Mass Research (Volume 4 - Issue 6)
About a year ago, Mike reviewed a systematic review discussing whether volume load or number of sets served as a better metric for equating the hypertrophy stimulus of different programs (2). In general, it seems that when comparing moderate-rep sets (sets of ~8-12 reps) and high-rep sets (sets of 20+ reps), the hypertrophy stimulus is equated if you match the number of sets performed. When comparing moderate-rep and low-rep sets (sets of <6 reps), there are some instances where matching volume load seems to do a good job of equating the hypertrophy stimulus, and other studies where matching the number of hard sets performed does a good job of equating the hypertrophy stimulus.
The presently reviewed study (1) matched volume load in three groups that performed the bench press twice per week for 10 weeks. One group did 7 sets of 4 reps each session, one group did 4 sets of 8, and one group did 3 sets of 12. Hypertrophy (changes in pec muscle volume) was similar in all three groups, while strength gains were greater in the groups doing sets of 4 or 8 compared to the group doing sets of 12. An intuitive interpretation of these results is that sets of 12 were the most efficient for promoting hypertrophy, and that sets of 4 were the least efficient. However, I’m not sure we can draw that conclusion. When interpreting research, we need to remember both what a study tests and what possibilities a study leaves unexplored.
The purpose of this study was to compare the strength and hypertrophy effects after 10 weeks of training with three different loads (4RM, 8RM, and 12RM) when matching for volume load.
The authors hypothesized that strength gains would increase as intensity increased (4RM > 8RM > 12RM) and that hypertrophy would increase as intensity decreased (4RM < 8RM < 12RM). They also hypothesized that hypertrophy would be most strongly correlated with strength gains when training with 12RM loads.
42 young, healthy men participated in this study. They had not been involved in resistance training in the year prior to the study.
This was a relatively straightforward longitudinal training study. All outcome measures were tested before and after the 10-week training protocol. Testing consisted of 1RM bench press (with grip width standardized to 165% of biacromial breadth, which is a pretty moderate grip width) and pec muscle volume assessed via MRI – arguably the gold standard for assessing changes in whole muscle size.
The subjects were split into four groups via a counterbalancing approach (which isn’t fully described, but which was intended to ensure all four groups had similar characteristics at baseline). One group trained the bench press with seven sets per workout using a 4RM load, one group trained using four sets with an 8RM load, one group performed three sets with a 12RM load, and the final group was a non-training control group (which will not be mentioned again). All groups trained twice per week for 10 weeks, and rested approximately three minutes between sets. Interestingly, the researchers ramped the subjects up to their training loads gradually over the first six training sessions (starting everyone at 3 sets of 10 at 60% of 1RM for the first week, and working up from there) to acclimate them to training. In the first session of week 4, the 4RM group tried to perform 7 sets of 4 reps at 90% 1RM, the 8RM group tried to perform 4 sets of 8 reps at 80% 1RM, and the 12RM group tried to perform 3 sets of 12 reps at 70% 1RM as their first “real” training session using the experimental protocol. If subjects were unable to complete all of the prescribed reps, they stuck with the same load until they were able to do so; if they did successfully complete all of the prescribed reps, the load was increased by 2.5kg for the next training session. Presumably, that’s how load increases worked for the rest of the study as well. So basically, the first three weeks involved non-failure introductory training, with the “real” training starting in week 4 and continuing until the end of the study.
Increases in pec muscle volume were similar in all three groups. However, bench press 1RM increased to a significantly greater degree (p < 0.03) in the 4RM and 8RM groups (28.4 ± 10.0% and 29.5 ± 11.6%, respectively) than in the 12RM group (18.7 ± 10.1%).
Relative (percentage) change in pec muscle volume was significantly associated with relative increases in bench press 1RM in the 12RM group (r = 0.684; p = 0.042), but not the 4RM or 8RM groups (|r| < 0.3; p > 0.75).
Volume load (sets × reps × load) was similar in all three groups, as intended. Training loads also increased to a similar degree in all three groups. The 4RM group averaged 3.4 ± 0.3 reps per set, the 8RM group averaged 7.1 ± 0.4 reps per set, and the 12RM group averaged 10.5 ± 0.4 reps per set.
