Does Muscle Growth Increase Your Potential for Strength Gains?

[01dragonslayer]

When the presently reviewed study (1) was published, my inbox was inundated with people telling me I needed to review it for MASS, since I’m on the record arguing that muscle growth does, in fact, contribute to strength gains (2). I always aim to please, so here we are. I’m pretty sure this was the longest MASS article to date, and as I was writing it, my computer crashed no fewer than a dozen times as I tried to wrangle with a spreadsheet with nearly 100,000 rows. I hope all of you that asked for this review are pleased with yourselves.

However, I do think this was a really neat study, and I appreciate the researchers’ experimental approach to the question. Briefly, subjects completed a two-part training protocol. During the first part of the protocol, one arm did hypertrophy training, while the other arm merely practiced 1RMs. During the second part of the protocol, both arms practiced 1RMs. In theory, if hypertrophy increases one’s potential for strength gains, you should expect the arms that initially did hypertrophy training to experience larger gains in strength than the 1RM practice arms during the second phase of the study. However, there weren’t meaningful differences in strength gains between conditions during the second part of the study. Is this study the nail in the coffin? Does this mean hypertrophy really doesn’t contribute to strength gains? In this article I’ll argue that no, it doesn’t, and we still have solid reasons for believing that muscle growth does, in fact, increase one’s “strength potential.”

Purpose and Hypotheses​

Purpose​

The purposes of this study were to:

1. Investigate the effects of hypertrophy training versus 1RM practice on biceps growth and strength gains.
2. Investigate whether previous hypertrophy promoted larger subsequent strength gains following a period of 1RM practice.

Hypotheses​

No hypotheses were directly stated. However, based on familiarity with the researchers, I presume they anticipated that hypertrophy training and 1RM practice would lead to similar strength gains, and that hypertrophy training would cause more muscle growth. Furthermore, I suspect that they anticipated that previous hypertrophy would not promote larger strength gains following a period of 1RM practice.

Subjects and Methods​

Subjects​

The presently reviewed paper presents the results of two separate experiments. The first experiment included 25 individuals who all had at least 6 months of experience with biceps training (13 males, 12 females). The second experiment included 18 of the subjects from the first experiment (10 males, 8 females).

Experimental Design​

As mentioned, this paper details the results of two experiments, but the experiments ran back-to-back, so I’ll describe the full experimental protocol of both studies together.

Subjects initially completed eight weeks of biceps training, using a within-subject, unilateral design. Each subject’s arms were randomized to two different training conditions. One arm performed hypertrophy training (four sets of dumbbell curls to failure with an 8-12RM load), and one arm performed 1RM practice (they were allowed up to five attempts to establish a dumbbell curl 1RM). They performed all of their training with their back against a wall, in order to keep their form strict. The subjects trained twice per week for eight weeks during this first phase.

During the second phase of the study, 18 of the original 25 subjects completed an additional four weeks of training. During this phase, they worked up to a 1RM with both arms on each training day, followed by two sets to failure with an 8-12RM load.

Testing took place before the start of the first eight weeks of training, 3-5 days after the first eight weeks of training, and 3-5 days after the last four weeks of training. Four outcome variables were assessed: unilateral biceps curl 1RM, unilateral isometric elbow flexion torque (at a 90° elbow angle), unilateral biceps strength endurance (reps to failure with 40% of the subjects’ baseline unilateral biceps curl 1RM), and elbow flexor muscle thickness (assessed via ultrasound, at 50%, 60%, and 70% of the distance between the shoulder and the elbow). Strength endurance was not assessed after the last four weeks of training.

It’s worth noting that subjects were allowed to continue their habitual resistance training during the duration of this study. They were asked to refrain from direct biceps training (any type of curl), but they were still allowed to perform indirect biceps training (like rows, pull-ups, or pull-downs). I think this is a fairly notable confounder.

The researchers used Bayesian statistics to analyze their results. I’m only mentioning that because the statistical evidence in the “Findings” section may be presented in a way you’re not familiar with. With Bayesian statistics, you’re interested in the relative level of support of a hypothesis, rather than the relative lack of support for the null hypothesis. For example, if you’re comparing muscle thicknesses in two groups, you could have three different hypotheses: 1) changes in muscle thickness did not differ between groups, 2) changes in muscle thickness are greater in group 1 than group 2, and 3) changes in muscle thickness were greater in group 2 and group 1. Based on your results, you could have some level of support for all three hypotheses. If results leaned in favor of the second hypothesis (larger gains in group 1 than group 2), but the difference wasn’t particularly large, you could wind up with results along the lines of:

Support for hypothesis 1 (no difference between groups ) = 0.18

Support for hypothesis 2 (group 1 > group 2) = 0.80

Support for hypothesis 3 (group 2 < group 1) = 0.02

This could be interpreted as probable support for hypothesis 2 – you’re 80% sure that gains were larger in group 1 than group 2, but there’s a 18% chance that there was no real difference. However, it’s quite unlikely (2% probability) that group 2 achieved better results.

