If I were to ask you to describe the ways you can perform resistance training, what would you say? I suspect it would probably go something along the lines of: “performing various exercises, either dynamically or statically, that target specific muscle groups with adequate training volume, high effort, and consistency.” And yeah, that would, on average, be a representative explanation of what the majority of people do day in and day out.
But if you look at how I phrased my answer, there’s one specific point that I’d like to highlight: “…either dynamically or statically…”. You see, while conventional resistance training is mainly performed dynamically, or as isotonic muscle actions (I’ll explain what that means in a bit), it can also be performed statically, or as isometric muscle actions.
When discussing which training modality, dynamic or isometric, builds more strength and size, a lot of people’s default position leans toward the former. Interestingly, it’s become somewhat of conventional wisdom that isometric training just isn’t as effective for building strength and size compared to dynamic training (one, two, three, four). Moreover, when people think of isometric training, planks or wall sits usually come to mind, since those are commonly prescribed. But as is the case with a lot of things, it’s not a problem with isometrics per se, but rather their application. So, while some muscle growth is likely to occur from performing planks, I wouldn’t bet my money on seeing much growth unless you’re really undertrained. However, I’m sure the association between these types of exercises and isometrics has at least something to do with the general misconceptions about their effectiveness (one, two, three). As it turns out, that’s just not entirely true or reflective of the evidence, hence me writing this article to set the record straight.
The main goal of this article will be to introduce: 1) different muscle actions and how they are usually defined throughout the literature, 2) the different types of isometrics, 3) the evidence on isometric training for strength and hypertrophy (as well as other applications such as tendon rehabilitation and pain management), 4) the most important programming variables to consider, 5) an overview of the literature comparing isometric vs. dynamic/isotonic training (as well as versus other muscle actions), and 6) how to practically apply isometrics in your own training.
Muscle actions
Before going any further, it is important to remind ourselves of the types of muscle actions we can perform: isotonic, isometric, and isokinetic. I’ll refer to these as muscle actions rather than contractions because, although “muscle contraction” is commonly used in exercise physiology, “contraction” can be interpreted as shortening (from contrahere, “to draw together”), which isn’t the case for all of these muscle actions.
First off, let’s start with the most common one. Isotonic muscle actions are defined as those during which muscles maintain constant tension while undergoing changes in length, and they occur during dynamic movements. Isotonic muscle actions consist of two components: concentric and eccentric muscle actions.
Concentric muscle actions involve active shortening of the muscle, bringing the two bones connected via a tendon closer together. This occurs due to myosin heads interacting with actin filaments by extending, binding, and pulling on them, resulting in shortening of the sarcomere (the smallest functional unit of the muscle). Using a dumbbell curl exercise as an example, curling the weight upward would constitute a concentric muscle action.
Conversely, eccentric muscle actions involve lengthening of the muscle because the forces acting on it (e.g., external load) are greater than those produced by the muscle fibers. Thus, the lowering phase in our dumbbell curl example would constitute an eccentric muscle action.
When the concentric and eccentric phases are performed in a dynamic, continuous fashion, you’re essentially moving through a joint range of motion under (ideally) isotonic conditions.
A lot of people define isometric muscle actions as involving no change in muscle length. However, that’s not entirely accurate because muscles and tendons can undergo slight changes in their length (e.g. fascicles shorten while the tendon elongates). Thus, it’s more accurate to state that isometric muscle actions occur without meaningful change in the muscle-tendon complex length.
Prior research suggests that different muscle actions may elicit specific morphological, neuromuscular, and performance adaptations. It is generally accepted that, in terms of maximal force capacity, muscle actions roughly follow the order eccentric > isometric > concentric. Eccentric muscle actions have been proposed to be more metabolically efficient (i.e., lower energy cost per unit force), in part because fewer cross-bridges are required and because of the contribution of titin, which can act as an internal spring that stores and releases elastic potential energy (one, two). Additionally, eccentric muscle actions often show lower voluntary activation (i.e., less motor unit recruitment and lower discharge rates) than both concentric and isometric muscle actions, likely reflecting spinal and/or supraspinal mechanisms that limit the neural drive.
Recently there’s been a series of claims on social media suggesting eccentric muscle actions don’t result in hypertrophy, and that the majority of gains come from concentric muscle actions. This goes against much of what we’ve read or been taught on the topic over the last few decades. However, rest easy, that’s not really the case. Prior meta-analyses (one, two) clearly show that both concentric and eccentric training can result in muscle hypertrophy. If you’re subscribed to MASS Research Review, you can read about this in more depth, but briefly: eccentric muscle actions are still plenty hypertrophic.
Finally, we have isokinetic muscle actions which, depending on who you ask, might be irrelevant or considered a pseudo muscle action because most people never get the chance to perform them. However, if you’ve (unfortunately) had the pleasure of doing any kind of isokinetic testing, there’s a good chance you’ve had a gnarly injury and did isokinetic testing to assess your strength levels, or you participated in lab research related to neuromuscular adaptations. Isokinetic muscle actions involve performing a movement at a constant angular velocity. This is mainly achieved using an isokinetic dynamometer where you set up an angular velocity (e.g., 60˚/s or 180˚/s) and perform the task at that constant velocity while the machine adjusts resistance to match your effort. Isokinetic testing is widely used in prehab/rehab and in research because it allows for objective and consistent assessment of strength, with some confounders accounted for (i.e., the varying velocity throughout a range of motion during traditional resistance training).
Brief history of isometric training
Some of the earliest references I could find to what we now know as isometric training relate to Russian-born strongman Alexander Zass, later known on stage as The Iron Samson. In a book, The Mystery of The Iron Samson, some of his early training as a child is described as trying to grab a large wooden tub with his small arms and move it.
“Pull as he may, the tub stood firmly rooted in one spot. The boy however was quite stubborn. Again and again, he tried to move the dreaded wooden edifice. Day after day there was a tug-o-war between the child and this big heavy tub in the basement. Even though the tub still sat motionless in the basement, Shura began noticing some very strange things happening during his workday. The heavy saddle that he could barely carry through the stables previously (in order to saddle Forsun) suddenly felt easier to carry. It was as if the sacks of grain got lighter too.”
Even in the early 20th century, as the book points out, people associated sport with visible movement (e.g., rapid jerks and presses with heavy loads). Zass, however, wondered whether simply “straining” his muscles against immovable objects (e.g., trying to bend an iron bar or tear a chain even if it did not bend or break) could make him stronger. Lo and behold, it did.
As an adult, Zass performed in the circus as a wrestler and strongman. He also served in World War I. He was captured multiple times, escaped, and eventually settled in the UK, where he spent the rest of his days. During his imprisonment, it’s said he maintained his strength by doing squats, stretches, and repeated “muscle tension efforts” against cuffs and braces, contracting his muscles maximally for 15–20 seconds at a time while chained. Supposedly, it was this developed strength that allowed him to break chains and bend iron bars during his escapes. Zass displayed his feats of strength in the circus ring by suspending a piano from his teeth and carrying a grand piano with a pianist and dancer on his shoulders (imagine that). There’s even a photograph of him carrying his wounded horse on his shoulders during his military service in World War I.
As noted in the book, Zass attributed his strength to a deliberate system of both static and dynamic exercise. Specifically, he believed static exercise should be performed alongside dynamic movements using a weighted sack and other loads, in order to build both muscle and tendon strength. Some of those guys really were ahead of their time, weren’t they?
Still, without turning this into a history lesson (you can read more about it in the book mentioned above), what does the scientific evidence actually say about isometric training?
From what I’ve gathered, the earliest known study on isometric training was performed by Hettinger and Müller in 1953. The full details are a bit unclear because the paper is written in German. However, researchers recruited nine untrained, male participants and had them perform daily 6-second isometrics at ~67% maximal voluntary contraction (MVC) for various muscle groups (e.g., elbow flexors and extensors, finger flexors but also shoulders, hips, knees, trunk, and ankle).
Loads were adjusted weekly to maintain the same relative intensity. Strength improved by about 5% per week, on average. The researchers repeatedly measured arm circumference with the biceps fully flexed. They corrected that value for skin/subcutaneous fat and used it to estimate biceps’ cross-sectional area (CSA). However, they only reported that strength and CSA increased proportionally throughout the training period. They didn’t provide much detail about the specific changes or their magnitude.
Nevertheless, this landmark study paved the way for subsequent research in the coming years.
Types of isometrics
When discussing isometric training, it’s important to highlight that there are two distinct types of isometric muscle actions: pushing (overcoming) and holding (yielding) (one, two). Pushing isometrics involve exerting force against an immovable object, whereas holding isometrics involve maintaining a set joint angle when resisting an external load. In other words, pushing isometrics involve an attempt to perform a concentric muscle action, but failing to do so due to the external load being too great to successfully overcome. Holding isometrics, on the other hand, imply an attempt to resist an eccentric muscle action by applying enough force to avoid joint movement.
Having a basic understanding of these different types of isometrics is important because it’s been reported that they share unique characteristics (e.g., may differ in their fatigability). As a result, they may induce different types of adaptations. For example, Hunter and colleagues compared time to task failure between “pushing” and “holding” isometrics at 15% MVC with the elbow flexors. They ensured the same net muscle torque was achieved in both conditions. Although MVC force was similar between conditions, time to task failure was almost twice as long for the pushing versus holding condition (1402 ± 728 seconds vs. 702 ± 582 seconds).
The average EMG of the elbow flexors increased progressively in both tasks. The rate of increase was similar when matched for the same absolute time. However, because the pushing task could be sustained for longer, the normalized average EMG at exhaustion (last 60 seconds) was higher in the pushing condition (22.4 ± 1.2%) than in the holding condition (14.9 ± 1.0%). Conversely, the amplitude of fluctuations in vertical and horizontal directions was greater during the holding task. The rates of increase in mean arterial pressure, heart rate, perceived exertion, anterior deltoid EMG, and EMG burst rate were also greater during the holding task.
Taken together, these findings suggest that the shorter time to task failure in the holding isometric may be associated with greater excitatory descending drive and inhibitory afferent input to the motor neurons, despite both tasks requiring the same net torque.
Indeed, these findings are supported by a recent meta-analysis by Oranchuk and colleagues who found significantly longer time to task failure at the same relative intensity with pushing versus holding isometrics.
Specifically, researchers reported moderate to large effects and an average mean difference of 22.4% between the two isometric types at lower training intensities (≤30% MVC). No statistically significant differences were observed at intensities ≥50% MVC. Interestingly, longer time to task failure was observed with holding isometrics for axial muscle groups. However, caution is needed here, as only two studies were included in the sub-analysis.
Similarly, greater fluctuations, RPE scores, mean arterial pressure, heart rate, EMG amplitudes, and burst rates were found for holding isometrics. One explanation for such findings may be that holding isometrics require greater central processing (e.g., greater supraspinal and spinal activation). This may result in greater motor-unit recruitment and increased rate coding (one, two, three, four, five, six).
Additionally, due to the nature of the task, holding isometrics likely rely on more complex control strategies (e.g., dynamic and reactive control to potential perturbations). That may require greater responsiveness and sensorimotor modulation. Conversely, pushing isometrics rely on preplanned, voluntary cortical control. This highlights different motor control strategies between the two isometric types. It may also be one of the reasons why they differ in the amount of fatigue they generate.
