Computed tomography angiography in the diagnosis of brain death: a systematic review and meta-analysis.

Abstract

Physiological instability and confounding factors may interfere with the clinical diagnosis of brain death. Computed tomography angiography (CTA) has been suggested as a potential ancillary test for confirmation of brain death, but its diagnostic accuracy remains unclear. We searched MEDLINE, EMBASE, and CENTRAL for studies comparing CTA with other accepted methods of diagnosing brain death (clinical or radiographic). Summary estimates of diagnostic accuracy were computed using random effects models. Subgroup analyses and meta-regression were performed to assess associations between CTA sensitivity and study or patient characteristics. Twelve studies, involving 541 patients, were included. If the CTA criterion for brain death was complete lack of opacification of intracranial vessels, then the pooled sensitivity was 62% (50-74%) for venous phase and 84% (75-94%) for arterial phase imaging. The sensitivity of CTA was higher when the criterion for brain death involved absence of opacification of internal cerebral veins, either alone (99%, 97-100%) or in combination with lack of flow to the distal middle cerebral artery branches (85%, 77-93%). CTA sensitivity was not influenced by different reference standards (clinical vs. radiographic) or predominant diagnostic category (stroke vs. brain trauma). Specificity of CTA could not be adequately determined from the existing data. Many patients who progress to brain death by accepted clinical or radiographic criteria have persistent opacification of proximal intracranial vessels when CTA is performed. The specificity of CTA in the diagnosis of brain death has not been adequately assessed. Routine use of CTA as an ancillary test in the diagnosis of brain death is therefore not recommended until diagnostic criteria have undergone further refinement and prospective validation. Absence of opacification of the internal cerebral veins appears to be the most promising angiographic criterion.