Methods

SEARCH STRATEGY

The search terms “bandages, ankle sprain and compression” were used to search the Cochrane Database, Medline (1966 onwards for all databases), CINAHL and EMBASE databases. All relevant abstracts were reviewed and papers relevant or possibly relevant were retrieved and reviewed in full. In addition, the bibliographies of all retrieved papers were searched for relevant papers and the manufacturers of Tubigrip were contacted directly.

STUDY DESIGN

Patients presenting with ankle injuries to the A&E department of the North Hampshire Hospital and the emergency department of Southampton General Hospital and given a final diagnosis of grade 1 or 2 lateral ankle sprain were eligible for entry into the study. Grade 3 sprains were rare (incidence not recorded separately) and were excluded from the trial. Table 1 shows the exclusion criteria used for grade 1 and 2 sprains.

After informed written consent was obtained a trial entry sheet was completed and the patient was randomised (see below) to receive either a DTG or no DTG. (Analgesia and rehabilitation advice were standardised between the two groups by means of an advice sheet which described exercises and advised simple analgesia if necessary). Patients were then telephoned by a member of the A&E reception staff one week after their attendance and a set of standardised questions was asked.

OUTCOME MEASURES

After reviewing the literature^{10, 11, 14} the following were chosen as the primary outcome measures:

1. Pain (sleep disturbance, need for analgesia)

2. Mobility (days until walking unaided, that is, no crutches, stick or holding onto furniture)

3. Need for and number of days off work

SAMPLE SIZE

This was based on the assumption (from previous studies^{2, 3, 20}) that grades 1 and 2 lateral ankle sprains take approximately 10 days to recover to a level where the patient can return to work. At a 5% significance level with 80% power, predicting a 10% difference in this outcome measure (and time to return to “normal” everyday activities (excluding sport)) led to an estimated sample size of 200 patients in each group (treatment with and without a DTG).

The results were analysed using the independent t test, χ² or Fisher's exact test and the Mann-Whitney U test (for number of days off work) as appropriate. The SPSS¹⁹ and Arcus²⁰ statistical packages were used for the statistical analyses.

RANDOMISATION

Randomisation was performed by a statistician from the Medical Statistics department of Southampton General Hospital using the SAS programme. Block randomisation was used. Four hundred sealed opaque brown envelopes with the trial number from 1 to 400 on the front were used, with a single piece of white paper stating either “DTG or “No DTG” placed inside. The randomisation code was broken after recruitment into the trial had ceased.

PATIENT ALLOCATION

Patients were enrolled consecutively by A&E medical staff and emergency nurse practitioners (ENPs). The trial entry sheet was numbered from 1 to 400, with the correspondingly numbered brown envelope attached. Once the trial entry sheet was completed by the enrolling doctor/ENP, it was left for the treatment nurse to open the envelope and administer treatment as appropriate.

Results

It was noted that groups 1 and 9 of the exclusion criteria above were the most common reasons for exclusion and these accounted for approximately 20% of patients presenting with ankle injuries.

Figure 1 shows a Consort diagram²¹ to summarise participant flow through the trial.

There was no significant difference in terms of age, sex and occupation between patients treated with and without DTG who completed the trial (table 2). There was also no significant difference with respect to these variables between patients who completed and those who did not complete the trial (age: t=0.75 (95% CI −2.12 to 4.69 years; sex: p=0.67, Fisher's exact test, 95% CI −12% to 8%; occupation: p=0.96, Fisher's exact test, 95% CI −1% to 20%).

Table 2 shows the demographic details for patients who completed the trial.

Tables 3 and 4 show the results for the main outcome measures.

There was no significant difference between the treatment groups for whether pain from the injury kept the patient awake at night (table 3) but there was a significant difference in the numbers who took painkillers (table 3). The difference seemed to be that patients treated with DTG were more likely to take analgesia. There was no significant difference in the need for or number of days off work between the two treatment groups, or in the number of days until walking unaided.

