Journal Logo

Clinical Sciences: Clinical Investigations

A preliminary examination of cryotherapy and secondary injury in skeletal muscle


Author Information
Medicine & Science in Sports & Exercise: November 1999 - Volume 31 - Issue 11 - p 1516
  • Free


Cryotherapy is the most commonly used modality for the treatment of acute musculoskeletal injuries (24,30–32). Cryotherapy has been shown to decrease skin, muscle, and intra-articular temperatures (16,22,23,26,31), decrease metabolism (17,18,34), decrease inflammation (24,25), decrease blood flow (1,3,7,22), decrease pain (9,24), decrease muscle spasm (4,15), and increase tissue stiffness (24). Unfortunately, the basic physiological mechanisms underlying these effects have not been well described (24,31,34).

The use of cryotherapy in the treatment of acute musculoskeletal injury has traditionally been based on metabolic inhibition and is described in the secondary injury model (21,24). In the secondary injury model, the initial trauma to a tissue is termed primary injury, whereas trauma that occurs subsequent to this primary injury is termed secondary injury. Secondary injury is thought to result from a period of posttrauma hypoxia (secondary hypoxic injury) and from posttrauma enzymatic activity (secondary enzymatic injury) (21,24).

Knight (21,24) proposed that secondary injury, particularly secondary hypoxic injury, is a significant problem after musculoskeletal trauma. As a result, the management of acute musculoskeletal trauma is based on the premise that cold reduces the metabolic rate of these hypoxic tissues, allowing them to better survive this period of hypoxia (21–24,31). This suggestion seems tenable because flux through chemical pathways is temperature dependent via the Q10 effect (14).

The secondary injury model as a basis for cryotherapy is dependent on several points. First, primary injury must produce hemodynamic changes of significant magnitude to result in ischemia. Second, the resulting ischemia must be of sufficient magnitude and duration to produce ischemic injury. Although not originally addressed in Knight ‘s model, reperfusion injury must also be considered here. Third, cold must somehow modify the secondary injury process, most likely by reducing metabolic demand and/or reducing the rate of damaging chemical reactions. Unfortunately, these premises have not been adequately examined in the literature. In a small number of studies concerned with tissue healing, investigators documented the presence of hypoxia after injury (18,19). Unfortunately, these studies were not directed toward examining the secondary injury model and did not attempt to determine whether this hypoxia is of sufficient magnitude or duration to cause cellular injury.

If cryotherapy is effective at reducing metabolic demand and secondary injury in muscle, then it is important that this phenomenon be documented experimentally. Therefore, the purpose of this study was to document secondary injury in skeletal muscle, quantify it, and to determine whether it is altered by acute cryotherapy.

Our attempts to investigate secondary injury were directed at an examination of oxidative phosphorylation. During hypoxia, when the mitochondria are damaged, mitochondrial enzymatic activity is diminished (27). An assay examining mitochondrial function through the reduction of triphenyltetrazolium chloride (TTC) to triphenylformazan (formazan red) has been previously used to examine mitochondrial injury occurring secondary to hypoxia (5). In the mitochondria, cytochrome oxidase normally reduces O2 to H2O. In the assay, cytochrome oxidase reduces TTC to triphenylformazan (a red dye), causing a quantifiable color change. Inability to reduce TTC is indicative of mitochondrial disruption, depletion of the enzymes of oxidative phosphorylation, or both (20).

It was hypothesized that secondary injury occurs after acute musculoskeletal trauma, that injured tissues would have less TTC reduction than controls (uninjured tissues), and that cryotherapy treated injured tissues would have more TTC reduction than untreated injured tissues.


Animal model.

Nineteen male Sprague-Dawley rats weighing between 275 and 320 g were used in this study. Animals were maintained in accordance with ACSM guidelines, and the procedures used were approved by the University of Toledo Institutional Animal Care and Use Committee before beginning the study. All animals were housed two per cage, maintained on a 12-h light/dark cycle, and fed and watered ad libitum. Animals were accommodated to the laboratory and human handling for approximately 1 wk before the beginning of experimentation.

Experimental treatment.

The animals were randomly assigned to either a cryotherapy group (N = 10) or a control group (N = 9). Each animal was anesthetized with sodium pentobarbital (40 mg·kg−1 of body weight, IP) and supplemented as necessary (20 mg·kg−1 of body weight, IP) for the duration of the experiment (approximately 5 h).

