What causes delayed onset muscle soreness (DOMS)?

18 min readJun 8, 2020

What are the possible mechanisms of DOMS?

Introduction

Over the years, many mechanisms have been proposed to explain how and why DOMS occurs after exercise. If you pick up any review paper written in the last thirty years, you will find a list of many proposed mechanisms. While previous reviewers have rarely tried to assess which mechanism is most likely, we can make some progress in figuring out the most likely explanation for DOMS by assessing the extent to which each mechanism can explain any or all of the following observations:

why DOMS is greater after exercise that is unaccustomed compared to exercise that is accustomed (this is called the repeated bout effect, and it lasts for at least 6 weeks and possibly even longer);
why DOMS is greater after eccentric than concentric contractions (but does still occur after concentric-only exercise, contrary to some statements on this subject);
why DOMS does not appear immediately after exercise, but appears gradually and reaches a peak between 24 and 72 hours after exercise (this is the biggest challenge for understanding DOMS, since it separates any damaging event during exercise from the subsequent pain);
why DOMS tends to occur more in certain regions of the exercised muscles than in others (indeed, some research has found that the most distal parts of the muscle experience the most DOMS, while other research suggests that the most affected sites can vary, and some studies have even found that the exact location varies between individuals).

The following sections include short analyses of all the mechanisms I could identify by searching through the literature (if you find any others for which there are good references, please let me know so that I can add them in). As you will see, there are many more mechanisms than are usually discussed in most review papers.

#1. Lactic acid (lactate and hydrogen ions)

Why it could theoretically be a mechanism: The accumulation of metabolites (including lactate and hydrogen ions) are at least partly responsible for the pain and fatiguing sensations that are experienced during high-intensity exercise, due to afferent feedback through group III/IV nerves. Consequently, lactate may also be responsible for DOMS after exercise as well.

Why it might/might not be a mechanism: (1) lactate does not differ substantially between unaccustomed and accustomed exercise, as shown by the lack of any difference in blood lactate between untrained and trained subjects; (2) lactate is actually higher during concentric exercise than in eccentric exercise; (3) lactate dissipates quickly after stopping exercise, whereas DOMS takes a long time to appear in the post-workout period; and (4) unless the fast twitch muscle fibers of a muscle are grouped in certain regions (which is quite rare), there is no plausible way in which lactate accumulation could directly cause regional effects within a muscle.

Conclusion: lactic acid is not a viable mechanism for DOMS.

#2. Reflex spasms

Why it could theoretically be a mechanism: The pain and fatiguing sensations incurred during exercise as a result of ischemia and metabolite (lactate and hydrogen ion) accumulation could lead to reflex spasms. These might then perpetuate a state of ischemia and elevated metabolite levels, which then causes discomfort in the same way as accumulated metabolites during exercise.

Why it might/might not be a mechanism: (1) pain and fatiguing sensations incurred during exercise probably do not differ that much between unaccustomed and accustomed exercise; (2) pain and fatiguing sensations are much higher during concentric exercise than in eccentric exercise (in which fatigue can occur without metabolite accumulation); (3) there is little evidence that reflex spasms occur in tandem with DOMS, which appears some time after stopping exercise; and (4) since reflex spasms would follow the size principle, they would only affect the most oxidative, slowest twitch muscle fibers of the muscle, and these would not be grouped in certain regions. Thus, there is no way in which reflex spasms could cause regional effects within a muscle (and the most oxidative muscle fibers would not cause ischemia and elevated metabolite levels anyway).

Conclusion: reflex spasms is not a viable mechanism for DOMS.

#3. Muscle damage

Why it could theoretically be a mechanism: After eccentric exercise, the insides of muscle fibers (the sarcolemma, myofibrils, and cytoskeleton) display structural changes. Such changes could reflect damage, which might be detected by receptors in the muscle and cause the pain of DOMS. Indeed, some studies have found (admittedly quite weak) associations between direct markers of damage and the degree of DOMS that is suffered.