Moreover, some of the differences in time to task failure between the pushing and holding isometrics may be due to relative intensity being lower with holding isometrics. In almost all studies, holding isometric MVC was derived from pushing isometric MVC and estimated to be ~77% of pushing isometric MVC.
The majority of research on isometrics, as well as most applied use, has focused on pushing isometrics. However, it may be worth considering in which scenarios holding isometrics might be a better option. For example, it is well established that many sports injuries, such as anterior cruciate ligament and hamstring tears or strains, occur via non-contact mechanisms (one, two). Typically, but not always, these injuries occur during muscle lengthening under load or when the demands are placed upon the muscle to hold or decelerate. Thus, practitioners may wish to consider including strength assessments, but also training that uses holding isometrics. This may potentially help in detecting muscle imbalances or deficits that could be relevant to injury risk. That said, the idea that holding-type assessments would or could directly predict injury risk is hypothetical.
At the time of writing this article, I’m only aware of a single study by Oranchuk and colleagues, which has not been published or preprinted yet, that explored the effects of pushing versus holding isometrics on knee extensor strength and hypertrophy. The findings were presented at a conference, so what I’m discussing has been publicly available for a while now. Using a within-participant design, 10 recreationally trained participants performed either pushing or holding 20-second isometric contractions with one of their limbs at 70% MVC for 6 weeks.
Strength assessments were performed for concentric, isometric, and eccentric torque changes. Morphological measurements included muscle thickness of all four heads of the quadriceps femoris, pennation angle, and fascicle length of the vastus lateralis and patellar tendon thickness. All the outcomes within both conditions increased following the 6-week training intervention. However, no significant differences were found between conditions. Strength and muscle thickness values (e.g., rectus femoris) did seem to slightly favor holding isometrics. I won’t go into interpretation of these findings at this time as I have no additional information about the study or the methods outside of what I mentioned here.
It’s worth bearing in mind that the majority of the studies included in a meta-analysis by Oranchuk and colleagues used lower loads than what is typically considered as an effective training stimulus for the goals most readers here are interested in (e.g., strength and hypertrophy).
When considering which specific type of isometrics to implement within your training routine, it may depend on the phase of the training cycle or the goal you’re pursuing. For example, pushing isometrics may be better suited for strength development due to a greater ability to produce maximal force. Conversely, holding isometrics may be more relevant in scenarios where local muscular endurance is important. They may also be relevant in sports/activities with greater stability requirements, where you’d hold a certain position (e.g., think iron cross in gymnastics). Finally, they may be useful at the very early stages of rehabilitation, as an introduction to implementing isometrics within a program.
Alongside pushing and holding isometrics, there’s recently been a growing interest in a hybrid form of isometric training termed eccentric quasi-isometrics (EQIs). Conceptually, EQIs start off as holding isometrics, where your intent is to resist an eccentric muscle action or further lengthening to occur. However, due to the onset of fatigue at a given joint angle (usually shorter muscle lengths), the demands to hold this isometric position are increased until “holding failure” is reached and a low-velocity eccentric muscle action commences whilst still being actively resisted.
EQIs, a term apparently coined by Verkhoshansky and Siff, have been described as “holding a position until isometric failure and maximally resisting the subsequent eccentric phase.” This additional eccentric component has been proposed to place the muscle under an additional loaded stretch. In turn, this may result in greater mechanical tension, and thus a greater hypertrophy response. In addition, EQIs have been proposed as an alternative training modality to traditional resistance training for hypertrophy.
Specifically, it’s been theorized that EQIs may result in greater motor unit recruitment, mechanical tension, time under tension and co-activation. In addition, they may result in less fatigue and muscle damage. Because EQIs are a low-velocity, submaximal exercise, it’s been suggested that any increases in strength are likely to be mediated by architectural or morphological changes (e.g., fascicle length and torque–angle shifts). It has also been hypothesized that EQIs might present a novel stimulus for increasing muscular endurance. This is thought to occur due to constant tension during the isometric portion of the exercise, resulting in temporary alterations in oxygenation (i.e., deoxygenation) that may lead to activation of different anabolic signaling factors.
While there’s slightly more evidence on pushing and holding isometrics, the evidence on EQIs is scarce. Much of their theorized benefits are based on indirect or limited data. To date, only a few studies (one, two, three, four, five, six) have explored the effects of EQIs on outcomes such as neuromuscular fatigue, force output, muscle damage, and range of motion, with the majority of them being acute studies.
For example, Oranchuk and colleagues found that when total angular impulse was matched, EQIs caused less muscle soreness and smaller reductions in peak torque and rate of torque development than eccentric-only training. Changes in muscle swelling and echo intensity of the quadriceps femoris were similar between conditions. Henderson and colleagues found that EQIs elicited greater time under tension and slower set-to-set fatigue than traditional isotonic elbow-flexor training. In that study, they also observed that females accumulated more time under tension and exhibited smaller performance drop-offs than males during EQI sets. This suggests that sex-specific responses may be particularly relevant when prescribing EQIs.
To date only one study has compared the long-term effects of EQIs versus traditional resistance training on strength and hypertrophy. Henderson and colleagues compared 8 weeks of EQIs with isotonic resistance training on strength and hypertrophy in 22 untrained young men and women. EQIs involved participants flexing their elbows to 130° with a supinated grip and holding a dumbbell in this position until failure. Once they could no longer hold the position, they were instructed to maximally resist the lengthening portion of the movement (eccentric muscle action) until reaching full elbow extension. Conversely, traditional resistance training involved performing a full ROM with a prescribed tempo (1-second concentric; 2-second eccentric). Loads corresponded to 70% of the estimated one-repetition maximum (1RM) in both conditions. Participants trained three times per week, with sets gradually increased after the 7th and 15th session. Total time under tension in the EQI condition was manually timed using a stopwatch, while the number of repetitions was counted by the researchers for the isotonic condition.
Traditional resistance training (6.7 ± 3.9%) resulted in greater muscle thickness changes compared to EQIs (4.0 ± 3.3%). Greater increases in estimated 1RM were also observed with traditional resistance training (19.6 ± 8.5%) compared to EQIs (12.8 ± 6.2%).
The researchers also explored whether there were sex differences in the outcomes. As mentioned, their prior work suggested females may be more fatigue-resistant during EQIs and able to accrue a greater number of repetitions and/or time under tension compared to males. However, no apparent sex differences were found in this study. This finding is consistent with a recent meta-analysis by Refalo and colleagues, which reported similar relative increases in hypertrophy between the sexes.
The greater effects observed with traditional resistance training may have been due, in part, to EQIs being performed, on average, at shorter muscle lengths. More importantly, they may reflect a lack of adequate progressive overload. Participants trained at home, and the primary progression model was an increase in the number of sets. While increases in training volume have been shown to result in robust hypertrophy throughout an intervention (one, two), the fact that they kept the relative load constant throughout the study may have resulted in relative intensity being too low for the EQI condition for meaningful growth to occur.
Tendons, pain, and rehabilitation
I’m sure you’ve heard people say that isometrics are good for “rehabilitating tendons,” “reducing pain,” or “improving a sticking point in a lift.” Indeed, previous research has reported reduction in pain perception following exercise, a phenomenon known as exercise-induced hypoalgesia. Moreover, resistance training is considered one of the most important interventions in the conservative management and rehabilitation of tendinopathy. Interestingly, isometric training has been suggested as an effective alternative to dynamic training (one, two). This is because it minimizes external mechanical stress that would otherwise result from dynamic movement and allows for a more controlled application of force across the joint. Management of various tendinopathies typically comes down to load management (one, two, three) and gradually building up exposure and tolerance. However, tendons can be prone to regression. Isometric training seems like an effective strategy to alleviate some of these concerns, at least in the early stages of rehabilitation.
Rio and colleagues reported significantly greater pain relief following isometric exercise compared to isotonic exercise in participants with patellar tendinopathy. Moreover, this reduction in pain was shown to be sustained at 45 minutes following the acute session with isometric exercise, but not isotonic exercise. However, these findings have been criticized for promoting isometrics as a form of acute pain relief, which may misguide many into thinking there’s a “hack” to tendinopathy management. Authors also emphasized prior evidence reporting functional deficits and altered muscle activation in participants who were pain-free but without being fully recovered. Importantly, the findings from Rio and colleagues haven’t been replicated, so they should be taken with a grain of salt.
A systematic review by Clifford and colleagues reported that isometric training does not appear to provide greater effects for managing chronic tendinopathy compared to other intervention modalities (e.g., ice therapy or isotonic training). Similarly, a 2021 systematic review by Bonello and colleagues reported lack of evidence for a clear exercise-induced hypoalgesia effect following acute isometric training. Out of 13 studies, only 5 studies found a significant effect on pain pressure threshold following isometric exercise, and positive effects were mostly observed with quadriceps exercises.
Interestingly, some studies reported that isometrics may even worsen symptoms in individuals with fibromyalgia. However, these negative effects were not observed in the review by Bonello and colleagues. Collectively, both reviews reported substantial heterogeneity between studies, limiting any further quantitative analyses. This included differences in tendinopathy type or pain assessment location, training protocols, comparison conditions (notably the lack of control groups), small sample sizes, and a limited overall number of studies.
The reviews by Clifford and colleagues and Bonello and colleagues included only 10 and 13 total studies, respectively. Moreover, studies were generally rated as being of low to moderate quality. Thus, there is a clear need for more high-quality research to better understand how isometric training should be programmed and applied in these contexts.
With all that said, isometrics are often prescribed in rehabilitation settings following injury or surgery. They provide an efficient training stimulus while transitioning from passive to a more dynamic phase of rehabilitation and are generally well tolerated. As such, isometrics can be incorporated as a form of specific progression once an individual has adapted to a more passive form of training. Importantly, as discussed earlier, both pushing and holding isometrics can be performed and at various muscle lengths.
Because isometrics at shorter muscle lengths have been reported to result in less muscle damage, at least acutely, until the repeated-bout effect kicks in, one approach may be to introduce isometrics at shorter muscle lengths first.
From there, you could gradually progress to longer muscle lengths, followed by the introduction of more dynamic movements as one builds up tolerance and recovery improves. Taken collectively, and based on the current evidence, it appears that any form of training (i.e., isometric or dynamic exercise) that allows for progressive increases in stimulus or load, regardless of the muscle action, can be an effective training strategy. However, training should be highly individualized and introduced based on pain progression or relief.
As far as tendons go, they are viscoelastic structures that transfer force from muscle to bone that allows joint rotation to occur. It’s well documented that tendons, similar to muscles, undergo a series of adaptations in response to mechanical overload. These adaptations are driven by changes in collagen synthesis and turnover within the tissue (one, two). Primarily, regular mechanical loading can lead to increases in tendon stiffness, Young’s (elastic) modulus, or CSA.
Some of these tendon properties are particularly important. For example, greater tendon stiffness has been associated with greater force production and rate of force development, faster sprint times, and reduced risk of muscle damage. Overly compliant tendons, on the other hand, may result in less efficient force transfer and could potentially increase injury risk.
