Objectives This review assesses prehospital selective immobilisation protocols across a range of outcomes, including neurological deterioration and characteristics of injured, non-immobilised patients.

Methods Six electronic reference databases and eight grey literature sources were systematically searched. We included studies that enrolled acute trauma patients in the prehospital setting who were assessed for spine injury according to predefined clinical criteria and either immobilised or not. Data items included instances of neurological deterioration among patients with spine injuries, as well as available characteristics of those who were injured and not immobilised. Available data and study heterogeneity prevented meta-analyses. Bias was assessed for both individual studies and across studies by outcome.

Results 604 unique articles were retrieved, of which 7 met inclusion criteria. There was moderate or high risk of bias across studies in all outcomes. Of 76 patients with spine injuries who were not immobilised, 72 had no neurological deficit that appeared after emergency medical services contact, and the remaining four were not followed. Within this group, there appears to be a trend towards elderly patients who suffered a thoracic or lumbar injury from a low-risk mechanism of injury. Among studies that report both the results of the protocol assessment and immobilisation status, there is variable correspondence between the two.

Conclusions Data limitations and study biases suggest caution when interpreting and applying the results of this review. Its findings are consistent with the conclusions of individual studies. The characteristics of injured, non-immobilised patients point to areas of future research to investigate apparent trends.