Cryotherapy is perhaps the simplest and oldest therapeu-tic modality in the treatment of acute soft-tissue injuries.It is proposed that by decreasing tissue temperature, icecan diminish pain, metabolism, and muscle spasm, mini-mizing the inflammatory process and thereby aidingrecovery after soft-tissue trauma.32 The majority ofresearch studies and reviews to date have used healthyhuman subjects to investigate these proposed physiologi-cal effects. Although there is evidence that cryotherapycan reduce deep-tissue temperature in both animal2 andhuman subjects,16,46,65 the degree of cooling seems todepend on the method and duration of application, the ini-tial temperature of the ice, and even the depth of subcuta-neous fat.40

Few literature reviews have considered the clinical evi-dence base. Kerr30 attempted to produce clear, evidence-based guidelines for an optimal cryotherapy protocol; how-ever, the majority of recommendations for practice werefinalized by expert consensus. A recent systematic reviewof the original literature provided preliminary recommen-dations for an optimal treatment protocol40; however, fewclinical studies were discussed, and conclusions werederived from studies using only animal or healthy humansubjects. To date, no review has measured the quality ofthe study methodology or considered the clinical appropri-ateness of applied treatments within cryotherapyresearch.

The current recommendations in standard textbooks onthe clinical use of ice also have many shortcomings,39 andmost physicians rely on empirical evidence. The selectionof parameters in a clinical environment continues to bemade pragmatically, and recommendations in review arti-cles range from 10 to 20 minutes 2 to 4 times per day,29 upto 20 to 30 minutes,60 or 30 to 45 minutes31,32 every 2hours. The most recent surveys of clinical practice have

The Use of Ice in the Treatment of Acute Soft-Tissue InjuryA Systematic Review of Randomized Controlled Trials

Chris Bleakley,*† BSc (Hons), MCSP, Suzanne McDonough,† PhD, MCSP, and Domhnall MacAuley,‡ MD, FISMFrom the †Rehabilitation Science Research Group, University of Ulster at Jordanstown, Antrim,Ireland, and the ‡Department of Epidemiology, Queens University, Belfast, Ireland.

Background: There are wide variations in the clinical use of cryotherapy, and guidelines continue to be made on an empiricalbasis.

Study Design: Systematic review assessing the evidence base for cryotherapy in the treatment of acute soft-tissue injuries.

Methods: A computerized literature search, citation tracking, and hand searching were carried out up to April 2002. Eligiblestudies were randomized-controlled trials describing human subjects recovering from acute soft-tissue injuries and employinga cryotherapy treatment in isolation or in combination with other therapies. Two reviewers independently assessed the validityof included trials using the Physiotherapy Evidence Database (PEDro) scale.

Results: Twenty-two trials met the inclusion criteria. There was a mean PEDro score of 3.4 out of of 10. There was marginal evi-dence that ice plus exercise is most effective, after ankle sprain and postsurgery. There was little evidence to suggest that theaddition of ice to compression had any significant effect, but this was restricted to treatment of hospital inpatients. Few studiesassessed the effectiveness of ice on closed soft-tissue injury, and there was no evidence of an optimal mode or duration of treat-ment.

Conclusion: Many more high-quality trials are needed to provide evidence-based guidelines in the treatment of acute soft-tis-sue injuries.

Keywords: ice; cryotherapy; soft-tissue injury; acute

251

* Address correspondence and reprint requests to Chris Bleakley,University of Ulster, Jordanstown, Rehabilitation Science ResearchGroup, Shore Road, Newtownabbey, Co. Antrim, BT37OQB Ireland.

The American Journal of Sports Medicine, Vol. 32, No. 1DOI: 10.1177/0363546503260757© 2004 American Orthopaedic Society for Sports Medicine

DOCTYPE = ARTICLE Team Physician’s Corner

252 Bleakley et al. The American Journal of Sports Medicine

identified variations on the optimal mode, duration, andfrequency of ice application,28,30 yet such factors dictatethe degree of cooling40 and the potential effectiveness oftreatment. In addition, ice is commonly combined withcompression and elevation, making it difficult to deter-mine the value of cryotherapy alone.45,55,61

Although cryotherapy has been promoted in the imme-diate18,32,41,43,45,52,60 and rehabilitative31,32,55 care of soft-tis-sue injury, the basis for its application at each stage isquite different. Immediately postinjury, ice is principallyused to reduce metabolism, thereby minimizing secondaryhypoxic injury and the degree of tissue damage.31,32 In con-trast, when applied for rehabilitative purposes, it is usedprimarily to relieve pain, which facilitates earlier andmore aggressive exercise.31,32 Currently, many cliniciansdo not fully understand the pathophysiological rationaleat each stage and may not be using it to its full advan-tage.31

Cryotherapy is an accessible and popular treatmentmodality for the physician and layman, and its use mustbe supported by high-quality research evidence. Therefore,the aim of this study is to explore the clinical evidencebase for cryotherapy, and the specific objectives are the fol-lowing:

1. to identify randomized-controlled studies assess-ing the effect of cryotherapy on acutely injuredhuman subjects;

2. to assess for the presence of confounding concomi-tant therapies;

3. to study the modes, duration, and frequency ofcryotherapy treatments employed and assess forevidence of an optimal treatment protocol;

4. to identify when cryotherapy was initiated in rela-tion to the injury and study the goals of treatmentin each study, that is, for immediate care or reha-bilitation; and

5. to make conclusions on the strength of the evi-dence supporting the use of cryotherapy in treat-ing acute soft-tissue injuries and make recommen-dations for future research.

METHODS

Search Strategy and Selection of Studies

Relevant studies were identified using a computer-basedliterature search on a total of eight databases: Medline onOvid (1966 to April 2002), Proquest (1986 to April 2002),ISI Web of Science (1981 to April 2002), Cumulative Indexto Nursing and Allied Health (CINAHL) on Ovid (1982 toApril 2002), the Allied and Complementary MedicineDatabase (AMED) on Ovid (1985 to April 2002), theCochrane Database of Systematic Reviews, the CochraneDatabase of Abstracts of Reviews of Effectiveness, and theCochrane Controlled Trials Register (Central) (last searchApril 2002). For the Medline, CINAHL, and AMED data-bases, the first two levels of the Medline optimal searchstrategy8 were combined with the following medical sub-

ject headings and free-text topic words: surgery,orthopaedics, sports injury, soft-tissue injury, sprains andstrains, contusion, athletic injury, acute, compression,cryotherapy, ice, RICE, and cold. The remaining five data-bases had less sophisticated interfaces; therefore, a differ-ent search strategy was performed. To maximize the yieldof relevant articles, this strategy sacrificed precision infavor of sensitivity.9 A series of 68 small searches were per-formed on each database by combining 13 keywords (sur-gery, orthopaedic, sport, injury, soft-tissue, sprain, strain,contusion, compression, cryotherapy, ice, RICE, and acute)using Boolean logic (AND). In addition, a smaller searchwas undertaken on the Physiotherapy Evidence Database(PEDro) (1966 to April 2002) using five keywords (ice,cryotherapy, cold, injury, and surgery). This was supple-mented with citation tracking of relevant primary andreview articles (n = 63) and all incoming full-text papers(n = 55). A convenience sample of 10 key journals was alsohand searched to identify articles that may have beenmissed in database and reference list searches (Table 1).

