1. DiStefano LJ et al, “Gluteal muscle activation during common therapeutic exercises.” JOSPT 2009.
2. Boren K et al, “Electromyographic analysis of gluteus medius and gluteus maximus during rehabilitation exercises.” IJSPT 2011.
3. Macadam P et al, “An Examination of the gluteal muscle activity associated with dynamic hip abduction and hip external rotation exercise: A systematic review.” IJSPT 2015.
4. Bishop BN et al, “Electromyographic Analysis of Gluteus Maximus, Gluteus Medius, and Tensor Fascia Latae during therapeutic exercises with and without elastic resistance.” IJSPT 2018.
Why this topic? I believe it is imperative for clinicians to be lifelong learners and that reading research is one approach to improve your skills as a clinician. I also think it's really hard to do this, which is why I stared writing this blog - a place where I could store things I've learned and write out my thought processes. The impetus behind this particular topic of learning boils down to two key points.
Key point number one: The Clamshell Debate.
If you've ever gone to outpatient orthopedic physical therapy, you've likely done the clamshell exercise. It's very common and frequently patients will say they've done them in the past. Thanks to Seattle Storm Sports Performance Coach Emily Blurton for her video demonstration:
Early in my PT career, I frequently recommended this exercise. I had pre-made exercise programs that I used repeatedly for multiple conditions. Now I make every program for the patient in front of me, often updating them at every session, and this exercise rarely appears. To me, the clamshell doesn't look like anything a person does in daily life or in sports. Maybe it looks like you're lifting your leg to get into a car. But that's standing up, so maybe not. Even worse, a lot of patients do them incorrectly and find them boring and too easy. I don't create exceptionally difficult exercise programs, but it's a lot easier to convince someone to work on something that's challenging than on something that's simple and mundane. And most of the time I don't really harp on having proper form - but with this exercise, is there any benefit to doing it wrong? Biomechanically speaking, an "incorrect" squat will still give you some strength gains, but potentially not where you may want to target them. I don't think shooting for the moon, missing, and landing amongst the stars is actually meaningful in this scenario.
The circumstances under which I will recommend the clamshell are very specific: the patient needs to 1) have a restricted weight bearing status from the physician due to a surgery or fracture and is therefore unable to do the exercises I prefer but would still benefit from strengthening or 2) have a significant deficit in their active hip external rotation movement with available passive mobility that I want them to purposefully control. (Translation - they're too weak to do any of the other exercises I like better, but I never actually tell patients they're weak!)
I share many of my patients with coworkers who like this exercise, so we've had conversations about clamshells for years. The biggest discussion came when I had a student who I practically forbade from assigning clams without excessive clinical reasoning of why they were a good choice for that patient. Now that I have some extra time to dig into the research, I'm looking to prove myself wrong. Does the evidence oppose my clinical bias? And how do you explore the efficacy of an exercise?
Point Number Two: What research should I be reading?
 |
| Levels of Evidence - Creative Commons CC BY-NC 3.0 |
This is a very simplistic demonstration, but sometimes simple demonstrates a major point. Right now with the Coronavirus Pandemic, small studies are being smeared all over the media without sufficient data and with incorrect descriptions. It's a good time to point out that you need to be cautious about how you interpret what you read. Notice at the top of the pyramid sits "systematic review." A systematic review takes multiple research papers into account to accumulate more data points to try to make conclusions. It is based on a researcher compiling data from other researcher's published works. Each individual study may not have done the same research, but their data overlaps and new conclusions are drawn from having more data points. I DID NOT USE SCIENTIFIC RIGOR in my collection of resources for this blog post and I have minimal experience in data analysis, so I'm certain there are flaws in what I've found by summarizing these four papers. However, I basically (unintentionally) created a mini systematic review by reading four different papers (including one which was also a systematic review!).
I collect articles to read in the future. This collection was one reason behind starting a journal club! I've previously written about the anti-library where you collect resources to learn about the things you know you don't know here. Along with my anti-library, I also have folders of papers I've previously read. I had already read Dr. Lindsay DiStefano's Glute EMG paper (the first in the list) which was probably the nail in the coffin for clamshells for me. Imagine assuming that all kids don't like peas because one kid said they didn't? I've basically assumed that clamshells are a poor quality exercise based on the findings of one paper that studied 21 participants doing 12 exercises. Dr. DiStefano was my research advisor in PT School so I almost always read her publications, or have them saved for the future. (Bias Alert!) So I need more data points! I searched for more recent Glute EMG papers to compare to hers and found three more, intentionally choosing papers that included the clam shell exercise and ultimately including one with a somewhat contrary view. Now the data I'm considering reflects over 500 participants and many more exercises including multiple variations of the clamshell.