Before we get into the interpretation, I want to point one thing out about the nomenclature used in this study: The authors use rep max terminology (4RM, 8RM, and 12RM) to label the groups, but I don’t think they were actually training with 4RM, 8RM, and 12RM loads. The paper isn’t completely clear about how load adjustments worked, but I’m pretty sure the subjects just aimed for sets of 4, 8, or 12 reps, and went up in weight once they completed all of their assigned sets without missing any reps. I don’t think they decreased weights during each workout when subjects started missing reps. If you can still get 4 reps on your 7th set of 4 without reducing the load on the bar, there’s no way you were using a 4RM. I recognize this is a somewhat pedantic point, but I think it’s worth pointing out. I’d suspect the subjects were actually training with ~5-6RM, ~9-10RM, and ~13-14RM loads, give or take.
There were two main reasons I wanted to review this study. It provides some evidence for a popular “in the trenches” method of gauging hypertrophy progress. It also gives me an excuse to discuss an aspect of research interpretation that sometimes goes undiscussed: You can’t draw strong inferences about something a study didn’t explicitly test.
One interesting finding in this study was that strength gains in the 12RM group correlated with changes in pec muscle volume, but strength gains in the 8RM and 4RM groups weren’t correlated with pec growth. On the surface, this seems to support a standard piece of “bro” advice – if you’re training in a typical “hypertrophy” rep range, you can use strength gains as a proxy for muscle growth. Now, the recommendation is generally that increases in, say, your 10RM will correlate with muscle growth if you’re training with ~10RM loads, but the basic logic should extend to 1RMs as well; if you’re not training with heavy, highly specific loads, increases in 1RM strength should primarily reflect increases in muscle mass, since you shouldn’t simultaneously be developing the additional neural adaptations that contribute to 1RM strength (at least, not to the same degree as you would be if you were training heavier). If that’s true, that’s a very useful association. Muscle growth occurs pretty slowly – slow enough that it’s tough to reliably detect it at home with a tape measure over the short-to-moderate term. However, if you can know that strength gains are reflecting muscle growth if you’re training in the “hypertrophy rep range,” you’d be in a pretty good spot, because measurable strength progress can occur over shorter time scales. A quarter inch (~6mm) increase in thigh circumference may be within the realm of measurement error, but you can know if your 1RM or 10RM squat has gone up 10 or 20 pounds. So, if you can be confident that your increased squat strength is reflecting muscle growth, it helps provide you with peace of mind, letting you know you’re actually growing, even if your growth isn’t yet detectable with a tape measure.
When this study was published (1), I saw several people on social media running with that interpretation of the data. However, I think we should pump the brakes on that interpretation. For starters, the 8RM group was also training in the “hypertrophy range,” but strength gains and hypertrophy weren’t significantly associated in the 8RM group. Second, the association in the 12RM group was barely significant; the p-value was 0.042; flukey significant p-values that barely duck below the magic cutoff of 0.05 happen by chance all the time. Third, the subjects in this study were untrained, so the 12RM group was almost certainly accruing beneficial neural adaptations for strength, though maybe not to quite the same degree as the other two groups. Finally, we consistently see that hypertrophy and strength gains aren’t strongly associated in untrained lifters, but the strength of the association between hypertrophy and strength gains increases in literature as training status increases. If these subjects had a couple years of training experience, I’d anticipate that hypertrophy and strength gains would be significantly associated in all three groups. So, for untrained lifters, strength gains may be more strongly associated with hypertrophy when subjects are training with lower loads, but I don’t know that we should generalize that inference to experienced lifters. For experienced lifters, I think consistency of training variables is the primary factor. In other words, if you normally train in “hypertrophy rep ranges,” you continue training in those same rep ranges, and your strength improves, I do think it’s reflective of growth. However, I think the same principle applies for powerlifters who usually train in lower rep ranges; as long as they keep training in lower rep ranges, strength gains are generally at least partially reflective of muscle growth. I think this relationship only breaks down when training variables change considerably; if you normally train in moderate rep ranges, but you decrease reps and increase intensity for a while, your 1RM strength will probably increase, but that increase may not be reflective of muscle growth. Conversely, if you normally train with heavy loads in low rep ranges, but you start training a bit lighter and in higher rep ranges, your 1RM strength may decrease a bit, but that doesn’t necessarily suggest that you’re losing muscle.