To be included within the review, studies had to fulfillthe following conditions: the study should be a random-ized-controlled trial of human subjects; it should bepublished in English as a full paper; subjects should berecovering from acute soft-tissue injuries or orthopaedicsurgical interventions; therapy should be inpatient, outpa-tient, or home-based cryotherapy treatment, used either inisolation or in combination with placebo or other thera-pies; comparisons should have been made to no treatment,placebo, a different mode or protocol of cryotherapy, orother physiotherapeutic interventions; and outcome meas-ures must have included at least one of the following: func-tion (subjective or objective), pain, swelling, or range ofmovement (ROM).

In the first stage of selection, the titles and abstracts ofall studies were assessed for the above eligibility criteria.If it was absolutely clear from information provided in the

TABLE 1Hand Search of Key Journals

Journal Year range of search

British Journal of Sports Medicine 1988 to December 2001International Journal

of Sports Medicine 1995 to December 2001Australian Journal

of Sports Medicine 1984 to December 2001Sports Medicine 1987 to December 2001Medicine and Science in

Sports and Exercise 1978 to December 2001Journal of Sports Medicine

and Physical Fitness 1988 to December 2001Journal of Orthopaedics and

Sports Physical Therapy 1986 to December 2001Archives of Physical Medicine

and Rehabilitation 1989 to December 2001Physiotherapy 1988 to December 2001Physical Therapy 1986 to December 2001

Vol. 32, No. 1, 2004 Ice for Acute Soft-Tissue Injury 253

title and/or abstract that the study was not relevant, itwas excluded. If it was unclear from the available abstractand/or title, the full-text article was retrieved. Full-textarticles were also retrieved for studies with a relevant titlebut no available online abstract. There was no blinding tostudy author, place of publication, or results. The primaryresearcher assessed the content of all full-text articles,making the final inclusion/exclusion decisions.

Assessment of Methodological Quality and Data Extraction

All eligible articles were rated for methodological quality,using the PEDro scale. Derived from the Delphi list,62 thisscale consists of an 11-item checklist, configured by expertconsensus to rate the quality of randomized-controlled tri-als50 (Table 2). It is routinely used to rate the quality ofrandomized-controlled trials on the PEDro(ptwww.cchs.usyd.edu.au/pedro). Reviewed studies wereawarded one point for each criterion that was clearly sat-isfied. As criterion 1 is a measure of the study’s externalvalidity, it was not included in the final PEDro score, giv-ing each study a possible maximum score of 10 on thePEDro scale. To increase the accuracy of the PEDro rat-ings, two independent reviewers assessed the quality ofeligible studies. Disagreement or ambiguous issues, whicharose between the first two raters, were resolved by eitherconsensus discussion or consultation with a third party.

Data Extraction and Analysis

The primary reviewer extracted all study characteristicsand data into summary tables. The type of acute soft-tis-sue injury and interventions applied was noted. For eachintervention, the mode, duration, frequency of cryotherapy,surface temperature of the cooling device, subjects’ tissuetemperature, and concomitant therapies were recorded.Attempts were also made to determine when cryotherapy

was initiated in relation to the time of the injury and thespecific purpose for its application, that is, for immediatecare or for rehabilitative purposes in conjunction withtherapeutic exercise. Means and standard deviations forthe four key outcome measures were extracted, and wherepossible, individual study-effect estimates were calculated.This took the form of standardized mean differences(SMD)24 for continuous data or risk ratios (RR) for dichoto-mous data, each with 95% confidence intervals (95% CI).25

RESULTS

From the initial examination of citations yielded from theliterature search, 55 studies were included. After review ofthe complete texts, 33 studies were excluded, leaving 22eligible randomized controlled trials to be included in thereview. Figure 1 shows the Quality of Reporting of Meta-Analysis (QUORUM) statement flow diagram,49 summa-rizing the process of study selection and the number ofstudies excluded at each stage, with reasons.

Study Quality

The 10 criteria and final scores assigned to each study arepresented in Table 3. Overall, the source of subjects andtheir eligibility criteria were well reported. Randomizationwas stringently performed, and only four studies1,15,63,64

employed unsatisfactory methods. In contrast, a verysmall number of studies provided adequate information onsubjects’ baseline data,11,15,20,35,38 and only three used con-cealed allocation during subject recruitment.34,36,38 In gen-eral, blinded application of treatment intervention wasalso poor; none of the studies blinded the therapistsadministering therapy, and just one group of subjects57

was blinded. In addition, in all but four trials6,19,36,59 therewas insufficient blinding of outcome assessment. Intentionto treat analysis was adequately performed in just onestudy,26 and eight10,11,27,34,35,47,51,57,64 supplied adequate

TABLE 2Physiotherapy Evidence Database (PEDro) Scoring Scale

1. Eligibility criteria were specified Yes/No2. Subjects were randomly allocated in groups 13. Allocation was concealed 14. The groups were similar at baseline regarding the most important prognostic indicators 15. There was blinding of all subjects 16. There was blinding of all therapists who administered the therapy 17. There was blinding of all assessors who measured at least one key outcome 18. Measures of at least one key outcome were obtained from more than 85% of the subjects initially

allocated to groups 19. All subjects from whom outcome measures were available received the treatment or control condition

as allocated or, when this was not the case, data for at least one key outcome were analyzed by “intention to treat” 1

10. The results of between-group statistical comparisons are reported for at least one key outcome measure 111. The study provides both point measures and measures of variability for at least one key outcome 1

Total points 10

254 Bleakley et al. The American Journal of Sports Medicine

information on patient dropout. Between-group statisticalcomparisons were well reported, however, and the majori-ty also included measures of group variability. Final val-ues were low, ranging from 1 to 5, with a mean PEDroscore of 3.4 of 10.

Study Characteristics

The study population, intervention, outcomes, follow-up,and reported results of the assessed trials were extractedand tabulated. Twenty-two trials were included, using atotal of 1469 subjects. The sample size ranged from 21 to143, and the mean number of subjects used was 66.7; how-ever, only one study26 undertook a power analysis prior tocommencement of the trial. Patients had a wide variety ofacute injuries. There were no studies using subjects withmuscle contusions or strains, and only five used subjectswith acute ligament sprains.11,35,47,59,64 The remaining 17studies used patients recovering from a range of operativeprocedures: ACL reconstruction,1,6,10,15,19,34,51,58 total kneearthroplasty (TKA),20,23,27,38,57a total hip arthroplasty,57b

knee arthroscopy,36,63 lateral retinacular release,4 andcarpal tunnel release (CTR).26

Table 4 summarizes the mode, duration, and frequencyof cryotherapy; the total cryotherapy treatment time (over-all dosage); the time cryotherapy was initiated in relationto the injury; and the number of days of treatment for eachincluded study. In total, five different modes were used:crushed or chipped ice, Cryocuff or cold compressivedevices, commercial ice machines, commercial/gel icepacks, and ice submersion. Five studies10,47,58,63,64 simplystated that an ice bag or pack was applied, and 8 stud-ies4,10,23,26,34,58,63,64 used more than one mode of cooling dur-ing the trial. Similarly, the duration and frequency ofcryotherapy treatments were not consistent across stud-ies. A total of 13 studies applied cryotherapy continuouslyafter injury, 7 studies employed an intermittent protocol,and 5 failed to specify the protocol. With such an array oficing protocols, the total treatment time subjects receivedwas extremely diverse. For one group of subjects, theentire course of cryotherapy treatment consisted of just 20minutes cooling35 compared to others whose treatmenttime ranged from 21657 to 336 hours.58

The temperature of the cooling device and the subject’stissue temperature reduction were poorly reported.Although a number of studies4,10,26,27,34,51,57 using commer-cial machines stated the temperature of water flowingthrough the machine, they failed to provide adequateinformation on the actual surface temperature of the cool-ing device. Skin temperature reduction during treatmentwas reported in just one study,34 with another measuringintra-articular knee temperature.51 Three studies4,20,26

No Hits.

