Antipsychotic Use for ICU Delirium and Associated Inflammatory Markers

Abstract

Purpose: We sought to evaluate critically ill patients with delirium to evaluate inflammatory cytokine production and delirium progression and the role of antipsychotics. Materials and Methods: Adult critically ill patients with confirmed delirium according to a positive CAM-ICU score were included and IL-6 and IL-8 levels were trended for 24 h in this single-center, prospective, observational cohort study. Results: A total of 23 patients were consented and had blood samples drawn for inclusion. There was no difference in IL-6 and IL-8 levels at baseline, 4 to 8 h, and 22 to 28 h after enrollment when comparing patients based on antipsychotic exposure. We identified 2 patient clusters based on age, APACHE III, need for mechanical ventilation, and concomitant infection. In cluster 1, 5 (33.3%) patients received antipsychotics versus 5 (62.5%) patients in cluster 2 (P = .18). Patients in cluster 1 had more co-inflammatory conditions (P < .0001), yet numerically lower baseline IL-6 (P = .18) and IL-8 levels (P = .80) compared to cluster 2. Patients in cluster 1 had a greater median number of delirium-free days compared to cluster 2 (17.0 vs 6.0 days; P = .05). Conclusions: In critically ill patients with delirium, IL-6 and IL-8 levels were variable and antipsychotics were not associated with improvements in delirium or inflammatory markers.

Keywords: antipsychotic; critically ill; delirium; inflammatory marker; interleukin.

Figure 1.

Inflammatory markers by patient. (A) IL-6 levels for antipsychotic and no antipsychotic exposure groups at baseline, 4 to 8 h, and 22 to 28 h; (B) IL-8 levels for antipsychotic and no antipsychotic exposure groups at baseline, 4 to 8 h, and 22 to 28 h.

Figure 2.

The principal component score plot. Through principal component analysis, we transformed high-dimensional data into lower-dimension while retaining as much information as possible. The score plot indicates the projection of the data onto the span of the principal components. The x-axis is the first principal component, while the y-axis denotes the second principal component. The graph shows that the patients are separated into 2 clusters by the 2 principal component scores.