceps, whereas triceps ex- with multiple heads – especially when
tensions seem to be awesome for grow- some heads only cross one joint and
ing the middle and long heads; and it other heads cross two joints – a variety
of exercises will probably lead to more
well-rounded growth than a single ex-
ercise (I’m primarily thinking about the
quads, triceps, and hamstrings, but this
THE RESULTS MAY BE may apply to the biceps as well). In a
vacuum, triceps extensions are proba-
DRIVEN BY THE FACT bly a better triceps exercise than bench
press, but since triceps extensions and
THAT THE GROUPS bench press target different heads of the
triceps, the combination of both exercis-
PERFORMING BOTH es gives you more bang for your buck
than just performing more triceps exten-
MULTI-JOINT AND SINGLE- sions would.

JOINT TRAINING ARE ALSO In a more general sense, I think it’s


worth emphasizing that single-joint
TRAINING WITH HIGHER exercises (or very specific multi-joint
exercises) are probably necessary to
OVERALL VOLUMES maximize growth of biarticular mus-
cles – muscles that cross two joints. For
FOR THE TARGET example, three of your four quads are
pure knee extensors because they only
MUSCLE GROUP(S). cross the knee joint, whereas the rectus
femoris is both a knee extensor and a
hip flexor since it crosses both the knee

104
is much higher when performing knee
extensions (11), which makes sense;
SINGLE-JOINT EXERCISES no hip extension is taking place when
you’re performing knee extensions.
ARE PROBABLY Thus, a diet of pure multi-joint train-
ing will probably leave your biarticular
NECESSARY TO MAXIMIZE muscles somewhat underdeveloped; you
need some single-joint triceps work for
GROWTH OF BIARTICULAR the long head of your triceps, some pure
knee flexion and hinge work for your
MUSCLES – MUSCLES hamstrings, and some knee extensions
or sissy squats for your rectus femo-
THAT CROSS TWO JOINTS. ris. As I mentioned, however, there are
some multi-joint exercises that would fit
the bill for biarticular muscle develop-
ment. For example, a combination pull-
and hip joints. The same is true of the over-triceps extension would train both
long head of your triceps; while the oth- functions of the long head of the triceps
er two heads are pure elbow extensors, (shoulder extension and elbow exten-
the long head is also a shoulder exten- sion), and so it’s probably an excellent
sor. For most multi-joint exercises, your exercise for developing the long head of
nervous system preferentially recruits your triceps. Glute-ham raises, which si-
monoarticular muscles (muscles that multaneously train knee flexion and hip
only cross one joint) because those mus- extension, would also fit the bill.
cles tend to be more efficient (9). If you Another topic addressed in the present
need to stand up from a squat, and you study is the effects of exercise order. The
can choose between your vasti muscles BP-TE and TE-BP groups had broadly
or your rectus femoris to bear the brunt similar results across all measures except
of the knee extension workload, the vas- for pec growth, which leaned in favor of
ti are the better option; you also need the BP-TE group. This is broadly in line
to extend your hips in a squat, and your with what we’ve seen in the literature
vasti don’t fight against your hip exten- up to this point. Exercise order doesn’t
sors, whereas your rectus femoris does. seem to make an absolute night-and-day
And that’s what we see when assessing difference, but when in doubt, you’re
muscle activation in the squat – really better off putting the exercises that are
high activation for all vasti muscles, but most important to you earlier in your
lower activation for your rectus femoris workout. For a more in-depth treatment
(10). However, rectus femoris activation of the topic, check out the meta-analysis

105
APPLICATION AND TAKEAWAYS
If you want to maximize triceps growth, neither bench press nor triceps extensions are
sufficient in isolation. Bench press does a better job of training the lateral head, while
triceps extensions do a better job of training the middle and long heads. As a general
rule, using a combination of exercises – both multi-joint and single-joint – is probably
ideal for maximizing muscle growth, especially for muscles that have both monoartic-
ular and biarticular heads.