Medline CINAHL Proquest WOS AMED Cochrane

3321 390 5039 4640 130 1819 (CDSR)183 (DARE)3644 (CCTR)

Potentially relevant studies retrieved for detailed evaluation of full text (n=55)

Studies excluded after evaluation of full text (n =33)

Animal subjects (n=3)Healthy human subjects (n=2)Experimentally induced injury (n=2)Non-acute injury (n=2)Controlled trial / Observational trial / Case study (n=14)Inadequate outcome measures (n=1)Variable subject inclusion criteria (n=3)No cryotherapy treatment applied (n= 4)Ice treatment standardized across groups (n=2)

Studies eligib le for inclusion in systematic review (n=22)

Figure 1. The Quality of Reporting of Meta-Analysis (QUO-RUM) statement flow diagram. CINAHL = Cumulative Indexto Nursing and Allied Health; WOS = Web of Science; AMED =Allied and Complementary Medicine Database; CDSR = Coch-rane Database of Systematic Reviews; DARE = CochraneDatabase of Abstracts of Reviews of Effectiveness; CCTR =Cochrane Controlled Trials Register.

TABLE 3Final Physiotherapy Evidence Database (PEDro)

Scores for Included Trials

Criterion no. TotalAuthor satisfieda score

Cote et al.11 2, 4, 8, 10, 11 5Michlovitz et al.47 2, 8, 10, 11 4Lessard et al.36 2, 3, 7, 10, 11 5Hochberg26 2, 9, 10, 11 4Healy et al.23 2, 10 2Schroder and Passler58 2, 10, 11 3Konrath et al.34 2, 3, 8, 10 4Whitelaw et al.63 10 1Laba35 2, 4, 8 3Sloan et al.59 2, 7, 10 3Edwards et al.19 2, 7, 10, 11 4Cohn et al.10 2, 8, 10, 11 4Wilkerson and Horn-Kingery64 8, 10, 11 3Ivey et al.27 2, 8, 10, 11 4Scarcella and Cohn57 2, 5, 8, 10, 11 5Dervin et al.15 4, 10, 11 3Barber et al.1 10 1Ohkoshi et al.51 2, 8, 10, 11 4Bert et al.4 2 1Levy and Marmar38 2, 3, 4, 10, 11 5Gibbons et al.20 2, 4, 10 3Brandsson et al.6 2, 7, 10, 11 4

a Criterion 1 is not included in final PEDro score.

Vol. 32, No. 1, 2004 Ice for Acute Soft-Tissue Injury 255

stated that cryotherapy was initiated immediately aftersurgery; however, they failed to provide a quantifiabletime period. Likewise, others stated that cooling beganprior to tourniquet release,58 in the recovery room,27 in theoperating theatre,10,19,57 or after wound closure and dress-ing application.1,6,7,38,51 Most studies using subjects postankle sprain initiated cryotherapy between day 1 and 3postinjury.11,35,47,59 One study64 initiated cryotherapy in the

acute phases of injury but again failed to state a definitetime period. Few studies reported the specific goals ofcryotherapy, and it is not clear whether cooling wasemployed for immediate care or for rehabilitative purpos-es. Only two studies11,36 stated that cryotherapy wasapplied in conjunction with exercise, for rehabilitative pur-poses. It seems that the majority of stud-ies,1,10,20,26,27,34,38,57,58 despite continuing cryotherapy for

TABLE 4Cryotherapy Protocol Employed Within Included Studiesa

Totalcryotherapy

time(overall

Rx duration No. days dosage) Time/place ofStudyb Mode (hours) No. Rx/day treated (hours) cryotherapy initiation

Cote et al.11 Water bath + ex’s 0.3 1 3 1 Third day postinjuryMichlovitz et al.47 Ice pack 0.5 1 3 1.5 1–28 hours postinjuryLessard et al.36 Gel pack + ex’s 0.3 4 7 9.3 At home after dischargeHochberg26a Commerical m 12 1 3 36 Immediately after surgeryHochberg26b Crushed ice;

commercial p 0.3 18 3 18 Immediately after surgeryHealy et al.23a Cryocuff — — — — UnclearHealy23b Crushed ice — — — — UnclearSchroder and Passler58a Cryocuff Continuous Continuous 14 336 Prior to tourniquet releaseSchroder et al.58b Ice bags — 3 — — UnclearKonrath et al.34a Commercial m — — 3—5 days

post D/C — UnclearKonrath34b Crushed ice — — — — UnclearWhitelaw et al.63a Cryocuff — — — — UnclearWhitelaw63b — — — — — UnclearLaba35 Crushed ice 0.3 1 1 0.3 Day 0–2 since injurySloan et al.59 Commercial p 0.5 1 1 0.5 Within 24 hours of injuryEdwards et al.19 Cryocuff Continuous Continuous 1.5 36 In operating theatreCohn et al.10a Commercial m Continuous Continuous 4 96 In operating theatreCohn10b Ice bag — 1 1 — In recovery roomWilkerson and Horn-Kingery64 Ice pack 0.5 1 Acute phasec 1.5 Acute stagesc

Wilkerson64 Commercial p Continuous Continuous Acute phasec 64 Acute stagesc

Ivey et al.27 Commercial m Continuous Continuous 3 64 In recovery roomScarcella and Cohn57a Commercial m Continuous Continuous 9 216 In operating theatreScarcella57b Commercial m Continuous Continuous 9 216 In operating theatreDervin et al.15 Cryocuff Continuous Continuous 2.5 55–60 UnclearBarber et al.1 Commercial m Continuous Continuous 3 (POD 1-3) 64 After application of postopera-

— — 3 (POD 4–6) 48 tive dressingaverage

Ohkoshi et al.51 Commercial m Continuous Continuous 2 48 After surgical wound was coveredBert et al.4 Commercial m/p Continuous Continuous 1–2 27 Immediately postsurgery in

recovery roomLevy and Marmar38 Cryocuff Continuous Continuous 3 64 After skin closure and dressing

were appliedGibbons et al.20 Cryocuff 6 (at least) 1 13 (at least) 78 Immediately after the surgical

procedureBrandsson et al.6 Cryocuff Continuous Continuous 1 24 After surgical wounds were

closed

a Rx duration = treatment duration; No. Rx/day = number of treatments per day; + ex’s = Exercises incorporated with cooling; commer-cial m = Commercial icing machine; Commercial p = Commercially produced ice pack; — = information not reported; D/C = discharge; POD= Postoperative day.

b The superscripts “a” and “b” after the study reference number depict that a single study applied more than one cryotherapy protocolc “Acute” stage of injury not specified.

256 Bleakley et al. The American Journal of Sports Medicine

days and even weeks after the immediate stages of trau-ma, chose not to incorporate functional movements orexercise.

