Acupuncture for treating acute ankle sprains in adults

Tae‐Hun Kim; Myeong Soo Lee; Kun Hyung Kim; Jung Won Kang; Tae‐Young Choi; Edzard Ernst

doi:10.1002/14651858.CD009065.pub2

. 2014 Jun 23;2014(6):CD009065. doi: 10.1002/14651858.CD009065.pub2

Acupuncture for treating acute ankle sprains in adults

Tae‐Hun Kim ¹, Myeong Soo Lee ^2,^✉, Kun Hyung Kim ³, Jung Won Kang ⁴, Tae‐Young Choi ², Edzard Ernst ⁵

Editor: Cochrane Bone, Joint and Muscle Trauma Group

PMCID: PMC10514753 PMID: 24953665

Abstract

Background

An acute ankle sprain is a sudden‐onset injury of one or more of the ankle ligaments. It is one of the most common musculoskeletal injuries in the general population as well as in athletes. In some countries, such as China and Korea, acupuncture is frequently used in the treatment of ankle sprains, either as a single treatment or a secondary intervention accompanied by standard medical treatment.

Objectives

To assess the effects (benefits and harms) of acupuncture for the treatment of ankle sprains in adults.

Search methods

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (May 2013), the Cochrane Central Register of Controlled Trials (The Cochrane Library 2013, Issue 4), MEDLINE (1948 to May week 2 2013), EMBASE (1980 to May week 2 2013), China National Knowledge Infrastructure databases (1994 to August week 4 2013), the Cumulative Index to Nursing and Allied Health Literature (1937 to May 2013), the Allied and Complementary Medicine Database (1985 to May 2013), Science Links Japan (1996 to August week 4 2013), several Korean medical databases (August week 4 2013), the World Health Organization International Clinical Trials Registry Platform (August week 4 2013), the bibliographic references of included trials and conference proceedings.

Selection criteria

We included randomised and quasi‐randomised controlled trials involving adults with acute ankle sprains. We included all types of acupuncture practices, such as needle acupuncture, electroacupuncture, laser acupuncture, pharmacoacupuncture, non‐penetrating acupuncture point stimulation (e.g. acupressure and magnets) and moxibustion. Acupuncture could be compared with control (no treatment or placebo) or another standard non‐surgical intervention.

Data collection and analysis

Two review authors independently screened the search results, assessed trial eligibility, assessed risk of bias and extracted data from the included trials. We calculated risk ratios (RRs) for dichotomous outcomes and mean differences for continuous outcomes. We conducted meta‐analyses using the fixed‐effect method or, where appropriate, the random‐effects method, and used 95% confidence intervals (CI) throughout.

Main results

We included a total of 20 heterogeneous studies (2012 participants with acute ankle sprains); three of which included more than one comparison. Seventeen trials were conducted in China. All of the studies had a high risk of bias due to lack of blinding. The results may also have been affected by selection bias, particularly as five studies were quasi‐randomised controlled trials and 12 studies gave no information on their method of randomisation. Of our three prespecified primary outcomes, only cure rate was reported by the majority of studies. No study reported on patient‐reported assessment of function and only one reported on adverse events (in which three participants receiving a control intervention experienced skin problems from over‐the‐counter Chinese herbal patches). The other 19 studies did not record or report on adverse events. We assessed the quality of evidence for cure rates as very low for all comparisons, which means we are very uncertain about the reliability of any of the estimates.

The single study comparing acupuncture treatment with no treatment found acupuncture to be more effective with regard to cure rate at five days (31/31 versus 1/30; RR 20.34, 95% CI 4.27 to 96.68). Acupuncture plus another standard treatment versus that standard treatment alone was tested in eight studies; with cure rate data available for seven. Most of these studies reported higher cure rates in the acupuncture plus another standard treatment group than in the standard treatment alone group. However, while the results of an exploratory meta‐analysis of cure rate data from eight trials testing acupuncture versus no acupuncture tended to favour acupuncture, the results were very inconsistent across the studies and the estimated effect was very imprecise (383/396 versus 272/355; RR 1.32, 95% CI 0.95 to 1.84; P value = 0.1; I² = 98%).

Fourteen studies compared acupuncture with a variety of other non‐surgical treatments, such as Chinese drug patches, hot and cold water, ice packs, oral Chinese herbal medicine and elastic bandage. Some studies found in favour of acupuncture, some in favour of the other treatment and some found a lack of evidence for a difference between the two interventions under test. The results of an exploratory meta‐analysis of cure rate data from 11 trials testing acupuncture versus another non‐surgical intervention tended to slightly favour acupuncture, but these were not statistically significant and the data were very heterogeneous (404/509 versus 416/497; RR 1.07, 95% CI 0.94 to 1.22; P value = 0.30; I² = 92%).

Authors' conclusions

The currently available evidence from a very heterogeneous group of randomised and quasi‐randomised controlled trials evaluating the effects of acupuncture for the treatment of acute ankle sprains does not provide reliable support for either the effectiveness or safety of acupuncture treatments, alone or in combination with other non‐surgical interventions; or in comparison with other non‐surgical interventions. Future rigorous randomised clinical trials with larger sample sizes will be necessary to establish robust clinical evidence concerning the effectiveness and safety of acupuncture treatment for acute ankle sprains.

Plain language summary

Acupuncture for sudden‐onset ankle sprains in adults

An acute ankle sprain is a sudden‐onset injury of the ankle ligaments (tough strands of tissue that connect and stabilise the bones at the ankle). It is one of the most common injuries in the general population as well as in athletes. Acupuncture is frequently used for treating ankle sprains in eastern Asian countries. This review aimed to assess the benefits and harms of acupuncture for the treatment of ankle sprains in adults. We searched the medical literature for studies up to May 2013.

Our review includes 20 studies involving 2012 people with ankle sprains. These studies differed from each other in many ways and compared various types of acupuncture with a variety of standard control interventions. Most studies reported only the 'cure rate' ‐ the number of participants who had recovered at a set time. No study reported on patient‐reported assessment of function. Only one study reported on adverse events and found skin problems in individuals receiving over‐the‐counter traditional Chinese herbal patches as a control intervention. Most trials had flaws in the way they were conducted, which makes their results less reliable; for example, most studies failed to ensure participants did not know which intervention they were receiving.

One study, which compared acupuncture with no treatment, found more people were cured with acupuncture. Most of the eight studies comparing acupuncture plus another standard treatment versus that standard treatment alone found higher cure rates in the acupuncture group. However, we found that pooling these results did not provide conclusive evidence that acupuncture resulted in a better cure rate.

Fourteen studies compared acupuncture with a variety of other non‐surgical treatments, such as Chinese herbal patches, hot and cold water, ice packs, Chinese oral herbal medicine and elastic bandages. Some studies found in favour of acupuncture, some in favour of the other treatment and some found a lack of evidence for a difference between the two interventions under test. The pooled results from 11 studies comparing acupuncture versus another non‐surgical intervention tended to favour acupuncture, but this evidence was not conclusive.

Currently, we are unable to conclude whether or not acupuncture is more effective than other standard methods for the treatment of ankle sprains in adults because of the very low quality of the available evidence. Because the adverse effects of acupuncture treatment were not described in most of the studies, we are also unable to draw any conclusions about the safety of acupuncture. Large, high quality studies of acupuncture for sudden‐onset ankle sprains in adults are needed.

Background

Description of the condition

An acute ankle sprain is a sudden‐onset injury of one or more of the ankle ligaments, which are tough strands of tissue that connect and stabilise the bones at the ankle. The most common mechanism of injury, which results in damage to the outer or lateral ankle ligaments, is inversion (internal rotation) of a plantar‐flexed foot (toes on ground and heel up) (Wolfe 2001).

Ankle sprain is one of the most common musculoskeletal injuries in the general population as well as in athletes (Junge 2009). Waterman 2010 reported an incidence of ankle sprains in the USA of 2.15 per 1000 person‐years. The higher annual incidence in athletes (about 7 sprains per 1000 person‐years) reflects the fact that ankle sprains are generally related to athletic activities (Beynnon 2001; Holmer 1994).

Sprains can be categorised according to the severity of the injury: grade 1 is a mild overstretching or slight tearing of a ligament that results in pain but with no major loss of function; grade 2 is incomplete tearing of a ligament with pain and some loss of function; and grade 3 indicates a complete rupture usually accompanied by severe pain and bruising, swelling and loss of function (Litt 1992). Physical and possibly radiological examinations are needed to evaluate the severity of the injury. To detect the rupture of an ankle ligament, delayed physical examination, four to five days after ankle trauma, may be more accurate than assessment in the acute stage (within 48 hours) (Van Dijk 1996 Van Dijk 1999). With the aim of avoiding unnecessary radiological examination, the 'Ottawa Ankle Rules' are increasingly being used to assess the likelihood that a foot or ankle fracture is present (Stiell 1994).

The primary goal of the management of an acute ankle sprain is to reduce swelling and pain. The PRICE (Protection, Rest, Ice, Compression and Elevation) technique is usually recommended for this purpose. Ice may reduce the local tissue temperature, which reduces swelling, inflammation and pain (Knight 1995). Compression and elevation limit the effusion of extracellular fluid and thus reduce swelling. Plaster cast immobilisation of the affected limb is one of the most frequently used conservative treatment options. However, systematic reviews have reported that, compared with immobilisation, functional treatment, a programme of early mobilisation along with external support, generates more favourable results than immobilisation (Kerkhoffs 2002). Surgical intervention, involving ligament repair or reconstruction, is often used to treat more severe ankle injuries. Though surgery is an effective way to reduce joint instability, there is no compelling evidence that surgery is superior to conservative management (Kerkhoffs 2007).

Description of the intervention

Acupuncture is an intervention that stimulates specific points on the body surface, generally using needles. It has been used for over 2000 years in eastern Asian countries such as China and Korea (Veith 2002). There are many varieties, such as traditional manual acupuncture, western acupuncture, electroacupuncture, non‐penetrating acupuncture point stimulation (e.g. acupressure) and moxibustion, which differ in the selection of treatment points and stimulating methods. Traditional manual acupuncture usually consists of needling at classic acupuncture points and stimulating the needles manually (i.e. twirling or lifting‐thrusting needles in order to obtain qi (de‐qi)). Electroacupuncture uses an electric current to stimulate acupuncture points. Western acupuncture selects non‐classic points, such as motor points and trigger points, for needling. Moxibustion means applying a burned moxa‐cone to stimulate an acupuncture point.