Nevertheless, to understand potential links between muscle damage and DOMS, it helps to set the research in context.

When they were first noted, the structural changes that arise after eccentric exercise were assumed to reflect muscle damage. Yet, it soon became clear that at least an element of these changes were actually the first step towards muscle fiber adaptations. Also, many early studies used electrically-stimulated contractions, and did not control for the damage done when taking a muscle biopsy. Since electrically-stimulated contractions cause more damage than voluntary contractions (but oddly do not cause more DOMS), and since muscle biopsies cause muscle damage even without exercise, these are severe limitations of the early work. Indeed, it seems likely that many early studies did overstate the amount of damage that is really happening after voluntary eccentric exercise (either by using electrical stimulation or by failing to deduct the damaging effects of the biopsy).

Furthermore, researchers have now presented evidence that sarcolemmal, myofibrillar, and cytoskeletal disruptions are quite mild after voluntary eccentric exercise, and have suggested that these disruptions may be the early signs of adaptations, such as sarcomerogenesis (the addition of sarcomeres in series). Indeed, the time course of changes does suggest an adaptive response, for two reasons. Firstly, overall, the disruptions do not peak immediately after exercise, but instead peak at 2–3 days afterwards. Secondly, the disruptions at 2–3 days after exercise comprise more misshapen sarcomeres and Z disks, while disruptions at 7–8 days after exercise include more supernumerary sarcomeres and double Z disks. This indicates a gradual shift towards an orderly arrangement with additional sarcomeres, which passes through a disordered phase that might easily be mistaken for damage.

Conventionally, it is accepted that long-lasting reductions in maximum strength after exercise reflect the magnitude of muscle damage. This assumption is very strongly ingrained in our collective understanding of fatigue and recovery. Indeed, some researchers have argued against muscle damage as being related to DOMS (in part) because losses in strength can vary, even while the magnitude of DOMS does not. Yet, it is likely that sarcolemmal, myofibrillar, and cytoskeletal disruptions are not responsible for the larger part of the reduction in maximum strength, especially immediately post-workout. In this phase, disruptions are minimal, while strength losses are almost always nearly maximal. In fact, much of the losses in strength after exercise likely occur due to excitation-contraction coupling failure (although there are other factors). This is very important for our understanding of the relationship between muscle damage and DOMS, because severe fatigue can be present even when muscle damage is not.

Even so, that is not to say that true muscle damage never happens after voluntary exercise, and that such damage is unrelated to the losses in strength that occur. Indeed, it is known that fatiguing voluntary exercise involving eccentric contractions can cause such severe muscle damage that fiber necrosis (cell death) results. This can be seen in studies of marathon runners and in recent, controlled studies involving subjects who are “responders” in terms of post-workout strength loss (they experience more severe strength loss than the average). In such individuals, not only is clear evidence of true muscle damage presented (necrosis cannot be mistaken for anything else), but the amount of muscle damage is related to post-workout strength loss.

Why it might/might not be a mechanism: (1) sarcolemmal, myofibrillar, and cytoskeletal disruptions certainly differ between unaccustomed and accustomed exercise; (2) such disruptions are also clearly larger after eccentric exercise than after concentric exercise; (3) although long-lasting reductions in maximum strength do not follow the same time course as DOMS (they are greatest immediately after the workout and on the following day, while the peak of DOMS occurs later), the time course of myofibrillar and cytoskeletal disruptions is actually quite similar, as the peak tends to occur at 2–3 days after exercise, which is exactly when DOMS is highest; (4) indirect measurements of cytoskeletal disruptions have shown that eccentric exercise can cause greater disruptions in certain regions of a muscle than in others, and that this regional disruption might be somewhat related to the regional nature of DOMS.

Conclusion: muscle damage (when measured by reference to myofibrillar and cytoskeletal disruptions, and not by reference to strength losses) is a possible mechanism for DOMS.