The second common interpretation I saw was that this study showed that 12RM loads were better for muscle growth than 8RM loads, and that both were better than 4RM loads. After all, all three groups experienced similar hypertrophy, but the 4RM group “needed” more sets, and the 8RM group “needed” one more set than the 12RM group.
That’s not a completely unreasonable interpretation, but I also don’t think it’s the correct interpretation. This is the second point I alluded to above: it’s dangerous to extrapolate too far beyond what a study explicitly tested. The present study is far from the first to equate volume load when comparing low-rep (≤5 reps per set) and moderate-rep (~8-15 reps per set) training; when you equate volume load, the low-rep group has to do more sets. And in all of the studies that matched for volume load or just allowed low-rep groups to perform more sets, hypertrophy was similar with both moderate-rep and low-rep training (1, 3, 4, 5, 6). If we just look at those studies, it certainly appears that you “need” to do more sets when training with low reps if you want to maximize hypertrophy. However, three studies have compared low-rep and moderate-rep training with the number of sets equated. Across those three studies, there’s one significant difference in favor of moderate-rep training, one significant difference in favor of low-rep training, and a bunch of comparisons where low-rep and moderate-rep training produced similar growth (7, 8, 9). So, when we see similar growth with different amounts of sets, that could suggest that the group doing more sets “needed” more sets in order to obtain similar results, or it could mean that the “extra” sets didn’t actually enhance the results much (if any). If we see that 7 sets of 4 and 3 sets of 12 produce similar hypertrophy, we can’t necessarily infer that 3 sets of 12 would therefore cause more growth than 3 sets of 4. We also frequently see that growth per set decreases as a function of the total number of sets performed; in other words, maybe you’ll grow more with 10 weekly sets instead of 5, but you probably won’t grow twice as much. Thus, 7 sets of 4 may look like it produces less growth per set than 3 sets of 12, but 7 sets of 12 would also likely produce less growth per set than 3 sets of 12.
I think it’s worth keeping this principle in mind, especially when dealing with linear interpolations. If, for example, a study found that a training frequency of five times per muscle group per week produced twice as much growth as a frequency of once per week, we can’t infer that a frequency of three times per week would have produced 50% more growth than a frequency of once per week. A frequency of three times per week may have been just as good as either once or five times per week. It’s also possible that it would have been better then either once or five times per week. There are possible dose-response scenarios where any of those possibilities could make sense. However, I think humans naturally gravitate to linear thinking, which can make it hard to consider other possibilities.
For what it’s worth, my personal hunch is that low-rep sets (of ~3-5 reps) are a bit less efficient for muscle growth on a per-set basis, but not quite to the point of needing to equate for volume load. In other words, if you just prefer low-rep training but you want a robust hypertrophic stimulus, and 3 sets of 12 and 7 sets of 4 produce similar volume loads, I’m not necessarily confident that you’d grow quite as well if you did 3 sets of 4 instead of 3 sets of 12, but I don’t think you need to do all 7 sets to equate the hypertrophic stimulus. I think 5ish sets would get the job done. I doubt that applies to singles or doubles, however; if you’re trying to maximize hypertrophy doing sets of just one or two reps, more power to you, but I doubt you’ll make it very far.
I’d be interested in research examining whether the dose-response relationship for the effects of training volume on hypertrophy differs between rep ranges. This wouldn’t be a feasible study because it would involve nine groups, but I’d like to see a study compare low, medium, and high set volumes (maybe 5, 15, and 25 sets per week) with low loads (25-35RM), moderate loads (8-12RM), and high loads (3-5RM). Most of the volume dose-response research uses exclusively moderate loads. I could easily envision a scenario where all three loads produce similar hypertrophy at low and moderate volumes (with maybe slightly less hypertrophy in the high-load groups), with diverging responses at high training volumes. Specifically, I suspect that very high set volumes of very high-rep training would quickly lead to diminishing returns.
If you want a concrete takeaway from this study, it’s that gains in strength may be more strongly associated with hypertrophy in new lifters when training with lower loads (12-15RM), though, as noted, I’m not incredibly confident in that takeaway. I think the biggest takeaway is just that you should be cautious about extrapolating beyond the findings of a study, especially if those extrapolations are based on linear assumptions.