A meta-analysis by Lazarczuk and colleagues explored: 1) the overall effects of training on tendon adaptations, 2) which tendon properties primarily adapt to training, and 3) moderating effects of variables such as training intervention, age, muscle action, volume, and intensity. For those interested, Greg wrote a detailed article about how tendons adapt to training stimulus and discussed this meta-analysis in more depth. To keep things brief, resistance training that places high strain on the tendons appears to have the largest effect on tendon adaptations, with tendon stiffness being the primary one.
Interestingly, increases in tendon stiffness seem to be driven mainly by increases in elastic modulus rather than CSA. As far as muscle actions, no significant moderating effect was observed. Similar tendon adaptations were seen across muscle actions, with the exception that eccentric muscle actions may result in greater increases in tendon CSA.
As for specific isometric training variables relevant to tendon adaptations, Oranchuk and colleagues reported tendon CSA and/or stiffness were positively impacted by high-intensity (≥ 70% of MVC) isometric training, contraction duration, and training at long muscle lengths. In contrast, low-intensity (≤ 70% of MVC) isometric training may result in greater tendon elongation. Moreover, Kubo and colleagues reported that distal tendon and deep aponeurosis elongation decreased following isometric training at long muscle lengths, whereas training at short muscle lengths resulted in a trivial increase in distal tendon and deep aponeurosis elongation. Thus, the idea that tendons respond best to heavy, but controlled, mechanical overload appears to hold true.
With rehabilitation and tendons covered, let’s zoom back out and finally look at how isometrics stack up for muscle hypertrophy and strength development.
Hypertrophy and strength
Even though isometrics aren’t commonly performed when training specifically for hypertrophy, there’s a substantial amount of evidence supporting their effect (one, two, three, four, five, six, seven, eight, nine, ten). For example, 6 to 14 weeks of isometric training have been reported to result in 5.4-23% increases in muscle CSA.
A 2019 systematic review by Oranchuk and colleagues explored the effects of isometric training on hypertrophy, tendon morphology, neuromuscular adaptations, and performance enhancement. Importantly, they also explored evidence on various training variables and their effects on these adaptations. This is important because, as is the case with traditional resistance training, manipulating training volume, joint angle, and contraction intensity have been shown to affect the degree of hypertrophy that occurs with isometric training.
Muscle lengths
We’ll start off with muscle lengths as they’ve been a hot topic recently, and they play a significant role in the magnitude of hypertrophy observed with isometric training.
At the time of the review, only three studies (one, two, three) compared isometric training at different joint angles on muscle volume or thickness. If you’ve been following the literature on muscle lengths closely, this finding likely isn’t too surprising. Isometric training at longer muscle lengths, on average, seems to result in greater hypertrophy (1.16 ± 0.46%) compared to isometric training at shorter muscle lengths (0.47 ± 0.48%). While three studies are not much, this does at least preliminarily give credence to the idea that training at longer muscle lengths matters.
Two new studies have been published since the paper by Oranchuk and colleagues in 2019. A study by Akagi and colleagues compared 8 weeks of isometric training of the tibialis anterior at short versus long muscle lengths on muscle architecture. Training involved 3 sets of 8, 9, or 10 isometric dorsiflexion MVCs performed three times per week. Five-second MVCs were performed at 0° (shorter muscle lengths) or 40° (longer muscle lengths) ankle joint angles. Participants were instructed to contract as fast and forcefully as possible. Researchers reported that both conditions increased muscle thickness and pennation angle throughout the intervention, with no statistically significant differences between them. However, greater fascicle length changes were observed only for the longer muscle length condition following 8 weeks of training. That isn’t too surprising, as evidence suggests fascicle length changes tend to be greater following training at longer muscle lengths (more on that later). The authors also assessed changes in optimum torque angles and observed similar changes between conditions. However, greater changes in fascicle length in the long muscle length condition corresponded with a wider plateau region of the torque–angle relationship.
Nakao and colleagues investigated 8 weeks of knee flexion training at shorter versus longer muscle lengths. Two groups of participants performed low-intensity unilateral training. One group trained at 30° knee flexion (longer muscle lengths) and the other trained at 90° knee flexion (shorter muscle lengths) three times per week. Participants trained in a supine position with their hip joints flexed to 90°. The targeted torque was 30% of MVC at the respective joint angles. Training involved 5 sets of 20 five-second contractions.
CSA of the biceps femoris (long head), semitendinosus, and semimembranosus was measured via ultrasound at proximal, intermediate, and distal regions. Passive torque and MVC torque at 30° and 90° were also assessed using an isokinetic dynamometer. The authors quantified passive torque because previous studies suggest that adaptations from training at longer muscle lengths may be influenced by passive as well as active tension. Both groups displayed similar, but small, increases in hamstring CSA across all three regions, with no significant differences between groups. The researchers attributed this to the relatively low training intensity.
MVC torque at both 30° and 90° increased in both groups. However, greater increases were observed at 90° for the group training at short muscle lengths. Active torque during training was significantly lower in the group that trained at longer muscle lengths. This is most likely due to differences in training specificity. Specifically, the authors noted that the larger increase in MVC at 90° may reflect greater neural adaptations in the group training at short muscle lengths, due to higher active torque production.
The biggest limitation of the literature at the moment is that studies generally compare training effects at relatively shorter versus longer muscle lengths. None of the studies performed training near maximum muscle lengths. In fact, in our recent meta-analysis (which also included a majority of the isometric studies mentioned here), we found that, across the muscle length literature on regional hypertrophy, the difference in average muscle length between the groups or conditions was only ~21.8%. That’s pretty small and may not be enough to actually make a practical difference.
In one of the sub-analyses (one, two, three, four), we explored the effects of specific muscle actions on regional adaptations. Although the estimated slope of muscle length tended to increase from proximal to distal sites, due to only a handful of studies being included in the analyses, and wide 95% quantile intervals associated with the estimates indicating high uncertainty, we advised caution and the need for more research on the topic.
If we consider the length-tension relationship, some have proposed that training might ideally be performed near the plateau region of the curve. This is because at those sarcomere lengths there is “optimal” overlap between actin and myosin (contractile filaments), which maximizes active force capacity at those lengths. In contrast, active force falls off at the very short and very long ends of the curve where overlap is either excessive or insufficient. Therefore, training at the plateau region for a specific muscle group might result in greater hypertrophy than training at shorter muscle lengths.
The active length-tension curve of an individual sarcomere
Would it actually outperform training at longer muscle lengths? We have no idea. At the moment, there is no published research that has directly compared “plateau-only” training with both shorter and longer muscle lengths under well-matched conditions. Additionally, I’d be cautious with making inferences from biomechanical models to real-world training, because the length–tension curve (and thus the exact plateau region) can differ quite a bit between individuals.
Architectural adaptations
As far as muscle architecture changes go, several studies (one, two, three, four, five) have explored the effects of isometric training on pennation angle and/or fascicle length changes. For example, Noorkõiv and colleagues compared isometric training at shorter versus longer muscle lengths. Researchers measured vastus lateralis fascicle length at proximal, middle, and distal regions and rectus femoris fascicle length in the middle region. Following a 6-week intervention, the leg performing isometrics at shorter muscle lengths showed greater vastus lateralis fascicle length changes in the middle region compared to the leg performing training at longer muscle lengths (5.6% versus 3.8%, respectively). Conversely, a significant increase in vastus lateralis fascicle length at the distal region was observed after training at longer (5.8%) but not shorter (-1.1%) muscle lengths. Rectus femoris fascicle length did not significantly change in either condition.
An 8-week study by Alegre and colleagues compared isometric knee extension training at shorter and longer muscle lengths. Researchers reported greater increases in vastus lateralis pennation only in the group training at longer muscle lengths (11.7 ± 14.7% vs. 7.3 ± 10.2%). Interestingly, although no significant differences were reported for vastus lateralis fascicle length changes, the long muscle length group experienced greater increases (4.2 ± 12.7%). Conversely, the short muscle length group experienced a decrease (-0.3 ± 12.5%).
As mentioned previously, Akagi and colleagues observed greater fascicle length changes in the condition that trained at longer muscle lengths. However, similar increases in pennation angle were observed for both conditions.
When discussing architectural adaptations between muscle actions, it’s often argued that eccentric muscle actions result in larger increases in fascicle length (an indirect proxy for the addition of sarcomeres in series), whereas concentric actions tend to result in greater changes in pennation angle (an indirect proxy for the addition of sarcomeres in parallel).
However, as discussed by Blazevich and colleagues recently, it seems that producing high force at longer muscle lengths, rather than eccentric muscle action itself, may be the main stimulus for sarcomerogenesis. That conclusion sparked somewhat of a back-and-forth with Power, Franchi, and Hinks (another highly prominent group of researchers on the topic) about whether eccentric training is, or is not, the primary driver of sarcomere addition. In their response, Blazevich and colleagues showed the difference in average serial sarcomere number (SSN) change across interventions versus controls. Typical eccentric training protocols in animal models produced ~1–5% SSN increases, whereas electrically stimulated isometrics performed at longer muscle lengths produced ~32% increases in SSN. In other words, electrically stimulated isometrics performed at long muscle lengths appear, at least in these animal models, to be a much more potent stimulus for the addition of sarcomeres in series than eccentric muscle actions.
So, it goes without saying: muscle length likely is an important factor to consider. Perhaps even more important than the muscle action itself. For example, prior studies (one, two) comparing eccentric training at shorter versus longer muscle lengths reported greater increases in fascicle length when training was performed at longer muscle lengths. That lines up with what we discussed earlier regarding isometric training at longer muscle lengths tending to result in greater fascicle length increases than training at shorter muscle lengths.
Training volume and intensity
In the context of regular, dynamic resistance training, we most frequently quantify volume as the number of hard sets performed for a muscle group per week. With isometrics, however, while number of sets can be used as one way to quantify volume, other ways include: a) total contraction duration (time under tension); b) total number of contractions; and c) the product of sets × repetitions × contraction duration.
Similar to what’s been recently demonstrated for dynamic resistance training by Pelland and colleagues, higher training volumes with isometric training may result in greater hypertrophy. If you want to go down the rabbit hole of training volume, strength, and hypertrophy, check out Greg’s article here. For example, Meyers compared low-volume (3 × 6-second contractions) to high-volume (20 × 6-second contractions) training at 100% MVC for 6 weeks. Larger increases in upper arm circumference were observed in the higher volume group. Measurements were taken with the arm relaxed, flexed at 90° for the trained arm, and also relaxed in the non-exercised arm. However, there were no significant between-group differences for isometric strength and muscular endurance, except that the higher volume group showed greater strength increases at 90° of elbow flexion.
Kanehisa and colleagues compared 10 weeks of unilateral isometric elbow-extension training using two volume-equated protocols. One group performed 12 × 6-seconds at 100% MVC per session. The other performed 4 × 30-second contractions at 60% MVC. Training volume was equated as %MVC × duration per set × sets per session. Muscle volume and anatomical CSA of the triceps brachii were assessed by magnetic resonance imaging (MRI). Fascicle pennation angle of the long head was measured by ultrasound. Isometric torque, as well as concentric and eccentric torque, were measured on an isokinetic dynamometer. Both groups had similar relative increases in CSA, fascicle angle, and dynamic torque. However, muscle volume increased more in the 100% MVC group (12.4%) than in the 60% MVC group (5.3%). Torque relative to muscle volume improved only in the 60% group.