#4. Connective tissue damage

Why it could theoretically be a mechanism: The connective tissue (particularly the endomysium) might be damaged by eccentric contractions. Such damage might be detected by receptors in the muscle, thereby causing the pain of DOMS. Indeed, muscle damage can vary between workout types, while markers of connective tissue and DOMS are similar.

Markers of connective tissue breakdown (hydroxyproline and hydroxylysine) are excreted after intense, military physical training programs, eccentric strength training exercise and plyometrics. However, they are not excreted after conventional strength training, concentric-only strength training, or downhill running, all of which are known to cause DOMS to some extent. Moreover, actual direct measurements of extracellular matrix disruptions have been recorded after eccentric contractions in both humans and rodents. Muscle collagen synthesis often increases after exercise, although whether this represents repair or an adaptation is not clear. Indeed, muscle collagen levels do increase after long-term training in rodents and in some populations of humans, which indicates that adaptations likely occur. Even so, the extent to which such changes are genuinely adaptive, and to what extent they are pathological, is unclear.

Why it might/might not be a mechanism: (1) markers of collagen damage, repair, or adaptation are indeed greater after unaccustomed exercise compared to after accustomed exercise; (2) markers of collagen damage, repair, or adaptation are sometimes (but not always) higher after eccentric exercise than after concentric exercise; (3) the time course of collagen damage and repair is not well-described, and it is unclear how it might relate to the time course of DOMS. Even so, preliminary findings suggest that the collagen response to eccentric exercise is more sustained than DOMS, and does not decay after 2–3 days as would be expected if it were the major trigger for exercise-induced soreness; (4) theoretically, collagen damage might occur in certain regions of a muscle to a greater extent than in others, simply due to the varying stresses and strains within the interlinked structure. Yet, evidence of this occurring in practice is difficult to find.

Conclusion: connective tissue damage is a possible mechanism for DOMS.

#5. Muscle fiber inflammation

Why it could theoretically be a mechanism: When muscle fibers are damaged after eccentric contractions, this leads to an inflammatory response. The inflammation within the muscle fibers could act upon receptors within the muscle, thereby causing the pain of DOMS. Inflammation has often been linked to pain, through a number of mechanisms.

Indeed, an inflammatory response occurs after any kind of tissue injury, and its purpose is to help repair any damage that has been caused. A key step in this repair process is the degradation and removal of any cellular debris. Inflammation first involves vasoconstriction (a narrowing of blood vessels), followed secondly by vasodilation (a widening of blood vessels). Along with the vasodilation, there is an infiltration of immune cells, which act mainly to remove cellular debris. A couple of hours after the injury, neutrophils (a type of white blood cell) arrive. Several hours after the arrival of the neutrophils, monocytes (another type of white blood cell) also appear. While neutrophils dissipate within hours, the monocytes remain for approximately 48 hours, often differentiating into macrophages.

The question of whether inflammation after eccentric exercise is responsible for DOMS is therefore inextricably linked to whether the same exercise causes damage. However, if myofibrillar and cytoskeletal disruptions are found to reflect the early phases of adaptations and not muscle damage, then we might expect inflammation after voluntary exercise to be quite rare (although we might still observe substantial inflammation after electrically-stimulated exercise or as a result of the muscle biopsy processes). In this respect, it is important to note that when studies have compared the effects of eccentric exercise with a no-exercise control condition, the same inflammation is found in both conditions, as a result of the damage caused by the biopsy.

Consequently, while several early studies found strong evidence to link certain systemic and local muscular markers of inflammation with DOMS, a small amount of recent work has cast doubt upon the idea, particularly since DOMS has been demonstrated to occur without any visible signs of inflammation. More specifically, neutrophil infiltration into muscles does not always occur after exercise in both humans and other animals.

Moreover, it has not always been a straightforward process to link the steps in the inflammatory response to DOMS. An early attempt was made to connect inflammation with soreness through the actions of prostaglandins (which do not directly cause pain but rather sensitize nociceptors, which leads to an increase in mechanical sensitivity, which is a key feature of DOMS). Yet, later work found that inhibiting their action did not prevent DOMS, which indicates that any link between the inflammatory response to DOMS must occur through other mediators.