How the intervention might work

For the treatment of ankle sprains, acupuncture is used as a stand‐alone treatment or a secondary intervention accompanying standard medical treatment (e.g. the PRICE technique or physiotherapy), depending on the clinical situation (Li 2008; Park 2004). Traditional acupuncture is applied according to the various concepts of the balance of Yin and Yang, Qi theory, five element theory, meridian theory and traditional diagnostic methods of oriental medicine (Berman 2010). Recently, acupuncture has been reinterpreted and is used increasingly as 'western medical acupuncture'; this revised technique is based on the knowledge of neurophysiology and scientific methodology (White 2009). Acupuncture is thought to generate analgesic effects through the local, segmental and central regulation of anti‐pain mechanisms (White 2008). In addition, it has been suggested that acupuncture modulates the anti‐inflammatory response and the recovery of soft tissue injuries (Li 2009; Zhang 2005). In particular, specific manipulation techniques (e.g. the winding of acupuncture needles) may be relevant to the cytoskeletal remodelling of connective tissue fibroblasts, which can contribute to the healing of the ligament injury (Langevin 2002; Langevin 2006).

Why it is important to do this review

Ankle sprains are very common and acupuncture is frequently used in their treatment in some countries. For instance, in Korea, sprains are ranked as the third most common cause for visits to Oriental Medicine (OM) clinics and approximately twice as many people visit OM clinics for sprains compared with conventional clinics (National Health Insurance Corporation 2009). As the efficacy of acupuncture for treating ankle sprains remains uncertain, a critical examination of the evidence for the use of acupuncture for ankle sprains is warranted.

Objectives

To assess the effects (benefits and harms) of acupuncture for the treatment of ankle sprains in adults.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised and quasi‐randomised (allocating participants to a treatment by a method that is not strictly random, e.g. by date of birth, hospital record number or alternation) controlled trials. All studies of ankle sprains were considered eligible for inclusion regardless of the reported method of diagnosis. We excluded cross‐over studies.

Types of participants

Adults with acute ankle sprains, irrespective of the method of diagnosis. We excluded studies that included only children. Mixed population studies including adults and children or people with acute and chronic injuries were included if the majority of the trial participants were adults with acute ankle sprains. In these circumstances, we sought separate data for the target population.

Types of interventions

We included all types of acupuncture practices, including needle acupuncture, electroacupuncture, laser acupuncture, pharmacoacupuncture, non‐penetrating acupuncture point stimulation (e.g. acupressure and magnets) and moxibustion.

There were two main comparisons.

1. Acupuncture versus no treatment or placebo (e.g. sham acupuncture). Acupuncture could be provided as the only treatment or in conjunction with another 'standard' non‐surgical treatment as long as the same standard treatment was provided to both groups. Non‐surgical 'standard' treatment interventions included the components of the PRICE technique; physical interventions, such as cast immobilisation, taping, bracing and other functional treatment; physiotherapy; and pharmacotherapy (e.g. oral or topical non‐steroidal anti‐inflammatory drugs (NSAIDs)).

2. Acupuncture versus another standard non‐surgical intervention (as described above). Again, acupuncture could be provided as the only treatment or in conjunction with another standard treatment as long as the same standard treatment was provided to both groups with the exception of the comparator intervention.

Types of outcome measures

Primary outcomes

Patient‐reported assessment of function (assessed using, e.g. the ankle activity score (Halasi 2004), Tegner activity level (Tegner 1985), Kaikkonen functional scale (Kaikkonen 1994), Karlsson ankle function score (Karlsson 1996) and Olerud and Molander ankle score (Rose 2000))
Cure rate (the ratio of the number of cured participants with acute ankle sprains to the number of participants: cure means improvement in both clinical symptoms, such as pain and swelling of the ankle joint, and recovery of ankle function to the participant's pre‐injury health status) (as defined by study authors)
Adverse events related to acupuncture treatment (short‐term events), e.g. severe nausea, fainting, dizziness and vomiting, unexpected and prolonged aggravation of existing problems, emotional reactions, prolonged pain and bruising, and infection (MacPherson 2001)

Secondary outcomes

Recurrence of ankle sprain, subsequent surgery or long‐term treatment
Pain: both dichotomous data (e.g. yes or no) and continuous data (e.g. visual analogue scale (VAS) scores or analgesic consumption)
Ankle instability: either subjective (e.g. giving way) or objective (e.g. talar tilt, anterior drawer test, inversion stress test, postural sway analysis)
Participant‐rated quality of life (assessed using, e.g. the Short‐Form (SF)‐36 health survey)
Swelling: both dichotomous data (e.g. yes or no) and continuous data (e.g. VAS scores)

Timing of outcome measurement

We considered both short‐term (within four weeks of injury) and longer‐term outcomes.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (May 2013), the Cochrane Central Register of Controlled Trials (The Cochrane Library 2013, Issue 4), MEDLINE (1948 to May week 2 2013), EMBASE (1980 to 2013 week 20), China National Knowledge Infrastructure databases (CNKI) (1994 to August week 4 2013), the Cumulative Index to Nursing and Allied Health Literature (CINAHL) (1937 to May 2013), the Allied and Complementary Medicine Database (AMED) (1985 to May 2013), Science Links Japan (J‐East) (1996 to August week 4 2013) and Korean medical databases (Korean Studies Information, DBPIA, Korean Institute of Science and Technology Information, Research Information Centre for Health Database, Korean National Assembly Library and Korean Traditional Knowledge Portal) (August week 4 2013). We also searched the World Health Organization (WHO) International Clinical Trials Registry Platform (August week 4 2013) for ongoing and recently completed studies. No restrictions were applied based on language or publication status.

Search strategies for The Cochrane Library, MEDLINE, EMBASE, AMED and CINAHL are shown in Appendix 1.

Searching other resources

We reviewed the bibliographic references of all included trials to identify other relevant studies. We searched the following unpublished conference proceedings: Korean Acupuncture & Moxibustion Society (1995 to August week 4 2013) and the Korean Academy of Oriental Rehabilitation Medicine (1995 to August week 4 2013).

Data collection and analysis

Selection of studies

Two review authors (Kim T‐H and Kang JW) independently screened all identified trials to assess their eligibility for inclusion. Disagreements were discussed and resolved, if necessary, by the other authors.

Data extraction and management

Two review authors (Kim T‐H and Lee MS) extracted data from the selected reports or studies by independently completing a common predefined data collection form. Any disagreements were resolved by discussion or, where necessary, by arbitration by a third author (Ernst E). We extracted the following characteristics of the reports or studies: study design, duration of the study, trial setting, ethical approval, demographic data of participants (age, sex, country and ethnic group), total number randomised to the acupuncture and control groups, grade of sprain, the number of dropouts, details of acupuncture treatment (type of acupuncture, combination treatments, acupuncture points used and frequency of acupuncture treatments), control interventions and all relevant outcomes. Where possible, we planned to contact the corresponding authors of the included trial reports where there were missing data.

Assessment of risk of bias in included studies

Two review authors (Kim T‐H and Kim KH) assessed the risk of bias for each included study in accordance to the recommendations of The Cochrane Collaboration (Higgins 2011). If there were any disagreements, one author (Lee MS) meditated the dispute and made the final decision. The following domains were evaluated for risk of bias: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting and other sources of bias (see Appendix 2).

Because it is difficult to develop a perfect placebo control that can blind both the acupuncturist and the participant, double‐blind clinical trials are rarely realisable in acupuncture studies (MacPherson 2008). It is necessary, therefore, that the assessment of key outcomes in acupuncture research should be conducted by blinded participants and blinded outcome assessors. Where this was not reported, we assumed it had not been done and assigned a 'high' risk of bias. We tried to contact the authors of included studies if any details necessary for the assessment of risk of bias were unclear. The review authors judged the risk of bias for each domain as either low risk, high risk or 'unclear' (where the risk was unclear or unknown), as based on the criteria listed in Table 8.5d in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Measures of treatment effect

After we compared all of the treatment effects of acupuncture with the matched control interventions in each trial, we calculated summary statistics for the treatment effects for each study. For dichotomous data, we presented the treatment effects as risk ratios (RRs) with 95% confidence intervals (CIs). For continuous data, we presented mean differences (MDs) with 95% CIs. We did not need to use standard mean differences, which are required when pooling outcome data assessed using different measures.

Unit of analysis issues

All included trials were parallel‐group trials; there were no cluster‐randomised or cross‐over trials.

If we had identified any studies reporting multiple observations for outcome variables, we would have classified the time frames of included studies as having short‐term (within four weeks) or long‐term (over four weeks) follow‐up and conducted meta‐analyses accordingly.

Dealing with missing data

We requested missing data from the original study investigators whenever possible. If missing data could not be explicitly determined from information provided by the original authors, we made the assumption that these outcomes were treatment failures and analysed only the data available. While we planned to conduct sensitivity analyses to test these assumptions and discuss their potential impact on our findings, there were insufficient data available to do this.

Assessment of heterogeneity

We used a visual inspection of the forest plot and the Chi² test, using a P value of 0.1 as the cut‐off value to determine statistical significance, to assess the heterogeneity between different studies. We calculated the I² statistic in order to quantify inconsistencies between included studies. If we identified any substantial heterogeneity (I² > 50%) in the primary outcomes between studies, we intended to conduct the sensitivity and subgroup analyses described below to identify possible contributing factors (Deeks 2008).

Assessment of reporting biases

In future updates of this review, we plan to generate funnel plots when more than 10 studies are included in an individual analysis and to inspect them for asymmetry, which could indicate a possible reporting bias (Egger 1997).

Data synthesis

Initially, we conducted meta‐analyses using the fixed‐effect model and 95% CIs. Where there was substantial heterogeneity, especially that which could not be readily explained, we looked at the results using a random‐effects model.

Subgroup analysis and investigation of heterogeneity

Should sufficient data be available in future, we will conduct subgroup analyses according to the severity of the ankle sprain (grade 1, 2 or 3), the type of acupuncture intervention (including manual acupuncture, moxibustion and electroacupuncture) and type of control intervention (placebo/sham acupuncture versus no treatment).

Sensitivity analysis

We conducted a sensitivity analysis to show that our results were robust with respect to the selection of the statistical model (fixed effect versus random effects). Had there been a sufficient number of trials, we would have assessed the influence of different sample sizes (large versus small), different methodological qualities (such as sequence generation, allocation concealment or blinding) or different diagnostic criteria for ankle sprains on the estimates of the overall effect on key outcomes using sensitivity analyses. A lack of contact author details precluded sensitivity analyses testing of the effects of missing data.