If, on the other hand, gains were WAY larger in group 1 than group 2, you could wind up with results along the lines of:

Support for hypothesis 1 (no difference between groups ) = 0.001

Support for hypothesis 2 (group 1 > group 2) = 0.999

Support for hypothesis 3 (group 2 < group 1) = 0.000

In this scenario, you have very strong support for hypothesis 2 – you’re 99.9% sure that gains were larger in group 1 than group 2.

In the present study, sometimes two hypotheses were tested (no difference, or larger gains in one condition versus the other), and sometimes three hypotheses were tested (no difference, larger gains with hypertrophy training, and larger gains with 1RM training). For the purposes of the “Findings” section, I’ll make statements along the lines of “muscle thickness increased to a greater degree in the hypertrophy training condition (probability = 0.82).” You can interpret that as “the hypothesis that hypertrophy training would produce larger increases in muscle thickness than 1RM training was most strongly supported by the results (posterior probability for this hypothesis = 0.82).” Also note, probabilities weren’t reported for all comparisons, so if I don’t report the largest probability, it’s probably because it wasn’t stated.

Findings​

During the first experiment, unilateral 1RM biceps curl strength and maximal isometric elbow flexion torque increased similarly in both conditions, while larger changes in strength endurance (probability > 0.9999) and muscle thicknesses occurred in the hypertrophy training condition.

The 18 subjects who completed both experiments experienced changes in muscle thicknesses and strength measures during the first study that were similar to the full 25-person cohort of the first study. Muscle thicknesses increased to a greater degree for arms in the hypertrophy training condition (probabilities = 0.773-0.999), while 1RM biceps curl strength (probability = 0.886) and maximal isometric torque (probability 0.858) increased to a similar degree in both conditions.


During the second experiment, muscle thicknesses changed to a similar extent (very little) in the arms that had previously completed hypertrophy training, and the arms that had previously completed 1RM training (probability = 0.646-0.677). Evidence also supported the null hypothesis (no difference between conditions) for changes in strength. For 1RM strength, the probability of no difference between conditions was 0.503, while the probability that strength gains were larger in the group previously completing hypertrophy training was 0.496. For isometric strength, the probability of no difference between conditions was 0.517, while the probability that strength gains were larger in the group previously completing hypertrophy training was 0.482.

---

Interpretation​

I’m forgoing a “criticisms and statistical musings” section for this article, because most of my discussion of this article will be statistical. I’m primarily reviewing this article based on popular demand. In 2019, I was involved in a written debate on this topic (does muscle growth contribute to strength gains?) with some of the authors of the presently reviewed study. Our side (headed up by Dr. Chris Taber) argued that muscle hypertrophy does contribute to strength gains, while the other side (headed by up Dr. Jeremy Loenneke) contended that hypertrophy does not contribute to strength gains (2, 3). Since I was involved in that back-and-forth, quite a few MASS readers requested that I review this study, and I was more than happy to comply.

In our 2019 back-and-forth, our side contended that there are multiple factors contributing to muscular strength, and muscle size is merely one such factor. Based on the influence of other factors (neural factors, technique, changes in connective tissue, etc.), changes in muscle size don’t necessarily increase strength on a 1:1 basis; however, we contended that a larger muscle has the potential to be a stronger muscle. In other words, if you increase your biceps thickness, your 1RM biceps curl may not automatically increase (and, similarly, you may be able to increase your 1RM biceps curl without increasing the thickness of your biceps); however, if you absolutely maxed the strength of your biceps at their current thickness, your biceps strength would be lower than it could be if you maxed out the strength of your biceps after they were considerably thicker. We also argued that the other factors contributing to muscle strength (everything beyond changes in muscle size) change to a greater degree in untrained lifters than trained lifters, overpowering the gains in strength that can primarily be attributed to hypertrophy. Therefore, we argued that the impact of hypertrophy on strength development would be easier to measure and establish in trained subjects than untrained subjects.

With this in mind, I first want to applaud Dr. Buckner and colleagues for conducting the presently reviewed study. Previous studies in the area either used untrained subjects, or didn’t use experimental models that would be adequate to determine whether larger muscles truly had greater strength potential than smaller muscles (they mostly looked at whether, in the short-to-moderate term, hypertrophy-style training increases strength to a greater extent than 1RM practice; 4, 5). The subjects in the presently reviewed study did have at least a bit of prior training experience, and the experimental model the researchers used was far more appropriate for addressing the “larger muscles have more strength potential” argument.

In short, here’s the logic of the experimental model:

One set of arms does hypertrophy training, which should increase both strength and muscle thickness.

Another set of arms does 1RM practice, which should increase strength, but not muscle thickness.

After enough training time has elapsed for these divergent adaptations to occur, the arms that had been doing 1RM practice should be closer to their strength potential (given their current biceps thickness), and the arms that had been doing hypertrophy training should be further from their strength potential (at their post-training biceps thickness), due to increases in biceps thickness and smaller changes in the “other factors” that contribute to maximal strength. Therefore, following a period of 1RM practice in both arms, you should expect the arms previously doing hypertrophy training to experience a larger increase in strength than the arms previously doing 1RM practice, if you assume that hypertrophy increases a muscle’s potential strength. In other words, if the arms previously doing hypertrophy training do gain more strength during this period of 1RM practice, that would be evidence that a larger muscle has the potential to be a stronger muscle; conversely, if both sets of arms increase strength to a similar degree following this period of 1RM practice, that would be evidence suggesting that larger muscles actually don’t have the potential to be stronger muscles.