EMG studies are used to understand muscle activity during movement. Not only are these papers valuable, but they can also influence patient care by helping to better understand what exercises target which muscles. I've previously hinted at reading EMG studies with regard to the shoulder here, and for sure that knowledge changed my PT practice for post-op shoulder patients. EMG studies aren't perfect, but by looking at a collection of them, you can assume similar risk of error for each study. Participants in these papers were first assessed for their maximal volitional isometric contraction (MVIC) (aka how strong they were for a specific muscle). Then, by placing sticky electrode sensors on the skin at certain locations, measurements of muscle activity were taken with each exercises and compared to the maximal strength possible for that person. For an exercise to improve strength, the exercise must elicit at least 40-60 % of maximal strength (MVIC) for gains to occur.
DiStefano et al: The top exercises for glute medius based on % MVIC were side-lying hip abduction (81%), single limb squat (64%), lateral band walk (61%) and for the glute maximus were single limb squat (59 %) and single limb deadlift (59%). Clamshell exercises (depending on hip flexion position) were shown to have glute medius activation of: 38-40% MVIC and glute maximus activation of 34-39% MVIC. (Clamshells don't pass the 40% minimum cut point which has been my rationale to discontinue using them in the clinic). Other exercises were examined in the study, but since they did not cross the 40% MVIC threshold, they are omitted here.
The Boren study used a cut-point of 70% MVIC as the minimum acceptable for strengthening and examined a larger variety of exercises. Glute med exercises with >70% (MVIC) were the side plank with hip abduction (103.11% bottom leg, 88.82% top leg), single leg squat (82.86%), clamshell variation #4 hold top leg in full hip extension while internally rotating (76.88%), and prone plank with hip extension (75.13%). For the gluteus maximus, plank with hip extension (106.22%), glute squeeze (80.72%), side plank with hip abduction top leg (72.87%), bottom leg (70.96%) single leg squat (70.31%). This paper suggested that the best exercises to target both glute med and glute max were the prone plank with hip extension, side-planks with hip abduction, and single limb squats.
The clamshell exercise had four variations in the Boren paper. Of these
variations, #4 described above was the only option to reach their defined threshold of >70%. If you use the criteria of 40-60% outlined in the DiStefano paper, all the clamshell variations would meet that target as sufficient for strenghtening for the glute med, but only the standard clamshell works for the glute max at 53.10% with other
variations all below 30% MVIC.
The Systematic Review by Macadam et al reviews multiple papers and breaks down exercises by body position. The cross-over step up and
lateral step up had the highest average glute max and glute med % MVIC. Pelvic drop, sidestepping with hip internally
rotated and band at the ankle, standing hip abduction variations, rotational
single leg squat and transverse lunges all had glute med % MVIC over 40%. Standing hip abduction, rotational single leg
squats and transverse lunges also had % MVIC over 40% for glute max.
In sidelying, side planks with hip abduction had the highest average glute max and
glute med % MVIC. (That exercise is really hard, though, and wouldn't actually be appropriate for the majority of my patients.) The paper examined 13 clamshell variations including the standard hip external rotation and others. In one study, three
clam variations (not the standard hip external rotation) had average glute med % MVIC over 60% but other
studies had clamshell averages in the 30% range. From this paper, I can start to ease up on my thoughts that patients do the clamshell incorrectly - and more that they're just creating their own variation.
And finally, Bishop et al compared glute med and glute max EMG to tensor fascia latae (TFL) EMG creating a "Glute to TFL Index." The TFL and gluteus maximus both insert on to the iliotibial band and contribute to hip abduction. The TFL is sometimes considered to be a contributor to low back and lower extremity pains and/or injuries and may increase in tone or "tightness" to compensate for gluteal deficits. Bishops "Glute to TFL Index" findings were that the clamshell with resistance was 99.54 and the clamshell without resistance was 87.89 meaning that the gluteal muscles were selectively used for these exercises much more than the TFL. The clamshell exercises far surpassed all the other tested exercises in this study with the next best exercise being the bridge with resistance at 48.86 and prone hip extension with 48.57, both about half of the clamshell values. Thus, this paper recommends use of the clamshell exercise as a glute-targeting exercise that does not incorporate the TFL. To these findings, I'll add another dose of bias. Does this really matter? Does the Glute to TFL Index have clinical meaning? I'm not sure because I don't know if I really want to "turn off" the TFL. I'll let the clams have this one.
In the end, I haven't been convinced. A healthy dose of confirmation bias potentially sprinkled with some self-serving bias may contribute to that decision, but to me, the evidence seems clear. Maybe I'll be nicer to those who know that the clamshell has the higher glute to TFL index, but without that rationale, these papers suggest many other glute strengthening exercises. For sure, I'm open to hearing alternative opinions. Are you team clamshell? Or team ANYTHING ELSE! I use prone glute squeezes more often than I use clamshells. And I'm incredibly grateful for my coworkers who challenge me to think about these topics. 
(Again - this is an abbreviated summary of the results sections of the papers. More information looking at the discussions and limitations and other aspects of the research is available if you'd like more! For the full written summary for Seattle Children's Journal Club, go ahead and email abby.gordon@seattlechildrens.org and I can send it any time).