Discussion

The management of sprained ankles with DTG has become a time honoured treatment for which good evidence from well conducted randomised controlled trials is lacking. Many studies fail to grade the severity of injuries they are studying, use inappropriate outcome measures or do not explain whether the assessor is blinded.²² Linde et al¹¹ compared elastic compression bandage with no treatment in 100 patients and found no significant difference in the course of pain, function, swelling or limitation of movement between the two groups. However, this study has a major flaw as the patients were not randomised. Brooks et al¹⁰ randomised 102 patients to one of four treatment groups: no treatment, early physiotherapy, double Tubigrip and below knee plaster of Paris cast. Patients given no treatment returned to work earlier and had a lower “clinical score” than the other groups, but this study suffers from small numbers and no statistical analysis. The patients in the non-intervention groups may also have had less severe injuries and 15 of 27 patients in the no treatment group said they would have preferred some sort of support for the ankle. This latter point may be important and is a frequently quoted reason for giving DTG to patients with ankle sprains. It would be very useful in any future study to include patients who have expressed a treatment preference, in order to compare their outcomes with patients who were randomised to that treatment.

Our aim was to conduct a well designed randomised controlled trial to answer the question of whether treatment of grade 1 and 2 lateral ankle sprains with DTG leads to a more rapid return to functional recovery (as evidenced by the numbers of days taken to return to normal activities, excluding sport) than no treatment. We wanted to look at the typical A&E population (as opposed to athletes or military recruits) using “average” modalities for assessment and treatment (that is, not using stress radiographs or arthrograms). We chose grade 1 or 2 sprains because these are the commonest grade of injury encountered and because some clinicians prefer to immobilise grade 3 sprains in a below knee plaster of Paris for a short period of time. All junior doctors and ENPs were given teaching sessions on how to grade these injuries and were assessed to ensure consistency.

Our study had a follow up of 197 of 400 patients—that is, just under 50%. For this reason we felt it was important to show that the groups who completed follow up were similar at the start of the trial (table 2) and similarly for those who did not complete follow up. The response rate was the same in the two treatment arms and is therefore unlikely to lead to an internally biased result, although we recognise that our power calculation required 400 subjects and our results therefore risk the introduction of a type II error. Although our follow up rate is comparable with other studies we feel that it could have been improved upon significantly by the use of a completely independent research assistant to perform the follow up telephone calls. A&E reception staff (despite their best efforts) are notoriously busy and a significant number of patients were lost to follow up either because of a failure to contact or incomplete data collection once contacted. Clearly this will have had an impact on how generalisable our study results are.

Our study was set up to detect a 10% difference in outcome between the two treatment groups, as such a difference is one that most clinicians would accept as being significant. An equivalence study may actually be the ideal trial design to answer this clinical question, however the sample size required would be prohibitively large. We felt that as we did not find evidence of a significant difference in outcome between the two groups (except that the DTG group required significantly more analgesia) by using a sample size calculated to detect a 10% difference (see above for limitations) then this study design is acceptable.

Our results show that treatment of grade 1 and 2 ankle sprains without a DTG compression bandage does not result in a longer time to functional recovery or increased need for or numbers of days off work. It is interesting that the patients in our study required an average of only 3.37 versus 3.21 (DTG versus no DTG) days off work, which contrasts with the longer times quoted in other studies.^{2, 3, 10} This should not reflect milder injuries, as the same grading system was used, but perhaps may be explained by a difference in the patient populations studied. The issues surrounding desire for a rapid return to work are complex and beyond the scope of this study.

The main conclusion from this study is that patients treated with DTG required significantly more analgesia for their sprains. This is perhaps surprising considering the claim that DTG provides analgesia by providing counter-irritation to the skin.⁶ The increased need for analgesia may have been because the application of a DTG made these patients more aware of their injury or perhaps it reflects a real effect that such bandages may increase the discomfort if inadequate reapplication by the patient causes a tourniquet effect, for example. The finding that the use of DTG may actually be harmful is important and warrants further investigation.

As highlighted above, our conclusions must be viewed with some significant statistical limitations in mind. Our power calculation required 400 subjects but only half completed the trial and we used 10 days as the endpoint for return to work but the average number of days off work was only a third of this in our study. This means that our conclusions cannot be firm and perhaps this study should be considered as a pilot study, giving an indication of what might be expected if a larger study were to be undertaken. However, as pointed out above, we feel that an equivalence study may actually be the way to answer this research question, in which case the number of patients required is very likely to be prohibitively large.

Conclusion

This study suggests that the use of DTG compression bandage in grade 1 and 2 anskle sprains does not shorten recovery time or number of days off work. Application of DTG seems to be associated with an increase in the need for analgesia. This may be because the patient is more aware of their injury or it may be a real effect of the DTG, possibly because of inadequate reapplication by the patient.