Under anesthesia, the right calf of each animal received a crush injury produced with a 13.97 cm Kelly hemostatic forceps as described by Stauber and colleagues (35). Injuries were made to the triceps surae muscles at roughly one-third of the length distal to the knee. Closing the hemostat to the last clasp notch then immediately releasing it caused a uniform injury. Care was taken to ensure that the hemostat was closed over the soft-tissue and not closed over the bone. The left calf was used as a control and was not injured.

After injury, the cryotherapy group received cold treatment in the form of a crushed ice pack applied to the injured calf with 2.5-cm wide elastic tape (Elastikon, Johnson & Johnson Inc., Skillman, NJ). Animals were not shaved before application of the ice pack. The ice packs were replaced as needed after melting and remained applied to the injured calf for the duration of the experiment (5 h). Animal core body temperature was monitored rectally with an electronic thermometer throughout the experiment. Warm water bottles were placed against the anesthetized animals and were replaced as needed to ensure that the animals did not become hypothermic during the experiment.

At 5 h post injury, the triceps surae were excised bilaterally, rinsed with ice cold Ringer’s solution, placed on a watch-glass imbedded in ice and dissected free of fascia, vessels, and nerves. The 5-h duration was based on the work of Belkin and colleagues (5), who demonstrated that ischemic injury, measurable with this assay, occurs after 3 h of total ischemia. Because our methodology did not produce complete ischemia, 5 h was chosen in an effort to maximize the likelihood of finding measurable secondary injury.

After dissection, the tissue samples were weighed, minced, and homogenized in 3 mL of 0.25 M sucrose by using a 7-mL Pyrex grinding tube. Additional sucrose was added to make a 20% homogenate by weight. The homogenate was filtered, and any remaining fascia fragments were weighed and subtracted from the sample weight. Protein content of the sample was determined using a modified version of the method of Lowry (Sigma, St. Louis, MO).

Biochemical assay.

Secondary injury was examined through an assay previously described by Belkin and colleagues (5). A 1-mL aliquot of homogenate was mixed with 1 mL of 0.15% TTC (Sigma) in 0.033 M phosphate buffer, pH = 7.4, (dibasic sodium phosphate). The mixture was incubated for 60 min at 39°C in a shaking water bath. After incubation, the reaction was stopped by adding 4 mL of acetone and vortexing. Reaction tubes were centrifuged for 10 min at 1500 RPM.

Absorbance of the formazan containing supernatant was measured at 485 nm. The triplicate measurements for each sample were averaged and converted to concentrations of triphenylformazan by using a regression equation calculated from a standard curve. For the averaged samples, the coefficient of variation was 0.15. Unfortunately, other studies using this relatively new assay (5,6,10,11) have not reported coefficients of variation with which comparisons can be drawn. Sample concentrations were expressed kinetically as micrograms TTC reduced per milligram of muscle protein per hour.

Statistical design.

The design for this study was a fixed model, 2 × 2 factorial with one between factor (treatment: ice or none) and one within factor (limb: control or injured). Many researchers have treated contralateral limbs as independent (between subjects) observations. In this study, limbs were treated as within subjects factors because the limbs are part of a single animal, and they cannot be assumed to be physiologically independent of one another.

The principle question in this study, whether cryotherapy affects postinjury tissue damage, was evaluated using an ANCOVA comparing the ice and no ice groups. There is a fairly large variability among the TTC reduction rates between animals because these rates depend on the initial mitochondrial density, which varies from animal to animal. To address this and to improve statistical power in the analysis, the paired control limb TTC reduction rates were treated as covariates in the ice versus no ice comparison for the injured limbs. ANCOVA was appropriate because the correlations between the controls (covariates), and the dependent values were similar across the two conditions.


Triphenylformazan concentrations were normalized to sample protein concentrations to determine TTC reduction rates, kinetically expressing cytochrome oxidase activity. Mean ± SEM triphenylformazan concentration, protein concentration, and TTC reduction rates at 5 h post injury are found in Table 1.

Table 1:
Means (± SEM) for selected variables by treatment.

