Use of BRASS in sedated critically-ill patients as a predictable mortality factor: BRASS-ICU

Objectives: To demonstrate that a BRASS score >= 3 at admission of intubated, ventilated and sedated patients is predictive of mortality Methods: we have realized an Observational prospective multicenter study. All Major patients without neurological history, admitted to ICU for a non-neurological cause, sedated and admitted under mechanical ventilation were included. Results: One hundred and ten patients were included, the BRASS score as well as the FOUR and RASS scores were collected. At day 28, patients with a BRASS score >= 3 had an excess mortality (OR 3.29 - CI 95% [1.42-7.63], p = 0.005) as well as day 90 (OR 2.65 - CI 95% [1.19-5.88], p = 0.02), without impact on the delirium measured by CAM-ICU (OR 1.8 - CI 95% [0.68-4.77], p = 0.023). After adjustment with SAPS II, FOUR and RASS, difference in mortality was not any more different. It is also noted that patients with BRASS >= 3 are more sedated (RASS: -5 [-5 - -5] vs -4 [-5 - -3], p < 0.0001) and more comatose (FOUR: 2 [1-4] vs 6 [4-9], p < 0.0001), and have higher doses of midazolam (10 mg/h [5-15] vs 7.5 mg/h [5-10], p = 0.02) and sufentanil (20 mu g/h [15-22.5] vs 10 [10-12.5], p = 0.01). Conclusions: The early alteration of brainstem reflexes measured by the BRASS score was not independently predictable in terms of mortality in the non-neurological ICU patients, admitted under sedation and mechanical ventilation.