Here’s an illustration to make this concept more tangible, assuming that increases in muscle size do increase strength potential. Let’s say you have two people who are identical in every way. At baseline, their biceps are 3cm thick, and they can DB curl 15kg. We can get an idea of how strong their biceps are, relative to their size, by dividing their DB curl 1RM by their biceps thickness: 15 ÷ 3 = 5kg per cm (in actuality, muscle cross-sectional area would be a slightly better size measure than muscle thickness for this purpose, but I’m using thickness in this example because thickness was used in the present study). Let’s assume that, when all non-hypertrophic factors are maximized, biceps are capable of curling 7kg per cm of thickness.

One person does hypertrophy training for a few months, and they increase their biceps thickness to 3.6cm. Because they weren’t training to maximize strength, their biceps curl increases proportionally to their muscle thickness: a 20% increase, for a new 1RM of 18kg. The other person does 1RM practice for a few months, and they also increase their curl 1RM to 18kg. However, they didn’t experience an increase in muscle thickness, so now they’re capable of curling 18 ÷ 3 = 6kg per cm of muscle thickness.

Then, both people do 1RM practice for a few months, until they wring every bit of strength out of their biceps, achieving a DB curl 1RM equal to 7kg per cm of muscle thickness. During this time period, neither person experiences an increase or decrease in biceps thickness. After this period of 1RM practice, the person previously doing hypertrophy training now curls 3.6cm × 7 kg per cm = 25.2kg, while the person who’d been doing 1RM practice the whole time now curls 3cm × 7 kg per cm = 21kg. During the final period of 1RM practice, the subject who previously did hypertrophy training increased their 1RM biceps curl by 7.2kg, while the subject who did 1RM practice the whole time only increased their 1RM biceps curl by 3kg. In total, the subject that started with hypertrophy training increased their 1RM by 10.2kg, while the subject that did 1RM practice the whole time increased their 1RM by 6kg. That would be a pretty striking difference, which would strongly suggest that hypertrophy increases the potential for strength gains. However, that’s not what was observed in the present study. The difference in 1RM strength gains between conditions during the second period of the study was just 0.37kg (with the 95% credible interval crossing zero). So, where did this illustration deviate from reality? And what does all of this mean about whether hypertrophy contributes to strength gains?

The first key difference between my illustration and the actual data in the presently reviewed study is the total amount of hypertrophy that took place. As anticipated, very little biceps growth occurred in the arms doing 1RM practice for the entire study (using a composite measure of biceps thicknesses at 50%, 60%, and 70% of humerus length, average biceps thickness only increased by a total of 0.63mm, on average). However, the arms that started with hypertrophy training didn’t grow very much either. Biceps thickness increased by 1.7mm during the period of hypertrophy training, and by an additional 0.47mm during the period where both sets of arms were doing 1RM practice. As a result, biceps thickness only increased by 2.2mm over the course of the entire study (about 7.4%, on average). Thus, even if we assume that hypertrophy increases strength potential, we should not anticipate that strength potential would have increased by very much in the group initially doing hypertrophy training, because the subjects simply didn’t grow very much.

The second key difference between my illustration and the actual data relates to the changes in 1RM strength per unit of muscle thickness. In my example, the capacity for change was fairly large (from 5kg per cm to 7kg per cm – a 40% increase), the person doing hypertrophy training did not increase their biceps curl 1RM strength per unit of biceps thickness during their period of hypertrophy training, and both theoretical individuals increased their biceps curl 1RM strength per unit of biceps thickness to a considerable degree during the final period of 1RM practice (from 5 to 7 kg per cm in the hypertrophy subject, and from 6 to 7 kg per cm in the subject who did 1RM practice the whole time). In reality, the arms doing hypertrophy training increased their biceps curl strength per unit of biceps thickness during the period of hypertrophy training by about 0.4 kg per cm (from about 5.7 to 6.1 kg per cm), which was a bit less than the arms doing 1RM practice (5.5 to 6.2 kg per cm), but it was still a notable increase. Furthermore, neither set of arms experienced much of an increase in strength per unit of muscle thickness during the second part of the study (when both sets of arms were doing 1RM practice) – the increase was only about 0.2kg per cm in both conditions. In total, the relative increase in 1RM strength per centimeter of muscle thickness was about 11.3% in the arms initially doing hypertrophy training, and about 16.0% in the arms that did 1RM practice for the entire study.

When you combine those first two deviations between the illustration and reality, the third deviation is predictable: the subjects in the present study simply didn’t get that much stronger. The arms doing 1RM practice the whole time increased 1RM strength by an average of 3.1kg over the entire course of the study, while the arms that started with hypertrophy training increased 1RM strength by an average of 3.3kg (compared to 6kg and 10.2kg in the illustration).

