Routine CT scanning of patients retrieved to a tertiary centre on veno-venous extracorporeal membrane oxygenation: a retrospective risk benefit analysis.

Introduction: Comprehensive clinical examination can be compromised in patients on veno-venous extracorporeal membrane oxygenation (VV-ECMO). Adjunctive diagnostic imaging strategies range from bedside imaging only to routine computed tomography (CT). The risk-benefit of either approach remains to be evaluated. Patients retrieved to the Royal Brompton Hospital (RBH) on VV-ECMO routinely undergo admission CT imaging of head, chest, abdomen and pelvis. This study aimed to identify how frequently changes in therapy or adverse events could be attributed to routine CT scanning.

Methods: Demographic and clinical data were gathered retrospectively from patients retrieved to RBH on VV-ECMO (January 2014-2016). Scans were categorized as 'routine' or requested to clarify a specific clinical uncertainty. Clinical records were reviewed to identify attributable management changes and CT- related adverse events. Seventy-two patients were retrieved on VV-ECMO (median age 44 years) and 65 scanned on admission (mean radiation dose 2344mGy-cm). Routine head CT head yielded novel clinical information in 11 patients, 10 of whom had unexpected intracranial haemorrhage and, subsequently, had their anticoagulation withheld. Routine thoracic CT identified unexpected positive findings in three patients (early fibrosis, pulmonary vasculitis, pneumomediastinum), eliciting management variation in one (steroid administration). Routine abdomen/pelvis CT identified new information in three patients (adrenal haemorrhage, hepatosteatosis, splenic infarction), changing the management in one (withholding anticoagulation).

Results: CT scanning was not associated with consequential adverse events (e.g. accidental decannulation, gas entrainment into the circuit, hypoxia, hypotension). Median transfer/scan time was 78 minutes, requiring five ITU staff-members. In our cohort, a policy of routine head CT changed the management in 17% of patients; the yield from routine chest, abdomen and pelvis CT was modest. CT transfer was safe, but resource intensive.

Conclusion: Prospective studies should evaluate whether routine CT impacts outcome.