Early onset delirium incidence and risk factors in hematology oncology patients admitted to the intensive care unit: A retrospective cohort study.

Abstract

Background: Delirium occurs frequently in intensive care unit (ICU) patients; however, there are limited data evaluating its impact on critically ill hematology-oncology patients. We aimed to determine the incidence and risk factors for early-onset delirium development in hematology-oncology patients admitted to the ICU.

Methods: This single-center, retrospective cohort study evaluated the primary outcome of incident delirium within 7 days of ICU admission in adults admitted to the hematology-oncology medical or surgical ICU. Patients with delirium (DEL) were compared to those without (No-DEL) for evaluation of secondary endpoints including hospital mortality, ICU, and hospital length of stay (LOS). Multivariable logistic regression modeling was performed to identify independent risk factors for delirium.

Results: Delirium occurred in 125 (51.2%) of 244 patients. Inhospital mortality was significantly higher in the DEL vs. No-DEL group (32.8% vs. 15.1%, P = 0.002). Median (1^st and 3^rd quartiles) ICU and hospital LOS were significantly longer in the delirium group, respectively (6 [4-10] days vs. 3 [2-5] days, P < 0.001, and 21 [14-36] days vs. 12 [8-22] days, P < 0.001). Higher Sequential Organ Failure Assessment score, high-dose corticosteroids, mechanical ventilation (MV), and brain metastases were each independently, associated with an increased delirium risk.

Conclusion: Hematology-oncology patients admitted to the ICU frequently develop delirium. Consistent with literature in nonhematology-oncology critically ill patients, identified independent risk factors for delirium were MV and organ dysfunction. Risk factors unique to the critically ill hematology-oncology patient population include high-dose corticosteroids and brain metastases. Further research is needed to evaluate strategies to mitigate delirium development in this population based on risk assessment.