Figure 1 shows the relationship between controls and injured limbs (no ice). The TTC reduction rate for the control limbs was significantly greater than the rate for injured limbs (F 1,14 = 9.584, P = 0.008, eta2 = 0.406, power = 0.821). Figure 2 similarly shows that controls demonstrated a significantly greater reduction of TTC than did injured tissues treated with ice (F 1,14 = 9.584, P = 0.008, eta2 = 0.406, power = 0.821). Figure 3 shows the relationship between injured limbs and injured limbs treated with ice. The ice treated limbs demonstrated a significantly higher TTC reduction rate than the no ice limbs (F 1,16 = 6.612, P = 0.02, eta2 = 0.903, power = 1.000).

Figure 1:
Mean ± SEM TTC reduction for injured tissues and paired controls.
Figure 2:
Mean ± SEM TTC reduction for ice treated injured tissues and paired controls.
Figure 3:
Mean ± SEM TTC reduction for injured and ice treated injured tissues.


The secondary injury model.

In the secondary injury model, injury can be caused by physical, chemical, thermal, metabolic, or biological trauma (24,36). Trauma caused by any of these will result in immediate, pathological, ultrastructural changes in the tissues (12,24,25). These ultrastructural changes include the disruption of the cell membrane and structural components of the cell, leading to a loss of homeostasis and therefore cell necrosis (2,12,24,36). The ultrastructural changes and the immediate consequence of cellular necrosis are referred to as “primary injury” (24,29,33). Primary injury typically affects several types of tissue simultaneously, including skeletal muscle, vasculature, connective and nervous tissues.

In this model, the physiological response to primary injury can lead to injury of otherwise uninjured cells. This resulting “secondary injury” is thought to occur from both enzymatic and hypoxic mechanisms (21,24). Knight (24) suggested that this secondary injury occurs in the cells on the periphery of the primary lesion.

Secondary enzymatic injury.

In secondary enzymatic injury, enzymes are released from dying and inflammatory cells. Although the enzymes were not specifically identified in the original model, they probably include phospholipases, acid hydrolases, and any of a number of human neurtrophil proteins. Some of these enzymes degrade the membranes of nearby cells by cleaving hydrocarbon chains from the lipid portion of membrane phospholipids (27). Changes in the structure of membrane phospholipids lead to the loss of resting membrane potential and increased hydropic swelling, eventually causing the death of the cell (27). There is evidence that hypoxia may contribute to the activation of such enzymes (8).

Secondary hypoxic injury.

Knight (24) suggested that in secondary hypoxic injury numerous factors cause a period of postinjury ischemia. These factors are thought to include hemorrhaging from damaged vessels, hemostasis from the clotting cascade, reduced blood flow from the inflammation induced increase in blood viscosity, and increased extravascular pressure from an expanding hematoma and muscle spasm. Additionally, hydropic swelling of cells secondary to membrane damage can occlude the vasculature, providing another source for ischemia (13,27).

The ischemia produces a resulting hypoxic period that prevents the use oxygen as the terminal electron acceptor in oxidative phosphorylation. This leads to a dependence on the glycolytic pathway for ATP production. The relative inefficiency of glycolysis coupled with low availability of fuel substrates (secondary to ischemia) would prevent adequate production of ATP. Likewise, it has been suggested that when the glycolytic pathway can no longer provide adequate ATP, membrane ion pumps fail, particularly the Na+/K+-ATPase pump (21,24,27). Such failure would result in cellular hydropic swelling leading to cellular necrosis (27). Researchers have suggested that evidence of hypoxic injury is first seen in the mitochondria and that mitochondrial injury occurs after anywhere from minutes to 6 h of hypoxia, depending on the tissue involved (5,28,34).

Examining secondary injury.

Our data provide some early evidence for the existence of secondary injury in skeletal muscle. It appears that secondary injury is a real and quantifiable phenomenon. The attempt to quantify secondary injury was guided by three hypotheses. The first was that secondary injury occurs after acute musculoskeletal trauma. The second hypothesis was that injured tissues (both treated and untreated) would have less cytochrome oxidase activity than controls (uninjured tissues). The third hypothesis was that cryotherapy-treated injured tissues would have more cytochrome oxidase activity than untreated injured tissues. The data from this study can be used to support all three of these hypotheses.

Cold retards injury.