Assessing the quality of the evidence and summarising the results

We used the GRADE (Grading of Recommendations Assessment, Development and Evaluation Working Group) approach to assess the quality of evidence relating to the primary outcome 'cure rate'. In future updates of the review, if sufficient data are available, we will create 'Summary of findings' tables, to convey key information about the quality of evidence, the magnitude of effect of the interventions examined and the sum of available data on the key comparisons and outcomes (Schünemann 2009).

Results

Description of studies

See: Characteristics of included studies and Characteristics of excluded studies.

Results of the search

We screened a total of 935 studies. We located 825 records from the following electronic databases: Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (0), Cochrane Central Register of Controlled Trials (35), MEDLINE (58), EMBASE (47), CNKI databases (424), CINAHL (45), AMED (48), J‐EAST (12), Korean medical databases (149) and the WHO International Clinical Trials Registry Platform (7). In addition, 110 records were screened through handsearching conference proceedings from the Korean Acupuncture & Moxibustion Society (56) and the Korean Academy of Oriental Rehabilitation Medicine (54).

The search resulted in the identification of 73 potentially eligible articles, for which (where possible) full reports were obtained. Upon study selection, we included 20 studies (Chen 2012; Cho 1977; Ge 2000; Hao 2006; Jian 2004; Jiang 2011; Ni 2010; Paris 1983; Ruan 1995; Shi 2013; Sun 2011; Wang 2009; Wei 2004; Wu 2006; Wu 2007; Yu 1996; Yu 1999; Zhang 2011; Zhang 2012; Zhou 2008). We found two articles by the same authors that reported identical results; we counted these as one study (Yu 1996). Among 20 studies, 17 were published journal articles (Chen 2012; Ge 2000; Hao 2006; Jian 2004; Jiang 2011; Ni 2010; Paris 1983; Ruan 1995; Shi 2013; Sun 2011; Wei 2004; Wu 2006; Yu 1996; Yu 1999; Zhang 2011; Zhang 2012; Zhou 2008), two were dissertations (Cho 1977; Wu 2007) and one was a conference abstract only (Wang 2009). We excluded 47 articles (Excluded studies), five studies await classification (ACTRN12605000483651; Cohen 2011; Feng 2008; Li 2002; Su 2009) and there are no ongoing studies.

A flow diagram summarising the study selection process is shown in Figure 1.

Included studies

We included a total of 20 studies, involving 2012 participants, in this review. Two studies were three‐armed studies (Wu 2007; Yu 1999), and one was a four‐armed study (Yu 1996). The remaining studies were two‐armed studies.

In one study, acupuncture was compared with no treatment (Wu 2007).

In eight studies, acupuncture plus various types of standard treatment was compared with the same standard treatment alone. Standard treatments included standard physical therapy (Paris 1983); external application of topical NSAIDs (Yu 1999); oral administration or external application of Chinese herbal drugs (Ge 2000; Hao 2006; Yu 1996; Zhang 2011); electrophysiotherapy (Zhang 2012); and a combination of TDP far‐infrared lamp radiation, oral administration and external application of over‐the‐counter (OTC) herbs (Jian 2004).

In 14 studies, acupuncture alone was compared with non‐surgical interventions, including OTC Chinese herbal patches (Jiang 2011;Wei 2004;Wu 2006); external application of Chinese herbal drugs (Yu 1996); external application of topical NSAIDs (Voltaren emulsion, i.e. diclofenac sodium topical gel (Chen 2012; Shi 2013), Dolobene gel, i.e. dimethyl sulphoxide gel (Yu 1999)); hot or cold water (Cho 1977); an ice pack (Yu 1996); traditional Chinese treatments (Ruan 1995; Wu 2007; Zhou 2008); immobilisation using an elastic bandage (Sun 2011); cold and hot external application and oral administration of a Chinese herbal medication (Ni 2010); and a Chinese herbal spray and partial immobilisation using an elastic bandage (Wang 2009).

Participants and diagnostic criteria

The mean number of participants per each included study was 100, ranging from 16 (Paris 1983) to 338 (Ruan 1995). Seventeen studies were conducted in China, and the other three were conducted in South Korea (Cho 1977), Canada (Paris 1983) and Singapore (Wu 2007). Two studies recruited only male participants (Cho 1977; Jian 2004), one study recruited army soldiers (Zhang 2012) and five studies recruited only college students (Chen 2012; Paris 1983;Ruan 1995;Yu 1999; Zhang 2011). Several studies did not describe the age range of participants but most of the participants were adults with the youngest participants being 13 years old in two studies (Ni 2010; Wei 2004). The oldest participant in the studies that provided an upper age limit was 63 years old (Ni 2010). Several studies included adolescent participants (Chen 2012; Ni 2010; Ruan 1995; Wei 2004; Yu 1996; Zhang 2011; Zhang 2012) but separate data on these were not available.

Symptom duration from the onset of ankle sprain was less than one week in the majority of studies but ranged from one to 30 days in one study (Jiang 2011); there was no detailed information on duration reported in three studies (Cho 1977; Wang 2009; Zhang 2012). The severity of ankle sprain was grade 2 in one study (Paris 1983) but most studies included ankle sprains of mixed severity or did not detail severity at all. Only two studies reported the type of ankle sprain clearly: lateral ankle sprains only (Wei 2004) and mixed types (23 lateral and 3 medial in Zhou 2008). The criteria used to diagnose ankle sprains were reported as the Chinese national guidelines for traditional Chinese medicine (State Administration 1994; Chen 2012; Ge 2000;Jian 2004;Jiang 2011;Ni 2010; Shi 2013; Wei 2004;Wu 2007;Zhang 2011), history taking and physical examination (Hao 2006;Ruan 1995;Sun 2011; Yu 1999;Zhou 2008) and stress roentgenography (X‐rays) (Paris 1983). There was no detailed information about the diagnostic criteria in four studies (Cho 1977;Yu 1996; Wang 2009; Zhang 2012). One study mentioned that the authors used diagnostic criteria from literature reviews, without providing any details (Wu 2006).

Interventions

Twelve studies used classic manual body acupuncture, in which acupuncture needles were inserted at the classic acupuncture points or at tender points (Ashi points) and stimulated manually (Cho 1977;Hao 2006;Jian 2004;Jiang 2011;Ni 2010; Ruan 1995;Shi 2013; Sun 2011; Wei 2004;Wu 2007;Yu 1999; Yu 1996). Ear acupuncture was used in two studies (Hao 2006;Paris 1983). Electroacupuncture was used in one study (Zhang 2011) and non‐penetrating electric acupuncture point stimulation was used in one study (Paris 1983). Warm needle therapy, in which burning dried mugwort (moxibustion) was added to the handle of the inserted acupuncture needle, was applied in three studies (Wu 2006;Zhang 2012; Zhou 2008). Fire needle acupuncture, in which thermal stimulation was applied to the body of the acupuncture needle, was used in one study (Wu 2006). Two studies gave no detailed description of the acupuncture stimulation method used (Ge 2000; Wang 2009). Acupressure, non‐penetrating acupuncture point stimulation, was used in one study (Chen 2012). The most frequently selected acupuncture points were BL60 (10 times); GB40 (9), KI3 and Ashi points (8); ST41, ST36 (7); and BL62, KI6 and GB39 (6). BL40, BL59, BL63, GB34, GB42, GB41, LI4, LR4, LR3, KI2, SP4, SP5, SP6, SP9, ST4, ST42, ST43, ST44, tender points and two non‐classic acupuncture points (ankle joint point and Xiaojiexue) were also used. The needle retention time was primarily 20 to 30 minutes (a minimum of 1 minute to a maximum of 40 minutes). The total number of acupuncture sessions ranged from 5 to 10 in most of the studies (a minimum of one to a maximum of 30). The methods used for the acupuncture treatments are provided in detail in the 'Characteristics of included studies' tables.

Several interventions were used as concomitant treatments: standard physical treatment, including cryotherapy, elevation of the extremity and adhesive ankle support (Paris 1983); electrophysiotherapy (Zhang 2012); and OTC herbs (used individually or as a mixture), administered externally (Hao 2006; Jian 2004; Zhang 2011) or administered orally (Ge 2000; Jian 2004). Various interventions were used in the control groups: ice packs (Yu 1996); hot or cold water (Cho 1977); OTC Chinese herbal patches (Jiang 2011;Wei 2004;Wu 2006); Chinese herbal drugs used externally (Hao 2006;Yu 1996; Zhang 2011), orally (Ge 2000) or both (Ni 2010); a Chinese herbal spray (Wang 2009); NSAIDs (Wei 2004); immobilisation using an elastic bandage (Sun 2011; Wang 2009); topical NSAIDs (Dolobene gel (Yu 1999), Voltaren emulsion (Chen 2012; Shi 2013)); and combinations of traditional Chinese treatments, including radiation with a TDP far‐infrared lamp, OTC oral herbal medications and OTC external herbal tinctures (Jian 2004); and traditional Chinese treatments, including Tuina (a manipulative therapy in traditional Eastern Asian medicine) (Ruan 1995), blood‐letting therapy (Wu 2007) and TDP far‐infrared lamp radiation (Zhou 2008).

Outcomes

Sixteen of the 17 studies that were conducted in China reported cure rates (Ge 2000;Hao 2006;Jian 2004;Jiang 2011;Ni 2010; Ruan 1995;Sun 2011; Wang 2009; Wei 2004;Wu 2006;Wu 2007;Yu 1996;Yu 1999;Zhang 2011; Zhang 2012; Zhou 2008). Individual pain, swelling, tenderness and function were assessed in four studies (Chen 2012; Jiang 2011; Shi 2013; Wu 2007). Although Wu 2007 assessed swelling, tenderness and function, we could not include these results because inappropriate score systems were used. Cho 1977 used a functional improvement score based on the percentage difference in the dorsiflexion angles of injured and normal ankle joints. Time needed for recovery was assessed in three studies (Paris 1983; Sun 2011; Wei 2004). Only one study reported adverse events related to the interventions (Wei 2004); the other 19 studies did not describe any adverse events related to either the acupuncture or the control interventions used. In most of the studies, only short‐term outcomes, measured post‐treatment, were reported. None of the included studies conducted follow‐up assessments.

Excluded studies

We excluded a total of 47 studies for a variety of reasons (see 'Characteristics of excluded studies' tables). The reasons were: comparisons of different types of acupuncture (Ahn 2003; Byeon 2003; Guo 2010; He 2006; He 2006a; Hong 2011; Jeong 2009; Ji 2004; Jung 2008; Kang 2008; Kim 2007; Lee 2004; Li 2001; Liao 2010; Luo 2002; Park 2004; Seo 2006; Song 2008; Sun 2011a; Wang 2006; Wang 2008; Woo 2003; Xu 2012; Yu 2007; Zhang 2012a; Zheng 2008); review article (Park 2013; Wimmer 2012; Zhao 2009); observational study (Zhang 2011a); acupuncture treatment not tested (Cheng 2012; Huang 2012; Wang 2011); inclusion of participants with both acute and chronic ankle sprains (Wang 2005), control group used non‐conventional treatments (Ceng 2006; Chen 2012a); comparison groups used different co‐interventions (Duan 2007; Liu 2012; Ru 2008; Shang 2010; Tan 2005; Wu 1995; Xu 2007; Yang 2011; Zhang 2008; Zheng 2010); and not a randomised controlled trial (Shao 1993).

Studies awaiting classification

We categorised five studies as 'Studies awaiting classification’ (ACTRN12605000483651; Cohen 2011; Feng 2008; Li 2002; Su 2009). The first two studies have been completed but not yet been published (ACTRN12605000483651; Cohen 2011). The other three trials (Feng 2008; Li 2002; Su 2009) are awaiting classification pending attempts to obtain clarification and missing data. One study (Feng 2008) had average scores only no standard deviations. There were no available data for assessing the effectiveness of acupuncture compared with a control intervention and only P values for the between‐group statistical analysis were provided in the article by Li 2002. In Su 2009 it is not clear whether individuals with acute ankle sprains only were included or whether those with chronic ankle sprains were also included.

Risk of bias in included studies

See Figure 2 and Figure 3.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study

Allocation

Only Jian 2004 described an appropriate method of random sequence generation, which consisted of coin tosses. Participants were allocated to groups according to the sequence of hospital visits in five quasi‐randomised controlled trials (Jiang 2011;Paris 1983;Shi 2013; Wei 2004; Zhang 2011); we judged all five to be at high risk of selection bias relating to inadequate methods of sequence generation and lack of allocation concealment. Aside from claiming that the participants were 'randomly allocated', none of the other 14 studies gave details of their methods of sequence generation.

A suitable method for allocation concealment was reported in two studies only: Sun 2011 used opaque envelopes and Wu 2007 used sealed envelopes. However, neither trial described adequate methods to secure allocation concealment (i.e. sequentially numbered opaque envelopes), and thus we judged these and all trials not describing the methods used to be at unclear risk of selection bias relating to allocation concealment.

Blinding

We judged all trials as being at high risk of performance bias, as blinding of participants and personnel was either not done or was not possible in any of the included studies. Blinding of outcome assessment was not reported in any of these trials, none of which tested a placebo intervention. We judged that all trials were at high risk of detection bias.

Incomplete outcome data

Most trials reported having no dropouts. We judged these trials to be at low risk of attrition bias (Cho 1977; Ge 2000; Hao 2006; Jian 2004; Ni 2010; Paris 1983; Ruan 1995; Shi 2013; Sun 2011; Wu 2006; Wu 2007; Yu 1996; Yu 1999; Zhang 2011; Zhang 2012; Zhou 2008) and Chen 2012, which had similar numbers of dropouts in both groups, was also considered to be at low risk . There was unclear risk of attrition bias in three trials: Jiang 2011 did not report participant numbers available for subjective outcomes; Wang 2009 did not report participants lost to follow‐up; and Wei 2004 had a large loss to follow‐up for one outcome.

Selective reporting

We rated all the trials as being at high risk of bias for selective outcome reporting because no study protocols were available. We chose this stringent approach to reflect the general lack of transparency about methods in these studies.

Other potential sources of bias

We rated the 15 studies in which there was no evident major imbalance in baseline characteristics between groups as being at low risk of bias. The remaining five studies (Cho 1977; Ge 2000; Jian 2004; Ni 2010; Wang 2009) were assessed as being at unclear risk as they did not report baseline characteristics.

Effects of interventions

Of the primary outcomes that we originally expected to assess in this review, only cure rate was reported in most of the studies. Patient‐reported assessment of the function of the sprained ankle joint was not reported in any of the studies. Of the secondary outcomes, only pain and swelling were assessed in some studies; recurrence rate, ankle instability and patient‐rated quality of life were not reported in any study.

Acupuncture versus no acupuncture

Only one study compared acupuncture with no treatment (Wu 2007). In this study, acupuncture was significantly more effective than no treatment after five days with regard to the global improvement in symptoms (cure rate: 31/31 versus 1/30; RR 20.34, 95% CI 4.27 to 96.98; Analysis 1.1) and pain measured on a 10‐cm VAS (MD ‐6.92 cm, 95% CI ‐7.33 to ‐6.51 cm; Analysis 1.2).

1.1 — Comparison 1 Acupuncture versus no treatment, Outcome 1 Cure rate (5 days).

1.2 — Comparison 1 Acupuncture versus no treatment, Outcome 2 Pain ‐ VAS (5 days/0 to 10, worst outcome).

In eight studies, acupuncture plus another standard treatment was compared with that standard treatment alone.

In one study (Paris 1983), electrical acupuncture points stimulation with standard physical therapy, consisting of cryotherapy, pressure wrap application, partial weight‐bearing exercise and adhesive ankle support application, according to the stage of the disease, was compared with the same standard physical therapy alone. Electrical acupuncture points stimulation with physical therapy was more effective than physical therapy alone in terms of the time to stop treatment (the recovery of the injured ankle took three days less with acupuncture: MD ‐3.00 days, 95% CI ‐5.48 to ‐0.52 days; Analysis 2.1) and the time to recover plantar flexion‐dorsiflexion range of movement (MD ‐7.25 days, 95% CI ‐10.41 to ‐4.09 days; Analysis 2.2). Although the results still favoured the acupuncture group, the difference between the two groups in the recovery times for three other characteristics of an injured ankle joint were not statistically significant: time to recover inversion‐eversion range of movement (MD ‐3.00 days, 95% CI ‐7.80 to 1.80 days; Analysis 2.3); time to reduction of oedema (MD ‐0.12 days, 95% CI ‐5.51 to 5.27 days; Analysis 2.4); and time to recovery from pain (MD ‐2.75, 95% CI ‐6.18 to 0.68; Analysis 2.5).

2.1 — Comparison 2 Acupuncture plus standard conventional physical therapy versus standard physical therapy alone, Outcome 1 Time to stop treatment (recovery of injured ankle; timing not known).

2.2 — Comparison 2 Acupuncture plus standard conventional physical therapy versus standard physical therapy alone, Outcome 2 Time of recovery for plantar flexion‐dorsiflexion ROM of injured ankle (timing not known).

2.3 — Comparison 2 Acupuncture plus standard conventional physical therapy versus standard physical therapy alone, Outcome 3 Time of recovery for inversion‐eversion ROM of injured ankle (timing not known).

2.4 — Comparison 2 Acupuncture plus standard conventional physical therapy versus standard physical therapy alone, Outcome 4 Time of recovery for oedema of injured ankle (timing not known).

2.5 — Comparison 2 Acupuncture plus standard conventional physical therapy versus standard physical therapy alone, Outcome 5 Time of recovery for pain of injured ankle (timing not known).

In one study (Yu 1999), acupuncture plus the external application of a topical NSAID (Dolobene gel) was compared with topical NSAID alone. Yu 1999 found a better cure rate at seven days in the group allocated a combination of acupuncture and topical NSAID (47/50 versus 40/50; RR 1.18, 95% CI 1.01 to 1.37; Analysis 3.1).

3.1 — Comparison 3 Acupuncture plus external application of topical NSAIDs versus external application of topical NSAIDs alone, Outcome 1 Cure rate (7 days).

Acupuncture plus the oral Chinese herbal drugs was compared with oral Chinese herbal drugs alone in Ge 2000. Ge 2000 found the acupuncture group had a greater cure rate at 10 days compared with oral herbal drugs alone (48/50 versus 22/30; RR 1.31, 95% CI 1.05 to 1.64; Analysis 4.1).

4.1 — Comparison 4 Acupuncture plus oral administration of Chinese herbal drugs versus oral herbal drugs alone, Outcome 1 Cure rate (10 days).

Three trials compared acupuncture plus external application of Chinese herbal drugs versus external application of herbal drugs alone (Hao 2006; Yu 1996; Zhang 2011). Two trials (Hao 2006; Yu 1996) found higher cure rates in the acupuncture group, whereas all participants in both groups were cured in Zhang 2011 at follow‐up. Pooled results using the fixed‐effect model showed a significant difference in cure rate in favour of acupuncture (177/183 versus 141/163; RR 1.11, 95% CI 1.04 to 1.18; not shown). However, these results were significantly heterogeneous (I² = 97%), and the results were no longer statistically significant when a sensitivity analysis using a random‐effects model was performed (RR 1.17, 95% CI 0.78 to 1.78; Analysis 5.1).

5.1 — Comparison 5 Acupuncture plus external application of mixed Chinese herbs versus external application of mixed Chinese herbs alone, Outcome 1 Cure rate (7 to 10 days).

Acupuncture with electrophysiotherapy was compared with electrophysiotherapy alone in Zhang 2012. This study found a greater cure rate at 30 days in the acupuncture group, a result that was of borderline statistical significance (32/34 versus 26/34; RR 1.23, 95% CI 1.00 to 1.51; Analysis 6.1).

6.1 — Comparison 6 Acupuncture plus electrophysiotherapy versus electrophysiotherapy alone, Outcome 1 Cure rate (30 days).

In one study (Jian 2004), manual acupuncture plus a combination of several treatments (TDP far‐infrared lamp radiation, oral administration of an OTC Chinese herbal drug Shujing Huoxue Pian and external application of an OTC Chinese topical herbal tincture) was compared with the combined treatments alone. The manual acupuncture group had better cure rates at seven days (48/48 versus 42/48; RR 1.14, 95% CI 1.02 to 1.28; Analysis 7.1).

7.1 — Comparison 7 Acupuncture plus TDP radiation, Chinese herbal drug external application and oral administration of Chinese herbal drug versus TDP radiation, Chinese herbal drug external application and oral administration of Chinese herbal drug only, Outcome 1 Cure rate (7 days).

Acupuncture versus no acupuncture: all comparisons

The pooled results of all eight studies (n = 751) comparing acupuncture versus no acupuncture and measuring cure rate are presented in Analysis 8.1. Although the pooled cure rate favoured acupuncture, these results were not statistically significant (383/396 versus 272/355; RR 1.32, CI 0.95 to 1.84; P value = 0.1). There was also very substantial heterogeneity, as evident visually and statistically (I² = 98%), and differences between subgroups (test for subgroup differences: P value = 0.02). We explored this heterogeneity by removing Wu 2007, which as well as being a clear outlier, was the only trial in which acupuncture was used alone rather than in combination with another treatment. Removing Wu 2007 barely reduced the heterogeneity (I² = 95%), but revealed that Wu 2007 was the source of the subgroup differences (new test for subgroup differences: P value = 0.86). The pooled cure rate continued to favour acupuncture but to a lesser extent (RR 1.19, 95% CI 0.96 to 1.47; P value = 0.11; analysis not shown).

8.1 — Comparison 8 Acupuncture versus no acupuncture, Outcome 1 Cure rate.

Acupuncture versus another non‐surgical intervention

Acupuncture treatment was compared with other types of non‐surgical interventions in 14 studies (Chen 2012; Cho 1977;Jiang 2011;Ni 2010; Ruan 1995;Shi 2013; Sun 2011; Wang 2009; Wei 2004, Wu 2006;Wu 2007;Yu 1996;Yu 1999; Zhou 2008).

Acupuncture was compared with hot‐ or cold‐water application in one study (Cho 1977); this found the functional score relating to ankle joint mobility at one day to be significantly improved in the acupuncture group (RR 36.30, 95% CI 27.97 to 44.63; Analysis 9.1).

9.1 — Comparison 9 Acupuncture versus hot or cold water application, Outcome 1 Functional improvement score (1 day; 0 to 100 = best outcome).

Acupuncture was compared with ice‐pack application in one study (Yu 1996); this found no significant difference between the two interventions in cure rate at seven days (15/30 versus 16/29; RR 0.91, 95% CI 0.56 to 1.47; Analysis 10.1).

10.1 — Comparison 10 Acupuncture versus ice pack application, Outcome 1 Cure rate (7 days).

Acupuncture was compared to OTC Chinese herbal patch applications in three studies (Jiang 2011; Wei 2004; Wu 2006). While two trials (Jiang 2011; Wei 2004) found little evidence for a difference between the two groups, Wu 2006 found higher cure rates in the acupuncture group. Pooled results using the fixed‐effect model showed a significant difference in cure rate at 14 to 21 days follow‐up in favour of acupuncture (103/107 versus 80/100; RR 1.20, 95% CI 1.08 to 1.33; analysis not shown). However, these results were significantly heterogeneous (I² = 83%), and the results were no longer statistically significant when a sensitivity analysis using the random‐effects model was performed (RR 1.17, 95% CI 0.94 to 1.45; Analysis 11.1). Jiang 2011 found better pain, tenderness, swelling and function outcomes in the acupuncture group (Analysis 11.2). Wei 2004 found acupuncture resulted in faster relief of pain and swelling (MD ‐5.80 days, 95% CI ‐6.69 to ‐4.91 days; Analysis 11.3) and shorter length of treatment (MD ‐5.40 days, 95% CI ‐6.60 to ‐4.20 days; Analysis 11.4) than OTC herbal patch application. One trial (Yu 1996), which compared acupuncture with an external application of mixed Chinese herbs, found no significant difference in cure rate at seven days (15/30 versus 19/30; RR 0.79, 95% CI 0.50 to 1.24; Analysis 14.1).

11.1 — Comparison 11 Acupuncture versus Chinese OTC herbal patch application, Outcome 1 Cure rate (14 to 21 days).

11.2 — Comparison 11 Acupuncture versus Chinese OTC herbal patch application, Outcome 2 Pain, tenderness, swelling and function scores (15 days).

11.3 — Comparison 11 Acupuncture versus Chinese OTC herbal patch application, Outcome 3 Time to the relief of pain and swelling (days).

11.4 — Comparison 11 Acupuncture versus Chinese OTC herbal patch application, Outcome 4 Average treatment time (days).

14.1 — Comparison 14 Acupuncture versus external application of mixed Chinese herbs, Outcome 1 Cure rate (7 days).

One study (Sun 2011) comparing acupuncture versus immobilisation with an elastic bandage found no significant difference between the two groups in the cure rate at 14 days (40/41 versus 38/41; RR 1.05, 95% CI 0.95 to 1.16; Analysis 12.1). However, it took a significantly shorter time for ankle pain to subside in the acupuncture group (MD ‐3.40 days, 95% CI ‐3.88 to ‐2.92 days; Analysis 12.2).

12.1 — Comparison 12 Acupuncture versus immobilisation with elastic bandage, Outcome 1 Cure rate (14 days).

12.2 — Comparison 12 Acupuncture versus immobilisation with elastic bandage, Outcome 2 Time of recovery for pain of injured ankle (days).

Acupuncture was compared with the external application of topical NSAIDs in three studies (Chen 2012; Shi 2013; Yu 1999). Yu 1999 found no statistically significant differences between acupuncture and Dolobene gel with regard to cure rate at seven days (33/50 versus 40/50; RR 0.82, 95% CI 0.65 to 1.05; Analysis 13.1). Results from the other two trials (Chen 2012; Shi 2013), based on VAS scores ranging from 0 to 6 (worst outcome), showed that participants in the acupuncture groups had less pain (MD ‐1.05, 95% CI ‐1.46 to ‐0.65; Analysis 13.2), swelling (MD ‐1.08, 95% CI ‐1.44 to ‐0.73; Analysis 13.3), tenderness (MD ‐1.01, 95% CI ‐1.47 to ‐0.55; Analysis 13.4) and better function (MD ‐0.38, 95% CI ‐0.61 to ‐0.15; Analysis 13.5) at 14 days than those in the group receiving external Voltaren emulsion.

13.1 — Comparison 13 Acupuncture versus external application of topical NSAIDs, Outcome 1 Cure rate (7 days).

13.2 — Comparison 13 Acupuncture versus external application of topical NSAIDs, Outcome 2 Pain score (14 days; 0 to 6, worst outcome).

13.3 — Comparison 13 Acupuncture versus external application of topical NSAIDs, Outcome 3 Swelling score (14 days; 0 to 6, worst outcome).

13.4 — Comparison 13 Acupuncture versus external application of topical NSAIDs, Outcome 4 Tenderness score (14 days; 0 to 6, worst outcome).

13.5 — Comparison 13 Acupuncture versus external application of topical NSAIDs, Outcome 5 Function score (14 days; 0 to 6, worst outcome).

Acupuncture was compared with three other types of Chinese non‐surgical treatments (i.e. Tuina (Ruan 1995), blood‐letting therapy (Wu 2007) and TDP far‐infrared lamp radiation (Zhou 2008)). All participants in both groups were cured in Ruan 1995 (period of follow‐up not reported): acupuncture versus Tuina (112/112 versus 110/110; RR 1.00, 95% CI 0.98 to 1.02; Analysis 15.1); and Wu 2007: acupuncture versus blood letting (31/31 versus 31/31; RR 1.00, 95% CI 0.94 to 1.06; Analysis 16.1). Zhou 2008 found a higher cure rate in the acupuncture group than in the TDP far‐infrared lamp radiation group (21/26 versus 12/23; RR 1.55, 95% CI 1.00 to 2.39; Analysis 17.1).

15.1 — Comparison 15 Acupuncture versus Tuina, Outcome 1 Cure rate (timing not known).

16.1 — Comparison 16 Acupuncture versus blood letting, Outcome 1 Cure rate (5 days).

17.1 — Comparison 17 Acupuncture versus TDP far‐infrared lamp radiation, Outcome 1 Cure rate (5 days).

In two studies, acupuncture was compared with complex interventions (Ni 2010; Wang 2009). Ni 2010 found a highly significant difference in cure rate (period of follow‐up not reported) in favour of cold and hot external applications and oral administration of Chinese herbal medication (1/64 versus 55/59; RR 0.02, 95% CI 0.00 to 0.12; Analysis 18.1). Conversely, Wang 2009 found a statistically significantly higher cure rate (period of follow‐up not reported) with acupuncture than with external application of a Chinese herbal spray and partial immobilisation using an elastic bandage (48/48 versus 15/24; RR 1.60, 95% CI 1.17 to 2.17; Analysis 19.1).

18.1 — Comparison 18 Acupuncture versus cold and hot application with Chinese herbal oral medication, Outcome 1 Cure rate (timing not knowing).

19.1 — Comparison 19 Acupuncture versus external application of Chinese herbal spray and partial immobilisation with elastic bandage, Outcome 1 Cure rate (timing not known).

Acupuncture versus another non‐surgical intervention: all comparisons

The pooled results of all 11 studies (n = 1006) comparing acupuncture versus another non‐surgical intervention and measuring cure rate are presented in Analysis 20.1. Although the pooled cure rate slightly favoured acupuncture, these results were not statistically significant (404/509 versus 416/497; RR 1.07, 95% CI 0.94 to 1.22; P value = 0.30). There was also very substantial heterogeneity, as evident visually and statistically (I² = 92%), and differences between the subgroups (test for subgroup differences: P value < 0.0001).

20.1 — Comparison 20 Acupuncture versus another non‐surgical intervention, Outcome 1 Cure rate.

Adverse events

There was no description of the assessment of adverse events related to acupuncture in most of the included studies. Adverse events were reported in Wei 2004: there were no adverse events in the acupuncture group but three participants in the control group had skin problems, such as erosion and papules, at the attachment sites of OTC Chinese herbal patches.

Discussion

Summary of main results

We included 20 studies (including 2012 participants with acute ankle sprains) in this review, three of which included more than one comparison. Five studies were quasi‐randomised trials and the other 15 were randomised trials, although this attribute was based solely on a claim of random allocation by the study authors in 12 cases. Of our three prespecified primary outcomes, only cure rate was reported by the majority of studies. No study reported on the patient‐reported assessment of function and only one reported on adverse events.

One study, which compared acupuncture with no treatment, found acupuncture to be more effective with regard to cure rate and pain. Acupuncture plus another active treatment versus that active treatment alone was compared in eight studies, with cure rate data available for seven. Most studies reported higher cure rates in the acupuncture plus another active treatment group than in the active treatment alone group. However, while the results of a pooled analysis of cure rate data from eight trials comparing acupuncture versus no acupuncture tended to favour acupuncture, the results were not statistically significant and the data were very heterogeneous.

Fourteen studies compared acupuncture with a variety of other non‐surgical treatments, such as Chinese drug patches, hot and cold water, ice packs, oral Chinese herbal medicine and elastic bandages. Some studies found in favour of acupuncture, some in favour of the other treatment and some found a lack of evidence for a difference between the two interventions under test. The pooled results from 11 studies comparing acupuncture with another non‐surgical intervention tended to favour acupuncture, but were not statistically significant and were associated with very substantial heterogeneity.