So, you may be wondering why I used the illustration in the first place, if it doesn’t comport particularly well with reality. I used it for two reasons. First, so everyone would be on the same page about why the experimental design was appropriate for answering the question at hand, and second, to illustrate a point about magnitudes. In reality, all of the results of this study were in line with what one would expect if hypertrophy does increase strength potential; the magnitudes of changes were all just very small. The arms that started with hypertrophy training grew more (but neither group grew by very much), and they gained more strength during the final four weeks of 1RM practice (but neither condition increased strength by very much). In fact, the researchers’ own Bayesian analysis provides some degree of support for hypertrophy increasing “strength potential.” The posterior probabilities for changes in strength during the final four weeks of 1RM practice basically represent a coin flip – about 50.3% probability in favor of the null hypothesis (prior hypertrophy training did not increase “strength potential”) and about 49.6% probability in favor of the alternate hypothesis (prior hypertrophy training did increase “strength potential”).

Let’s tweak the illustration above, using the actual data from the study (muscle thickness can increase from 2.97 to 3.19cm instead of 3 to 3.6cm, strength per unit of muscle thickness can increase from 5.5 to 6.4kg per cm, instead of 5 to 7kg per cm), while still making idealized assumptions (strength per unit of thickness won’t increase from hypertrophy training, no hypertrophy will occur when doing 1RM practice, and both people will wind up with the same relative increase in strength per unit of thickness). The person starting with hypertrophy training would increase their 1RM from 2.97cm × 5.5kg per cm = 16.3kg to 3.19 × 5.5 = 17.5kg following hypertrophy training, and then to 3.19 × 6.4 = 20.5kg following 1RM practice, while the person who did 1RM practice the whole time would increase their 1RM from 16.3 to 2.97 × 6.4 = 19kg over the entire course of the study, for a net difference between groups of just 1.5kg. In other words, even if we explicitly assume that hypertrophy is directly and causally linked with strength gains, and that hypertrophy training doesn’t contribute to strength development in any way other than via the actual hypertrophy it induces, the overall magnitude of improvements observed in this study were so small that the difference in strength gains that could be attributed to differences in hypertrophy still would have been quite small. Just to illustrate this point further, composite biceps thickness increased by only 2.2mm in the hypertrophy group; reliability studies suggest that the limits of agreement for ultrasound biceps thickness measurements (test-retest and interrater reliability) is approximately 3mm (6, 7). Quite simply, it’s hard to know if hypertrophy increases a muscle’s strength potential if you don’t observe much hypertrophy in the first place. This reminds me a bit of a human study designed to see whether myonuclear addition from resistance training would aid the re-growth of muscle following a period of de-training (8), inspired by the design of this rodent study by Egner and colleagues (9). The researchers designed a good study and executed it well, but they had a problem: the subjects simply didn’t accrue more myonuclei during their initial period of resistance training, so the researchers couldn’t really test whether myonuclear addition aided muscle regrowth. There’s a similar issue here (though not quite as extreme) – it’s hard to know if hypertrophy increases strength potential, unless a pretty fair amount of hypertrophy actually takes place (unless you either assume “strength potential” and “hypertrophy” scale perfectly on a 1:1 basis and your measurements are outrageously accurate and precise, or you have an enormous sample size and insane statistical power).

In short, I think the researchers used a very interesting experimental model, but our ability to take much away from this study was hamstrung by four major factors:

1. Very little muscle growth actually occurred.
2. The hypertrophy training still increased the subjects’ 1RM strength per unit of biceps thickness. As such, the conditions didn’t produce truly divergent responses during the first phase of the study. While the hypertrophy condition did produce more hypertrophy, both conditions seem to have improved the “other” factors that contribute to maximal strength (neural adaptations, the skill of maxing, etc.).
3. The second phase of the study likely didn’t last long enough to truly ascertain whether hypertrophy increased “strength potential.” As far as I can tell, there was no procedure in place to ensure that the subjects actually attained their “strength potential” at their given level of muscle mass by the end of the study. Arms in the 1RM practice condition seemed to still be getting stronger after 12 straight weeks of 1RM practice. How can we know that the arms in the hypertrophy condition had truly reached their “strength potential” after just four weeks of 1RM practice?
4. The subjects were still allowed to perform their normal resistance training, with the exception of direct biceps training. However, we know that pull-downs and rows can increase biceps strength and thickness (10, 11; there’s every reason to suspect that pull-ups also increase biceps strength and thickness. I just don’t know of a citation to back up that claim). That’s a major confounding factor.

I think the study could be improved with a few simple (though not easy) tweaks:

1. Let the first phase of the study run for a longer period of time, so that more total hypertrophy can occur in the hypertrophy condition. I recognize this would be a substantial burden (the 12-week total length of the present study was probably chosen to fit neatly within a semester), but I really don’t see any way around it when using trained subjects.
2. For the first phase of the study, use low-load training (30-50% of 1RM) to induce hypertrophy in the hypertrophy condition, instead of moderate-load (8-12RM) training. We know that low-load training can be just as effective for promoting hypertrophy (12), and it shouldn’t cause the substantial increases in strength per unit of size during the period of the initial period of the study, when the goal is to induce divergent adaptations (primarily “neural” adaptations with minimal hypertrophy in the 1RM practice condition, and primarily hypertrophy with minimal “neural” adaptations in the hypertrophy condition).
3. Use some sort of objective cut-off for the second phase of the study to establish that subjects have truly reached their “strength potential” at their current level of muscle size. For example, they could do 1RM practice twice per week, until they fail to achieve a new 1RM for three weeks in a row.
4. Use a more expansive training program, such that you can require an abstinence from all upper body training outside of the study. For example, instead of just training biceps curls, subjects could train curls, triceps extensions, unilateral chest press, and unilateral rows. That would have the added benefit of allowing you to test whether hypertrophy contributes to strength gains in multiple different muscles (biceps, triceps, and potentially pecs), instead of just one.