that Mike reviewed last month. ticular body of literature are even less
Finally, I’d like to end on a word of generalizable than individual studies in
caution. Studies intend to test broad con- most other bodies of literature within
cepts, but ultimately, they’re constrained the resistance training research canon.
by the discrete design choices the re-
searchers made. In the present study,
Next Steps
bench press was compared against lying We need more studies in general, but
triceps extensions, range of motion was I’d love to see studies comparing sin-
constrained to 90 degrees of elbow flex- gle-joint exercises against multi-joint
ion for the triceps extensions, and the exercises that are theorized to be ideal
bench press was performed with a wide for developing a certain muscle; for ex-
grip. Would the triceps extensions lead ample, comparing triceps extensions to
to more muscle growth if they were per- close grip bench, or flyes to dumbbell
formed through a longer range of mo- presses, rather than comparing the ef-
tion? Would the results be different if the fects of wide-grip bench and triceps ex-
researchers used push-downs instead of tensions on triceps growth. I’d also love
lying triceps extensions? Or close-grip for more studies to examine the indi-
bench instead of wide-grip bench? Or vidual heads of muscles like the triceps,
incline press instead of bench press? All quads, and hamstrings that have both
of those tweaks may have dramatically mono-articular and bi-articular heads.
changed the outcomes of the study, or
they may have only made trivial dif-
ferences. We’ll need more research to
fully explore the concept of multi-joint
versus single-joint training, and we can
never extrapolate too much from a sin-
gle study, but I think studies in this par-

106
References

1. Brandão L, de Salles Painelli V, Lasevicius T, Silva-Batista C, Brendon H, Schoenfeld BJ,


Aihara AY, Cardoso FN, de Almeida Peres B, Teixeira EL. Varying the Order of Combina-
tions of Single- and Multi-Joint Exercises Differentially Affects Resistance Training Adap-
tations. J Strength Cond Res. 2020 Mar 2.
2. Mannarino P, Matta T, Lima J, Simão R, Freitas de Salles B. Single-Joint Exercise Results in
Higher Hypertrophy of Elbow Flexors Than Multijoint Exercise. J Strength Cond Res. 2019
Jul 1.
3. Paoli A, Gentil P, Moro T, Marcolin G, Bianco A. Resistance Training with Single vs. Multi-
joint Exercises at Equal Total Load Volume: Effects on Body Composition, Cardiorespirato-
ry Fitness, and Muscle Strength. Front Physiol. 2017 Dec 22;8:1105.
4. Barbalho M, Coswig V, Raiol R, Fisher J, Steele J, Bianco A, Gentil P. Single joint exercises
do not provide benefits in performance and anthropometric changes in recreational body-
builders. Eur J Sport Sci. 2020 Feb;20(1):72-79.
5. Barbalho M, Gentil P, Raiol R, Fisher J, Steele J, Coswig V. Influence of Adding Single-Joint
Exercise to a Multijoint Resistance Training Program in Untrained Young Women. J Strength
Cond Res. 2018 May 15.
6. de França HS, Branco PA, Guedes Junior DP, Gentil P, Steele J, Teixeira CV. The effects of
adding single-joint exercises to a multi-joint exercise resistance training program on upper body
muscle strength and size in trained men. Appl Physiol Nutr Metab. 2015 Aug;40(8):822-6.
7. Gentil P, Soares SR, Pereira MC, da Cunha RR, Martorelli SS, Martorelli AS, Bottaro M.
Effect of adding single-joint exercises to a multi-joint exercise resistance-training pro-
gram on strength and hypertrophy in untrained subjects. Appl Physiol Nutr Metab. 2013
Mar;38(3):341-4.
8. Fonseca RM, Roschel H, Tricoli V, de Souza EO, Wilson JM, Laurentino GC, Aihara AY, de
Souza Leão AR, Ugrinowitsch C. Changes in exercises are more effective than in loading
schemes to improve muscle strength. J Strength Cond Res. 2014 Nov;28(11):3085-92.
9. Bryanton MA, Carey JP, Kennedy MD, Chiu LZ. Quadriceps effort during squat exercise
depends on hip extensor muscle strategy. Sports Biomech. 2015 Mar;14(1):122-38.
10. Escamilla RF, Fleisig GS, Zheng N, Lander JE, Barrentine SW, Andrews JR, Bergemann
BW, Moorman CT 3rd. Effects of technique variations on knee biomechanics during the
squat and leg press. Med Sci Sports Exerc. 2001 Sep;33(9):1552-66.
11. Salzman A, Torburn L, Perry J. Contribution of rectus femoris and vasti to knee exten-
sion. An electromyographic study. Clin Orthop Relat Res. 1993 May;(290):236-43.