The hypothesis that cryotherapy-treated injured tissues would have more TTC reduction than untreated injured tissues is well supported. The injured tissues treated with cold had mean TTC reduction activity of 6.59 ± 1.0 μg·mg−1·h−1, whereas the injured but untreated tissues demonstrated significantly less activity of only 4.47 ± 0.79 μg·mg−1·h−1. Stated more simply, the no-ice TTC reduction was only two-thirds that of the ice group. From these data, it can be suggested that cooling tissues with ice packs causes some mechanism by which mitochondrial damage is retarded. It is likely that this mechanism is a reduction in metabolic demand of the tissue. Although the physiological mechanisms for the retarding of mitochondrial damage cannot be determined here, some support for the use of cryotherapy in acute settings is implied.

In this study, continuous application of ice packs with elastic wraps for 5 h was used. This differs somewhat from the cryotherapy treatments used in clinical practice (24). Typically, acute musculoskeletal injuries are treated with 20- or 30-min cold applications (24,33). This duration is based on the work of Mlynarczyk (Masters’ Thesis. Dept. of Physical Education, Indiana State University, Terre Haute, IN, 1984), who measured ankle skin rewarming after ice application of different durations. Because no direct measure of the physiological efficacy of ice was used, it is difficult to speculate on the best duration for cold application. Therefore, we used continuous cold application in an effort to maximize any potential injury-reducing effect of the cold.

Injured tissues reduce less TTC than do uninjured tissues.

The second hypothesis, that injured tissues (both treated and untreated) would have less TTC reduction than controls (uninjured tissues), was also supported. For the ice treated group, the uninjured tissues produced a mean TTC reduction value of 7.94 ± 1.49 μg·mg−1·h−1, whereas the injured value was significantly lower at only 5.59 ± 1.01 μg·mg−1·h−1. Similarly, in the no-ice group, the control limbs TTC reduction value was 6.62 ± 0.75 μg·mg−1·h−1, whereas the injured value was significantly lower at 4.48 ± 0.79 μg·mg−1·h−1.

Although these findings would be expected, the effect of crush injury to skeletal muscle on TTC reduction has not been previously reported. In the previous studies where this TTC assay was used, the purpose of the assay was to examine ischemic-reperfusive injury after complete occlusion of the blood supply (5,6,10,11). Crush injury would not be expected to completely prevent blood flow, and therefore might not have produced similar results.

The 5-h postinjury values for TTC reduction in injured tissues (regardless of treatment) are substantially higher than in previous studies using this assay (5,6,10,11). At 5 h post injury, the injured tissue mean TTC reduction value was 5.59 ± 0.68 μg·mg−1·h−1. Belkin and colleagues (5), the first to use this assay, reported 5-h postischemia values of less than 0.50 μg·mg−1·h−1 (extrapolated from graphed data). The disparity between the results of the present study and the results of Belkin and colleagues is large and important. The difference indicates that the crush injury combined with secondary injury seen here is much less severe than ischemia-reperfusion injury produced by complete vessel occlusion. It is unlikely that blood supply was completely occluded in this study. It appears that the ischemia-reperfusion model is not a good source of comparison for crushing musculoskeletal injury.

Secondary injury to skeletal muscle is a real phenomenon.

The final hypothesis examined in this study was that secondary injury occurs after acute musculoskeletal trauma. This hypothesis can also be supported with the results of this study. The evidence for secondary injury is not as obvious as for the previous two hypotheses and therefore requires a degree of explanation. The injured tissues received relatively uniform crush injuries as their primary injuries. If only primary injury had occurred, all injured tissues would be expected to display similar TTC reduction levels, and these levels should have differed from the control group levels. If TTC reduction levels differed between the treated versus untreated injured groups, then the differences must be the result of an event that occurred subsequent to the primary injury. In other words, differing quantities of secondary injury would explain the differences because the primary injuries were the same.

In this study, there were differences between the two injured groups as noted earlier. Likewise, the injured groups differed from the noninjured control group. From these findings, it can be suggested that something subsequent to the primary injury caused the difference in TTC reduction between the injured groups. Therefore, by definition, secondary injury must have occurred. The only difference in treatment was the application of cold and compression. Therefore, cold and compression applied continuously, retarded secondary injury.

