To evaluate the sensitivity of ancillary testing in patients with death by neurological criteria (DNC).

Ancillary tests are used to support the diagnosis of DNC when required by legislation, or when the clinical examination is inconclusive or confounded. Despite widespread use, their diagnostic accuracy is currently unclear.

We performed a systematic review by searching MEDLINE, EMBASE and Cochrane databases from their inception to April 21, 2020. We selected published abstracts and manuscripts reporting cohort and case-control studies, as well as case series, including patients with DNC or deeply comatose patients who underwent ancillary testing (neuroimaging, neurophysiology) for DNC. As per PRISMA methodology, two individuals independently extracted data and assessed study risk of bias using the QUADAS-2 tool. We estimated sensitivities from studies that included only DNC patients using a Bayesian hierarchical model with studies nested within ancillary test types.

Overall, 134 publications met our eligibility criteria. Ninety-three studies included DNC patients only, of which 92 (99%) used clinical examination as the reference standard. In total, 9,435 ancillary tests were applied. No study had low risk of bias on all QUADAS-2 criteria. Pooled ancillary test sensitivities ranged from 0.81 (CT-angiography, 7-point scale) to 0.92 (4-vessel angiography). Other reported sensitivities included electroencephalography (0.86), radionuclide imaging (0.86-0.88), evoked potentials (0.87-0.88), transcranial Doppler (0.88), and CT-perfusion (0.89). The standard deviations between sensitivities assessed within each test type were larger (sigma=0.11-0.16) than the standard deviation between pooled ancillary test sensitivities (sigma=0.04), suggesting heterogeneity within each ancillary test type was considerably higher than it was between tests.