As is likely clear by now, I don’t agree with the researchers’ contention that hypertrophy is not a contributory factor for strength gains. However, instead of just arguing against their position, I’d like to present the case for why hypertrophy is a contributing factor for strength gains.

First, as I recently discussed on the Iron Culture podcast, there’s a reasonably straightforward case for why you should expect hypertrophy to increase the potential for strength gains. In short, we know that contractile force is produced via actin-myosin crossbridges within muscles. Ultimately, that is the mechanistic basis for muscle contraction. As such, it’s very reasonable to simply assume that muscle hypertrophy will contribute to strength gains, merely by increasing the sheer amount of basic contractile units in each cross-section of the muscle. Thus, for hypertrophy to not contribute to strength gains (or at least increase the potential for strength gains), one of two things would need to be true.

1. As muscle size increased, the density of contractile proteins within the muscle would need to decrease substantially. In fact, virtually 100% of muscle growth would need to occur via sarcoplasmic hypertrophy, so that the absolute amount of contractile proteins in the muscle’s cross-section remained constant.
2. As muscle size increased, the maximal possible level of the “other factors” contributing to strength would need to decrease. For example, if you were necessarily able to master motor skills to a greater degree with smaller muscles, or if you were necessarily able to activate motor units better with smaller muscles, or if lateral force transmission to connective tissue was necessarily greater with smaller muscles than with larger muscles, that could theoretically mean that hypertrophy wouldn’t increase strength potential, in spite of increased absolute amounts of contractile proteins.

As far as I’m aware, there’s no strong evidence for any of these possible complicating factors. There is some evidence suggesting that sarcoplasmic hypertrophy can occur (13), but as far as I know, most evidence still leans firmly in favor of myofibrillar hypertrophy being the dominant factor in long-term muscle growth. In fact, there’s some evidence against the existence (or importance) of these complicating factors, at least collectively. If these complicating factors exist and are relevant, you’d anticipate that whole-muscle and muscle fiber strength per unit of cross-section area would decrease as hypertrophy occurred. However, a 2018 meta-analysis (which included the same senior author as the presently reviewed study) found that strength per unit of whole-muscle size and per unit of fiber size tended to increase with hypertrophy (14). That being the case, there’s a very strong logical case one could make for why hypertrophy should be a contributory cause of strength gains.

However, logic can only get you but so far. Plenty of logical things don’t pan out in the real world. So, what sort of empirical evidence suggests that hypertrophy may contribute to strength gains? Well, for starters, muscle size and lean mass are strongly associated with muscle strength on a cross-sectional basis. In fact, among lifters, the association is very strong: r = 0.8+ (15, 16, 17). That’s not the strongest evidence in the world, however; as we know, you can’t infer causation from correlation. We also see that changes in muscle size or lean mass are associated with changes in strength, especially in trained subjects (18, 19, 20). The association is much weaker in untrained subjects, suggesting that non-hypertrophic factors are driving strength gains early in one’s training career (21, 22). However, that also doesn’t necessarily imply causation. So, what else do we have beyond associations?

Well, in a perfect world, we’d have a study like the presently reviewed study where more hypertrophy occurred in the hypertrophy condition, or we’d have a randomized controlled trial where “hypertrophy” could be randomized. For example, if hypertrophy could be induced in one group but not another, while equating all other factors that might independently influence strength gains (training volume, intensity, frequency, etc.), a difference in strength gains between groups would suggest that hypertrophy independently contributes to strength gains. I don’t think such a study exists, but I can think of a relevant body of literature that comes reasonably close. In studies where subjects are given exogenous androgens (steroids), they build more muscle than people not on steroids, and they also increase their strength to a greater degree. That applies when a steroid group and steroid-free group aren’t lifting and when they are lifting (on the same training program). There’s even a close dose-response relationship between blood testosterone levels following testosterone administration and changes in strength, without any resistance training stimulus. In a 2001 study by Bhasin and colleagues, subjects given 300 or 600 mg of testosterone per week increased their leg press 1RMs by 70-80 kg, without any resistance training stimulus (23). One could potentially argue that steroids simultaneously increase muscle mass and increase strength, with those two adaptations being completely independent and unrelated, but that feels like a pretty big stretch to me. I’m unaware of any non-hypertrophic mechanisms by which steroids could independently increase strength to such a notable degree. Unless you’re going to contend that steroids cause neural adaptations that rival those of exercise (for example, in a study in trained lifters, resistance training drug-free led to a 10kg increase in bench press 1RM; steroid administration without any lifting caused a 9kg increase; 24), it certainly seems that the hypertrophy induced by steroids contributes to steroids’ effects on strength gains.