Overall completeness and applicability of evidence

The included trials are very heterogeneous and while we have performed exploratory analyses of cure rates, these do not provide a basis for estimating the efficacy or effectiveness of acupuncture treatment for ankle sprains. There are several reasons for this. First, the acupuncture intervention showed considerable clinical heterogeneity among the included studies, as did the control interventions. Second, different settings and trial designs were adopted. Third, we cannot describe the review population with any confidence since most studies did not report the type or severity of ankle sprains. Although it is likely that individuals with different lesions and severities constituted the review population, the high cure rates reported in a week or less in many trials suggest injuries were less severe. Fourth, most of the included studies reported only cure rate as the main outcome and did not record or report outcomes including validated functional scores, adverse events, recurrence of ankle sprain or ankle instability. Fifth, apart from issues relating to high or unclear risks of bias in the included studies, the small sample sizes might have contributed to imprecision in this review. In all, this means that we are unable to provide conclusive evidence for the effectiveness or safety of acupuncture treatment for acute ankle sprains in adults.

Acupuncture is a common non‐drug intervention in Asian countries. Given that most of the included studies in this review were conducted in Asian countries, including China (17 studies), Singapore (one study) and South Korea (one study), where acupuncture is regularly practised with cultural familiarity for a broad range of health‐related conditions, it is questionable whether the findings from these trials can be generalised to other countries and or be applicable in different contexts. This aspect is discussed in Hawe 2004, which describes how the effects of non‐pharmacological treatments can be affected by the local context, including community‐based specialties.

Quality of the evidence

All of the included studies had a high or unclear risk of bias in almost all domains. In particular, none of the studies adopted sham controls to blind the participants and practitioners, which contributed to high risks of performance and detection bias. There was insufficient information for us to judge whether sequence generation and allocation concealment were conducted appropriately in most of the studies. These findings are compatible with those of empirical studies that reveal an inappropriate understanding and performance of randomisation in Chinese trials (Vickers 1998; Wu 2009). It is therefore uncertain that the randomisation processes used in the individual studies were appropriate. Study protocols were not available for any of the included studies; such protocols are needed to accurately evaluate selective outcome reporting.

Aside from limitations related to methodological aspects, other limitations in the quality of the evidence were the lack of data for primary outcomes (patient‐reported assessment of function and adverse events), variations in the definition of the key outcome (i.e. cure rate) for which data were available, the heterogeneity related to the settings and interventions, the small sample sizes of the individual studies and the potential for publication bias. Using the GRADE approach to assess the quality of evidence relating to cure rate, we downgraded the quality of evidence for this outcome the maximum of three levels (nominally, two levels for limitations in design and implementation that related to potential risk of bias; and one level for inconsistency in the results) to 'very low', meaning that we are very uncertain about the estimates of effect.

Potential biases in the review process

Our search was comprehensive but we may have missed trials published in the grey literature. Our intention to contact trial authors for more information, such as on the method of randomisation, was frustrated by the frequent lack of author information in the majority of studies that were published in China. While we planned to include only randomised or quasi‐randomised controlled trials in this review, this lack of author contact details meant that we were unable to confirm whether or not all the included studies were actually randomised trials.

Agreements and disagreements with other studies or reviews

A systematic review on the effect of acupuncture for ankle sprains was published in 2013 (Park 2013). In Park 2013, the authors adopted different criteria for the study selection, in which participants with chronic ankle sprains were also included, and limited the intervention to penetrating needle acupuncture only. In addition, the authors searched up to February 2012, whereas our search was up to May 2013. Despite the differences in the included trials resulting from these inclusion criteria and search dates, both reviews agree that the available evidence is flawed and insufficient to draw conclusions on the efficacy of acupuncture treatment for treating ankle sprains.

Authors' conclusions

Implications for practice.

The currently available evidence from a very heterogeneous group of randomised and quasi‐randomised controlled trials evaluating the effects of acupuncture for treating acute ankle sprains in adults does not provide reliable evidence on either the effectiveness or safety of acupuncture treatments, alone or in combination with other non‐surgical interventions, or in comparison with other non‐surgical interventions.

Implications for research.

To evaluate the effectiveness and safety of acupuncture for acute ankle sprains in adults, rigorous randomised controlled trials with sufficient sample sizes are necessary. In particular, studies using sham acupuncture as the control condition are required. In addition, validated specific outcome measures for ankle sprains are necessary. Adverse events relating to acupuncture should be measured in the trials. Clinical trials involving diverse ethnic groups would help ensure the wide relevance of the study results. Standard protocols for acupuncture treatment should be established to improve the reproducibility of results in clinical practice, as well as to assure safety. Finally, Consolidated Standards of Reporting Trials statements (CONSORT statements) should be used for designing and reporting the clinical trials, all of which should be prospectively registered in publicly available trial registries.

Acknowledgements

The authors appreciate the input of Mrs Lindsey Elstub, Dr Helen Handoll, Dr Mario Lenza, Ms Laura MacDonald, Dr Alasdair MacSween, Professor Cathie Sherrington and Mr Peter Struijs who provided encouraging suggestions and feedback about the protocol and review. We would also like to thank Dr Joanne Elliott of the Cochrane Bone, Joint and Muscle Trauma Group for help with developing the search strategies and running searches.

No external funding was received.

Appendices

Appendix 1. Search strategies

The Cochrane Library (Wiley Online Library)

#1 MeSH descriptor: [Ankle Injuries] this term only (454)  #2 MeSH descriptor: [Ankle] this term only (319)  #3 MeSH descriptor: [Ankle Joint] this term only (415)  #4 MeSH descriptor: [Lateral Ligament, Ankle] this term only (23)  #5 ankle*:ti,ab,kw (3333)  #6 ((deltoid or tibiotalar or tibiocalcaneal or tibionavicular or tibiofibular) near/3 ligament*):ti,ab,kw (2)  #7 #1 or #2 or #3 or #4 or #5 or #6 (3333)  #8 MeSH descriptor: [Acupuncture] this term only (133)  #9 MeSH descriptor: [Acupuncture Therapy] explode all trees (2688)  #10 (acupuncture or electro‐acupuncture or electroacupuncture or pharmaco‐acupuncture or pharmacoacupuncture or acupressure or moxibustion):ti,ab,kw (6423)  #11 #8 or #9 or #10 (6429)  #12 #7 and #11 (35)

MEDLINE (OvidSP)

1 Ankle Injuries/ (7576)  2 Ankle/ (6582)  3 Ankle Joint/ (10382)  4 Lateral Ligament, Ankle/ (329)  5 ankle*1.tw. (33828)  6 ((deltoid or tibiotalar or tibiocalcaneal or tibionavicular or tibiofibular) adj3 ligament*).tw. (447)  7 or/1‐6 (40076)  8 Acupuncture/ (1132)  9 exp Acupuncture Therapy/ (15436)  10 (acupuncture or electro‐acupuncture or electroacupuncture or pharmaco‐acupuncture or pharmacoacupuncture or acupressure or moxibustion).tw. (14401)  11 or/8‐10 (17948)  12 and/7,11 (79)  13 exp Animals/ not Humans/ (3847813)  14 12 not 13 (58)

EMBASE (OvidSP)

1 Ankle Sprain/ (1841)  2 Ankle Injury/ (5770)  3 Ankle/ (17707)  4 Ankle Lateral Ligament/ (727)  5 Ankle Pain/ (1249)  6 ankle*1.tw. (43855)  7 ((deltoid or tibiotalar or tibiocalcaneal or tibionavicular or tibiofibular) adj3 ligament$).tw. (454)  8 or/1‐7 (49954)  9 exp Acupuncture/ (28859)  10 Moxibustion/ (1251)  11 (acupuncture or electro‐acupuncture or electroacupuncture or pharmaco‐acupuncture or pharmacoacupuncture or acupressure or moxibustion).tw. (21233)  12 or/9‐11 (30558)  13 and/8,12 (134)  14 exp Randomized Controlled trial/ (342470)  15 exp Double Blind Procedure/ (114620)  16 exp Single Blind Procedure/ (17374)  17 exp Crossover Procedure/ (36896)  18 Controlled Study/ (4033874)  19 or/14‐18 (4115896)  20 ((clinical or controlled or comparative or placebo or prospective* or randomi#ed) adj3 (trial or study)).tw. (687990)  21 (random* adj7 (allocat* or allot* or assign* or basis* or divid* or order*)).tw. (167468)  22 ((singl* or doubl* or trebl* or tripl*) adj7 (blind* or mask*)).tw. (153575)  23 (cross?over* or (cross adj1 over*)).tw. (65676)  24 ((allocat* or allot* or assign* or divid*) adj3 (condition* or experiment* or intervention* or treatment* or therap* or control* or group*)).tw. (212658)  25 or/20‐24 (1025835)  26 or/19,25 (4628822)  27 limit 26 to human (2814272)  28 and/13,27 (47)

AMED (OvidSP)

1 Ankle/ (998)  2 Ankle Injuries/ (939)  3 Ankle Joint/ (1562)  4 ankle*1.tw. (6100)  5 ((deltoid or tibiotalar or tibiocalcaneal or tibionavicular or tibiofibular) adj3 ligament*).tw. (95)  6 or/1‐5 (6112)  7 Acupuncture/ (3194)  8 exp Acupuncture Therapy/ (6626)  9 (acupuncture or electro‐acupuncture or electroacupuncture or pharmaco‐acupuncture or pharmacoacupuncture or acupressure or moxibustion).tw. (9166)  10 or/7‐9 (9573)  11 and/6,10 (48)

CINAHL (Ebsco)

S1 (MH "Ankle") (3,028)  S2 (MH "Ankle Joint") (2,620)  S3 (MH "Ankle Injuries+") (3,333)  S4 (MH "Lateral Ligament, Ankle") (67)  S5 TI ankle* or AB ankle* (10,756)  S6 TI ( (deltoid N3 ligament*) or (tibiotalar N3 ligament*) or (tibiocalcaneal N3 ligament*) or (tibionavicular N3 ligament*) or (tibiofibular N3 ligament*) ) or AB ( (deltoid N3 ligament*) or (tibiotalar N3 ligament*) or (tibiocalcaneal N3 ligament*) or (tibionavicular N3 ligament*) or (tibiofibular N3 ligament*) ) (111)  S7 S1 or S2 or S3 or S4 or S5 or S6 (13,162)  S8 (MH "Acupuncture+") (8,967)  S9 (MH "Acupressure+") (829)  S10 (MH "Moxibustion") (250)  S11 TI ( acupuncture or electro‐acupuncture or electroacupuncture or pharmaco‐acupuncture or pharmacoacupuncture or acupressure or moxibustion ) and AB ( acupuncture or electro‐acupuncture or electroacupuncture or pharmaco‐acupuncture or pharmacoacupuncture or acupressure or moxibustion ) (2,108)  S12 S8 or S9 or S10 or S11 (9,686)  S13 S7 and S12 (45)