With regard to training intensity, in isometric training it’s typically expressed relative to one’s MVC. When considering the available research, a similar pattern to dynamic training emerges. As long as intensity is high enough, or time under tension is equated, you could expect very similar hypertrophy to occur.
As mentioned, Kanehisa and colleagues found similar relative increases in CSA and fascicle angle when comparing low (60% MVC) versus high (100% MVC) training intensities. Greater muscle volume was observed in the 100% MVC group (12.4%) than in the 60% MVC group. Torque relative to muscle volume improved more in the lower-intensity condition. Oranchuk and colleagues highlighted that training intensities at ≤70% MVC resulted in 0.77 ± 0.26% hypertrophy per week, compared to 0.70 ± 0.55% per week when training was performed at >70% MVC, under broadly comparable loading.
They noted that for hypertrophy and strength, increasing contraction duration, total volume, or performing isometrics at longer muscle lengths are likely more important variables than small shifts in %MVC alone.
For strength, studies comparing shorter versus sustained contractions at lower versus higher intensities suggest that both shorter and sustained contractions can increase strength (one, two, three, four). However, shorter-duration, high-intensity contractions may be a more time-efficient option than long sustained contractions. This matters as many people cite lack of time as one of the main barriers to engaging in physical activity (one, two, three). If we consider the dynamic resistance training literature, given that high loads are usually the primary training variable when strength increase is the goal, it makes sense to perform shorter, high-intensity contractions.
One more thing to keep in mind is that we generally have fewer studies on the effects of training volume and intensity than we do with dynamic training. Thus, any inferences we make regarding them are a bit more tentative.
Contraction duration and intent
When discussing how long each contraction should be, we can generally categorize them as short (<10 seconds), moderate (20–40 seconds), and long (>40 seconds) contraction durations. The evidence on whether shorter contractions or those sustained for longer are better for strength and hypertrophy is not entirely clear.
Schott and colleagues compared short MVCs (4 sets × 10 contractions × 3 seconds) with long MVCs (4 × 30 seconds) for the knee extensors. Total time under tension (120 seconds) was equated between groups. The sustained-contraction group showed greater proximal and distal vastus lateralis hypertrophy and isometric strength compared to the shorter-contraction group. Even though the total time under tension was equated, the group that sustained their contractions for longer saw greater growth.
Conversely, Kubo and colleagues compared isometric knee extensions consisting of 3 sets of 50 rapid contractions (non-sustained; ~1 second each) versus four 20-second sustained contractions at 70% MVC. They found similar strength (31.8% vs. 33.9%) and hypertrophy (7.4% vs. 7.6%) changes. Although the group performing non-sustained contractions had a greater total number of contractions and a longer total contraction time than the sustained group (150 contractions and ~150 seconds vs. 4 contractions and 80 seconds), the changes were similar between groups.
One reason sustained contractions may sometimes result in greater adaptations is reduced blood flow to the muscle and acute ischemia. This occurs as blood vessels are compressed during active contraction. That may also result in slightly greater metabolite accumulation and increase the overall training stimulus. Thus, in terms of volume, it seems the magnitude of strength and hypertrophy depends on both the total number and duration of contractions within a session. It also depends on how much of that time is actually spent at a sufficiently high intensity.
Balshaw and colleagues compared sustained and explosive isometric contractions. Sustained isometrics involved performing contractions at 75% MVC by ramping the torque up to the target over 1 second. Participants then held that torque target for 3 seconds. Conversely, explosive isometrics involved performing short, explosive 1-second contractions to ~80% MVC as fast as possible and then resting. Both groups performed 4 sets of 10 contractions. The results showed greater increases in quadriceps muscle volume following sustained versus explosive isometric training (8.1% vs. 2.6%). Additionally, both groups experienced strength increases. However, greater changes were observed in the group performing sustained isometrics (23% vs. 17%). The greater hypertrophy observed in the sustained-contraction group may be due to greater total time under tension compared to the explosive-contraction group.
Massey and colleagues assigned 42 healthy, young, recreationally active men to an explosive-contraction training group (ECT, n = 14), a sustained-contraction training group (SCT, n = 15), or a non-training control group (CON, n = 13). Both training groups completed unilateral isometric knee-extension training three times per week for 12 weeks. Each session involved 4 sets of 10 contractions per leg (40 contractions per leg per session). In SCT, each contraction followed a target. Torque ramped linearly from rest to 75% of maximum voluntary torque (MVT) over 1 second. Participants then held that torque output for an additional 3 seconds, with 2 seconds of rest between contractions. In ECT, participants were instructed to contract “as fast and hard as possible” to reach about 80% MVT. They had 5 seconds of rest between contractions. After 12 weeks, both ECT and SCT increased knee-extensor MVT compared with control. However, strength gains were larger with sustained contractions (SCT: ~24%; ECT: ~17%). Quadriceps muscle volume increased significantly only in SCT (~8%). No clear hypertrophy was observed in ECT or in the control group.
Rest periods
Interestingly, there’s actually not much research to go off as it pertains to rest periods. They’re mostly considered only as part of comparisons alongside other training variables.
Waugh and colleagues explored whether manipulating only the rest period between contractions during heavy isometric plantarflexion would impact Achilles tendon adaptations. 14 healthy adults trained both legs three times per week for 12 weeks. Each session, they performed 5 sets of 10 3-second isometric plantarflexion contractions at ~90% MVC on an isokinetic dynamometer. One leg was assigned to a short-rest condition with 3 seconds of rest between contractions. The other leg was assigned to a long-rest condition with 10 seconds between contractions. Rest between sets was 90 seconds for both legs.
The researchers assessed plantar flexor strength, tendon stiffness, stress, strain, and Young’s modulus using an isokinetic dynamometer and ultrasound. They also assessed tendon collagen organization using ultrasound tissue characterization. Both conditions showed similar increases in maximal torque, Achilles tendon stiffness, and modulus after 12 weeks. The only apparent difference was that the proportion of well-organized collagen decreased in the short-rest leg. It was maintained or slightly increased in the long-rest leg. The authors suggested that while short rests do not seem to impact strength or tendon morphology, longer rests may favor maintaining tendon collagen organization in already healthy tendons.
This is also the only study I’m aware of that manipulated rest periods as a primary variable of interest. Overall, if we take a look at studies that perform isometric training, rest periods between individual contractions are usually somewhere between 2 and 10 seconds. That’s especially true when using shorter contraction durations (e.g., 3-5 seconds per contraction). Longer contraction durations are often a whole set anyways (i.e., a single repetition). If someone is doing something in the middle, say, 15-20 second contractions, perhaps double that would be needed for sufficient rest.
Between sets, most studies use around 60 to 180 seconds of rest. That’s typical and reflective of regular resistance training. As mentioned, given the lack of direct research on the topic, concrete practical recommendations are limited. Hopefully this provides you with a general idea and a good starting point which you can adjust based upon the rest of the training variables (e.g., volume, intensity, intent).
The biggest problem when trying to discuss specific training variables within the isometric literature is that there isn’t much to go off when only a single variable is manipulated while the others are held constant. This inevitably introduces confounders into the mix. The methods used to manipulate any one variable (e.g., volume, intensity, contraction duration, intent) also vary between studies.
From a practical standpoint, it’s probably safe to say that similar principles to regular resistance training apply. Muscle hypertrophy can be achieved using moderate to high training intensities as long as you accumulate sufficient “time under tension.” On average, performing contractions at longer muscle lengths also seems beneficial. If total time under tension and intensity are equated, longer sustained holds may be slightly more hypertrophic.
For strength gains, similar to traditional resistance training, higher intensities (80% to 100% MVC) with shorter repetition duration (1–5 seconds) do seem to, on average, result in greater strength increases compared to lower training intensities. As for total duration of contractions per session, Lum and Barbosa recommend aiming for roughly 80–150 seconds and 30–90 seconds when focusing on hypertrophy or strength, respectively.
A few things on strength development
When discussing training for maximal strength development, most are aware that it’s highly task specific. In line with the SAID (Specific Adaptations to Imposed Demand) principle, performing training that is more specific to your goals is usually suggested as a better modality for achieving them. However, that doesn’t mean less-specific tasks are inherently useless. For those interested, I highly recommend reading this article by Eric Helms going over some of the nuances here.
In regard to isometric training, you’ve probably heard that isometric training is very joint-angle specific (one, two). And indeed, previous studies suggest greater strength changes are observed when training is performed around the same joint angles that are used in the measured task.
However, there’s also a high degree of transferability to other joint angles. This seems primarily predicated on which joint angles isometric training is being performed at. For example, in a study by Alegre and colleagues, the group that performed training at longer muscle lengths showed greater increases in isokinetic peak torque following an eight-week intervention. Moreover, a shift in optimum angle toward longer muscle lengths was only found in the longer muscle lengths group (pre: 77.5 ± 7.9°; post: 88.5 ± 8.0°). The opposite was true for the group training at shorter muscle lengths (pre: 70.4 ± 6.1°; post: 65.1 ± 5.6°).
This is probably due, in part, to the addition of sarcomeres in series during training at longer muscle lengths. Shifts in the optimal torque-angle toward longer muscle lengths have previously been suggested to lower injury risk of the knee flexor and extensor muscles. The hypothesis is that sarcomeres may operate near the plateau region of the length-tension curve over a broader joint range of motion.
Similar to hypertrophy, greater strength increases across a broader range of joint angles were reported with isometric training at longer versus shorter muscle lengths. For example, Kubo and colleagues found greater strength increases across multiple joint angles when unilateral knee extension training was performed at longer (100° of knee flexion) versus shorter (50° of knee flexion) muscle lengths for 12 weeks. Specifically, training at shorter muscle lengths resulted in significant MVC increases from 40° to 80° of knee flexion. Training at longer muscle lengths led to greater MVC increases from 40° to 110° of knee flexion.
Conversely, Noorkõiv and colleagues found that training at shorter muscle lengths resulted in greater MVC increase near the training joint angle (30–50°). In contrast, there were no significant MVC increases at any joint angle when training was performed at long muscle lengths, despite greater observed hypertrophy. The reason for observed statistically significant findings may just be due to less variability in the group of participants training at short muscle lengths compared to the group training at long muscle lengths.
What about the transfer to dynamic performance?
While isometrics are interesting, the majority of your training is still likely to be performed dynamically. A recent meta-analysis compared the effects of isometric training versus non-training controls and dynamic resistance training on isometric and isokinetic muscular strength. Isometric training resulted in moderately larger increases in isometric and isokinetic strength (pooled outcomes) compared to non-training controls (SMD = 0.65; 95% CI [0.52, 0.77]) and dynamic training (SMD = 0.35; 95% CI [0.21, 0.48]). Additionally, greater isometric strength gains were reported for isometric training compared to dynamic training (SMD = 0.43; 95% CI [0.27, 0.59]). In contrast, there were no statistically significant differences between isometric and dynamic training for isokinetic strength (SMD = −0.20; 95% CI [−0.55, 0.14]).
Greater strength transfers were also observed when isometric training was performed at relatively longer muscle lengths, which may be due to shifts in torque-angle curve and thus, increased force capacity across a broader range of motion.