I wanted to look into this a bit deeper, so I downloaded the OpenPowerlifting IPF dataset (downloaded 4/6/2021). I wanted to see whether changes in body mass between meets were predictive of changes in people’s powerlifting totals between meets. Again, such an association wouldn’t conclusively establish a causal relationship, but it would provide some support for the idea that hypertrophy contributes to strength gains. To be clear, if we observed a positive association between changes in body mass and changes in strength among powerlifters, that would suggest that one (or more) of these possibilities may be true:

1. Muscle growth contributes to strength gains, and muscle loss contributes to strength loss.
2. Adipose tissue has contractile properties that have hitherto remained elusive.
3. Hypertrophy-focused training is more effective for strength development than strength-focused training that’s not intended to cause hypertrophy, but the reasons it’s more effective for strength development are not related to hypertrophy.
4. Training that’s effective for strength development happens to increase body mass, in a manner where it’s completely irrelevant if hypertrophy occurs in the process.
5. Merely being in a sustained calorie deficit mechanistically decreases strength, and merely being in a sustained calorie surplus mechanistically increases strength, in a clear dose-response relationship. Basically, if you’ve been in a cumulative calorie surplus, you’ll get stronger (regardless of whether you gained muscle or fat), and if you’ve been in a cumulative calorie deficit, you’ll get weaker (regardless of whether you lost muscle or fat).
6. Changes in body mass increase or decrease your strength due to changes in leverages, independent of whether you gain or lose muscle in the process.
7. Changes in body mass generally reflect how hard someone is training. When people are training hard, they gain strength, and also happen to gain body mass (though it would be better if they didn’t, since not gaining body mass would improve their competitiveness), and when people aren’t training hard, they lose strength, and also happen to lose body mass.

Of those seven possibilities, I think the first is clearly the most plausible. I actually think the fifth and sixth possibilities also contribute to some degree, but I don’t think they’re sufficient explanations to explain the magnitude of effect we’re about to see in the data.

Once I downloaded the dataset, I did some light cleaning. First, I removed lifters who competed in single-ply powerlifting gear or knee wraps, filtering it down to the raw division. Then, I removed lifters who competed in bench-only, deadlift-only, or push-pull meets, filtering it down to lifters who competed in all three lifts (squat, bench press, and deadlift). Then, I removed lifters who were competing before or after the typical “prime years” for strength (20-35 years old), so that age effect wouldn’t affect strength changes too much, and puberty wouldn’t affect body weight changes. Then, I removed lifters for whom actual body mass wasn’t recorded (some entries just had the lifter’s weight class, but not their actual body mass). Then, I removed lifters who hadn’t competed in at least two competitions, since I was interested in changes between meets. Finally, I removed lifters whose change in body mass or change in strength were at least four standard deviations above or below the mean (these were mostly obvious typos, or lifters who clearly took some token lifts, probably due to entering a meet injured or injuring themselves during the squat), and meets that took place less than 28 days apart (since we want to see “true” changes in strength; if lifters compete twice in a month, you’re probably just seeing normal fluctuations in strength, rather than true increases or decreases).

I was left with 86,084 meet results from 27,871 individual lifters (about two-thirds males and one-third females). People improved their totals by 15.0 ± 27.1 kg between meets, gained 0.72 ± 3.61 kg of body mass between meets, and went 233 ± 226 days between meets, on average.

I started by seeing whether changes in strength were associated with changes in body mass between meets. They were: r = 0.375 (this means that changes in body weight explain about 14.1% of the variability of changes in strength). Then, I wanted to see how predictive every other factor in the OpenPowerlifting dataset was. So, I crammed them all into one big multiple regression model: r = 0.284 (this means that literally every other factor – sex, time between meets, age, body weight on meet day, powerlifting total, how many meets the person had competed in, and the person’s IPF points – cumulatively explained 8.1% of the variability of changes in strength). In other words, merely knowing how much someone’s body mass changes between meets tells you more about how their total will change between meets than knowing how strong or competitive they already are, how many meets they’ve done, how long they’ve been training between meets, their sex, and their age … combined.

However, those are pretty “naive” analyses. The biggest problem with them is that subjects who competed more times were given more weight. For example, if someone competed twice, they were included once (changes in their total and their body mass between meets one and two), but if someone competed 15 times, they were included 14 times (changes in their total and their body mass between each consecutive pair of meets). How many meets someone had done was also predictive of their change in strength. The association wasn’t as strong as the association between changes in body mass and changes in strength, but people did tend to gain more strength between their first and second meets than between, say, their 9th and 10th meets. So, I filtered the data down further to folks who had competed at least four times and tried to analyze the data using linear mixed models, with change in weight versus change in strength nested within subject. Trex and I spent an afternoon on it and couldn’t get a model to converge, so I went with an alternate approach.

Between each pair of meets – meet 1 to meet 2, meet 2 to meet 3, meet 3 to meet 4, etc. (out to 10 meets) – I wanted to see if there was still an association between changes in body mass and changes in strength. One possibility is that newer lifters are more likely to gain a lot of strength and experience a lot of hypertrophy (though hypertrophy isn’t actually contributing to those strength gains), while more seasoned lifters are likely to experience little hypertrophy and small strength gains – that would have the effect of creating a positive association between gains in body mass and gains in strength within the full dataset, which would break down when looking at each pair of meets individually. However, that’s not what I found. The strength of the association between changes in body mass and changes in strength was fairly stable between each pair of meets (r = 0.29-0.40), and the slope of the relationship (the degree to which you’d expect a lifter’s total to change for each kilogram that body mass increases) was also pretty stable between each pair of meets (slope = 2.2-3.1). This is basically the same result I got from the initial linear regression analysis I ran on the full dataset, but it’s more robust and statistically justifiable.


