Delirium is associated with increased risk of death, prolonged stay, higher cost of care, and likely long-term brain deficits in survivors. This form of brain dysfunction occurs in intensive care unit (ICU) patients in epidemic proportions, and the scope of this problem is likely to worsen in upcoming years due to the aging of our population and increased utilization of the ICU. Currently, delirium goes unrecognized and untreated in the vast majority of circumstances in the ICU unless the patient presents with hyperactive delirium and agitation. In the latter circumstance, a commonly used typical antipsychotic called haloperidol is considered the principal agent for treating delirium based largely on anecdotal evidence to support its usefulness, though no placebo controlled trials exist. There are no FDA approved medications for delirium. The atypical antipsychotics provide a promising alternative for the treatment of delirium due to their enhanced beneficial effects on positive (agitated) and negative (quiet) symptoms proven in mania and schizophrenia, reduced risk for side effects common to haloperidol such as extrapyramidal symptomatology, and less potentially lethal heart rhythm disturbances. It is imperative that well-designed phase II studies to determine proof of principle be conducted. A pilot study of feasibility to begin assessing the role of antipsychotics in the management of ICU delirium.

This investigation will be the first placebo controlled trial of delirium prevention/treatment, in or out of the ICU. As mentioned above, clinical practice guidelines for medical management of pain, anxiety, and delirium (major determinants of patient comfort) are endorsed by the major critical care societies. These guidelines will form the template for this investigation. Pain management is prioritized as a clinicians' first concern. The assessment and treatment algorithm in the guidelines then places anxiety and delirium, respectively, as sequential tiers of priority. While delirium monitoring is now available, recent data indicate that less than 5% of practicing ICU healthcare professionals use a specific delirium monitoring instrument. Thus, as outlined here, most delirium is not recognized or treated, which serves as the rationale for this placebo-controlled investigation. Anxiety is currently treated with drugs such as benzodiazepines. Such anxiety, however, may be due to delirium, in which case treatment with anxiolytics such as benzodiazepines might exacerbate this form of brain dysfunction. On the other hand, it is possible that treatment with antipsychotics will reduce the duration and severity of delirium, result in less breakthrough sedatives (due to the sedating effects of the antipsychotics), and improve clinical outcomes. Alternatively, treatment with antipsychotics may not alter or worsen clinical outcomes.

The specific aims of this study are as follows:

Aim 1: To determine whether antipsychotics reduce the incidence and duration of delirium in high risk mechanically ventilated patients.

Aim 2: To determine whether antipsychotics reduce the severity of neuropsychological dysfunction at hospital discharge in high risk mechanically ventilated patients.

Hypothesis 1: Our primary hypothesis is that in mechanically ventilated patients, the duration of delirium and the days alive and free of delirium - as measured using the Confusion Assessment Method for the ICU (CAM-ICU)- will be significantly improved by early treatment with antipsychotics (haloperidol or ziprasidone) as compared to placebo. Furthermore, we hypothesize that delirium duration will be comparable between the two intervention groups (haloperidol and ziprasidone). To test the primary hypothesis, we propose to perform a randomized, double-blind, placebo-controlled trial of the prevention/treatment of delirium in ICU patients using oral liquid formulations of haloperidol versus ziprasidone versus placebo. This study is powered to show a 50% improvement in the duration of delirium (CAM-ICU positive days) and will enroll 102 patients (34 in each group) over a two-year period. In addition, we will compare between groups the overall incidence of delirium and the number of delirium free days (DFDs) - defined as days alive and free of coma and delirium to day 21.

Hypothesis 2: We hypothesize that scores on a neuropsychological testing battery administered at the time of hospital discharge will be better in patients treated with antipsychotics (either haloperidol or ziprasidone) than those treated with placebo. Furthermore, we hypothesize that neuropsychological test scores will be comparable between the two intervention groups (haloperidol and ziprasidone).

Inclusion Criteria:

• Medical or surgical ICU patients on mechanical ventilation who are receiving sedatives or analgesics or displaying an abnormal level of consciousness (delirium or coma).

Exclusion Criteria:

• Subjects expected to have a short time on mechanical ventilation. That is, those in whom the likelihood for the need for mechanical ventilation is less than 24 hours.
• Subjects who have been on mechanical ventilation for more than 72 hours.
• Subjects in whom gastric access is not available (i.e., no enteral feeding tube or NG/OG tube)and is not anticipated to be available for 48 hours.
• Subjects younger than 18 years old.
• Subjects who are pregnant (a pregnancy test will be performed on all women of child bearing age) or breastfeeding.
• Inability to obtain informed consent from the subject or the subject's authorized representative.
• Documented history of allergic reaction to ziprasidone or haloperidol.
• Subjects admitted to the ICU for drug/alcohol overdose, suicide attempts, alcohol withdrawal/delirium tremens.
• Subjects with active seizures or cerebrovascular accident within the last 2 weeks.
• Subjects who are benzodiazepine dependent at the time of index hospitalization (i.e., patients on benzodiazepines as outpatient and whose attending judges it unsafe to withhold these medications due to risk for withdrawal syndrome).
• Subjects with chronic pain syndromes or who are on maintenance narcotics.
• Subjects with a history of torsades de pointes, known history of QT prolongation (e.g., congenital long QT syndrome), a QTc at baseline of 500 ms or over in the absence of bundle branch block, documented myocardial infarction within the previous 2 weeks, or uncompensated NYHA IV heart failure (dyspnea or anginal syndrome present at rest due to CHF). [NOTE: ICU patients who have an incidental rise in troponin in the absence of definitive ischemic ECG changes remain eligible]
• Subjects who are on neuroleptic therapy as an outpatient maintenance drug (e.g., haloperidol, mesoridazine, thorazine, chlorpromazine, trifluoperazine, droperidol, risperidone, quetiapine, olanzapine, or ziprasidone).
• Subjects who are receiving and will continue to receive other drugs that prolong the QT interval such as sotalol, quinidine, other Class Ia or III anti-arrhythmics, dofetilide (Tikosyn for arrhythmias), pimozide (for Tourette's), gatifloxacin, moxifloxacin (levofloxacin permissible), pentamidine, tacrolimus (Prograf), dolasetron (Anzemet). Azithromycin is an acceptable medication for study patients, and anyone slated to receive (or receiving) either clarithromycin or erythromycin can be switched to azithromycin by their primary team and be enrolled into the study the following day. Patients receiving clindamycin or clotrimazole will be excluded from the study.
• Subjects who have a history of neuroleptic malignant syndrome.
• Subjects with potassium levels below 3.0 mg/dl or magnesium levels below 1.8 mg/dl. NOTE: If the patient is receiving replacement of K+ or Mg+, then he/she would be eligible unless there is reason to suspect that these electrolyte abnormalities will be refractory.
• Subjects with moderate/severe dementia (e.g., Alzheimer's type, vascular origin, or HIV-related) as documented by medical history or modified Blessed dementia rating scale (mBDRS) 4 or more or Informant Questionnaire of Cognitive Dysfunction in the Elderly (IQCODE) over 3.6.
• Subjects who have suspected anoxic brain injury or documented cerebral edema at the time of screening.
• Subjects who are moribund and not expected to survive 24 hours from the time of study enrollment, or who have a "Do Not Resuscitate" order, or whose family or medical team have not committed to aggressive support (e.g., not going to use vasopressors or mechanical ventilation or likely to have withdrawal of support within 24 hours).