Outcome Measures

Pain (visual analogue scale and/or analgesic consumption)was the most common outcome measure, but only twostudies provided adequate data for function.57,64 Fewdichotomous measures were used, and the majority ofstudies recorded continuous measures over short timeperiods (for example, 1-week postinjury). The longestreported follow-ups were measures of pain, swelling, andROM recorded at 4 weeks postinjury58; however, insuffi-cient data were available for effect size calculation. Thelongest follow-up data from which an effect size could becalculated was a measurement of knee ROM recorded 2weeks postsurgery.38 In total, nine studies6,19,20,23,34,57-59,63

failed to provide sufficient data for any of the key outcomemeasures, and it was not possible to calculate individualstudy effect estimates (either SMD or RR).

Effectiveness of Treatment

A total of 12 treatment comparisons were made. Table 5subgroups the studies according to treatment comparisonand provides the sample size, overall PEDro score, andeffect size estimates for individual studies (SMD, RR).Fourteen studies1,4,10,11,15,26,27,35,36,38,47,51,57,64 provided suffi-cient data for calculation of individual effect sizes (SMD,RR) for at least one of the key outcome measures. Thesevalues are provided in the right-hand columns, with a pos-itive SMD or RR representing an effect in favor of thetreatment group (for example, group A if the groups arecompared as A versus B). Any significant differencesbetween groups reported in the text are based on the Pvalues (P < 0.05) provided in the original studies.

Although it is evident that a number of studies carriedout the same treatment comparisons, the effect sizes fromindividual trials could not be pooled for statistical analy-sis. This was due to heterogeneity of the study population,intervention mode and dosage, timing and type of outcomemeasures, or insufficient reporting of data.

Ice Versus Heat/Contrast Bath. There was some evi-dence that cryotherapy was more effective than ther-motherapy after ankle injury. A single study11 found thatice submersion with simultaneous exercises was signifi-cantly more effective than heat (SMD, 1.38; 95% CI, 0.35to 2.29) and contrast therapy (SMD, 2.35; 95% CI, 1.13 to3.37) plus simultaneous exercises, at reducing swellingbetween 3 and 5 days post ankle sprain.

Ice Versus Ice and Electrical Stimulation. A singlestudy47 compared the effect of ice alone to ice and simulta-neous high-voltage electrical stimulation after acute anklesprains. There was no significant difference when compar-ing ice alone and ice combined with low-frequency electri-cal stimulation (28 pulses per second) in terms of swelling(SMD, –0.47; 95% CI, –1.34 to 0.44), pain (SMD, –0.64; 95%CI, –1.51 to 0.28), and ROM (SMD, –0.69; 95% CI, –1.56 to

0.24). Similarly, there was no significant difference com-paring ice alone and ice combined with higher frequencyelectrical stimulation (80 pulses per second) in terms ofswelling (SMD, –1.39; 95% CI, –2.3 to 0.36), pain (SMD,–0.62; 95% CI, –1.5 to 0.31), and ROM (SMD, –1.36; 95%CI, –2.3 to –0.3).

Ice Versus No Ice. Ice alone seems to be more effectivethan applying no form of cryotherapy after minor kneesurgery. A single study36 compared the effect of an inter-mittent icing protocol combined with knee exercises toexercises alone, after minor arthroscopic knee surgery. Theapplication of ice immediately before a rehabilitation pro-gram significantly decreased pain as measured by theaffective component of the McGill Pain Questionnaire(SMD, 0.59; 95 CI, –0.02 to 1.17). The study also reportedthat subjects applying cryotherapy used significantly lessprescription and nonprescription analgesia and had a sig-nificantly better weightbearing status; however, insuffi-cient data are provided for the calculation of an effect size.In contrast, there were no significant differences betweengroups in terms of knee girth (SMD, 0.35; CI, –0.24 to 0.93)and knee ROM (SMD, 0.38; CI, –0.21 to 0.97) 1 week post-surgery.

Ice (Continuous) Versus Ice (Intermittent). Using sub-jects post-CTR, Hochberg26 compared the effect of continu-ous cryotherapy to intermittent 20-minute ice applicationsover the first 3 postoperative days. Subjects applying con-tinuous cryotherapy had a significantly greater decreasein pain (SMD, 1.09; CI, 0.4 to 1.7) and wrist circumference(SMD, 2.2; CI, 1.43 to 2.9) in comparison to those usingcryotherapy intermittently. This was the only study tocompare the effectiveness of two different cryotherapy pro-tocols, and although it appears that continuous cryothera-py should be the treatment of choice after surgery, themodes of cryotherapy application were not consistentacross the two groups.

Ice and Compression Versus Ice and Compression. Fourstudies23,34,58,63 compared two different methods of apply-ing simultaneous compression and cryotherapy, but fewconclusions could be reached. Poor reporting of data meantthat individual effect size could not be calculated for anyof these studies. Furthermore, two studies58,63 did not pro-vide adequate information on the mode of cryotherapy, andall failed to specify the duration and frequency of the iceapplication.

Ice and Compression Versus No Ice. There is marginalevidence that a single simultaneous treatment with iceand compression is no more effective than no cryotherapyafter an ankle sprain. Laba35 found that a single applica-tion of ice and compression, in addition to standard reha-bilitation treatment (ultrasound, mobility, and propriocep-tive exercises), produced similar levels of swelling (RR,0.76; CI, 0.5 to 1.02) and pain immediately posttreatment(RR, 1.5; CI, 1.24 to 1.76) and at discharge (RR, 0.88; CI,0.62 to 1.14) when compared to those receiving standardtreatment only. Sloan59 also found that a single applicationof simultaneous ice and compression was as effective as notreatment in terms of reducing pain, swelling, and ROMpost ankle sprain. Similarly, Edwards19 found that the con-

Vol. 32, No. 1, 2004 Ice for Acute Soft-Tissue Injury 257T

AB

LE

5E

ffec

t S

ize

Est

imat

es f

or I

ndi

vidu

al S

tudi

esa

Eff

ect

size

(95

% C

I)

Inte

rven

tion

Inju

ryb

NF

un

ctio

nP

ain

Sw

elli

ng

RO

MP

ED

ro (

10)

Ice

vs h

eat

An

kle11

30—

—1.

38 (

0.35

–2.2

9)—

5Ic

e vs

con

tras

tA

nkl

e1130

——

2.35

(1.

13–3

.37)

—5

Ice

vs i

ce +

E-S

tim

(F

req

28 p

ps)

An

kle47

a30

—–0

.64

(–1.

51 t

o 0.

28)

–0.4

7 (–

1.34

to

0.44

) (d

ay 1

);–0

.69

(–1.

56 t

o 0.

24)

(day

1);

–0.1

4 (–

1.01

to

0.75

) (d

ay 3

)–0

.58

(–1.

45 t

o 0.

24)

(day

3)

4Ic

e vs

ice

+ E

-Sti

m (

Fre

q 80

pps

)A

nkl

e47b

30—

–0.6

2 (–

1.5

to 0

.3)

–1.3

9 (–

2.3

to 0

.36)

(da

y 1)

;–1

.36

(–2.

3 to

–0.

3) (

day

1);

–0.0

9 (–

0.96

to

0.8)

(da

y 3)

–0.3

9 (–

1.3

to 0

.5)

(day

3)

4Ic

e vs

No

Rx

Art

h36

45—

0.24

(–0

.35

to 0

.82)

(T

);0.

59 (

–0.0

2 to

1.1

7) (

A)

0.35

(–0

.24

to 0

.93)

0.38

(–0

.21

to 0

.97)

5Ic

e (c

onti

nu

ous)

vs

ice

(in

term

itte

nt)

CT

R26

48—

1.09

(0.

4–1.