For strength specifically, there’s a practical use case for implementing isometric training. For example, lifters perform pause squats at the bottom or deadlift holds around mid-shin to overcome sticking regions where they typically fail a lift. If you wanted to increase your force production capacity at a specific joint angle more directly, you could set the bar against the pins and perform maximal pushing isometrics at that position. But that doesn’t mean you’re limited to a single joint angle. While the degree of strength transfer to broader joint angles seems greater when training at longer muscle lengths, you could also consider performing isometrics at multiple joint angles.
For example, Lum and colleagues randomized 16 male athletes into two groups. One group performed isometric bench press at 90° of elbow flexion. The other group performed isometric bench press at 60°, 90°, and 120° of elbow flexion. Participants performed a full-body routine, two times per week for six weeks. Isometrics were integrated into the training routine and made up half of the total bench press training volume. For example, three sets of isometrics and one set of dynamic bench press were performed during the first weekly session, while four sets of dynamic bench press were performed later in the week. This is due to a previous study reporting greater strength increases when isometric and dynamic training were combined versus dynamic training alone. Researchers assessed overhead throw, ballistic push-up performance, and bench press 1RM pre- and post-intervention. Both groups significantly increased bench press 1RM, both in absolute terms and relative to body mass. However, no significant between-group differences were observed.
Greater peak power during ballistic push-up performance was observed in the group performing isometrics at multiple joint angles. However, there were no statistically significant differences between the groups. The researchers hypothesized that these marginal improvements may have been due to greater improvements in force production capacity across not only the trained joint angles, but also a broader range of motion. Because the push-ups were performed from a fixed bottom position without the use of a stretch-shortening cycle, participants had to continuously accelerate and push themselves off the ground. Finally, neither group significantly improved their overhead throwing performance (~1%) which may just be due to training being less specific to the task being tested.
Based on the above, isometrics can provide some versatility in training. Interestingly, some have also suggested isometrics may be less fatiguing compared to dynamic training (one, two, three). However, does that mean you should replace all of your training with isometrics? Probably not. While joint-angle specificity is one of the primary advantages of isometric training, it’s also not a be-all-and-end-all type of situation. It’s worth keeping in mind what kind of adaptations you’re aiming to achieve.
For example, Lum and colleagues compared six weeks of isometric and plyometric training on countermovement jump and isometric-mid thigh pull performance. Both groups increased jumping performance (PLYO: ~11% versus ISO: ~9%). However, the isometric group showed greater strength increases (11.5% versus 4.8%). The plyometric group improved countermovement jump height alongside a moderate increase in reactive strength index modified (jump height / time-to-takeoff). They also showed a shorter unweighting phase and longer propulsion phase, suggesting a more efficient use of the stretch-shortening cycle.
Conversely, the isometric group displayed a longer propulsion phase and slightly greater increase in countermovement depth. Combined with greater isometric mid-thigh pull peak forces, this may have resulted in a greater propulsive impulse.
While some prior evidence suggests that isometric training may result in smaller jump performance increases compared to plyometric or dynamic training (one, two, three), context matters. You need to consider: a) the type of task you’re performing, and b) how you’re implementing isometrics. For example, you could use isometrics to improve your force capacity of the hip, knee, and ankle extensors from a squatting position. Ideally, this would closely mimic squat jump starting position (~90° of knee flexion). Indeed, previous studies (one, two) have shown squat jump improvements following isometric training.
However, if we consider countermovement jump, which relies on the stretch-shortening cycle, dynamic training might be more goal-specific. That’s not to say that isometric training can’t also improve ballistic type tasks like the countermovement jump. Those that failed to observe improvements often used single-joint exercises. Conversely, those using multi-joint exercises (e.g., isometric squat) did. Thus, how effective they are in practice depends on how isometrics are performed (e.g., explosively) and exercise selection (single-joint vs. multi-joint).
This also doesn’t mean you can’t combine the two. As mentioned earlier, combining isometric and dynamic training has been shown to result in greater jump performance compared to dynamic training alone. However, if your primary goals revolve around compound lifts like the squat, bench press, or deadlift, you’re likely still better off using dynamic training as the main variation of the lift. That said, it might be worth giving isometrics, or even multi-joint isometrics, a shot if you’re struggling with a specific sticking point during a lift.
Isometric training versus isotonic (dynamic) training
Now that we’ve cleared the air about isometric training being an effective modality for increasing muscle strength and size, the last question that remains is: how comparable is isometric training to dynamic training for muscle growth? Spoiler alert: plenty.
To date, there have been several studies that directly compared isometric to dynamic training. One of the first human studies to compare isometric and dynamic training on muscle growth was conducted by Rasch and Moorehouse in 1957. Researchers had 49 physically active young men perform unilateral isometric or isotonic elbow flexion and shoulder press for six weeks. Hypertrophy of the upper arm was assessed via circumference. Strength of the elbow flexors and shoulder muscles was assessed via a strain gauge dynamometer. Researchers also assessed non-specific strength changes by having participants perform an isometric strength test in a supine position at 100˚ of elbow flexion.
The isotonic group performed 3 sets of 5 elbow flexions and shoulder presses with a three-minute rest interval. The isotonic group took, on average, 15 seconds to complete a set. Loads corresponded to two-thirds of their 1RM. A similar approach was applied for the isometric group. Participants exerted force for 15 seconds at two-thirds of their MVC. They performed the elbow flexion task at about ~100° of elbow flexion. They performed the shoulder press task in the same shoulder-height starting position used during the strength tests (upper arm down by the side, elbow flexed, neutral grip).
Reported strength changes in direct strength tasks were greater in the isotonic group compared to the isometric group following a six-week intervention and a training cessation period. Greater arm circumference change was reported in the exercised arm of isotonic versus isometric group (1.22cm vs 0.59cm) after six weeks. Interestingly, following six weeks of training cessation, there was a greater decrease in arm circumference in the isotonic group (-0.60cm). However, there were no statistically significant changes in the isometric group (0.09cm). As a result, total gains maintained were similar in both groups (0.62 for isotonic vs. 0.68 for isometric).
Fukunaga and Sugiyama (1978) had 14 untrained young men randomized to either an isometric or isotonic group. Participants performed unilateral elbow flexion with their dominant arm for 12 weeks. The isotonic group trained from 75˚ to 105˚ elbow flexion. The isometric group trained at 90˚ of elbow flexion. Maximal isometric strength similarly increased in both groups (isometric: 25.9 ± 4.2%; dynamic: 30.8 ± 3.8%). However, greater 1RM strength gains were reported in the dynamic group (+33.1 ± 4.0% vs. 22.7 ± 1.1%). Elbow flexor CSA was measured via B-mode ultrasound. The authors reported that CSA increased in both groups (isometric: 5.4 ± 1.4%; dynamic: 3.2 ± 1.3%). No statistically significant differences were observed between groups.
The next study was published almost 30 years later by Kubo and colleagues. Researchers had 10 healthy young men perform isometric training with one leg at 90˚ of knee flexion. They performed full range of motion knee extension with the other leg, from 90˚ to 0˚ of knee flexion, for 12 weeks. Isometrics involved performing ten 15-second duration contractions at 70% MVC. The dynamic condition performed 5 sets of 10 repetitions with 80% 1RM. Quadriceps muscle volume was estimated via CSA measurements taken with MRI.
While changes in muscle volume slightly favored the dynamic condition (5.6% vs. 4.5%), there were no statistically significant differences between the conditions. Changes in isometric strength tended to favor the isometric condition (49% vs. 31.5%). However, this was not statistically significant (p = 0.056). This is not surprising given the strength task more closely reflected the isometric protocol. The findings are very much in line with the principle of specificity.
A study by Malas and colleagues might be less relevant to this audience, but it’s still necessary for discussing broader literature on the topic. Researchers had 61 older adults (51 women and 10 men) with diagnosed bilateral tricompartmental knee osteoarthritis randomized into six groups. They performed unilateral isometric, isotonic, or isokinetic training. Participants attended supervised sessions five times per week for three weeks. Each session included 20 minutes of hot packs and 10 minutes of therapeutic ultrasound to both knees, followed by the unilateral training protocol.
Strength was assessed via isokinetic bilateral knee extension and flexion at 60˚/s. Architectural changes of the vastus lateralis (muscle thickness, pennation angle, and fascicle length) were assessed via B-mode ultrasound. Strength and architectural changes were assessed in both exercised and non-exercised legs to explore potential contralateral effects. Knee extension strength increased just ever so slightly in all three groups. However, there were no statistically significant differences between them. Researchers reported that only the isometric group had a significant pre- to post-intervention strength increase. However, that may have been due to unbalanced baseline strength levels, since the isometric group started off weaker.
A two-way ANOVA was used to assess these changes. A more fitting statistical model here might be ANCOVA with baseline strength as a covariate, to take these baseline differences into consideration. Knee flexion strength remained about the same. Muscle thickness increased in all three groups for both legs, except for the non-exercised one in the isometric group. No change to very minimal changes were reported for the pennation angle.
One interesting finding, however, was that fascicle length changes were much greater in the isometric group. Very small to no changes reported in the other two groups for the exercised leg. What makes these findings interesting is that isometrics were performed at shorter muscle lengths, with participants contracting at full knee extension (0˚ knee flexion). As mentioned earlier, if you’ve been keeping up with the muscle length research these last few years, you might have heard that training at short muscle lengths may predominantly cause radial hypertrophy, while training at long muscle lengths may cause more longitudinal hypertrophy. Specifically, pennation angles and fascicle lengths are often used as proxies for these distinct types of hypertrophy, respectively.
A recent systematic review by Wolf and colleagues concluded that the evidence on whether training at shorter versus longer muscle lengths induces differential architectural adaptations is mixed. Studies report conflicting findings and also have some methodological drawbacks (e.g., using linear extrapolation to estimate fascicle lengths). Those were also a part of the study by Malas and colleagues. For those interested, there’s a Stronger By Science article that breaks it down in far more detail.
Overall, I’d be very cautious when interpreting this study. Again, it’s less relevant to healthy populations, it’s only three weeks in duration, and effort wasn’t clearly described. I’d be very surprised if training in any of the groups was reflective of how you’d usually train (based on the description, it wasn’t).
A 2017 study by Lee and colleagues compared isometric, isotonic, and isokinetic training on lower body strength and lean mass gains. Participants were distributed into three groups according to their baseline MVC values. 31 untrained participants completed an eight-week intervention. The authors performed a series of strength tasks, specific to each muscle action, pre- and post-intervention. They used dual-energy x-ray absorptiometry (DXA) to assess lower body lean mass changes.
Participants trained only with their dominant limb. Training sessions included 4 sets of 10 repetitions and 1-second contractions performed with 75% MVC (relative load equated between groups). Isometric training was performed at four different knee joint angles (90˚, 70˚, 50˚, 30˚ knee flexion). Participants had a minimal torque target they were required to achieve. The isotonic group performed their repetitions in a 1-second concentric/eccentric fashion, while the isokinetic group trained at 90˚/s angular velocity.
The authors reported that all groups increased their maximal isometric torque throughout the intervention. Significant findings were only observed when torque values of all angles were pooled together. In that case, the isometric group displayed greater strength increase compared to the other two groups. Isotonic and isokinetic strength tests also revealed similar increases between the groups, with values favoring the respective muscle actions.