To make the relationship easier to intuitively grasp, between each pair of meets, I calculated the average change in strength if someone lost at least 10kg between meets, if they lost 5-10kg, if they lost 2-5kg, if they lost 0-2kg, if they gained 0-2kg, if they gained 2-5kg, if they gained 5-10kg, and if they gained 10+ kg. From meet 5 to meet 6 onward, I pooled the two largest weight loss buckets (people who lost at least 5kg), and from meet 7 to meet 8 onward, I pooled the two largest weight gains buckets (people who gained at least 5kg) in order to ensure each “bucket” had at least 30 lifters in it. You can see the results in Figure 3. As you can see, as rates of weight loss decrease or rates of weight gain increase, strength gains get larger and larger.


Finally, just as a bit of evidence against the seventh hypothesis I outlined above (“changes in body mass generally reflect how hard someone is training. When people are training hard, they gain strength, and also happen to gain body mass”), I also looked at the strength of the association between changes in body mass and changes in IPF points (a scaled scoring system that accounts of body mass and sex; two lifters of different body masses with the same amount of IPF points should be similarly skilled lifters). There was virtually no relationship (r = -0.03). In other words, it doesn’t seem that people are training particularly well and becoming objectively more competitive as they gain weight, or that people are slacking off and becoming less competitive as they lose weight, on average.

If you’re inclined to be skeptical about the proposition that hypertrophy contributes to strength gains, you’re probably not impressed by any of these analyses of powerlifter data. “It’s just more associations (which don’t necessarily imply causation),” you might say, “and you’re looking at changes in body mass, not even changes in muscle mass.” And, while those are both very fair points, I think this is stronger evidence than mere associations. If hypertrophy truly didn’t contribute to strength gains, this dataset would be the place for such a lack of relationship to shine. Powerlifters are incentivized to be as strong as possible while also being as light as possible – that’s literally the entire point of the sport. If we could maximize our strength gains without building more muscle (or even while being able to lose muscle), that would be the holy grail for powerlifting training. Furthermore, powerlifters have generally been training for at least a few years before they step on the platform (and certainly before they compete in their 10th meet) – long enough that hypertrophy no longer occurs by accident; it’s hard to chalk gains in body mass and simultaneous gains in strength up to the effects of strength training that also happens to cause a bunch of hypertrophy as an unnecessary byproduct. Powerlifters also employ a wide array of training styles; if heavy, low-volume, non-hypertrophic strength training year-round was truly capable of maximizing long-term strength gains, you’d expect that people who maintained weight would gain as much strength as lifters who gained weight. Furthermore, as simple as the sport is, and as highly specific as most powerlifting training tends to be, you shouldn’t anticipate major improvements in neural or technique factors (on the whole) between someone’s 9th and 10th meets, after someone’s been competing in the sport for years. If gains in body mass are still associated with gains in strength in that context, but you want to contend that hypertrophy has nothing to do with strength gains, you’re going to have a hard time explaining this relationship, unless you chalk it all up to leverages or some inherent, causal effect of calorie deficits or surpluses (or hitherto unobserved contractile properties of fat mass).

With all of that being said, I do have to acknowledge that it’s never been conclusively, empirically proven that hypertrophy either causes or increases the potential for strength gains. I also genuinely appreciate the authors of the presently reviewed study for conducting a very interesting study using a very clever experimental model. I just wish more hypertrophy would have actually occurred so that the results would have been clearer and more easily interpretable.

So, what should we do with all of this?

For starters, I do think that if you want to maximize strength gains over time, you should also be interested in maximizing hypertrophy over time. That could take one of two forms: 1) interspersing strength-focused training blocks and hypertrophy-focused training blocks, or 2) doing hybrid training (or, if you insist, “powerbuilding”), with a decent emphasis on hypertrophy work, while including enough heavy (>80% 1RM) strength work to hone and improve your skills with heavy loads. I find that two approaches to hybrid training can work really well. First, you can basically just do hypertrophy-focused training, but add in 1-3 heavy but sub-maximal single-rep sets (with ~85-95% of your max, or at RPE 7-9ish) before your main working sets for your most important compound exercises. Second, you can train your primary compound lifts like a powerlifter (mostly pretty heavy sets of <8 reps), but make sure you do a fair bit of additional “bodybuilding” training within the same workout. For example, you could start a workout with 4 sets of 3 squats at 85% of 1RM, followed by leg press for sets of 10-15 reps, and walking lunges for sets of 20-25 reps.

