MASS Research Review

Study Reviewed: Hip Thrust and Back Squat Training Elicit Similar Gluteus Muscle Hypertrophy and Transfer Similarly to the Deadlift. Plotkin et al. (2023)

This new preprint by Plotkin and colleagues compared hip thrust to back squat training for nine weeks in 34 (16 in the squat group, 18 in the hip thrust group) untrained university-aged men and women. It’s generating a lot of discussion online, for good reason. But before we get into those reasons, you might ask “what is a preprint?” 

A preprint is just like any other scientific paper you might find in an academic journal, but it’s put online before it’s printed in a peer reviewed journal. Preprints facilitate multiple goals of the open science movement, including faster dissemination of research and improved research quality. The traditional publication process can take ages, and preliminary publication of data allows the scientific community to discuss, comment, critique, and make suggestions the authors can implement before the paper gets to formal peer review (which can be hit or miss in terms of quality). Thus, the fact that this preprint is generating a lot of chatter (and that I’m reviewing it in MASS) is a good thing – it’s a great example of the preprint concept working as intended. 

So, what is interesting about this study? Well, Plotkin and colleagues addressed a lot of questions of interest by not only comparing hip thrusts against squats for their impact on glute hypertrophy (and with top-of-the-line methods, as they assessed hypertrophy with MRI which is important because assessing the glutes with ultrasound is very challenging), but also in seeing whether acute, baseline EMG glute measurements predicted longitudinal hypertrophy. On top of that, they also compared the strength transfer of hip thrusts and squats to deadlifts. Finally (and I’ll elaborate on this in the interpretation), many are viewing this study as a comparison of long versus short muscle length training for the glutes – although I think taking that viewpoint is a mistake. 

To not bury the lede, the main finding was that hip thrusts and squats produced similar hypertrophy in all four glute regions assessed, the upper, middle and lower glute max, and the glute medius and minimus (although these muscle grew minimally), as shown in Figure 1.

In addition, the quadriceps and adductors grew more following squats, while the hamstrings, like the glute medius and minimus, did not meaningfully hypertrophy (Figure 2).

Thus, it appears that both exercises are similarly effective at growing the glute max, neither is effective at growing the hamstrings or glute medius or minimus, and the squat is better for both growing the adductors and quadriceps. Most of this is unsurprising with an understanding of functional anatomy. For example, along with the glutes, despite its name, the adductor magnus is a principal hip extensor. The glutes and adductors are the two largest hip extensors by cross sectional area and both have strong moment arms for hip extension (2). But, unlike the glutes, the hip extensor moment of the adductor increases as hip flexion increases (what happens on the descent of a squat), which is likely part of the reason why deeper squats produce more adductor hypertrophy than shallower squats (3). Thus, given that subjects in the squat group achieved a reasonably deep position where their femurs were parallel to the floor, and considering that hip thrusts involve less hip flexion than sufficiently deep squats (4), greater adductor growth in the squat group makes sense. Likewise, it makes sense that quadriceps growth favored squats, as they involve far more knee flexion than hip thrusts, and it also makes sense that the biarticular hamstrings did not grow in response to squats, as we have previously discussed. The interesting hypertrophy finding in my mind, is just how targeted the hip thrust appears to be for glute hypertrophy. You could create a negative narrative about how the squat is unequivocally superior because it produces similar glute growth in addition to producing quadriceps and adductor growth, but there’s no need to create false dichotomies. Exercises are not mutually exclusive and they should be seen as tools. What are the features of the hip thrust as a tool? It requires less technical expertise than the squat, doesn’t axially load the spine, avoids hypertrophy in the hamstrings, and produces minimal hypertrophy in the adductors and quads. These features should open your eyes to a lot of applications. For example, what if I have an athlete with an injury who can’t squat or deadlift? What if I have a physique competitor who has quads (or adductors or hamstrings) that are overly dominant for their division, but they need more glute mass? I think you see my point. 

Moving onto the strength data, quite unsurprisingly, squat three-repetition max (3RM) increased more in the squat group, and hip thrust 3RM increased more in the hip thrust group, and both had similar transfer to a deadlift 3RM and the wall push (which I won’t spend time on given the MASS audience probably has more interest in deadlifts) as shown in Figure 3.

I found the strength data interesting, as depending on whether you take a movement-centric view or a muscle-centric view of strength, you might expect different outcomes. For example, both the deadlift and squat are multi-joint lifts performed with free weights, where you stand on the ground and produce vertical force through the combined actions of hip and knee extension. Certainly, the deadlift is more hip extension dominant, and the squat is more knee extension dominant, but they have a lot of kinematic overlap, so someone thinking about transfer from a movement-centric lens would predict better deadlift improvements from squatting. During the hip thrust, on the other hand, you are partially lying down with your back on a bench, and isolating hip extension while producing horizontal force – it’s a very different movement. But from a muscle-centric perspective, you might argue the hip thrust would produce better strength transfer as you’re specifically targeting the hip extensors, the prime movers in a deadlift. However, both exercises had similar strength transfer, at the group level. Notably, you are not a group, so don’t take this as a sweeping generalization about either exercise’s potential as a deadlift accessory. There are assuredly individuals who would get more transfer to the deadlift from one or the other, depending on their individual strengths and weaknesses.