Lower body lean mass increased in the isometric (3.1%) and isotonic (3.9%) groups but not in the isokinetic group (not reported). While the authors stated that isotonic training resulted in greater changes in lean mass, caution is warranted. They used typical measurement error established in another study, and then multiplied it by two, to argue that any change greater than that value was considered a “real” change. Moreover, based on how they described their procedures, it seems lean mass was derived from the whole lower leg rather than the upper leg compartment which can confound the results. Lean mass outcomes are also reported very briefly in text, without additional data presented in tables or figures to verify those changes. I’m not implying anything suspicious occurred. However, I’d still urge caution when interpreting the results just due to those reasons.
Kruszewski and colleagues had 20 young, resistance-trained men (at least two years of training experience) perform unilateral elbow flexion training for seven weeks. Strength was assessed via MVC at 30˚, 60˚, and 90˚ of elbow flexion, as well as dumbbell preacher curl 1RM on a Scott bench. Body composition was assessed via multi-frequency bioelectric impedance (BIA). All measurements were performed pre- and post-intervention.
Participants performed two training sessions per week. Both conditions were equated for total set duration. Isometric training involved performing 75% MVC for 6 sets of 10-second duration at 30˚, 60˚, and 90˚ of elbow flexion. Contractions were performed twice at each joint angle (total of 60 seconds). Conversely, participants in the isotonic condition performed 6 sets of 10 repetitions with 75% 1RM, using a three-second concentric and eccentric tempo.
Researchers reported that the only statistically significant between-group differences were observed for MVC values at 30˚ and 90˚ of elbow flexion. Both groups significantly increased their isometric strength at 60˚ from pre- to post-intervention, with no statistically significant differences between them. Researchers also reported no statistically significant differences between the groups for 1RM values. However, if we look at the reported data, neither isometric (pre: 22.85 ± 3.32; post: 25.55 ± 2.75) nor isotonic (pre: 22.80 ± 5.24; post: 25.75 ± 3.34) training really resulted in meaningful strength increase over seven weeks. Greater strength changes observed in isometric tests with isometric training is not surprising given the greater task specificity.
However, body composition data should be taken with a grain of salt. First, researchers used BIA. That’s an indirect method of estimating body composition and is often used as a proxy for muscle hypertrophy. These devices work by sending a weak electrical current through the body (often hand-to-foot or foot-to-foot, depending on the device). They use the measured impedance to estimate lean mass and fat mass. But keep in mind it’s an estimate, and hydration status and food intake can impact the results.
Thus, one should be very cautious when interpreting the results, especially when discussing hypertrophy. BIA isn’t able to detect extremely small magnitudes of change at the muscle fiber level. With all that said, compared to strength, body composition data is just not that clearly reported, outside of a brief in-text mention. Researchers reported that “active body mass”, which I believe refers to the segmental mass of the upper limbs, didn’t significantly change from pre- to post-intervention in either condition. The same was true for body fat percentage. All in all, I’d say this is one of the methodologically weaker studies on the topic. I’d urge caution when interpreting the results.
We recently published the first study to explore isometric training at longer muscle lengths versus isotonic, full range of motion training. This study is a part of my PhD thesis looking at the effects of varying muscle lengths and muscle actions on regional hypertrophy. Outside of that, it plays an important role in filling in the gaps in muscle length research.
Over the last few years, training at longer muscle lengths – and, as an extension of that, lengthened partials – has become extremely popular, but also controversial. People disagree about whether lengthened partials result in greater hypertrophy than full range of motion training. One hypothesis suggests that this advantage stems from the increased passive – and potentially total – tension the muscle experiences at longer lengths.
As mentioned earlier, a muscle’s ability to produce active force depends on its length. This is described by the length-tension relationship (one, two). In short, it describes how force production varies across different muscle lengths (and, in practice, across different joint angles). This occurs because the arrangement of the contractile filaments changes as sarcomeres shorten or lengthen. Specifically, the length-tension relationship is often described as having three segments: a) an ascending limb, where sarcomeres are relatively short due to greater overlap between contractile filaments, b) a plateau region, where there’s an optimal overlap between contractile filaments, and c) a descending limb, where sarcomeres are lengthened and there’s progressively less overlap between the contractile filaments.
In general, muscle fibers produce the most active force near the plateau region (around their resting length) due to optimal overlap between contractile filaments. At much shorter or much longer sarcomere lengths, fewer or less effective cross-bridges can form. As a result, active force decreases. But this considers only active tension. There’s also a passive component to tension which increases with muscle lengthening. Many mention titin, a giant protein molecule, as one contributor to passive tension (one, two, three). It’s been suggested that this rapid increase in passive tension at longer muscle lengths may increase total tension (active + passive). This may partially explain why we observe greater hypertrophy when training at longer muscle lengths.
Additive effects of active and passive tension on total mechanical tension
However, training at the longest muscle lengths isn’t necessarily always reflective of greater total tension. For discussion surrounding that, check out a previous Stronger By Science article by Milo and Greg. It may also be that it’s not total tension per se that’s driving these effects. Instead, it may be that a different stimulus (passive tension) is imposed on the fibers and is sensed by different mechanoreceptors (e.g., titin). Indeed, different mechanoreceptors appear to respond to different stimuli. Some may be more sensitive to active tension (e.g., costameres, filamins), while others may be more sensitive to passive tension (e.g., titin).
Thus, mechanotransduction that occurs during training at longer muscle lengths may trigger additional distinct signaling pathways. At minimum, it may augment those already active from regular resistance training. Their additive enzymatic activity may ultimately increase rates of muscle protein synthesis and hypertrophy compared to regular resistance training alone.
Up until this point, a few studies have directly compared lengthened partials and full range of motion. They found either similar or slightly greater muscle growth with lengthened partials (one, two, three, four, five, six, seven, eight). The most up-to-date published meta-analysis comparing lengthened partials to full range of motion is one by Wolf and colleagues. Recently, I did my own Bayesian meta-analysis to check how newly accumulated evidence has shifted the initial estimates reported by Wolf and colleagues (standardized mean difference: –0.28; 95% CI: –0.81, 0.16). If you want to read about some of the methods and studies that I included or excluded, check out my Instagram post. But here are the overall results from studies that directly compared the two training interventions head to head: Pooled estimates essentially suggest negligible effects (standardized mean difference: 0.04) in favor of lengthened partials versus full range of motion. However, 95% credible intervals included zero and were also pretty narrow. This means you are likely to observe similar hypertrophy with one or the other.
Even though lengthened partials include, on average, training at longer muscle lengths, there’s still some shortening of muscle length occurring with those repetitions. This raises the question: Would isometric training performed at longer muscle lengths offer an even greater stimulus compared to full range of motion or lengthened partials?
Additionally, while there is a relationship between changes in joint angle and muscle length, the two likely do not correlate perfectly. Therefore, our study also was informative of whether range of motion or muscle length itself is the primary training variable we should be manipulating.
I know I get far too in-depth when the muscle length discussion starts, but this was a necessary detour. With that out of the way, what did we do?
We recruited 23 resistance-trained individuals (13 males and 10 females) and had their lower limbs randomized into either isometric training at long muscle lengths or isotonic, full range of motion training conditions. Participants trained two times per week for six weeks. Their training volumes gradually increased throughout the intervention (from 6 to 10 weekly sets). All training was performed unilaterally on an isokinetic dynamometer.
The isometric condition performed 30-second MVCs at long muscle lengths. Specifically, average knee joint angle was ~125°. While this likely isn’t maximal knee flexion angle and maximal muscle length, it’s pretty close. We originally planned to have participants perform both sessions to a maximum individualized knee flexion angle. However, due to dynamometer construction, we were limited in how far we could push their knees into flexion.
The isotonic condition was set to isotonic mode to reflect traditional resistance training. This leg performed full range of motion knee extensions for 30 seconds. Range of motion was set from ~125° of knee flexion to 10° of knee flexion. Thus, both conditions were equated for set volume and total time-integral (i.e., duration of exertion during the set).
The biggest challenge when comparing isometric and isotonic training is properly equating effort between them. There are multiple ways to approach this. We chose what we thought would be close enough between conditions and similar to what’s traditionally done in the gym. We instructed participants to perform isotonic training to an RPE of 7–10. We increased their torque threshold (load) once they surpassed the minimum RPE threshold. Regional hypertrophy of the quadriceps was measured using B-mode ultrasound at proximal, mid-belly, and distal sites. Measurements were taken for the anterior thigh (rectus femoris and vastus intermedius) and lateral thigh (vastus lateralis and vastus intermedius) composites.
We found that both conditions resulted in similar muscle hypertrophy for both the summed anterior (ISOM: 2.7% vs. 0.8%) and lateral thigh (ISOM: 1.1% vs. ISOT: 0.8%) regions as well as at individual muscle sites. One potential reason for the lack of a meaningful lateral thigh growth in either condition may be exercise selection. Several studies (one, two, three) reported, on average, greater growth in the rectus femoris and smaller in the vasti muscles following leg extension training.
For the anterior thigh, almost all regional measurements favored the isometric condition. We hypothesized we would observe greater distal growth in the anterior thigh following isometric training. This was based on previous studies reporting greater distal growth when training at longer muscle lengths. Interestingly, we found greater effects in the proximal region following isometric training (3% vs. 0.7%). Although the 95% high-density credible interval included zero, the posterior distribution of between condition estimates leaned toward ISOM with a high probability (82%). This suggests a practically relevant effect may be possible.
I know what you’re thinking: only a six-week study, who cares? Hear me out. This study was never intended to provide a definitive answer. It served more as a proof-of-concept study. The goal was to explore whether training exclusively with isometrics at longer muscle lengths can result in hypertrophy, adding to the other studies mentioned earlier. The decision to run a six-week study was mainly due to timing constraints. We wrapped up the last set of measurements just three days before Christmas.
I won’t really go much into head-gymnastics as to why we observed similar growth. That could just be due to usual concerns with regard to sampling variance. I’ll just briefly restate what we wrote in the paper and share a few of my thoughts on the matter relevant to the discussion.
It may be that there’s a duration and magnitude component to all this. I’m fairly confident that’s the case. While participants in the isometric group stayed at longer muscle lengths 100% of the time (on average), participants in the full range of motion condition only spent about half the time at longer muscle lengths. That may have resulted in a sufficient exposure to tension at those lengths. This may suggest a ceiling effect. Once sufficient tension is achieved at long muscle lengths, additional exposure may not proportionally increase hypertrophy.
This repeated exposure in the full range of motion condition may have been sufficient to surpass the threshold needed to stimulate muscle hypertrophy, leading to similar outcomes. If we consider lengthened partials versus full range of motion, similar findings were recently observed by Wolf and colleagues. Despite greater average exposure to longer muscle lengths in the partial ROM, both conditions emphasized long muscle lengths. Both resulted in similar hypertrophy of elbow flexors and extensors. Thus, it’s pretty safe to say that training at relatively longer muscle lengths is important, and those lengths should be sufficiently challenging. But it’s probably not worth going into absolutely maximal muscle lengths with the hope of maximizing tension and thus the hypertrophic response.
Finally, a just-published study by Ismail and colleagues compared 12 weeks of isometric versus high-intensity dynamic training on quadriceps hypertrophy and strength in obese adults with knee osteoarthritis. 80 participants were randomized into two groups. Dynamic training involved performing knee extension exercise for three sets with loads increasing from 75 to 90% 1 RM and repetitions decreasing from eight to four throughout the intervention, respectively. Isometric training involved performing 3–5 sets of 5 to 10 repetitions of 5-second unilateral knee extension. Isometrics were performed at 50% and 60% MVC in weeks one and two, respectively. Both number of sets and repetitions performed per session increased throughout the intervention.