It is important to note, however, that a typical acute cryotherapy application of 30 min may not have the same effect. It was not the purpose of this study nor is it possible to determine from the results of this study the efficacy of a 20- or 30-min·h−1 application of cryotherapy. Likewise, it was not the purpose of this study nor is it possible to determine from the results of this study the time course of secondary injury in musculoskeletal trauma. It was the purpose of this study to attempt to verify the existence of secondary injury and to make some suggestion that continuous application of cold and compression retarded secondary injury at 5 h post primary injury.

In conclusion, secondary injury in skeletal muscle was examined by inducing a crush injury to the hindlimb of rats. From these data it may be concluded that 1) crushing muscle tissue produces injury measurable with the TTC reduction assay, 2) continuous cryotherapy application for 5 h retarded injury, and 3) secondary injury occurred after primary injury.

Implications and recommendations.

This was the first study to attempt to verify the existence of secondary injury in muscle tissue. The documentation of secondary injury lends some credence to the 20-yr-old but previously unexamined secondary injury model (21). Likewise, the hypothesis that cold can retard secondary injury was supported. In fact, this is the first empirical evidence that cryotherapy is effective at the cellular level when used to treat musculoskeletal injuries.

Because this was a preliminary investigation, the clinical relevance of these findings is somewhat limited. To improve the clinical applicability, additional research is required in the following areas: 1) The time course for the progression of secondary injury needs to be examined. 2) The effect of application duration of the cryotherapy needs to be examined. 3) A correlation between tissue temperature and secondary injury needs to be established. 4) The interaction of application duration, tissue temperature, and compression force needs to be examined.


