Seventeen case control studies were included in the review (n=1,109; 538 endovascular and 571 open surgery). Overall, the quality of the included studies was judged to be moderate: quality scores ranged from 9 to 20.
Perioperative mortality: there were 30 deaths (5.57 per cent) in the stenting group and 94 deaths (16.5 per cent) in the open arm group. Endovascular stenting was associated with a reduction in perioperative mortality (OR 0.36, 95% CI: 0.228, 0.578).
Major neurological injury (13 studies, n=958): major complications occurred in 26 (5.4 per cent) of stent patients and 67 (14 per cent) of open patients. A significant reduction in major neurological injury was found in the endovascular stenting group compared with the open arm group (OR 0.39, 95% CI: 0.25, 0.62); no evidence of statistical heterogeneity was found. When the subgroups (perioperative stroke and paraplegia) were analysed, stenting was associated with a statistically significant reduction in risk of paraplegia (OR 0.33, 95% CI: 0.18, 0.63) but not perioperative stroke.
Major reintervention (nine studies, n=769): major re-interventions were required in 29 (7 per cent) of stent patients and 30 (8.4 per cent) of open patients. No statistically significant between-group difference was found; there was no evidence of statistical heterogeneity.
Length of hospital and ITU stay (seven studies, n=468): a reduction in length of ITU stay (ES -1.26 days, 95% CI: -1.77, -0.76) and hospital stay (ES -1.00 days, 95% CI: -1.58, -0.45) was found in the stenting arm. Evidence of significant heterogeneity was found for both outcomes.
Emergency versus elective surgery (three studies, n=187): a reduction in perioperative mortality was found in the endovascular stenting group compared with the open arm group (OR 0.23, 955 CI: 0.09, 0.59). No statistically significant between-group differences were found for major neurological injury or major reintervention.
Subgroups by pathology
Stenting versus open surgery (thoracic aortic aneurysms) (three studies, n=358): mortality (OR 0.25, 95% CI: 0.09, 0.66) and major neurological injury (OR 0.28, 95% CI: 0.13, 0.61) were significantly lower in the endovascular group. No statistically significant between-group differences were found for major reinterventions. Unstable patients were excluded from analysis.
Stenting versus open surgery (descending thoracic trauma) (seven studies, n=274): endovascular repair was associated with a borderline statistically significant reduction in neurological injury compared with open repair (OR 0.17, 95% CI: 0.03, 1.03). No statistically significant difference was found for mortality or major re-interventions.
Stenting versus open surgery (thoracic aortic rupture) (two studies, n=83): no statistically significant difference was found for mortality, major neurological injury or major re-interventions when endovascular repair was compared with open repair in patients with acute thoracic aortic rupture.
Stenting versus open surgery (thoracic aortic dissection): one study was found, results not reported.
No evidence of statistical heterogeneity was found and there was no evidence of publication bias.