Use of imaging studies for determination of brain death in South Australian intensive care units

Objectives: To describe the use of imaging studies (four-vessel angiography or radionuclide scan) for brain death determination in South Australian intensive care units, and to determine the rates of adherence with The ANZICS statement on death and organ donation of the Australian and New Zealand Intensive Care Society (ANZICS). Design, patients and setting: Retrospective case-note review of 190 South Australian adult patients (≥ 18 years) who were brain dead and were organ donors (actual and intended), from 1 January 2008 to 31 December 2014. Main outcome measures: We compared brain death determination by clinical examination and by imaging, and identified, using logistic regression, the independent predictors of brain death determination by imaging (and for imaging without a documented indication). Results: Brain death determination by imaging occurred for 79 patients who were brain-dead donors (41.6%), with a documented indication in only 38 patients (48.1%), of whom 35 had an indication which adhered to ANZICS recommendations. The group who had brain death determined by imaging were younger (P < 0.001), with a higher proportion of hypoxic brain injury (P = 0.01) and therapeutic hypothermia (P = 0.02). Independent predictors of brain death determination by imaging were female sex (β = 3.101, P = 0.03), age (β = 0.964, P = 0.01), brain death determination between 5 pm and 8 am (β = 0.332, P = 0.04), cause of death (β = 1.833, P = 0.04), therapeutic hypothermia (β = 0.162, P = 0.04) and terminal serum sodium level = 150 mmol/L (β = 0.131, P = 0.005); Nagelkerke R² = 0.669. Hypoxia was the only independent predictor of imaging without a documented ANZICS indication (β = 0.071, P = 0.032; Nagelkerke R² = 0.581). Conclusions: Therapeutic hypothermia, terminal serum sodium level ≥ 150 mmol/L and cause of death were independent predictors of brain death determination by imaging study. Documentation of imaging indication was poor, particularly after hypoxic brain injury. This may reflect emerging indications for imaging, poor adherence to ANZICS recommendations, or simple omissions.