Why it might/might not be a mechanism: (1) while the inflammatory response is sometimes greater after unaccustomed exercise than after accustomed exercise, it is not always different, and is sometimes actually greater after accustomed exercise; (2) is higher after concentric exercise than after eccentric exercise, as is expected due to the much greater likelihood of damage; (3) whether the inflammatory responses follows the same time course as DOMS is very difficult to assess, due to the multiple phases that are involved (an initial increase in neutrophils, followed later by a secondary infiltration of monocytes). There are elements of the inflammatory process that can be linked to the arrival and dissipation of DOMS, but there have also been reports showing that inflammation is not linked to DOMS; (4) while it is not clear whether inflammation can occur in certain regions after exercise, it seems plausible that if muscle damage can occur locally, then inflammation must also be able to occur locally.

Conclusion: muscle inflammation is a possible mechanism for DOMS.

#6. Connective tissue inflammation

Why it could theoretically be a mechanism: Given that connective tissues can be damaged by muscular contractions, it is reasonable to expect that an inflammatory response also occurs there. Indeed, collagen can trigger an immune function response, and inflammation has been recorded within the epimysium and perimysium after eccentric exercise. This inflammation could act upon key receptors within the muscle to produce DOMS, particularly since the extracellular matrix surrounding muscle fibers is known to contain a large number of afferent nerve endings, and thus fascia is very sensitive to insults. Moreover, some studies have shown that fascia becomes more sensitive after eccentric exercise that leads to DOMS, which could result from inflammation in that area.

Why it might/might not be a mechanism: (1) whether connective tissue inflammation differs between unaccustomed and accustomed exercise is unclear, but it may well do if connective tissue damage also differs; (2) we should also expect connective tissue inflammation to be greater after eccentric vs. concentric exercise, although this too is uncertain; (3) whether connective tissue inflammation follows the same time course as DOMS is unknown, but it is doubtful given that connective tissue damage and repair does not; (4) connective tissue inflammation can likely occur in certain regions of a muscle, in the same way as connective tissue damage, but research in this area is limited.

Conclusion: connective tissue inflammation is poorly-understood, but remains a possible mechanism for DOMS.

#7. Muscle fiber swelling

Why it could theoretically be a mechanism: Muscle fiber swelling often occurs after strength training exercise. This muscle fiber swelling is believed to occur due cell membrane damage, and subsequent inflows of calcium ions, which in turn increase muscle fiber water content. Muscle fiber swelling causes an increase in intramuscular pressure both at rest and during muscular contractions, and it is feasible that this increase in intramuscular pressure might activate group IV afferent nerves.

Why it might/might not be a mechanism: (1) muscle fiber swelling is indeed less after repeated exercise than after an initial bout, and this is associated with reduced post-workout sarcolemmal permeability; (2) muscle fiber swelling is indeed greater after eccentrics than after concentrics; (3) muscle fiber swelling does not follow the same time course as DOMS, because it peaks several days after DOMS peaks; (4) whether muscle fiber swelling might occur preferentially in certain regions (and might cause more DOMS in certain regions) is unclear.

Conclusion: muscle fiber swelling is a possible mechanism for DOMS.

#8. Connective tissue swelling

Why it could theoretically be a mechanism: After muscle-damaging exercise, there is often an increase in total body water, which seems to involve a temporary increase in the fluid content of the extracellular space around muscle fibers (as well as muscle fiber swelling), possibly due to the actions of proteoglycans in the extracellular matrix. This increase in fluid content could activate sensory receptors in the surrounding fascia or skin.

Why it might/might not be a mechanism: (1) the extent to which the increase in the fluid content of the extracellular space differs between unaccustomed and accustomed exercise is unclear; (2) similarly, it is not known whether eccentric causes more extracellular fluid accumulation than concentric exercise, although this is very likely; (3) whether changes in the fluid content of the extracellular space follows the same time course as DOMS is unknown; (4) how fluid content of the extracellular space might differ between regions of a muscle is not clear, and likely depends on the underlying mechanism of its effect (which is uncertain).