7)2.

2 (1

.43–

2.9)

—4

I/C

vs

I/C

TK

A23

76—

N/A

N/A

N/A

2A

CL

5844

—N

/AN

/AN

/A3

AC

L34

a10

0—

N/A

—N

/A4

Art

h63

102

—N

/AN

/AN

/A1

I/C

vs

No

Rx

An

kle35

30N

/A1.

5 (1

.24–

1.76

);0.

88 (

0.62

–1.1

4)0.

76 (

0.5–

1.02

)—

3A

nkl

e5914

3—

N/A

N/A

N/A

3A

CL

19a

63—

N/A

—N

/A4

I/C

vs

ice

AC

L10

54N

/A4.

43 (

3.3–

5.24

);4.

49 (

3.41

–5.4

)—

—4

I/C

vs

C (

sam

e m

ode)

An

kle64

a34

–0.1

4 (–

0.97

to

0.7)

——

—3

TK

A27

90—

–0.4

3 (–

0.95

to

0.1)

——

4T

KA

5724

–0.7

5 (–

1.55

to

0.1)

——

0.39

(–0

.44

to 1

.18)

5A

CL

34b

100

N/A

N/A

—N

/A4

AC

L19

b63

—N

/A—

N/A

4A

CL

1578

N/A

–0.3

3 (–

0.7

to 0

.12)

(V

AS

);–0

.17

(–0.

6 to

0.3

) (A

/ges

ic);

–0.0

9 (–

0.5

to 0

.4)

(IV

)—

—3

AC

L1

99—

N/A

—1.

14 (

1.0;

1.28

)1

AC

L51

a21

—–0

.6 (

–1.6

4 to

0.5

) (V

AS

);0.

3 (–

0.75

to

1.36

) (A

/ges

ic)

—1.

02 (

–0.1

6 to

2.0

5)4

AC

L51

b21

—1.

21 (

0.0–

2.2)

(V

AS

)0.

89 (

–0.2

6 to

1.9

2)0.

8 (–

0.27

to

1.9)

(A

/ges

ic)

4T

HA

5750

—N

/A—

—5

I/C

vs

C (

Dif

f M

ode)

LR

R4

110

Ove

rall

sco

re:

0.35

(0.

27 t

o 0.

42)

2A

nkl

e64b

340.

55 (

–0.3

2 to

1.3

8)—

——

3T

KA

3880

—0.

75 (

0.3–

1.2)

(V

AS

[D2]

);0.

64 (

0.19

–1.0

8) (

day

7);

0.41

(–0

.04

to 0

.85)

(A

/ges

ic)

—0.

89 (

0.42

–1.3

4) (

day

14)

5T

KA

2060

—N

/A—

N/A

3I/

C +

P v

s P

vs

I/C

an

d I/

A i

nj

AC

L6

50—

N/A

——

4

aS

tudi

es a

re g

rou

ped

acco

rdin

g to

th

e tr

eatm

ent

com

pari

son

s em

ploy

ed.A

pos

itiv

e st

anda

rdiz

ed m

ean

dif

fere

nce

or

risk

rat

io r

epre

sen

ts a

n e

ffec

t in

fav

or o

f th

e tr

eatm

ent

grou

p (f

or e

xam

ple,

grou

p A

if t

he g

roup

s ar

e co

mpa

red

as A

vs

B).

Eff

ect

size

= r

elat

ive

risk

rat

io;R

OM

= r

ange

of m

ovem

ent;

PE

Dro

= P

hysi

othe

rapy

Evi

denc

e D

atab

ase;

Ank

le =

an

kle

spra

in;—

= o

utc

ome

not

mea

sure

d;E

Sti

m =

ele

ctri

cal s

tim

ula

tion

;pps

= p

uls

es p

er s

econ

d;N

o R

x =

no

trea

tmen

t;T

= t

otal

McG

ill Q

ues

tion

nai

re s

core

;A =

aff

ecti

ve c

ompo

nen

tM

cGil

l qu

esti

onn

aire

sco

re;C

TR

= c

arpa

l tu

nn

el r

elea

se;I

/C =

sim

ult

aneo

us

ice

and

com

pres

sion

;Sam

e =

mod

e of

com

pres

sion

con

stan

t ac

ross

gro

ups

;TK

A =

tot

al k

nee

art

hro

-pl

asty

;N/A

= d

ata

not

ava

ilab

le;A

CL

= a

nte

rior

cru

ciat

e li

gam

ent

reco

nst

ruct

ion

;Art

h =

art

hro

scop

y;I

= ic

e tr

eatm

ent;

C =

com

pres

sion

;VA

S =

vis

ual

an

alog

ue

scal

e;A

/ges

ic =

oral

an

alge

sic

con

sum

ptio

n;I

/V =

in

trav

enou

s an

alge

sic

con

sum

ptio

n;T

HA

= t

otal

hip

art

hro

plas

ty;D

iff

= m

ode

of c

ompr

essi

on d

iffe

red

acro

ss g

rou

ps;L

RR

= l

ater

al r

etin

acu

lar

rele

ase;

P =

pla

cebo

;I/A

in

j = i

ntr

a-ar

ticu

lar

anal

gesi

c in

ject

ion

.b

Th

e su

pers

crip

ts “

a”an

d “b

”af

ter

the

stu

dy r

efer

ence

nu

mbe

r de

pict

th

at a

sin

gle

stu

dy a

ppli

ed m

ore

than

on

e cr

yoth

erap

y pr

otoc

ol

258 Bleakley et al. The American Journal of Sports Medicine

tinuous use of ice and compression had similar benefits tono treatment in terms of improving pain and ROM whenapplied postsurgically; however, insufficient data wereprovided for these later two studies.19,59

Ice and Compression Versus Ice. Only one clinical studyhas compared ice and compression to ice alone.10 The com-bination of treatments appeared to be significantly moreeffective than ice in terms of reducing the amount of intra-muscular (SMD, 4.43; 95% CI, 3.3 to 5.24) and oral anal-gesia (SMD, 4.49; 95% CI, 3.4 to 5.4) administeredpost–ACL reconstruction. These results must be interpret-ed with caution, however, as the mode and duration of icetreatment were not controlled for across groups.

Ice and Compression Versus Compression. The majorityof included studies have tried to disentangle the effects ofice from compression by comparing a variety of treatmentcombinations. In four studies, it was difficult to comparethe efficacy of each modality4,20,38,64 as the mode of com-pression differed between the intervention and controlgroups. On the contrary, eight studies strictly controlledfor the type of compressive bandages used across compar-ison groups1,15,19,27,34,51,57,64; however, there seemed to be lit-tle difference in the effectiveness of ice and compressionand compression alone.

Wilkerson64 found no significant difference in the time ofrestricted activity after ankle sprain in subjects treatedwith compression alone and simultaneous ice and com-pression (SMD, –0.14; 95% CI, –0.97 to 0.7). Using subjectspost–ACL reconstruction, others reported no significantdifferences between groups in terms of function,34 pain,19,34

and swelling1,19,34; however, insufficient data were report-ed and effect size could not be calculated for these out-comes. Similarly, Dervin15 found no significant differencesin subjective pain scores (SMD, –0.33; 95% CI, –0.77 to0.12) and the amounts of intravenous (SMD, –0.09; 95%CI, –0.53 to 0.35) and oral analgesics (SMD, –0.17; 95% CI,–0.62 to 0.27). In a group of subjects post-TKA, Ivey27

found no significant difference between groups withregard to the amount of injected morphine (SMD, –0.43;95% CI, –0.95 to 0.1) postsurgery. Scarcella57 found no sig-nificant difference in subjects post-TKA in terms of ROM(SMD, 0.39; 95% CI, –0.44 to 1.18) and the time to inde-pendent ambulation (SMD, –0.75; 95% CI, –1.55 to 0.1).The study57 also reported that the analgesic consumptionin each group was almost identical. Correspondingly, in asubgroup of patients recovering from total hip arthroplas-ty, Scarcella57 reported no significant differences in anal-gesic consumption postsurgery; however, insufficient datawere provided and effect size could not be calculated.