The more interesting question is whether hypertrophy training will maximize your competitiveness in powerlifting. As I mentioned earlier, change in body mass between meets was not associated with change in IPF points between meets in the OpenPowerlifting dataset. Does that mean that, while muscle growth may contribute to maximizing total strength, it doesn’t actually make you a better powerlifter? That’s certainly one possibility, but I think the nature of how people tend to gain and lose weight in powerlifting clouds the true relationship between hypertrophy and competitiveness in powerlifting over time. Powerlifting has weight classes, and a lot of lifters try to move up or down a full weight class between meets. For example, if someone is competing in the 83kg class, and they try to move up to the 93kg class, they’ll often try to “fill out” their new weight class between meets. Unless they’re waiting at least two or three years between meets, they probably aren’t gaining 10kg of pure muscle. More likely, they’ll gain 3-4kg of muscle and 6-7kg of fat between their last meet at 83kg and their first meet at 93kg. In the process, they’ll probably get stronger, but they may actually be less competitive. This would register as no change in their IPF points, or maybe even a slight decrease. However, as they fill out their class over time, they’ll lose some fat, gain some muscle, and improve their IPF score, which would register as an increase in IPF points with no change in body weight between meets. The opposite can occur when losing weight; if you were near the top of the 93kg weight class, and tried to cut to 83kg between meets, a relatively speedy 10kg loss of weight could lead to some muscle loss, leading to a decrease in IPF points. However, as you stabilize in your new weight class, you’ll probably start building strength again, improving your IPF points. These relationships I’ve just described would create an inverted-U relationship – increases in IPF score at weight maintenance, and decreases when gaining or losing weight. However, there’s more to it than that. If someone does gain weight gradually, such that most of the weight gain is muscle, their IPF score should improve as they gain weight. Conversely, if someone has a fair amount of fat to lose, and they cut gradually so that they don’t sacrifice muscle mass, their IPF score should improve as they lose weight. There are also a lot of people in the middle, who stay in the same weight class, but don’t have meaningful changes in strength between meets. Put all of that together, and it’s a bit unsurprising that changes in body weight don’t predict changes in IPF score from meet to meet.

However, I do think there’s a relatively straightforward case for hypertrophy improving competitiveness in powerlifting over time. If we assume that contractile force per unit of muscle cross-sectional area doesn’t decrease with hypertrophy, muscle growth should increase relative strength (which is what IPF points are assessing; strength scaled to body mass). As long as your level of body fat remains constant, hypertrophy increases the fraction of your total body mass composed of muscle. Just using some round numbers, the average untrained person has about 30kg of skeletal muscle mass. If they can total 300kg (10kg per kg of muscle mass), and they weigh 80kg, their total would be worth 42.31 IPF points. If they can add 5kg of muscle mass without increasing body fat, even if their strength per unit of muscle mass doesn’t increase, their total would increase to 350kg at a body mass of 85kg, which would be worth 47.87 IPF points. This works for any set of assumptions I’m aware of. Instead of an untrained lifter, let’s assume a high-level 105kg lifter has 48kg of muscle mass at 105kg, and they can total 864kg (18kg per kg of muscle mass), for 106.67 IPF points. They decide they’re comfortable water cutting 3kg, so they gain 3kg of muscle mass. Now they should be able to total 918 in the gym at 108kg, for 111.87 IPF points. If they can hit that same total on the platform after weighing in at 105, their 918kg total would be worth 113.34 IPF points. Now, things would clearly be a bit more complicated in the real world. Training may increase bone mass a bit, your strength per unit of muscle mass will increase over time, etc., but as long as fat mass either stays constant or decreases, hypertrophy should increase your competitiveness in powerlifting, unless it can be shown that the contractile force of muscle necessarily decreases following hypertrophy. As I mentioned previously, there’s evidence that the opposite actually occurs – strength per unit of muscle actually increases following hypertrophy (14).

Anyway, that’s all I’ve got. To wrap up, let me admit that I may just be blinded by my biases on this topic. I’m well aware that I turned the discussion of a study purporting to show no impact of hypertrophy on “strength potential” into a very long-winded defense of hypertrophy’s impact on strength development. If I can psychoanalyze myself for a moment, I think I’m inclined to be verbose on this topic, because I’m slightly insecure about how discussing it forces me to adopt an argumentative style that’s pretty foreign to me. I generally default to empirical reasoning (“what does the data show us?”), but the strongest argument in favor of hypertrophy contributing to strength gains is a rationalist argument (actin-myosin cross-bridges are what cause muscle contraction, and they scale in number with increases in muscle size). I’m also generally happy to assume the typical null hypothesis (“there’s no difference between these things” or “there’s no relationship between these things” or “this thing doesn’t cause that thing”), but in this case, the relationship between hypertrophy and strength development seems so self-evident, that I tacitly accept this positive relationship as the null hypothesis that needs to be disproven. So, since I feel like I’m arguing from my back foot to some degree, I may cross a few more “T”s and dot a few more “I”s than I typically would. If you made it to the end of this article, I appreciate your patience, and I hope that at least a bit of this rambling was interesting or informative.

Next Steps​

I already laid out my ideal next step in the “Interpretation” section. Namely, I’d like to see a similarly designed study, where hypertrophy is initially induced via low-load training, the initial phase runs long enough for more hypertrophy to occur, the second phase has some sort of objective cutoff to ascertain if subjects have truly reached their “strength potential” at their given level of muscle mass, and the subjects aren’t allowed to train the muscle(s) being assessed outside of the study’s training program.