Conclusion: connective tissue swelling is a possible mechanism for DOMS.

#9. Neurotrophic signaling

Why it could theoretically be a mechanism: It is possible that neurotrophic signaling during or shortly after exercise might be the cause of the DOMS. Nerve growth factor (NGF) and glial cell line-derived neurotrophic factor (GDNF) are released by muscle fibers after exercise, and could lead to the sensitization of group III/IV afferent nerves, which could cause DOMS.

Why it might/might not be a mechanism: (1) the amount of NGF release seems to be smaller when exercise is accustomed rather than unaccustomed; (2) it is not entirely clear if NGF and GDNF release is greater after eccentric than after concentric contractions; (3) NGF expression seems to be delayed until 12 hours after exercise, which would be appropriate for triggering DOMS shortly thereafter; (4) whether NGF and GDNF are released in some muscle fibers more than others, and whether this could contribute to regional effects within a muscle is not clear.

Conclusion: neurotrophic signaling is a possible mechanism for DOMS.

#10. Oxidative stress (including nitric oxide)

Why it could theoretically be a mechanism: Oxidative stress (the formation of reactive oxygen species [ROS]) and signs of ROS-induced damage are quite commonly (but not always) observed after eccentric exercise. And, given that ROS contribute to the sensations of fatigue and discomfort during exercise, they may cause similar sensations after exercise as well.

ROS are commonly produced during (fatiguing) exercise, and include free radicals of oxygen (such as superoxide, hydroxyl, and nitric oxide) as well as derivatives of these free radicals (such as hydrogen peroxide). Free radicals are molecules that contain at least one unpaired electron. They are chemically very reactive, since the unpaired electron of the free radical is always trying to join with electrons belonging to nearby molecules. ROS often cause damage when they react with the contents of muscle fibers, but they are most effective at damaging anything that contains lipids, such as the muscle cell membrane. It is not known exactly what causes ROS to be generated during exercise, as there are a number of pathways. Yet, the post-exercise inflammatory response (particularly neutrophil generation) can also stimulate the generation of ROS. Indeed, it is likely the generation of ROS in the post-exercise period that is relevant for an understanding of DOMS.

The role of ROS in causing DOMS has been investigated by the administration of antioxidant supplements around the exercise bout, with varying results. Some studies have found evidence that antioxidants might reduce DOMS, while others have not. A recent systematic review concluded that any effect of antioxidant supplements on DOMS was unlikely to be meaningful.

Why it might/might not be a mechanism: (1) oxidative stress likely differs between unaccustomed and accustomed exercise, and is indeed expected, because of the reduced likelihood of damage and subsequent inflammation (since neutrophils trigger ROS formation). Also, some evidence suggests that muscle cells do become more resistant to the damaging effects of ROS formation with repeated bouts, which might result in similar effects, even if actual ROS formation were similar; (2) oxidative stress after exercise is higher when using eccentric vs. concentric contractions; (3) markers of lipid peroxidation (which may result from ROS production) follow a delayed but slightly different time course to DOMS, as they peak after 6 hours, and disappear after 72 hours. Conversely, more direct measurements of ROS show a gradual rise to a peak at 72 hours, which is also dissimilar to the time course of DOMS, which peaks much earlier; (4) whether ROS production occurs in certain regions of a muscle is hard to identify, but since muscle damage does occur regionally, it is theoretically possible.

Conclusion: oxidative stress is a possible mechanism for DOMS.

#11. Muscle spindle nerve damage

Why it could theoretically be a mechanism: Eccentric contractions cause a high degree of muscle mechanical tension that compresses the fluid cavity of the muscle spindles that lie alongside muscle fibers. Compression of the fluid cavity might then entrap the nerve endings of the muscle spindle, damaging them and thereby causing DOMS.