Only two studies reported significant differencesbetween subjects treated with ice and compression andcompression alone. Although Barber1 found no differencesbetween groups in knee ROM after ACL reconstruction(RR, 1.14; 95% CI, 1.0 to 1.28), a significantly decreasedanalgesic consumption was reported in favor of the ice andcompression group; however, inadequate data were provid-ed. Again using subjects post–ACL reconstruction,Ohkoshi51 treated two groups with simultaneous ice andcompression and a third with compression only. The iceand compression groups were cooled to slightly different

temperatures using a commercial ice machine (5°C and10°C). Subjects using less extreme cooling (10°C group)with concomitant compression had significantly lower sub-jective pain scores (SMD, 1.21; 95% CI, 0.00 to 2.2) andanalgesic consumption (SMD, 0.88; 95% CI, –0.27 to 1.91)compared to those using compression alone. In contrast,there were no significant differences in subjects treatedwith simultaneous cooling (5°C group) and compressionand those treated with compression only in terms of sub-jective pain scores (SMD, –0.6, 95% CI, –1.64 to 0.5) andanalgesic consumption (SMD, 0.3, 95% CI, –0.75 to 1.36). Abetter improvement in ROM was observed in the 5°C(SMD, 1.02; 95% CI, –0.16 to 2.05) and 10°C groups (SMD,0.89; 95% CI, –0.26; 1.92) when compared to compressiongroup; however, these differences were not significant.

Therefore, despite eight trials comparing the effective-ness of ice and compression to compression alone, onlytwo1,51 reported significant differences in favor of ice andcompression. Both Barber’s1 and Ohkoshi et al.’s51 studieswere of low quality, scoring just 1 out of 10 and 4 out of 10on the PEDro scoring scale, respectively, and therefore thestrength of their conclusions is limited. Generally, therewas very little evidence to suggest that the addition of iceto compression has any significant effect. It must be noted,however, that all but one of the studies64 were undertakenpostsurgery, and any conclusions are restricted to hospitalinpatients with postsurgical wound dressings.

Ice and Compression Plus Placebo Injection Versus Iceand Compression Plus Injection Versus Placebo Injection.Brandsson et al.6 found that ice and compression plus aplacebo injection were significantly more effective thanplacebo injection alone at reducing postoperative pain. Theaddition of a pain-killing injection to ice and compressiontherapy significantly improved the analgesic effect fur-ther; however again, no data were provided and effect sizecould not be calculated.

DISCUSSION

Cryotherapy continues to be employed in both the clinicaland sporting environments to treat acute soft-tissueinjuries as well as postsurgical patients within a hospitalsetting.42 A number of review articles have advocated theuse of cryotherapy in both of these contexts,31,41,42,55 andothers have scrutinized its physiological and clinicaleffects.30,40,43,45,52,61 This is the first study to systematicallyreview the literature, assessing the clinical evidence basesupporting the use of cryotherapy based on the highestquality research evidence. The review is restricted toEnglish language, however, and as the inclusion criteriafor study population were broad, some of the informationcontained was difficult to compare and synthesize. Theincluded randomized controlled trials scored an averagePEDro score of only 3.4, and the contrast in treatment pro-tocols means that comparison within and across studies isoften impossible. Moreover, persistent methodologicalproblems and the failure of the majority of studies to carryout a power analysis may prevent wider extrapolation ofevidence.

Vol. 32, No. 1, 2004 Ice for Acute Soft-Tissue Injury 259

Soft-tissue injuries such as contusions, strains, andsprains are the most common injuries in Gaelic football,12

soccer,22 and rugby.21 To date, however, no randomizedstudies have assessed the efficacy of ice in the treatmentof muscle contusions or strains, and only five studies haveassessed the effect of ice on acute ankle sprains. Singleapplications of combined ice and compression seem to beas effective as no treatment after an acute sprain; howev-er, these conclusions must be taken with caution. Asidefrom the paucity of high-quality studies undertaken, thisparticular research question may also be subject to aunique set of problems inherent to cryotherapy research.Given the strong empirical evidence base and the popular-ity of cryotherapy treatment with the layman, it may bedifficult to randomize a subject to a “no ice” group. This isparticularly evident in Laba’s35 study, in which 60% of sub-jects randomized to the no ice group had already appliedice as a self-treatment prior to recruitment.

Evidence from a recent systematic review suggestedthat intermittent 10-minute ice treatments are most effec-tive at cooling injured animal tissue and healthy humantissue.40 The effectiveness of this particular protocol hasnot yet been tested on injured human subjects; however,Hochberg26 found that intermittent 20-minute applica-tions are less effective than continuous ice treatment afterCTR surgery. The strength of the study’s conclusions isgreatly limited, however, as Hochberg26 crucially failed tocontrol for the mode of cryotherapy across the continuousand intermittent groups. No individual study has rigor-ously compared the efficacy of different modes, durations,or frequencies of ice treatment, and preliminary recom-mendations for an optimal cryotherapy protocol cannot bemade.

Other systematic reviews5 have provided evidence-based guidelines on optimal treatment parameters by sub-grouping trials to highlight a dose-dependant pattern.Although it was the author’s intention to carry this out,subgrouping trials according to the mode or duration ofcryotherapy was impossible due to clinical heterogeneityand the large number of trials supplying insufficient treat-ment detail.

Within clinical practice, ice is commonly combined withcompression and elevation, making it difficult to deter-mine the value of cryotherapy alone.45,55,61 A number ofstudies have compared a wide range of combinations of iceand compression in a bid to try and disentangle their rel-ative efficacy. Only one study10 compared simultaneous iceand compression to ice alone. This study does little to sep-arate and quantify the individual effects of ice and com-pression as both the modes and durations of cold treat-ments applied across groups were starkly contrasting.

Twelve studies compared the effectiveness of concomi-tant ice and compression to compression alone, but onlyeight well-controlled studies1,18,22,30,37,55,61,64 used the samemode of compression between intervention and controlgroups. The initial consensus seems to be that the additionof ice to compression is no more effective than compressionalone. However, such a conclusion is limited, as in all eightof these studies, postsurgical dressings or socks were usedto separate the injured area of the body and the cooling

device. The thickness of dressings varied from gauze57 tocast padding and an elastic bandage27; such barriers havethe potential to mitigate the cooling effect of cold com-press.

To maximize the therapeutic effects of cryotherapy, anoptimal tissue temperature reduction of 10° to 15°C maybe necessary.40 Skin temperature reductions to below13.6°C may be needed to achieve local analgesia,7 and per-haps lower tissue temperatures of between 10°C and 15°Cmay be required to maximally lower metabolism.33,56

Generally, the surface temperature of the cooling deviceand the subject’s tissue temperature reduction duringtreatment were poorly reported in this review. The onlystudy34 that monitored skin tissue temperatures duringtreatment reported a maximum reduction to just 28°C.Correspondingly, there is evidence from many stud-ies13,14,37,53 that it is difficult to achieve optimal tissue tem-perature reductions when cooling is applied over postoper-ative dressings. The interaction between the cooling sur-face and the subjects’ tissue is vital in determining theeffectiveness of treatment and must be considered infuture studies, particularly within a postsurgical environ-ment.