Appendix 2. The 'Risk of bias' assessment tool

Domain	Description	Judgement question	Note
Sequence generation	The process of sequence generation should be described sufficiently	Was the allocation sequence adequately generated?
Allocation concealment	Investigators and participants should not predict the assignment before the allocation process	Was allocation adequately concealed?
Blinding of participants and personnel	Participants should not know about the allocation result during the study period	Was knowledge of the allocated interventions adequately prevented from participants during the study?
Blinding of outcome assessment	Outcome assessors should not know about the allocation result during the study period	Was knowledge of the allocated interventions adequately prevented from outcome assessors during the study?
Incomplete outcome data	Any missing data with reasons or imputation of the data with a reasonable statistical method should be described properly	Were incomplete outcome data adequately addressed?
Selective outcome reporting	The outcomes of the trial should be described completely according to the protocol in a pre‐defined way	Are reports of the study free of suggestion of selective outcome reporting?	The study was regarded as high risk if there is no protocol available
Other sources of bias	The trial should be free of other sources of bias	Was the study apparently free of other problems that could put it at a high risk of bias?
Answers to the judgement question were: 'yes' for low risk of bias, 'no' for high risk of bias or 'unclear' for unclear or unknown risk of bias

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cure rate (5 days)	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2 Pain ‐ VAS (5 days/0 to 10, worst outcome)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Time to stop treatment (recovery of injured ankle; timing not known)	1	16	Mean Difference (IV, Fixed, 95% CI)	‐3.0 [‐5.48, ‐0.52]
2 Time of recovery for plantar flexion‐dorsiflexion ROM of injured ankle (timing not known)	1	16	Mean Difference (IV, Fixed, 95% CI)	‐7.25 [‐10.41, ‐4.09]
3 Time of recovery for inversion‐eversion ROM of injured ankle (timing not known)	1	16	Mean Difference (IV, Fixed, 95% CI)	‐3.0 [‐7.80, 1.80]
4 Time of recovery for oedema of injured ankle (timing not known)	1	16	Mean Difference (IV, Fixed, 95% CI)	‐0.12 [‐5.51, 5.27]
5 Time of recovery for pain of injured ankle (timing not known)	1	16	Mean Difference (IV, Fixed, 95% CI)	‐2.75 [‐6.18, 0.68]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cure rate (7 days)	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
1.1 Dolobene gel	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cure rate	8	751	Risk Ratio (M‐H, Random, 95% CI)	1.32 [0.95, 1.84]
1.1 Acupuncture versus no treatment (5 days)	1	61	Risk Ratio (M‐H, Random, 95% CI)	20.34 [4.27, 96.98]
1.2 Acupuncture + topical NSAIDs versus topical NSAIDs alone (7 days)	1	100	Risk Ratio (M‐H, Random, 95% CI)	1.18 [1.01, 1.37]
1.3 Acupuncture + oral Chinese herbal drugs versus oral Chinese herbal drugs alone (10 days)	1	80	Risk Ratio (M‐H, Random, 95% CI)	1.31 [1.05, 1.64]
1.4 Acupuncture + external mixed Chinese herbs versus external mixed Chinese herbs alone (7 to 10 days)	3	346	Risk Ratio (M‐H, Random, 95% CI)	1.17 [0.78, 1.78]
1.5 Acupuncture plus electrophysiotherapy versus electrophysiotherapy alone (30 days)	1	68	Risk Ratio (M‐H, Random, 95% CI)	1.23 [1.00, 1.51]
1.6 Acupuncture + TDP radiation, external and oral herbal drugs versus these 3 interventions only (7 days)	1	96	Risk Ratio (M‐H, Random, 95% CI)	1.14 [1.02, 1.28]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cure rate (14 to 21 days)	3		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected
2 Pain, tenderness, swelling and function scores (15 days)	1	240	Mean Difference (IV, Fixed, 95% CI)	‐0.59 [‐0.73, ‐0.45]
2.1 Pain score (1 to 3 = worst outcome)	1	60	Mean Difference (IV, Fixed, 95% CI)	‐0.4 [‐0.73, ‐0.07]
2.2 Tenderness score (1 to 3 = worst outcome)	1	60	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.65, 0.05]
2.3 Swelling score (1 to 3 = worst outcome)	1	60	Mean Difference (IV, Fixed, 95% CI)	‐0.4 [‐0.71, ‐0.09]
2.4 Function score (1 to 3 = worst outcome)	1	60	Mean Difference (IV, Fixed, 95% CI)	‐0.85 [‐1.06, ‐0.64]
3 Time to the relief of pain and swelling (days)	1	67	Mean Difference (IV, Fixed, 95% CI)	‐5.80 [‐6.69, ‐4.91]
4 Average treatment time (days)	1	45	Mean Difference (IV, Fixed, 95% CI)	‐5.4 [‐6.60, ‐4.20]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cure rate (14 days)	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2 Time of recovery for pain of injured ankle (days)	1	82	Mean Difference (IV, Fixed, 95% CI)	‐3.40 [‐3.88, ‐2.92]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Cure rate	11	1006	Risk Ratio (M‐H, Random, 95% CI)	1.07 [0.94, 1.22]
1.1 Acupuncture versus ice pack application (7 days)	1	44	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.45, 1.59]
1.2 Acupuncture versus Chinese OTC herbal patch application (14 to 21 days)	3	207	Risk Ratio (M‐H, Random, 95% CI)	1.17 [0.94, 1.45]
1.3 Acupuncture versus immobilisation with elastic bandage (14 days)	1	82	Risk Ratio (M‐H, Random, 95% CI)	1.05 [0.95, 1.16]
1.4 Acupuncture versus external application of topical NSAIDs (7 days)	1	100	Risk Ratio (M‐H, Random, 95% CI)	0.83 [0.65, 1.05]
1.5 Acupuncture versus external application of mixed Chinese herbs (7 days)	1	45	Risk Ratio (M‐H, Random, 95% CI)	0.84 [0.49, 1.45]
1.6 Acupuncture versus Tuina (Timing not known)	1	222	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.98, 1.02]
1.7 Acupuncture versus blood letting (5 days)	1	62	Risk Ratio (M‐H, Random, 95% CI)	1.0 [0.94, 1.06]
1.8 Acupuncture versus TDP far‐infrared lamp radiation (5 days)	1	49	Risk Ratio (M‐H, Random, 95% CI)	1.55 [1.00, 2.39]
1.9 Acupuncture versus cold and hot application with Chinese herbal oral medication (timing unknown)	1	123	Risk Ratio (M‐H, Random, 95% CI)	0.02 [0.00, 0.12]
1.10 Acupuncture versus external application Chinese herbal spray + elastic bandage (timing unknown)	1	72	Risk Ratio (M‐H, Random, 95% CI)	1.60 [1.17, 2.17]

Methods	Randomised controlled trial Comparisons: acupressure versus external application of Voltaren emulsion, an NSAID
Participants	Setting: China 108 Wushu athletes with acute ankle sprains aged 15 to 35 years old were allocated to either acupressure (male 32/female 20) or external application of Voltaren emulsion (male 32/female 19) Disease duration: within 2 days Type of ankle sprains and severity: not specified Diagnostic criteria: clinical guideline for clinical trials on Chinese herbal medicines (Zhang 2008a) and textbook on the osteopathy of traditional Chinese medicine (Fan 2008)
Interventions	(1) Acupressure group: Style ‐ semi fixed style Number of points used ‐ over 8 points Location of points ‐ unilateral (LR3, LR4, GB40, SP5, ST41, ST44, BL62 and KI6, LI4 and BL40, if necessary) Depth of stimulation ‐ not reported Stimulation sensation ‐ de‐qi sensation Stimulation method ‐ manual stimulation (pressing the points with pressure of 0.5 kg and frequency of 100 times per minute) Needle retention time ‐ not reported Needle type (e.g. manufacture, material) ‐ no needle used Total number of sessions ‐ 14 sessions Frequency/duration ‐ once a day Concomitant treatment ‐ no concomitant treatment (2) External application of Voltaren emulsion group: Voltaren emulsion was applied once a day for 2 weeks
Outcomes	Length of follow‐up: 2 weeks Pain (0 to 6: 0 = no pain; 6 = worst pain) Swelling (0 to 6: 0 = no swelling; 6 = worst swelling) Tenderness (0 to 6: 0 = no tenderness; 6 = worst tenderness) Ankle function (0 to 6: 0 = no symptoms; 6 = worst ankle function) Adverse events: not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported. "108 patients, who were members of Wushu team of Shanghai university of traditional medicine or of Shanghai academy of traditional Chinese medicine, were randomly allocated into treatment group and control group."
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   outcome	High risk	Open trial
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Similar numbers of participants (two in treatment group and three in control group) dropped out during the study
Selective reporting (reporting bias)	High risk	No study protocol was identified
Other bias	Low risk	No baseline imbalance

Methods	Quasi‐randomised trial Comparisons: acupuncture versus Chinese drug patch application
Participants	Setting: China 60 adults (48 men/12 women) with acute ankle sprains were allocated to an acupuncture group (total 30, aged 19 to 59 years) or to Chinese drug patch application (total 30, aged 18 to 55 years) Disease duration: 1 to 30 days Type of ankle sprain and severity: not specified Diagnostic criteria: Chinese government guideline on rational Chinese medicine 1994 (State Administration 1994)
Interventions	(1) Acupuncture group: Style ‐ semi‐standardised acupuncture Number of points used ‐ over 6 points Location of points ‐ unilateral BL62, GB40, ST41, KI6, BL60, KI3 and Ashi‐points Depth of insertion ‐ not reported Needle sensation ‐ de‐qi sensation Needle stimulation ‐ manual Needle retention time ‐ 30 minutes Needle type (e.g. manufacture, material) ‐ not reported Total number of sessions ‐ 1 treatment course (15 sessions) Frequency/duration ‐ 1 session a day for 15 days Concomitant treatment ‐ none (2) Chinese drug patch application group: OTC external patch (ShenYanghongyaotiegao, which is made up of several Chinese herbal drugs) was applied at GB40, ST41, KI6, BL60 and KI3 once every 1 to 2 days for 15 days
Outcomes	Length of follow‐up: 15 days Cure rate Pain (1 to 3: 1 = no pain; 3 = worst pain) Swelling (1 to 3: 1 = no swelling; 3 = worst swelling) Tenderness (1 to 3: 1 = no swelling; 3 = worst swelling) Function (1 to 3: 1 = no symptom; 3 = worst ankle function) Adverse events: not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quasi‐randomised trial. Participants were allocated according to the visiting sequence. "According to the visiting sequence, patients were randomly allocated into two groups."
Allocation concealment (selection bias)	High risk	Quasi‐randomised trial
Blinding of participants and personnel (performance bias)   outcome	High risk	Open trial
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Cure rate ‐ low risk of bias because no dropouts Pain, swelling, tenderness and function ‐ unclear risk of bias because no report on the numbers of participants available for these outcomes
Selective reporting (reporting bias)	High risk	No study protocol was available
Other bias	Low risk	No baseline imbalance between groups