Vastus lateralis muscle thickness, pennation angle, and fascicle length were assessed via ultrasound at the middle site. MVC was measured at 70° of knee flexion. Researchers also used a visual analog scale to assess pain changes and The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire to evaluate pain, joint stiffness, and physical function.
Conditions in both groups resulted in increased quadriceps architecture and strength. However, greater statistically significant effects across all of the measurements were reported for the dynamic versus the isometric group. Conversely, no significant between-group differences were observed for pain assessment or WOMAC scores.
One detail worth mentioning here is that the groups were not well-equated in terms of effort. Based on the description, isometrics were performed at fairly low relative intensities. The joint angle used was also not described. Even though volume (rather than intensity) might be the primary variable impacting hypertrophy from isometric training, these isometrics were performed at higher volumes. However, when compared to dynamic training – especially involving higher loads – intensity (i.e., % MVC) may begin to play a more significant role. Although isometric training at <70% MVC can result in hypertrophy, growth experienced may be less than usual when training with low intensities. Moreover, greater strength increases with dynamic versus isometric training might be an indication of that. Even though the task being tested was more specific to the training (isometric), the group that performed dynamic training generally did more “outcome specific” training (e.g. training for greater force capacity). I’m basing this on what is reported in the paper. They also didn’t clearly state whether later weeks were performed as 5-second MVCs or at a % of MVC.
So where does this leave us? The table below summarizes the research comparing isometric versus dynamic training on hypertrophy-related outcomes. Broadly, across eight studies, similar changes between the isometric and dynamic training were observed. Two studies slightly favored isometric training, and three favored dynamic training.
Participants
49 physically active young men
Study Design
Between-participant
Exercise(s)
Elbow flexion, overhead press
Training Detail
ISOM: 3 × 15-s contractions at 2/3rds MVC at ~100° of elbow flexion.
ISOT: 3 × 5 reps at 2/3rds 1RM; each set took ~15 s.
Hypertrophy Measure
Upper-arm circumference
Summary of Findings
Greater arm circumference in ISOT vs ISOM (1.22 cm vs 0.59 cm). After 6 weeks detraining, ISOT lost more size (-0.60 cm) while ISOM maintained (+0.09 cm). Net retained gains were similar between groups (~0.6-0.7 cm).
Participants
14 untrained young men
Study Design
Between-participant
Training Detail
ISOM: 3 sets/day of 10-s contractions at 90° elbow flexions at 2/3rds MVC.
ISOT: 3 × 5 elbow flexions from ~75° to 105° at 2/3rds 1RM.
Hypertrophy Measure
Elbow flexor CSA
Summary of Findings
Strength gains on specific tests favored ISOT, but both groups increased their strength levels. ISOM and ISOT resulted in similar increases in isometric strength (26% vs 31%) and 1RM gains were greater in the ISOT (33% vs 23%). Both groups increased elbow flexor CSA (5.4% vs 3.2%), with no statistically significant differences between them.
Participants
10 healthy untrained men
Study Design
Within-participant
Exercise(s)
Knee extension
Training Detail
ISOM: 10 × 15-s knee extensions at 70% MVC at 90° knee flexion.
ISOT: 5 × 10 knee extensions from 90° to 0° at 80% 1RM.
Hypertrophy Measure
Quadriceps femoris muscle volume
Summary of Findings
Quadriceps muscle volume increased slightly more with ISOT than ISOM (5.6% vs 4.5%), with no statistically significant group differences. Isometric strength gains favored ISOM (49% vs 31.5%).
Participants
61 older adults with bilateral tricompartmental KOA (51 women, 10 men)
Study Design
Between-participant
Exercise(s)
Knee extension
Training Detail
ISOM: 10-s 100% MVC at full knee extension.
ISOT: 1.5-kg isotonic knee extensions.
ISOK: concentric repetitions at 60-240°/s.
Hypertrophy Measure
VL muscle thickness, pennation angle, fascicle length
Summary of Findings
ISOM produced bilateral strength gains and increased VL thickness and fascicle length in the trained limb and fascicle length increases contralaterally. ISOK increased VL thickness bilaterally and fascicle length contralaterally, whereas ISOT increased VL thickness bilaterally without clear fascicle-length changes.
Participants
31 untrained young men
Study Design
Between-participant
Exercise(s)
Knee extension
Training Detail
ISOM: 1-s contraction at four knee angles (90°, 70°, 50°, 30° knee flexion).
ISOT: dynamic 1-s concentric and eccentric muscle actions.
ISOK: dynamic training at 90°/s over a set ROM.
Hypertrophy Measure
Lower-limb lean mass (DXA)
Summary of Findings
Lean mass of the trained limb increased in ISOM (~3.1%) and ISOT (~3.9%) groups, but not in ISOK. All groups improved isometric and isokinetic torque, with strength gains favoring the specific training group. ISOK improved triple-hop distance.
Participants
20 recreationally resistance-trained men
Study Design
Between-participant
Exercise(s)
Elbow flexion (preacher curl)
Training Detail
ISOM: 6 × 10-s contractions at 75% MVC at 30°, 60°, and 90° elbow flexion (two contractions per angle per set; total 60 s of TUT per session).
ISOT: 6 × 10 reps at 75% 1RM preacher curl with a controlled 3-s concentric and eccentric tempo.
Hypertrophy Measure
Segmental upper-limb lean mass and body fat (BIA)
Summary of Findings
ISOM and ISOT produced similar changes in 1RM and similar angle-specific MVC changes, with some joint-angle advantages for ISOM. No significant changes in segmental upper-limb “active mass” or body fat were detected by BIA in either condition.
Participants
23 resistance-trained men and women
Study Design
Within-participant
Exercise(s)
Knee extension
Training Detail
ISOM: 30-s MVC holds per set at a long muscle length (~125° knee flexion).
ISOT: 30-s full-ROM knee extensions (~125° to 10°) per set with load adjusted so RPE was ~7-10. Set duration and time-integral were matched between conditions.
Hypertrophy Measure
Regional quadriceps muscle thickness (anterior and lateral thigh composites at 30, 50, 70% length)
Summary of Findings
Both ISOM and ISOT resulted in similar increases in anterior and lateral quadriceps muscle thickness. Most regional estimates slightly favored ISOM condition for the anterior thigh, with a trend toward greater rectus femoris growth at the proximal site.
Participants
71 obese men with grade 1-2 KOA
Study Design
Between-participant
Exercise(s)
Knee extension
Training Detail
ISOM: 3-5 sets of 5-10 reps of 5-s unilateral knee-extension holds per session.
ISOT: full range of motion training at 65-90% 1RM, 3 sets of ~8 to 4 reps, progressed over 12 weeks.
Hypertrophy Measure
VL muscle thickness, pennation angle, fascicle length
Summary of Findings
ISOT resulted in significantly greater gains in all architectural measurements and MVC (~35% vs ~14%), while pain and function improvements were similar between groups.
Collectively, we can state that both isometric and dynamic training are very likely to result in similar hypertrophy overall, based on the available evidence. However, keep in mind that many studies on the topic have notable methodological limitations – including the use of indirect measurements, questionable training protocols, clinical populations, and basic reporting limitations within the papers themselves.
There is a need for more research on the topic using more modern equipment (i.e., ultrasound, MRI, CT scan), in addition to more refined training protocols. Moreover, no other study besides ours has performed isometric training at longer muscle lengths. Given that studies comparing isometric training at shorter versus longer muscle lengths report greater hypertrophy at longer lengths, it remains unclear how much more effective (or similar) those are when compared to dynamic training. Right now, we only really have our study to go off.
Additionally, due to the heterogeneity of the studies, it’s currently difficult to even attempt quantifying these effects via meta-analysis to paint a clearer picture. For that, we’ll have to wait until enough evidence builds up.
Even though this isn’t a traditional isotonic resistance training study, it’s worth mentioning. Prior studies have reported that plyometric training (a ballistic form of isotonic training) can result in similar hypertrophy to traditional resistance training. In a study by Kubo and colleagues, researchers compared 12 weeks of plyometric and isometric training on muscle and tendon properties. 11 untrained men performed plyometric training with one leg. This involved single-leg hopping and drop jumps. The other leg performed plantar flexion in a prone position.
In the plyometric condition, participants performed five sets of each exercise (hopping and drop jumps), with 30 seconds of rest between sets. The sets consisted of unilateral plantar flexion at 40% of 1RM, with 10 repetitions per set. Isometric training involved performing ten 15-second contractions at 80% MVC, with 30 seconds of rest between contractions. Both conditions resulted in significant muscle thickness increases in the medial and lateral gastrocnemius, soleus, and summed plantar flexor thickness (ISO: 5.7 ± 2.6% versus 5.5 ± 2.3%). No significant differences were found between conditions. Moreover, while both conditions increased isometric strength, greater increases were reported with isometric training (22.1 ± 14.2% versus 4.4 ± 5.0%).
Isometric training versus other muscle actions
In the previous section, we discussed differences between isometric and dynamic training. But if we look at direct longitudinal evidence in humans comparing isometrics to other muscle actions (e.g., concentric or eccentric), what do we observe?
I mentioned there was a major gap between the studies by Fukunaga and Sugiyama and the work by Kubo and colleagues. However, there was one additional foundational quadriceps study published by Jones and Rutherford in 1987. 12 healthy adults were assigned to either: 1) a group that performed unilateral isometric training, with the other leg serving as a control; or 2) a group that performed concentric-only training with one leg, and eccentric-only training with the other.
Participants trained three times per week for 12 weeks. Isometric training involved performing 4 sets of 6 contractions held for four seconds, performed at 80% of MVC. MVC was measured weekly to reassess training targets. Concentric and eccentric training involved 4 sets of 6 repetitions at a load they could complete for 6 repetitions (~80% of 1RM) for each muscle action. Loads in the eccentric condition averaged about 145% of those used in the concentric condition.
Strength was measured using maximal voluntary isometric contraction pre- and post-intervention, and every two weeks during the intervention period. Quadriceps CSA was measured via CT scan at the midpoint between the greater trochanter and tibiofemoral joint space. Isometric strength changes were significantly greater in the isometric condition (35 ± 19%) compared to both the concentric (15 ± 8%) and eccentric (11 ± 3.6%) conditions. All three conditions resulted in small CSA increases (isometric: 5%; concentric and eccentric: 4–6%), with no statistically significant differences between them.
Carmichael and colleagues compared isometric versus eccentric hip extension training on hamstring architectural, strength, and morphological adaptations. 24 recreationally trained men performed 6 weeks of unilateral training, twice per week. One leg performed isometric training on a glute-ham raise bench with hips and knees at ~0° of flexion. The other performed eccentric-only hip extension on a Roman chair set 30° relative to the floor. That condition started from 0° of hip and knee flexion and descended under a controlled tempo (5 seconds) until reaching ~90° of hip flexion.