There have been some deleterious side effects ofcryotherapy previously documented. A number of casestudies have reported the occurrence of skin burns54 andnerve damage3,17,44,48 after as little as 20 to 30 minutes ofcooling. Within this review, there was just one reportedcase of cold-induced nerve palsy, possibly caused by a con-tinuous 40-minute ice application in the recovery roompostsurgery.10 None of the other studies reported any inci-dences of skin burns or nerve palsies, despite applying con-tinuous ice treatments for between 6 and 226 hours.

Cryotherapy is a versatile modality and may be used inthe immediate18,32,41,43,45,52,60 and rehabilitative31,55 phasesof injury management. However, a common source of con-fusion is the basis for its application at each phase.Immediately postinjury, ice reduces tissue metabolism,thereby minimizing secondary hypoxic injury, cell debris,and edema. The sooner after injury cryotherapy is initiat-ed, the more beneficial this reduction in metabolism willbe.31 A number of studies35,47,59 began cryotherapy between24 and 48 hours after injury and therefore may not haveoptimized this positive physiological effect. It may be eas-ier to initiate early cryotherapy in studies using surgicalpatients. Although most surgical studies stated thatcryotherapy was initiated either immediately after sur-gery, in the operating theatre, or after dressing and woundclosure, few significant differences were reported. Again,this may be due to concomitant compression or woundbandaging mitigating the cooling effect and preventingadequate metabolic reduction.

Outside the immediate stages of injury management,cryotherapy may be most effective when combined withexercise.31,32 Adequate cooling can reduce pain, spasm, andneural inhibition, thereby allowing for earlier and moreaggressive exercises. In the current review, many stud-ies1,10,20,26,27,34,38,57,58 continued cryotherapy treatment fordays and even weeks after injury but chose not to inte-grate therapeutic exercise. Although cryotherapy in isola-

260 Bleakley et al. The American Journal of Sports Medicine

tion may reduce the need for analgesics in the subacutephases of recovery, early exercise may be the most impor-tant component of treatment.31,32 Only two studies11,36

incorporated exercise with cryotherapy, during the suba-cute phases of recovery, and both recorded results signifi-cantly in favor of cryotherapy. Nonetheless, it seems thatthe majority of studies in this review have not fully con-sidered the pathophysiological basis of cryotherapy andmay not have used it to its full potential. Future studiesmust seek to optimize cryotherapy’s effects at each phaseof injury management to provide clinicians with clearerevidence of its potential effectiveness and versatility.

CONCLUSION

Many more high-quality studies are needed to ensure thatclinicians and sportsmen are following evidence-basedguidelines in the treatment of acute soft-tissue injuries.Primarily, these must focus on developing modes, dura-tions, and frequencies of ice application, which will opti-mize cryotherapy during immediate and rehabilitativecare. Similarly, an optimal mode and duration of compres-sion treatment must be highlighted. This evidence willhighlight the respective value of each individual modalityand if appropriate provide the basis of an optimal methodfor treatment combination.

REFERENCES

1. Barber AF, McGuire DA, Click S: Continuous-flow cold therapy foroutpatient anterior cruciate ligament reconstruction. Arthroscopy 14:130–135, 1998

2. Barlas D, Homan CS, Thode HC Jr: In vivo tissue temperature com-parison of cryotherapy with and without external compression. AnnEmerg Med 28: 436–439, 1996

3. Bassett FH III, Kirkpatrick JS, Engelhardt DL, et al: Cryotherapy-induced nerve injury. Am J Sports Med 20: 516–518, 1992

4. Bert JM, Stark JG, Maschka K, et al: The effect of cold therapy onmorbidity subsequent to arthroscopic lateral retinacular release.Orthopaedic Review 20: 755–758, 1991

5. Bjordal JM, Couppe C, Ljunggren AE: Low level laser therapy fortendinopathy: Evidence of a dose response pattern. Physical TherapyReviews 6: 91–99, 2001

6. Brandsson S, Rydgren B, Hedner T, et al: Postoperative analgesiceffects of an external cooling system and intra-articular bupiva-caine/morphine after arthroscopic cruciate ligament surgery. KneeSurg Sports Traumatol Arthrosc 4: 200–205, 1996

7. Bugaj R: The cooling, analgesic and rewarming effects of ice mas-sage on localized skin. Phys Ther 55: 11–19, 1975

8. Clarke M, Oxman AD, eds: Optimal search strategy for RCT’s.Cochrane Reviewers Handbook 4.1.4 [updated October 2001];Appendix 5c. In: The Cochrane Library, Issue 4, 2001. Oxford:Update Software

9. Clarke M, Oxman AD, eds: Locating and selecting studies. CochraneReviewers Handbook 4.1.4 [updated October 2001]; Section 5.6. In:The Cochrane Library, Issue 4, 2001. Oxford: Update Software

10. Cohn BT, Draeger RI, Jackson DW: The effects of cold therapy in thepostoperative management of pain in patients undergoing anteriorcruciate ligament reconstruction. Am J Sports Med 17: 344–349,1989

11. Cote DJ, Prentice WE, Hooker DN, et al: Comparison of three treat-ment procedures for minimizing ankle sprain swelling. Phys Ther 68:1064–1076, 1988

12. Cromwell F, Walsh J, Gormley J: A pilot study examining injuries inelite Gaelic footballers. Br J Sports Med 34: 104–108, 2000

13. Culp RW, Taras JS: The effect of ice application versus controlledcold therapy on skin temperature when used with postoperativebulky hand and wrist dressings: A preliminary study. J Hand Ther 8:249–251, 1995

14. Daniel DM, Stone ML, Arendt DL: The effect of cold therapy on pain,swelling, and range of motion after anterior cruciate ligament recon-structive surgery. Arthroscopy 10: 530–533, 1994

15. Dervin GF, Taylor DE, Keene GC: Effects of cold and compressiondressings on early postoperative outcomes for the arthroscopic ante-rior cruciate ligament reconstruction patient. J Orthop Sports PhysTher 27: 403–406, 1998

16. Draper DO, Schulthies S, Sorvisto P, et al: Temperature changes indeep muscles of humans during ice and ultrasound therapy: An invivo study. J Orthop Sports Phys Ther 21: 153–157, 1995

17. Drez D, Faust DC, Evans JP: Cryotherapy and nerve palsy. Am JSports Med 9: 256–257, 1981

18. Dyment PG: Management of minor soft tissue trauma in adolescentathletes. J Adolesc Health Care 7: 133S–135S, 1986

19. Edwards DJ, Rimmer M, Kenne GC: The use of cold therapy in thepost operative management of patients undergoing arthroscopicanterior cruciate ligament reconstruction. Am J Sports Med 24:193–195, 1996

20. Gibbons CER, Solan MC, Ricketts DM, et al: Cryotherapy comparedwith Robert Jones bandage after total knee replacement: A prospec-tive randomized trial. Int Orthop 25: 250–252, 2001

21. Gissane C, Jennings D, Kerr K, et al: A pooled analysis of injury inci-dence in rugby league football. Sports Med 32: 211–216, 2002

