Sigurd Liavaag and coauthors are to be commended for their manuscript, “Immobilization in External Rotation After Primary Shoulder Dislocation Did Not Reduce the Risk of Recurrence. A Randomized Controlled Trial.” This multicenter trial in Norway demonstrated no apparent clinical benefit of three weeks of immobilization in external rotation following primary anterior shoulder dislocation in 188 subjects. The authors of this clinical trial focused on a clinically relevant age group (range, sixteen to forty years), executed an effective stratified block randomization, monitored treatment compliance, and reported excellent follow-up (98% at two years).
The findings of Liavaag et al. mirror those of Finestone et al.1, who also reported no apparent benefit of immobilization in external rotation. The patients in the study by Finestone et al. were all male, were younger (mean age, twenty years), were predominantly military personnel, and were managed with immobilization in a greater degree of external rotation (15° to 20°) and for a longer period of time (four weeks), with excellent reported treatment compliance.
Most notable in the published literature is the discrepancy between the reported absence of a treatment effect in these studies as compared with the dramatic effect reported by Itoi et al.2. In a similar clinical trial, Itoi et al. reported a significantly diminished rate of recurrence following immobilization in external rotation as compared with internal rotation (26% compared with 42%).
What explanations might account for the discrepant findings between these studies? A possible explanation is that the studies inherently included different patient populations. The population studied by Itoi et al. was older than the others, with an average age of thirty-seven years and with one patient who was ninety years old. Despite this age difference, Itoi et al. also reported a significant treatment effect favoring immobilization in external rotation in a subgroup analysis of subjects younger than thirty years of age. In addition to age differences between the study populations, it is also possible that the discrepant recurrence rates reflect different cultural or recreational habits that might contribute to differing risks of recurrent instability.
Compliance also might be implicated as a factor in the discrepant findings. Both Liavaag et al. and Itoi et al. reported relatively poor compliance rates. It is difficult to know how to interpret the studies’ findings in light of reported poor compliance. However, both of these studies included intent-to-treat, per-protocol, and sensitivity analyses to support their conclusions.
Last, although recurrent dislocation is a well-defined primary outcome measure, it may not be sensitive enough to detect symptomatic instability that is short of full dislocation.
Perhaps three weeks is not long enough to expect meaningful healing at the bone-soft tissue interface. Perhaps 15° of external rotation at the shoulder is not enough external rotation. The magnitude of external rotation and the ideal duration of immobilization need to be more clearly defined. Furthermore, although immobilization in external rotation might permit coaptation of the glenoid-labrum complex, it does not address, and might even be deleterious to, the component of capsular stretching that is associated with anterior dislocations of the shoulder. The discrepant findings in this emerging body of evidence certainly warrant further investigation.