Biceps femoris long head fascicle length, pennation angle, and muscle thickness were assessed via ultrasound. Muscle volumes of the individual hamstring muscles were derived from CSA measured via MRI. Strength assessment included isokinetic eccentric and concentric knee flexion at 60°/s and 180°/s angular velocities, isometric knee flexion at 30° of knee flexion, maximal isometric hip extension, maximal bilateral eccentric knee flexor strength during the Nordic hamstring exercise, and a strength endurance test during a single-leg hamstring bridge exercise. Measurements were performed pre-intervention, mid-intervention, post-intervention, and ~4 weeks post-detraining.
Significant increases in biceps femoris fascicle length were only observed in the leg that performed eccentric training post-intervention. Those changes decreased close to baseline values after the four-week detraining period (pre: 7.7 ± 0.9; post: 9.2 ± 1.1; post-detraining: 7.8 ± 1.1).
No significant increases were observed for biceps femoris long head pennation angle or muscle thickness in either condition from pre- to post-intervention. These differences may reflect the muscle length trained at. Eccentrics were performed, on average, at longer muscle lengths whereas isometrics were performed at relatively shorter muscle lengths. As mentioned earlier, several studies report fascicle length increases tend to be greater following training at longer muscle lengths.
Greater increases in biceps femoris long head and semimembranosus muscle volumes were observed with eccentric compared to isometric training from pre- to post-intervention. However, following detraining, only biceps femoris long head volume remained slightly elevated. Semimembranosus lost muscle volume. Conversely, no significant increases were observed in the isometric condition for either muscle. Both conditions resulted in similar increases in semitendinosus and biceps femoris short head muscle volume from pre- to post-intervention, with slightly greater growth following isometric training. Interestingly, muscle volume decreased to a greater extent in the isometric condition during detraining. In contrast, it remained increased for the eccentric condition for both muscles.
As far as strength changes, they followed the principle of specificity, more or less. For example, similar increases in average peak eccentric torque at 60°/s were reported in both conditions (12%). Only eccentric training resulted in a significant increase from pre- to post-intervention for the average peak eccentric torque at 180°/s. Greater increases in average peak isometric torque were reported for the isometric condition (10%). While changes tended to favor specific training, no significant between-group differences were reported. Greater isometric hip extension strength increases were observed with isometric training (12%). Similarly, no significant within- or between-group differences were found for average peak concentric torque, eccentric strength during Nordic hamstring exercise, or hip extension strength endurance. Overall, strength gains were maintained during a four-week detraining period.
There’s only one other study that I’m aware of that directly compared the three muscle actions head to head. Sato and colleagues randomized 49 untrained participants into four groups (concentric-only, eccentric-only, isometric, and control). They performed daily three-second MVCs of the elbow flexors for 20 days. Concentric-only and eccentric-only groups performed repetitions from 10° to 100° and 100° to 10° of elbow flexion, respectively, at 30°/s. The isometric group performed MVCs at 55° of elbow flexion.
Strength changes were assessed via MVCs for each respective muscle action. Isometric MVCs were performed at 20°, 55°, and 90° of elbow flexion. Concentric and eccentric MVCs were measured at 30°/s and 180°/s with range of motion set from 10° to 100°. Elbow flexor muscle thickness was measured at proximal, middle, and distal sites via ultrasound. No changes in muscle thickness were observed for any of the groups which is not surprising given the extremely low training volume.
Overall, the greatest increases in MVC were reported for the eccentric-only group. Similar increases in eccentric MVC at 30°/s and 180°/s were reported for the eccentric-only and isometric groups. Interestingly, no significant changes in MVC variables were observed for the concentric-only group.
Researchers hypothesized that greater increases in MVCs may have been due to greater force production with eccentric muscle actions compared to concentric and isometric muscle actions. Indeed, the authors reported that average peak torque over the 20 sessions was 54% and 39% greater for the eccentric compared to the concentric and isometric groups.
As it stands, with only three studies that directly compare isolated isometric to concentric-only or eccentric-only training, the evidence is far less clear than it’s being made out to be regarding one’s superiority versus the other.
To be honest, I’m skeptical that eccentric muscle actions, by themselves, are a substantially more potent hypertrophy stimulus. If that were the case, we’d see obviously greater hypertrophy when work-matched eccentric muscle actions are compared with concentric muscle actions (Note: I’m specifically not talking about overloaded eccentrics here). Although, I will admit there’s a lack of studies that directly explored the question. Studies typically tend to equate for relative intensity but usually end up performing more work with the eccentrics.
Given the discussion surrounding muscle lengths and mechanotransduction occurring from both active and passive elements, my working hypothesis is: If all the muscle actions were properly equated (e.g., time-integral) and performed at longer muscle lengths, I think we’d see pretty similar hypertrophy occur. We’re very likely to also observe some specific morphological adaptations with one muscle action versus the other (e.g., fascicle length and pennation angle).
In fact, a recent study by Karyofyllidou and colleagues compared concentric versus eccentric muscle actions when both were performed at longer muscle lengths. Participants performed 5 sets of 15 contractions (75 total) with maximal voluntary effort using their elbow flexors at an angular velocity of 45°/s. Range of motion was restricted to 45° (5°–50°, where 0° indicates full extension). This trained the elbow flexors at longer muscle lengths. Researchers found that both conditions resulted in very similar overall muscle damage. They attributed most changes to unfamiliar exercise rather than the muscle action per se.
So, my bets are on the muscle length being the primary mediator of the adaptations we’re observing when discussing differences between muscle actions. For you as a lifter, this doesn’t change much. I don’t think many lifters perform concentric-only or eccentric-only training anyway. This section is mostly relevant (and hopefully interesting) if you’re really invested in the science, exercise physiology, and the potential mechanisms, or at least, an attempt to explain the adaptations we observe in studies.
Putting it all together
If you were not previously convinced by the evidence behind isometrics and their efficacy in increasing muscle strength and size, hopefully this article changed that for you to some extent. Before wrapping up, I think it’s worth discussing a few pragmatic considerations when it comes to applying isometrics in your own training.
First off, where do you start? As alluded to several times throughout the article, manipulating specific training variables is a key necessity if your goal is to optimize gains (just as it is with traditional dynamic resistance training). One point of contention people often have with isometrics is tracking progress. And indeed, due to the lack of visible movement at a given joint angle, it can be difficult to gauge whether you’re actually making progress over time. With advancements in technology, that’s become much easier. For example, force plates and strain gauge dynamometers are becoming more widely available and are no longer reserved solely for lab or team-sport settings.
Another option is to track time to task failure for a given set. For example, you could select an exercise with a given load and perform a pushing isometric to a specific joint angle. You’d apply just enough force that there’s still visible joint displacement, but you’re close to or at the joint angle you actually want to train. You would then hold that load and resist it until you fail. This essentially becomes a hybrid transition from pushing to holding to EQI. From there, you could progress by either increasing the time it takes to reach task failure across weeks or increasing the load while trying to maintain the same time.
A simpler approach, likely more relevant for situations where isometrics are commonly performed (e.g., lack of equipment, hotel gyms, injuries), is to simply increase the number of sets or repetitions. Alternatively, you could increase the duration of each repetition performed at perceived MVC across weeks. After that, you could bring the duration back down, increase the number of reps, and again build rep duration over subsequent weeks.
Finally, when should you implement isometrics within a traditional training session? At the start, in the middle, or at the end as an advanced technique? In studies that combined isometrics with dynamic training, they were typically implemented at the beginning or somewhere in the middle of the session. I can see a rationale for placing them at the start. You’re unfatigued, which puts you in a better position for maximal force production and recruitment of high threshold motor units.
That said, this doesn’t mean they can’t or won’t work if performed after dynamic training or as an extended-set technique. For example, you could reach full range of motion failure and then perform isometrics at longer muscle lengths. In fact, recent studies by Larsen and colleagues (one, two) showed there may be a slight benefit for calf hypertrophy by extending the set beyond full range of motion failure. Thus, it’s reasonable to hypothesize that isometrics might produce similar effects.
Practical applications
To make any measurable progress, you will need to progressively increase loading demands and manipulate key training variables. Here are some of my suggestions for where to start, based on the available evidence. Keep in mind that all of this is assuming you won’t be jumping on the isometrics train 100%, all the time, but rather supplementing or replacing a relatively small fraction of your regular training.
Type of isometric contraction
As of now, we have limited evidence directly comparing pushing versus holding isometrics. Pushing isometrics typically allow for greater force output and thus may be more appropriate for maximal strength development. Holding isometrics may be useful for developing local muscular endurance, improving joint stability, or in some early rehabilitation contexts.
Muscle length / joint angle
In regard to specific variables, you should most likely perform isometrics at longer muscle lengths when the goal is strength or hypertrophy. While the number of studies comparing training at shorter versus longer muscle lengths is limited, those that do exist tend to report slightly greater hypertrophy with training at longer muscle lengths. Training at longer muscle lengths may also shift optimal torque angles or increase fascicle length. For strength, training at longer muscle lengths may also result in greater strength transfer across a broader range of motion.
Training volume
While volume can be quantified in several ways, total contraction duration appears to be an important variable to consider. Some studies report greater hypertrophy with longer contraction durations. Based on the available literature, aiming for roughly 80–150 seconds of total contractions per session when hypertrophy is the primary goal and 30–90 seconds for maximal strength appears to be a reasonable starting point. This would translate to roughly ~10–40-second contractions for hypertrophy and ~1–5-second maximal contractions for strength per set, typically performed for around 3–5 sets.
Training intensity
This is typically expressed relative to one’s MVC. However, studies comparing different training intensities reported similar hypertrophy outcomes when total contraction duration was equated. As such, moderate-to-high (~50-70% MVC) intensities will likely be sufficient. For maximal strength development, higher intensities (~80–100% MVC) combined with shorter contractions (~1–5 seconds) may also be a more time-efficient strategy.
Contraction intent
While both explosive and longer sustained (e.g., ~10–40 seconds) isometrics can work, when contraction duration was equated, sustained contractions sometimes resulted in greater hypertrophy. However, when the goal is maximal strength or power development, short and explosive isometrics may be the better option from a rate-of-force-development standpoint.
Rest intervals
Very few studies have directly manipulated rest intervals. Most protocols use short rest between contractions (especially when shorter durations are used) and longer rest between sets. As a practical starting point, use ~2–10 seconds rest between contractions and ~60–180 seconds rest between sets.
Exercise selection and specificity
Finally, multi-joint isometric exercises that resemble the targeted task (or an outcome) may result in a better transfer to dynamic movements than single-joint exercises. As mentioned, isometrics at multiple joint angles may be useful when broader strength transfer is desired, whereas training at more “specific angles” may be more appropriate when trying to address a specific sticking point. From the available literature, strength gains tend to occur within ~20–50° of the trained joint angle, although this may be influenced by the muscle length (joint angle) at which training occurs.
We went through a lot. That said, there’s still much more to explore. Hopefully, this article provided a solid introduction to the topic and a useful overview of the more recent and relevant evidence. There are still plenty of unanswered questions, some of which I hinted at throughout the article, but cool things are on the horizon over the next few years.
I’ll finish by saying isometrics are a fun and extremely useful tool to have and experiment with. But, keep your expectations in check and try not to blow them out of proportion.
Finally, I’d like to sincerely thank Greg for providing me with the opportunity and platform to share the evidence and really cool area of research with all of you. I hope you enjoyed the article and found it interesting.
