Ankle instability

Last evidence check May 2014

Principal authors Jim Barrie, Rebecca Hope and Jo Lishman

Most ankle ligament injuries are stable and clearly should be managed conservatively. The nature of this conservative management varies from simple advice and symptomatic management to prolonged rehabilitation programmes with or without bracing. Many patients with chronic injuries consider themselves to have “had physiotherapy”, but this often amounts to some exercises and a bit of ultrasound.

Rehabilitation and bracing

Early bracing (Leanderson and Wredmark 1995) does not alter the final outcome but allows quicker recovery. A number of studies (Goldie et al (1994), Wester et al (1996), Holme et al (1999)) suggest that active rehabilitation of ankle ligament injuries produces some improvement of results. The treatment regimes aim at improving flexibility, peroneal strength and proprioception. However, the differences are quite small, and the treatment in the control limbs of the studies were very different: control groups in some studies received more treatment than treatment groups in other studies. Recruitment methods may have introduced some bias. Bleakley (2010) reported a RCT comparing exercises within the first week after a mild-moderate ankle sprain (defined as negative clinical stress tests) against PRICE for the first week. All patients had a three-week ankle rehab programme for weeks 2-4. In weeks 1 and 2, the lower extremity functional score was significantly higher in the early exercise group and this group were more active. There were no differences in pain and at 3 months the Karlsson score was the same in both groups.

Lamb (2009) reported a RCT comparing below-knee cast, Aircast brace, Bledsoe boot and tubular bandage in 584 patients with severe ankle sprains, as defined by the inability to weightbear for at least 3 days after injury (not the usual definition of severe sprain). At 3 months the mean Foot and Ankle Outcome score was better in the patients treated with cast or Aircast brace, with a slightly larger treatment effect in cast. A total of 54 patients refused to have a cast as compared with 3 for the brace, which fits with clinical experience; the cost of the brace was higher.

Rehabilitation versus surgery

There have been a number of randomised controlled trials of surgical versus non-surgical treatment of acute ligament tears, most of which were meta-analysed by Pijnenberg (2000) – although this paper has implicit flaws. Almost all trials treated their non-surgical patients in cast. Pijnenberg et al (2003) reported a trial in which patients were randomised to surgical repair with early mobilisation and rehabilitation or mobilisation and rehabilitation without repair. The repaired patients reported less late instability and pain, but apparently had not required further treatment for this. Pijnenberg considered the clinical difference too small to warrant routine acute repair.

Kitaoka et al (1997) reported a retrospective study of 53 injured ankles in 48 patients followed up for 12-33 years. 22 ankles in 21 patients underwent primary repair. 31 ankles in 27 patients underwent either an Evans or a Watson-Jones procedure for late instability. Treatment was at the discretion of the admitting physician and, of course, ankles which were successfully treated without surgery were not included. There was no difference in objective stability on stress radiography and structurally unstable ankle at final follow-up were often asymptomatic. Outcome using an ankle score and patient satisfaction showed no difference between the groups. As this was not a RCT there are many possible confounding factors. However, this study shows similar outcomes whether ankle ligament injuries are all treated by repair, or reconstruction is performed only when non-surgical treatment fails.


Our practice for acute injuries is: