A multimodal approach to prognostication of outcome after cardiac arrest (CA) is recommended. Evidence for combinations of methods is low. In this post-hoc analysis we described findings on head computed tomography (CT) after CA. We also examined whether generalised oedema on CT alone or together with the biomarker Neuron-specific enolase (NSE) could predict poor outcome.

Patients participating in the Target Temperature Management after out-of-hospital-cardiac-arrest-trial underwent CT based on clinical indications. Findings were divided into pre-specified categories according to local radiologists descriptions. Generalised oedema alone and in combination with peak NSE at either 48?h or 72?h was correlated with poor outcome at 6 months follow-up using the Cerebral Performance Category (CPC 3–5).

356/939 (37.9%) of patients underwent head CT. Initial CT???24?h after CA was normal in 174/218 (79.8%), whilst generalised oedema was diagnosed in 21/218 (9.6%). Between days 1–7, generalised oedema was seen in 65/143 (45.5%), acute/subacute infarction in 27/143 (18.9%) and bleeding in 9/143 (6.3%). Overall, generalised oedema predicted poor outcome with 33.6% sensitivity (95%CI:28.1–39.5) and 98.4% specificity (95%CI:94.3–99.6), whilst peak NSE demonstrated sensitivities of 61.5–64.8% and specificity 95.7% (95%CI:89.5–98.4). The combination of peak NSE?>?38?ng/l and generalised oedema on CT predicted poor outcome with 46.0% sensitivity (95%CI:36.5–55.8) with no false positives. NSE was significantly higher in patients with generalised oedema.

In this study, generalised oedema was more common >24h???7d after CA. The combination of CT and NSE improved sensitivity and specificity compared to CT alone, with no false positives in this limited population.