Methods	Randomised controlled trial Comparison: acupuncture versus cold and hot applications of Chinese herbal oral medication
Participants	Setting: China 123 participants (53 men/70 women, aged 13 to 63 years) with acute ankle sprains were allocated to an acupuncture group (total 64) or to receive cold and hot applications of Chinese herbal oral medication (total 59) Disease duration: 1 day to 5 days Type of ankle sprain and severity: not specified Diagnostic criteria: Chinese government guideline of traditional Chinese medicine 1994 (State Administration 1994)
Interventions	(1) Acupuncture group: Style ‐ fixed points Number of points used ‐ 2 points Location of points ‐ Xiaojiexue (non‐classic acupuncture point) in the non‐affected side and tender points at the wrist joint in the affected side Depth of insertion ‐ not reported Needle sensation ‐ de‐qi sensation Needle stimulation ‐ not reported Needle retention time ‐ 40 minutes Needle type (e.g. manufacture, material) ‐ 0.3 mm x 2 Cun (diameter x length) acupuncture needle Total number of sessions ‐ 3 sessions Frequency/duration ‐ everyday for 3 days Concomitant treatment: none (2) Cold and hot application with Chinese herbal oral medication: cold pack of Chinese herbal oral medication within 24 hours after ankle injury and hot pack after 24 hours were applied to the ankle. In addition, DuyiWei Jiaonang (独一味胶囊) 0.9 mg three times daily administered for 3 days
Outcomes	Length of follow‐up: not known Cure rate Adverse events: not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported. "Randomly allocated into two groups."
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   outcome	High risk	Open trial
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	High risk	No protocol was available
Other bias	Unclear risk	Baseline characteristics of the participants were not reported

Methods	Quasi‐randomised trial Comparisons: electrical acupuncture points stimulation plus standard physical therapy versus standard physical therapy alone
Participants	Setting: Canada  16 college students with acute ankle sprains were allocated into acupuncture group (total 8: 4 men/4 women) versus standard physical treatment (total 8: 7 men/1 women) Disease duration: less than 48 hours Type of ankle sprain and severity: grade 2 Diagnostic criteria: through subjective evaluation and stress roentgenograms (X‐rays) to ascertain severity of the injury
Interventions	(1) Electrical acupuncture point stimulation: Style ‐ fixed points Number of points used ‐ 24 points Location of points ‐ ear acupuncture points (Shen men, Ankle, Vertebral Innervation, Thalamus, Tragus and Endocrine) and body acupuncture points (SP6, BL59, ST41, BL60, KI3 and SP5) Depth of insertion ‐ not inserted Needle sensation ‐ pain within subjective tolerance Needle stimulation ‐ electrical stimulation (Neuroprobe systems 2 NP 200 by Medical Research Labs, Inc.) Needle retention time ‐ about 1 minute Needle type (e.g. manufacture, material) ‐ not reported Total number of sessions ‐ not reported Frequency/duration ‐ everyday for 5 days for the first week and after that three days a week for 5 days Concomitant treatment: Cryotherapy and elevation of the extremity for the first 48 hours after trauma When swelling was profuse, a pressure wrap above the malleoli was applied After 48 hours participants were encouraged to bear partial weight with crutches When oedema receded sufficiently, the pressure wrap was replaced by an adhesive ankle support applied like Gibney closed basket weave support (2) Standard physical treatment: Cryotherapy and elevation of the extremity for the first 48 hours after trauma When swelling was profuse, a pressure wrap above the malleoli was applied After 48 hours participants were encouraged to bear partial weight with crutches When oedema receded sufficiently, the pressure wrap was replaced by an adhesive ankle support applied like Gibney closed basket weave support
Outcomes	Length of follow‐up: timing not known Time to stop treatment (day) Time to recovery of plantar flexion‐dorsiflexion ROM of injured ankle (day) Time to recovery of inversion‐eversion ROM of injured ankle (day) Time to recovery of oedema of injured ankle (day) Time to recovery of pain of injured ankle (day) Adverse events: not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quasi‐randomised trial. Participants were allocated alternatively according to the visiting sequence. "When each patient was seen for the first time, he was alternatively assigned to one of two groups by one of the authors (DLP): the standard physical therapy group, designated group 1 (4 men and 4 women) and the standard physical therapy plus Neuroprobe group, designated Group 2 (7 men and 1 woman)."
Allocation concealment (selection bias)	High risk	Quasi‐randomised trial
Blinding of participants and personnel (performance bias)   outcome	High risk	Open trial
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	High risk	No protocol was available
Other bias	Low risk	No baseline imbalance between groups

Methods	Quasi‐randomised controlled trial Comparison: acupuncture versus external application of Voltaren emulsion, an NSAID
Participants	Setting: China 87 adults with acute ankle sprains (28 men, 59 women; aged 19 to 46 years) were allocated to an acupuncture group (total 43) and an external application of Voltaren emulsion group (total 44) Disease duration: 2 to 24 hours Type of ankle sprain and severity: not reported Diagnostic criteria: clinical guideline for clinical trials on Chinese herbal medicines (Zhang 2008a) and textbook on the osteopathy of traditional Chinese medicine (Fan 2008)
Interventions	(1) Acupuncture group: Style ‐ individual acupuncture Number of points used ‐ 2 points Location of points ‐ Ashi points around the affected ankle joint Depth of insertion ‐ not reported Needle sensation ‐ not reported Needle stimulation ‐ manual acupuncture (rapid rotation of acupuncture needle ‐ 80 to 120 times/minute) Needle retention time ‐ 20 minutes Needle type (e.g. manufacture, material) ‐ 0.3 mm x 25 mm (diameter x length) Total number of sessions ‐ about 7 times Frequency/duration ‐ every other day for 2 weeks Concomitant treatment ‐ none (2) external application of Voltaren emulsion group: Voltaren emulsion was applied three times a day for 2 weeks
Outcomes	Length of follow‐up: 2 weeks Pain (0 to 6: 0 = no pain; 6 = worst pain) Swelling (0 to 6: 0 = no swelling; 6 = worst swelling) Ankle function (0 to 6: 0 = no symptom; 6 = worst ankle function) Adverse events: not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	"According to the order of hospital visit, 87 patients were randomly allocated into two groups, 43 patients in acupuncture group and 44 patients in Voltaren emulsion group."
Allocation concealment (selection bias)	High risk	Quasi‐randomised trial
Blinding of participants and personnel (performance bias)   outcome	High risk	Open trial
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	High risk	No protocol was available
Other bias	Low risk	No baseline imbalance

Study	Reason for exclusion
Ahn 2003	Two different types of acupuncture treatments were tested
Byeon 2003	Two different types of acupuncture treatments were tested
Ceng 2006	Non‐conventional treatment (combination with TDP radiation, blood‐letting acupuncture and cupping therapy) was used in the control group
Chen 2012a	Non‐conventional treatment (herbal fumigation) was used in the control group
Cheng 2012	Acupuncture was not used as an intervention
Duan 2007	Different co‐interventions were used between groups
Guo 2010	Acupuncture was used in both groups
He 2006	Acupuncture was used in both groups
He 2006a	Two different types of acupuncture treatments were tested
Hong 2011	Acupuncture plus intermediate frequency therapy were compared with single acupuncture treatment
Huang 2012	Acupuncture was not used as an intervention
Jeong 2009	Two different types of acupuncture treatments were tested
Ji 2004	Acupuncture was used in both groups
Jung 2008	Two different types of acupuncture treatments were tested
Kang 2008	Two different types of acupuncture treatments were tested
Kim 2007	Two difference types of acupuncture treatments were tested
Lee 2004	Two different types of acupuncture treatments were tested
Li 2001	Acupuncture was used in both groups
Liao 2010	Two different types of acupuncture treatments were tested
Liu 2012	Different cointerventions were used in the groups
Luo 2002	Acupuncture was used in both groups
Park 2004	Two different types of acupuncture treatments were tested
Park 2013	This was a systematic review
Ru 2008	Different cointerventions were used in the groups
Seo 2006	Two different types of acupuncture treatments were tested
Shang 2010	Different cointerventions were used in the groups
Shao 1993	This was not a randomised controlled trial
Song 2008	Two different types of acupuncture treatments were tested
Sun 2011a	Acupuncture was used in both groups
Tan 2005	Different co‐interventions were used in the groups
Wang 2005	Participants with acute or chronic ankle sprains were included
Wang 2006	Both intervention group and control group received acupuncture treatment
Wang 2008	Two different types of acupuncture treatments were tested
Wang 2011	Acupuncture was not used as an intervention
Wimmer 2012	This was a review
Woo 2003	Acupuncture was used in both groups
Wu 1995	Different co‐interventions were used in the groups
Xu 2007	Different co‐interventions were used in the groups
Xu 2012	Acupuncture was used in both groups
Yang 2011	Different co‐interventions were used in the groups
Yu 2007	Both groups had acupuncture treatments
Zhang 2008	Different co‐interventions were used in the groups
Zhang 2011a	This study was an observational study
Zhang 2012a	Acupuncture was used in both groups
Zhao 2009	This was a systematic review
Zheng 2008	Acupuncture was used in both groups
Zheng 2010	Different co‐interventions were used in the groups

PERMALINK

Acupuncture for treating acute ankle sprains in adults

Tae‐Hun Kim

Myeong Soo Lee

Kun Hyung Kim

Jung Won Kang

Tae‐Young Choi

Edzard Ernst

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Timing of outcome measurement

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Assessing the quality of the evidence and summarising the results

Results

Description of studies

Results of the search

1.

Included studies

Participants and diagnostic criteria

Interventions

Outcomes

Excluded studies

Studies awaiting classification

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Acupuncture versus no acupuncture

1.1. Analysis.

1.2. Analysis.

2.1. Analysis.

2.2. Analysis.

2.3. Analysis.

2.4. Analysis.

2.5. Analysis.

3.1. Analysis.

4.1. Analysis.

5.1. Analysis.

6.1. Analysis.

7.1. Analysis.