22. Hawkins RD, Hulse MA, Wilkinson C, et al: The association of foot-ball medical research programme: An audit of injuries in professionalfootball. Br J Sports Med 35: 43–47, 2001

23. Healy WL, Seidman J, Pfeiffer BA, et al: Cold compressive dressingafter total knee arthroplasty. Clin Orthop 299: 143–146, 1994

24. Herbert RD: How to estimate effects from reports of clinical trials. I:Continuous outcomes. Aust J Physiother 46: 229–235, 2000

25. Herbert RD: How to estimate effects from reports of clinical trials. II:Dichotomous outcomes. Aust J Physiother 46: 309–313, 2000

26. Hochberg J: A randomized prospective study to assess the efficacyof two cold therapy treatments following carpal tunnel release. JHand Ther 14: 208–215, 2001

27. Ivey M, Johnston RV, Uchida T: Cryotherapy for postoperative painrelief following knee arthroplasty. J Arthroplasty 9: 285–290, 1994

28. Johannsen F, Langberg H: The treatment of acute soft tissue traumain Danish emergency rooms. Scand J Med Sci Sports 7: 178–181,1997

29. Kellnet J: Acute soft tissue injuries: A review of the literature. Med SciSports Exerc 18: 489–500, 1986

30. Kerr KM, Daily L, Booth L: Guidelines for the Management of SoftTissue (Musculoskeletal) Injury with Protection, Rest, Ice,Compression and Elevation (PRICE) during the First 64 Hours.London, Chartered Society of Physiotherapy, 1999

31. Knight KL, Brucker JB, Stoneman PD, et al: Muscle injury manage-ment with cryotherapy. Athletic Therapy Today 5: 26–30, 2000

32. Knight K: Cryotherapy in sports injury management. Int PerspectPhysiother 4: 163–185, 1989

33. Knight KL: Effects of hypothermia on inflammation and swelling. JAthletic Train 11: 7–10, 1976

34. Konrath GA, Lock T, Goitz HT, et al: The use of cold therapy afteranterior cruciate ligament reconstruction. Am J Sports Med 24:629–633, 1996

35. Laba E: Clinical evaluation of ice therapy for acute ankle spraininjuries. NZ J Physiother 17: 7–9, 1989

36. Lessard LA, Scudds RA, Amendola A, et al: The efficacy of cryother-apy following arthroscopic knee surgery. J Orthop Sports Phys Ther26: 14–22, 1997

37. Levy AS, Kelly B, Lintner S, et al: Penetration of cryotherapy in treat-ment after shoulder arthroscopy. Arthroscopy 13: 461–464, 1997

Vol. 32, No. 1, 2004 Ice for Acute Soft-Tissue Injury 261

38. Levy AS, Marmar E: The role of cold compression dressings in thepostoperative treatment of total knee arthroplasty. Clin Orthop 297:174–178, 1993

39. MacAuley D: Do textbooks agree on their advice on ice? Clin J SportMed 11: 67–64, 2001

40. MacAuley D: Ice therapy: How good is the evidence? Int J SportsMed 22: 379–384, 2001

41. McClean DA: The use of cold and superficial heat in the treatment ofsoft tissue injuries. Br J Sports Med 23: 53–54, 1989

42. McDowell JH, McFarland EG, Nalli BJ: Use of cryotherapy fororthopaedic patients. Orthop Nurs 13: 21–30, 1994

43. McMaster WC: A literary review on ice therapy in injuries. Am JSports Med 5: 124–126, 1977

44. Malone TR, Engelhardt DL, Kirpatrick JS, et al: Nerve injury in ath-letes caused by cryotherapy. J Athletic Train (Dallas) 27: 235–237,1992

45. Meeusen R, Lievens P: The use of cryotherapy in sports injuries.Sports Med 3: 398–414, 1986

46. Merrick MA, Knight KL, Ingersoll CD, et al: The effects of ice andcompression wraps on intramuscular temperatures at variousdepths. J Athletic Train 28: 236–245, 1993

47. Michlovitz S, Smith W, Watkins M: Ice and high voltage pulsed stim-ulation in treatment of acute lateral ankle sprains. J Orthop SportsPhys Ther 9: 301–304, 1988

48. Moeller JL, Monroe J, McKeag DB: Cryotherapy-induced commonperoneal nerve palsy. Clin J Sport Med 7: 212–216, 1997

49. Moher D, Cook DJ, Eastwood S, et al: Improving the quality ofreports of meta-analyses of randomized controlled trials: The QUO-RUM statement. Lancet 354: 1896–1900, 1999

50. Moseley A: Evidence for physiotherapy practice: A survey of thePhysiotherapy Evidence Database (PEDro). Aust J Physiother 48:43–49, 2002

51. Ohkoshi Y, Ohkoshi M, Nagasaki S, et al: The effect of cryotherapyon intraarticular temperature and postoperative care after anteriorcruciate ligament reconstruction. Am J Sports Med 27: 357–362,1999

52. Olson JE, Stravino VD: A review of cryotherapy. Phys Ther 52:840–853, 1964

53. Osbahr DC, Cawley PW, Speer KP: The effect of continuouscryotherapy on glenohumeral joint and subacromial space tempera-tures in the postoperative shoulder. Arthroscopy 18: 748–754, 2002

54. O’Toole G, Rayatt S: Frostbite at the gym: A case report of an icepack burn. Br J Sports Med 33: 278–279, 1999

55. Rivenberg DW: Physical modalities in the treatment of tendoninjuries. Clin Sports Med 11: 645–659, 1992

56. Sapega AA, Heppenstall B, Sokolow DP, et al: The bioenergetics ofpreservation of limbs before replantation. J Bone and Joint Surg 70A:1500–1513, 1988

57. Scarcella JB, Cohn BT: The effect of cold therapy on the postopera-tive course of total hip and knee arthroplasty patients. Am J ofOrthop Nov: 847–852, 1997

58. Schroder D, Passler HH: Combination of cold and compression afterknee surgery: A prospective randomized study. Knee Surg SportsTraumatol Arthrosc 2: 158–165, 1994

59. Sloan JP, Hain R, Pownall R: Clinical benefits of early cold therapy inaccident and emergency following ankle sprain. Arch Emerg Med 6:1–6, 1989

60. Swenson C, Sward L, Karlsson J: Cryotherapy in sports medicine.Scand J Med Sci Sports 6: 193–200, 1996

61. Thorsson O: Cold therapy of athletic injuries: A literature review.Lakartidningen 98: 1512–1513, 2001

62. Verhagen AP, de Vet HCW, de Bie RA, et al: The Delphi List: A crite-ria list for quality assessment of randomized clinical trials for con-ducting systematic reviews developed by Delphi consensus. J ClinEpidemiol 51: 1235–1241, 1998

63. Whitelaw GP, DeMuth KA, Demos HA, et al: The use of cryocuff ver-sus ice and elastic wrap in postoperative care of knee arthroscopypatients. Am J Knee Surg 8: 28–31, 1995

64. Wilkerson GB, Horn-Kingery HM: Treatment of the inversion anklesprain: Comparison of different modes of compression and cryother-apy. J Orthop Sports Phys Ther 17: 240–246, 1993

65. Zemke JE, Anderson JC, Guion WK, et al: Intramuscular temperatureresponses in the human leg to two forms of cryotherapy: Ice mas-sage and ice bag. J Orthop Sports Phys Ther 27: 301–307, 1998