107
VIDEO: Bodyweight and Home
Training Progressions
BY MIC HAE L C . ZO URD O S

In these uncertain times, you may not have gym access. This video first presents
scientific data demonstrating that bodyweight training can help to preserve muscle
mass and strength. Second, this video presents a nine-week at-home training
program along with specific examples of household items that can be used. Further,
the video demonstrates various examples of how progressive overload can be
achieved with this approach.
Click to watch Michael's presentation.

108
Program
1. MASS Bodyweight Home Example Program

References
1. Kikuchi N, Nakazato K. Low-load bench press and push-up induce similar muscle hypertro-
phy and strength gain. Journal of Exercise Science & Fitness. 2017 Jun 1;15(1):37-42.
2. Calatayud J, Borreani S, Colado JC, Martin F, Tella V, Andersen LL. Bench press and push-
up at comparable levels of muscle activity results in similar strength gains. The Journal of
Strength & Conditioning Research. 2015 Jan 1;29(1):246-53.
3. Sánchez-Moreno M, Cornejo-Daza PJ, González-Badillo JJ, Pareja-Blanco F. Effects of Ve-
locity Loss During Body Mass Prone-Grip Pull-up Training on Strength and Endurance Per-
formance. The Journal of Strength & Conditioning Research. 2020 Apr 1;34(4):911-7.
4. Staron RS, Leonardi MJ, Karapondo DL, Malicky ES, Falkel JE, Hagerman FC, Hikida RS.
Strength and skeletal muscle adaptations in heavy-resistance-trained women after detraining
and retraining. Journal of Applied Physiology. 1991 Feb 1;70(2):631-40.

109
VIDEO: Nutrition on Lockdown
BY E RI C HE LMS

Obviously it’s less than ideal to be stuck at home, and you or your clients may be
struggling with eating due to boredom, depression, or emotional reasons. However,
there are other aspects of being stuck at home that may be influencing your food
behaviors. Likewise, there may be some silver lining to this change in environment
which could be used to establish better nutrition habits that outlast lockdown.
Click to watch Eric's presentation.

110
References
1. Hopkins M, Blundell JE. Energy balance, body composition, sedentariness and appetite reg-
ulation: pathways to obesity. Clinical Science. 2016 Sep 1;130(18):1615-28.
2. McCrory MA, Harbaugh AG, Appeadu S, Roberts SB. Fast-food offerings in the United
States in 1986, 1991, and 2016 show large increases in food variety, portion size, dietary
energy, and selected micronutrients. Journal of the Academy of Nutrition and Dietetics. 2019
Jun 1;119(6):923-33.
3. An R. Fast-food and full-service restaurant consumption and daily energy and nutrient in-
takes in US adults. European journal of clinical nutrition. 2016 Jan;70(1):97.
4. Bhutani S, Schoeller DA, Walsh MC, McWilliams C. Frequency of eating out at both fast-food
and sit-down restaurants was associated with high body mass index in non-large metropolitan
communities in Midwest. American Journal of Health Promotion. 2018 Jan;32(1):75-83.
5. Seguin RA, Aggarwal A, Vermeylen F, Drewnowski A. Consumption frequency of foods
away from home linked with higher body mass index and lower fruit and vegetable intake
among adults: a cross-sectional study. Journal of environmental and public health. 2016.
6. Wansink B, Johnson KA. The clean plate club: About 92% of self-served food is eaten. In-
ternational Journal of Obesity. 2015 Feb;39(2):371-4.
7. Rolls BJ, Morris EL, Roe LS. Portion size of food affects energy intake in normal-weight
and overweight men and women. The American journal of clinical nutrition. 2002 Dec
1;76(6):1207-13.
8. Steenhuis I, Poelman M. Portion Size: Latest Developments and Interventions. Current Obe-
sity Reports. 2017;6(1):10.

111
Just Missed the Cut
Every month, we consider hundreds of new papers, and they can’t all be included in
MASS. Therefore, we’re happy to share a few pieces of research that just missed the
cut. It’s our hope that with the knowledge gained from reading MASS, along with our
interpreting research guide, you’ll be able to tackle these on your own.

