Pediatric delirium is a common though underrecognized and understudied phenomenon in pediatric critical care. The unique combination of the vulnerability and resiliency of children's developing brains makes it imperative that research on prevention, evaluation, treatment, and follow-up of this complication of pediatric medical illness and treatment be vigorously pursued. Although antipsychotics are rarely used for treatment, a small but growing body of research supports the efficacy and short-term safety of using both conventional and atypical antipsychotics. Further understanding of the complex neurophysiology of delirium may lead to more refined treatments. Developmentally valid screening tools will make the assessment and treatment of pediatric delirium more feasible when expert consultation is not available.

The DSM-IV diagnostic criteria for delirium are 1) an acute onset and often fluctuating course of 2) disturbances in arousal and 3) cognition, in the presence of 4) an underlying general medical condition triggering the syndrome (1). Children have been found to have a pattern of symptoms similar to that seen in adults (2).

Clinical subtypes of delirium, recognized in both adult and pediatric patients, are characterized by disturbances in arousal: hyperactive, hypoactive, and mixed delirium (3—5). Hypoactive delirium is underrecognized and has low referral rates because of its less disruptive clinical picture (5—7).

Associated symptoms of delirium include impairments in arousal, attention (thought to be a core symptom), cognition, thought (delusions), language, memory, orientation, perception (hallucinations), sleep-wake cycle, behavior, mood, and affect. Pediatric and adult delirium have similar high rates of impaired alertness, orientation, anxiety, and hallucinations (2). However, pediatric and particularly infant delirium may have subtle developmentally specific symptoms and signs not yet accounted for in available screening measures.

Adult delirium is associated with more self-extubations, longer requirement of mechanical ventilation, inadvertent removal of catheters, prolonged hospital stay, higher health care costs, and increased mortality (6). Adult survivors of delirium have been shown to be at risk for long-term cognitive impairment and posttraumatic stress disorder (PTSD) (6). Children have reported delusional memories of intensive care unit (ICU) experiences, which can become a basis for PTSD later.

Children, like elderly patients, have an elevated vulnerability to delirium (7). The "delirium cascade" can originate in any number of pathways associated with systemic illness or iatrogenic intervention: oxygen deprivation, infiamma-tory cytokine release, modification of blood-brain barrier permeability, sick euthyroid syndrome, up-regulation of the hypothalamic-pituitary-adrenal axis, and disruption of intracellular second messenger systems (6, 8). Dysregulation of multiple neurotransmitter systems, particularly of acetylcholine and dopamine, probably plays a role in symptom generation. From research in traumatic brain injury, epilepsy, and meningitis, it is clear that the immature brain has different responses to oxygen deprivation and cytokine release compared to the adult brain (9). Because delirium may represent "end organ damage" of the brain in critical illness, long-term neurocognitive outcome studies in child survivors of untreated delirium are needed.

Clinical Evaluation

Psychiatric evaluation of the patient in the pediatric intensive care unit (PICU) includes review of the medical history and hospital course, with special attention to the current medication history, pain, itching, and other disturbing medical symptoms; corroborative history, including from multidisciplinary staff and caregivers; and examination of the child.

Screening Instruments

There is a need for screening instruments that could help identify patients with delirium and the subset who need treatment. Some experts have recommended that screening be done during each shift in the PICU, because of the fluctuating course of delirium and the potential for prevention and reversal when treated (10).

The Delirium Rating Scale, the most comprehensive scale used to date, has been shown to have cutoff scores for a pediatric cohort comparable to those for adults (11). However, this scale includes questions that are not applicable to younger children, including self-report of symptoms such as hallucinations and delusions. No validation studies for the Delirium Rating Scale have been conducted with pediatric samples.

The Pediatric Confusion Assessment Method for the ICU (6) is the first validated screening instrument for critically ill children that can be used by staff for the rapid assessment of ventilated and nonventilated patients over age 5. Through the use of yes/no questions, pictures, and hand signals, this instrument allows the evaluation of a child's "inattention" and "disorganized thinking." Sensitivity and specificity are currently being evaluated in U.S. and European pilot studies.