1. Abramson, D. I. Physiologic basis for the use of physical agents in peripheral vascular disorders. Arch. Phys. Med. Rehabil. 46:216–244, 1965.
2. Armstrong, R. B. Initial events in exercise-induced muscular injury. Med. Sci. Sports Exerc. 22:429–435, 1990.
3. Barcroft, H., and O. G. Edholm. The effect of temperature on blood flow and deep temperature in the human forearm. J. Physiol. 102:5–20, 1943.
4. Basset, S. W., and B. M. Lake. Use of cold applications in the management of spasticity. Phys. Ther. Rev. 38:333–334, 1958.
5. Belkin, M., R. D. Brown, J. G. Wright, W. W. Lamonte, and R. W. Hobson. A new quantitative spectrophotometric assay of ischemia-reperfusion injury in skeletal muscle. Am. J. Surg. 156:83–86, 1988.
6. Belkin, M., W. L. Lamorte, J. G. Wright, and R. W. Hobson. The role of leukocytes in the pathophysiology of skeletal muscle ischemic injury. J. Vasc. Surg. 10:14–18, 1989.
7. Bennett, D. Water at 67° to 69° Fahrenheit to control hemorrhage and swelling encountered in athletic injuries. Athl. Train. 1:12–14, 1961.
8. Block, E. R., M. P. Jawaharlal, and D. Edwards. Mechanism of hypoxic injury to pulmonary artery endothelial cell membranes. Am. J. Physiol. (Cell Physiol. 26) 257:C223–C231, 1989.
9. Bugaz, R. The cooling, analgesic, and rewarming effects of ice massage on localized skin. Phys. Ther. 55:11–19, 1975.
10. Cheng, H. T., and H. Chang. Reduction of reperfusion injury in rat skeletal muscle following administration of cinnamophilin, a novel dual inhibitor or thromboxane synthase and thromboxane A2 receptor. Thor. Cardiovasc. Surg. 43:73–76, 1995.
11. Colburn, M. D., W. J. Quinones-Baldrich, H. A. Gelabert, H. Nowara, and W. S. Moore. Standardization of skeletal muscle ischemic injury. J. Surg. Res. 52:309–313, 1992.
12. Fisher, B. D., V. E. Baracos, T. K. Shnitka, S. W. Mendryk, and D. C. Reid. Ultrastructural events following acute muscle trauma. Med. Sci. Sports Exerc. 22:185–193, 1990.
13. Glynn, L. E., and H. P. Himsworth. The intralobular circulation in acute liver injury by carbon tetrachloride. Clin. Sci. 6:235–245, 1948.
14. Guyton, A. C. Textbook of Medical Physiology, 8th Ed. Philadelphia: W. B. Saunders Co., 1991, pp. 149–203.
15. Hartviksen, K. Ice therapy in spasticity. Acta Neurol. Scand. Suppl. 3:79–84, 1962.
16. Hobbs, K. T. Results of intramuscular temperature changes at various levels after the application of ice. Sport Health: Official Gazette of Australian Sports Med. 1:15, 1983.
17. Jones, D. P. Renal metabolism during normoxia, hypoxia, and ischemic injury. Ann. Rev. Physiol. 48:33–50, 1986.
18. Jozsa, L., A. Reffy, B. J. Balint, M. Jarvinen, and M. Kvist. Ischemic-hypoxic changes in muscles with tendon injuries. Magy. Traumatol. Orthop. Helyreallito Sebesz. 24:218–226, 1981.
19. Jozsa, L., A. Reffy, Z. Demel, and I. Szilagyi. Changes of the tissue CO2 and O2 tension in the healing process of crushed muscles. Magy. Traumatol. Orthop. Helyreallito Sebesz. 24:48–51, 1981.
20. Klein, H., P. Puschmann, J. Schaper, and W. Schaper. The mechanism of the tetrazolium reaction in identifying experimental myocardial infarction. Virch. Arch. 393:287–291, 1981.
21. Knight, K. L. Effects of hypothermia on inflammation and swelling. Athl. Train. 11:7–10, 1976.
22. Knight, K. L., J. Aquino, S. M. Johannes, and C. D. Urban. A re-examination of Lewis’ cold-induced vasodilation in the finger and the ankle. Athl. Train. 15:238–250, 1980.
23. Knight, K. L., and J. F. Elam. Rewarming of the ankle, finger, and forearm after cryotherapy: further investigation of Lewis’ CIVD. J. Can. Athl. Ther. Assoc. 8:17–18, 1981.
24. Knight, K. L. Cryotherapy in Sports Injury Management. Champaign, IL: Human Kinetics, 1995, pp. 3–98.
25. Leadbetter, W. B. An introduction to sports-induced soft-tissue inflammation. In: Sports Induced Inflammation, W. B. Leadbetter, J. A. Buckwalter, and S. L. Gordon (Eds.). Chicago: American Academy of Orthopedic Surgeons, 1990, pp. 3–23.
26. Lowden, B. J., and R. J. Moore. Determinants and nature of intramuscular temperature changes during cold therapy. Am. J. Phys. Med. 54:223–233, 1975.
27. Majno, G., and I. Joris. Cells, Tissues, and Disease: Principles of General Pathology. Cambridge, MA: Blackwell Science, 1996, pp. 71–227.
28. Makitie, J., and H. Teravainen. Histochemical studies of striated muscle after temporary ischemia in the rat. Acta Neuropathol. 37:101–109, 1977.
29. Marmur, J. D., M. Poon, M. Rossikhina, and M. B. Taubman. Induction of PDGF-responsive genes in vascular smooth muscle: Implications for the early response to vessel injury. Circulation 3:(Suppl.)III-53–III-60, 1992.
30. Meeusen, R., and P. Lievens. The use of cryotherapy in sports injuries. Sports Med. 3:398–414, 1986.
31. Merrick, M. A., K. L. Knight, C. D. Ingersoll, and J. A. Potteiger. The effects of ice and compression wraps on intramuscular temperatures at various depths. J. Athl. Train. 28:236–245, 1993.
32. Reed, B. Wound healing and the use of thermal agents. In: Thermal Agents in Rehabilitation, 3rd Ed., S. Michlovitz (Ed.). Philadelphia: F. A. Davis Co., 1996,
33. Olsson, R. A. Local factors regulating cardiac and skeletal muscle blood flow. Ann. Rev. Physiol. 43:385, 1981.
34. Sapega, A. A., R. B. Heppenstall, D. P. Sokolow, et al. The bioenergetics of preservation of limbs before replantation: the rationale for intermediate hypothermia. J. Bone Joint Surg. 70-A:1500–1513, 1988.
35. Stauber, W. T., V. K. Fritz, T. E. Burkovskaya, and I. Ilyina-Kakueva. Effect of injury on mast cells of rat gastrocnemius muscle with respect to gravitational exposure. Exp. Mol. Pathol. 59:87–94, 1993.
36. Tidball, J. G. Inflammatory cell response to acute muscle injury. Med. Sci. Sports Exerc. 27:1022–1032, 1995.


© 1999 Lippincott Williams & Wilkins, Inc.