1. Bartolomei et al. A Comparison Between Total Body and Split Routine Resistance
Training Programs in Trained Men
2. Cornish et al. A focused review of myokines as a potential contributor to muscle
hypertrophy from resistance-based exercise
3. Valenzuela et al. Acute Ketone Supplementation and Exercise Performance: A
Systematic Review and Meta-Analysis of Randomized Controlled Trials
4. Angleri et al. Are resistance training systems necessary to avoid a stagnation and
maximize the gains muscle strength and hypertrophy?
5. Cheng et al. Carbohydrates do not accelerate force recovery after glycogen‐
depleting followed by high‐intensity exercise in humans
6. Rial-Vázquez et al. Cluster vs. traditional training programmes: changes in the
force–velocity relationship
7. Stahn et al. Combined protein and calcium β-hydroxy-β-methylbutyrate induced
gains in leg fat free mass: a double-blinded, placebo-controlled study
8. Motameni et al. Comparing the effects of resistance exercise type on serum levels
of oxidative stress and muscle damage markers in resistance-trained women
9. Soriano et al. Comparison of 1-Repetition-Maximum Performance Across 3
Weightlifting Overhead Pressing Exercises and Sport Groups.
10. Toth et al. Does mental practice still enhance performance? A 24 Year follow-up
and meta-analytic replication and extension
11. Pearson et al. Effect of Competition Frequency on Strength Performance of
Powerlifting Athletes
12. Nogueira et al. Effects of local cryotherapy for recovery of delayed onset muscle
soreness and strength following exercise-induced muscle damage: systematic
review and meta-analysis
13. Chaves et al. Effects of resistance training with controlled versus self-selected
repetition duration on muscle mass and strength in untrained men
14. Vachon et al. Effects of tapering on neuromuscular and metabolic fitness in team
sports: a systematic review and meta-analysis

112
15. Martín-Fuentes et al. Electromyographic activity in deadlift exercise and its variants.
A systematic review
16. Stuntz et al. Exert more and feel better, not worse?: Examining links among changes
in exertion, feelings of accomplishment, and feeling states
17. Parry et al. Human Skeletal Muscle Mitochondrial Adaptations Following Resistance
Exercise Training
18. Teng et al. Impact of coconut oil consumption on cardiovascular health: a systematic
review and meta-analysis
19. Wilk et al. Impact of Duration of Eccentric Movement in the One-Repetition Maximum
Test Result in the Bench Press among Women
20. Pérez-Castilla et al. Influence of Grip Width and Anthropometric Characteristics on
the Bench-Press Load-Velocity Relationship.
21. Romero-Parra et al. Influence of the Menstrual Cycle on Blood Markers of Muscle
Damage and Inflammation Following Eccentric Exercise.
22. Gomes et al. Is there a relationship between back squat depth, ankle flexibility, and
Achilles tendon stiffness?
23. Hussain et al. Muscle Fatigue in the Three Heads of Triceps Brachii During Intensity
and Speed Variations of Triceps Push-Down Exercise
24. Reidy. Muscle or Nothing! Where Is the Excess Protein Going in Men with High
Protein Intakes Engaged in Strength Training?
25. Carvalho and Barroso. Muscle thickness and strength adaptations in dominant and
non-dominant upper limbs
26. Joffe and Tallent. Neuromuscular predictors of competition performance in advanced
international female weightlifters: a cross-sectional and longitudinal analysis
27. Grandou et al. Overtraining in Resistance Exercise: An Exploratory Systematic
Review and Methodological Appraisal of the Literature
28. Alizadeh et al. Push-Ups vs. Bench Press Differences in Repetitions and Muscle
Activation between Sexes
29. Snijders et al. The concept of skeletal muscle memory: Evidence from animal and
human studies
30. Akinci et al. The effectiveness of three different recovery methods on blood lactate,
acute muscle performance, and delayed-onset muscle soreness: a randomized
comparative study
31. Ertel et al. The effects of training status and exercise intensity on exercise-induced
muscle damage
32. Androulakis-Korakakis et al. The Minimum Effective Training Dose Required to
Increase 1RM Strength in Resistance-Trained Men: A Systematic Review and Meta-
Analysis

113
Thanks for
reading MASS.
The next issue will be released to
subscribers on June 1, 2020.

Graphics and layout by Kat Whitfield

114

You might also like