The final take home from this study (but not my final point of discussion) comes from the EMG data. I’ve discussed the limitations of EMG previously (here, and here), and two quotes are directly relevant: “If you want to know with confidence which exercises are best for developing specific muscles, you have to compare them longitudinally and measure hypertrophy. This is similar to the relationship between muscle protein synthesis and hypertrophy in studies on protein timing or type, or between body fat loss and acute fat oxidation in studies on cardio or macronutrient manipulation. In almost any acute study in MASS, we bring up the importance of longitudinal assessments to confirm acute findings, and EMG is no exception.” My second quote provides more specifics: “Joint angle differences between exercises acutely alter muscle morphology and position, subsequently impacting EMG signal strength and acquisition. In addition, not all muscles have the same recruitment pattern. Some follow a superficial to deep pattern, while others deep to superficial, which can understandably impact surface EMG (5). Therefore, comparisons between different exercises which train the same muscle, even when based on MVICs [maximum voluntary isometric contractions], may not be valid representations of relative signal strengths, and therefore of expected longitudinal hypertrophy differences (6).” Ultimately, EMG provides narrow information about the hypertrophy potential of an exercise, especially relative to another dissimilar exercise. For this reason, I wasn’t surprised that the hip thrust produced higher peak and mean EMG values (most of which were significant) than the squat for all glute regions and muscles (Figure 4), despite glute hypertrophy being similar. Thus, we have to be very skeptical about any hypertrophy predictions made, especially between very different exercises, based primarily on EMG data. 

The final point of discussion is what you should not conclude from this study. I’ve seen people say that this study shows that the glutes, unlike other muscles, don’t grow better when trained at longer muscle lengths. The rationale behind such a conclusion is that the squat trains the glutes at a longer muscle length than the hip thrust, yet both resulted in similar glute hypertrophy. The problem with this conclusion is that this study did not isolate muscle length as the independent variable. The independent variable was the choice of exercise, and hip thrusts and squats differ in many ways, not just the length at which they train the glutes. Just as one example of how these exercises differ, these untrained participants trained to failure, yet the hip thrusts are a much more isolated exercise than squats. Given that the hypertrophy observed following hip thrusts was almost exclusively in the glutes, you’d expect hip thrust failure to occur because the glutes specifically could not keep producing sufficient force to meet the loading demands. However, squats are a multi-joint lift, so failure could be attributed to any number of the contributing muscles being the proverbial “weakest link,” depending on each individual’s unique biomechanics. If one actually wanted to determine if the glutes responded better to being trained at longer or shorter relative muscle lengths, you’d have to compare the same exercise performed at different muscle lengths. Fortunately, that data exists. Specifically, Kubo and colleagues (3) compared half squats (which train the glutes at a shorter length) to deep squats (which train the glutes at a longer length) and reported that deep squats produced significantly more glute (and adductor) hypertrophy. Further, in a recent conference presentation, Maeo reported that long muscle length unilateral hip extension partials produced more glute and biceps femoris growth than their full range of motion version (7). Therefore, not only is there even more data forthcoming that supports the beneficial effect of long muscle length training for hypertrophy, it seems the glutes are no exception.  

To wrap up, this study by Plotkin compares two exercises which differ in multiple ways, with several findings that are interesting and practically applicable. But, this is not a study designed to inform us about longer versus shorter muscle length training for the glutes.

References

1. Plotkin DL, Rodas MA, Vigotsky A, McIntosh MC, Breeze E, Ubrik R, et al. Hip thrust and back squat training elicit similar gluteus muscle hypertrophy and transfer similarly to the deadlift. bioRxiv. 2023:2023-06.
2. Neumann DA. Kinesiology of the hip: a focus on muscular actions. J Orthop Sports Phys Ther. 2010 Feb;40(2):82-94.
3. Kubo K, Ikebukuro T, Yata H. Effects of squat training with different depths on lower limb muscle volumes. Eur J Appl Physiol. 2019 Sep;119(9):1933-1942.
4. Brazil A, Needham L, Palmer JL, Bezodis IN. A comprehensive biomechanical analysis of the barbell hip thrust. PLoS One. 2021 Mar 29;16(3):e0249307.
5. Vigotsky AD, Halperin I, Lehman GJ, Trajano GS, Vieira TM. Interpreting Signal Amplitudes in Surface Electromyography Studies in Sport and Rehabilitation Sciences. Front Physiol. 2018 Jan 4;8:985.
6. Vigotsky AD, Halperin I, Trajano GS, Vieira TM. Longing for a Longitudinal Proxy: Acutely Measured Surface EMG Amplitude is not a Validated Predictor of Muscle Hypertrophy. Sports Med. 2022 Feb;52(2):193-199.
7. Maseo S. Effects of hip extension training performed with full versus partial range of motion at long muscle lengths on muscle hypertrophy and sprint performance. 27th Annual Congress of the ECSS. 06.07.2023.