Muscle spindles are structures that lie adjacent to muscle fibers inside muscles. They are made up of a small group of tiny muscle fibers, surrounded by a collagen layer. They are attached to the outer extracellular matrix of the muscle at each end. Their main function is to send signals along afferent nerves from the muscle to the spinal cord, and they have been linked to pain in rodent models. Compression of muscle spindle might be exacerbated by (1) the heightened sympathetic nervous system (SNS) activity that tends to occur during high-intensity exercise bouts (because increased SNS activity could activate the ɣ efferent nerves inside muscle spindles, causing them to contract), and (2) by the increased muscle fluid content that usually occurs during strength training exercise.

One interesting benefit of this explanation is that the same mechanism could explain both DOMS and the reduced range of motion that tends to occur in the exercised muscles in the same time period. Indeed, damage to muscle spindles would be expected to cause changes in the extent to which muscle stretch would be perceived as excessive. Another interesting benefit of this explanation is that it could also explain some of the long-lasting fatigue that occurs after a bout of exercise (although it cannot explain the early reduction in strength in the first 24 hours, which is actually where strength is reduced to the greatest extent).

In contrast, a major challenge for this explanation is that the mechanism rests upon damage that occurs during exercise. Thus, we need to suppose that an additional effect prevents DOMS from being experienced during and immediately after exercise. The researchers who originally proposed this explanation suggested that SNS activity during exercise could prevent the muscle spindle damage from being experienced immediately. Indeed, reactivation of the parasympathetic nervous system after exercise can take 24–48 hours, which makes this a valid hypothesis. Even so, requiring such an adjustment weakens the overall explanation somewhat.

Why it might/might not be a mechanism: (1) it is not known whether the amount of muscle spindle damage that occurs is affected by whether the exercise is accustomed or unaccustomed, since no direct measurements have been taken of this outcome, but SNS activity would be expected to be greater during unaccustomed exercise and this might cause greater muscle spindle compression; (2) similarly, it is not known whether the amount of muscle spindle damage that occurs is affected by whether the exercise involves eccentric or concentric contractions, but it is again assumed that eccentrics would cause greater muscle spindle compression due to the higher level of mechanical tension; (3) the assumptions inherent to this model suggest that muscle spindle damage should be immediate, but the resulting DOMS should occur later than expected, due to the mitigating effects of SNS activity, but whether this is the case is currently unknown; (4) it seems plausible that muscle spindles in different regions of a muscle could be affected differently depending on which regions experience the highest forces during an exercise, which would be a function of their internal moment arm lengths.

Conclusion: muscle spindle damage is a possible mechanism for DOMS.

What are the likely mechanisms of DOMS?

When I first set out to write this article, I fully expected to be able to reject a number of mechanisms out of hand, and zero in one one or two others that were very promising. In fact, aside from the oldest hypotheses (lactic acid and reflex spasms), there seems to be a reason to consider most of the suggestions that have been put forward to explain DOMS.

During future reviews, it might be useful to divide these mechanisms into three categories: (1) mechanisms involving muscle fibers (muscle fiber damage, inflammation, swelling, and oxidative stress); (2) mechanisms involving connective tissue (connective tissue damage, inflammation, and swelling; and (3) those mechanisms involving muscle spindle (and associated nerve) damage. Exactly where neurotrophic signaling might fit is less clear. Currently, many researchers conflate these mechanisms within and across these categories, but they are clearly quite distinct from one another.

What is the takeaway?

Delayed onset muscle soreness (DOMS) is pain or discomfort associated with a previously-exercised muscle, which usually appears the day after exercising. In addition to any pain or discomfort that is present at rest, the muscle is more sensitive to touch and to movement. DOMS has been proposed to occur due to mechanisms inside (1) muscle fibers, (2) connective tissues, and (3) muscle spindles and their associated nerves. It might result from damage to any of these structures, as well as from inflammation or swelling of muscle fibers and connective tissues, and oxidative stress within muscle fibers. Ultimately, what causes DOMS to occur is currently a complete mystery.