The Pediatric Anesthesia Emergence Delirium Scale was developed in studies of emergence delirium, which occurs when children are awakening from anesthesia (6, 12). This instrument includes items about the child's eye contact with caregivers, awareness of surroundings, purposeful nature of movements, ability to be consoled, and restlessness. It is validated in children from 19 months to 6 years of age in postoperative settings and is being studied for use in PICU settings. As part of a pediatric delirium diagnostic algorithm, which includes caregiver and multidisciplinary team consensus, the Pediatric Anesthesia Emergence Delirium Scale's focus on neurobehavioral symptoms may make it a useful tool in this younger population and may guide us in developing tools for assessing delirium in the very youngest patients, in whom neurocognitive symptoms are difficult or impossible to assess (10).

Sedation in Pediatric Critical Care

Sedative and analgesic agents are integral to the care of critically ill children. Although nonpharmacologic methods, such as intervention by child life specialists and parental presence at the bedside, are routine in the PICU, medications are universally required for tolerance of mechanical ventilation and painful procedures. Fear and anxiety in this stressful environment are also addressed by medications. The medications routinely used in the PICU are presented in Table 1 (13, 14).

Treatment of Pediatric Delirium

When considering treatment for delirium, it is imperative that one consider the risks and benefits of intervention while evaluating 1) the underlying illnesses and treatments, 2) offending agents, such as medications, 3) environmental factors, and 4) indication for antipsychotics. Here we focus on the last three factors.

Potential offending agents include benzodiazepines, opioids, and steroids, as well as other sedatives, including propofol and ketamine. These agents have potentially neurotoxic effects. Recent biochemical research has shown that the toxicity is mediated through complex pathways that may have pro-apoptotic mechanisms as well as through the enhancement of γ-aminobutyric acid (15). These toxic effects have been found to be dependent on dose and duration of exposure, and there is evidence suggesting that they are minimized during exposure to noxious neurostimulation, such as surgery. PICU patients are often exposed to these agents for days, weeks, or months, raising concern about potential long-term neurocognitive harm despite the lack of long-term outcome studies in children (16). Timely dosage tapering minimizes withdrawal symptoms in patients exposed to benzodiazepines, barbiturates, and opioids. Research exploring the association between Akt phosphorylation and anti-apoptotic activity suggests that antipsychotics may be neuroprotective. Coupled with their enhancement of cellular glucose uptake and metabolism, the atypical antipsychotics may make cells more resistant to cellular stress and restore levels of nerve growth factor and brain-derived neurotrophic factor (17). Although this research is preliminary, the potential for neuroprotection and substantial clinical benefits further supports the consideration of antipsychotics.

Environmental stressors do not cause delirium but are an exacerbating factor. They can be minimized by having an active parental (or other familiar caregiver) present at bedside; having familiar, orienting objects in view, such as favorite toys and pictures of home and pets; limiting noxious noise and light; limiting room and staff changes; maintaining approximately normal diurnal schedules; and using one-on-one nursing observation as necessary (5, 18).

Pharmacologic Strategies

Pharmacotherapy may be used to resolve delirium; to reduce agitation, allowing removal of potentially offending agents and improving respiratory management; to limit the time the patient is delirious, decreasing any associated trauma, distress, and potential toxicity; and to reduce the length of hospital stay. The medications recommended are conventional and atypical antipsychotics, which are listed with dosage recommendations in Table 2 (3, 5, 6, 19, 20). There is a nascent pediatric delirium literature indicating safe and rapid resolution of delirium with these agents, even in very young children (2, 3, 5, 6, 21, 22). As in adults, a model of high dopamine for hyperactive delirium and low dopamine/cholinergic imbalance for hypoactive and mixed delirium may suggest the choice of antipsychotic (3). In the case of hyperactive delirium, haloperidol may be more effective because of its narrower targeting of the dopamine D₂ receptor. Because of their wider receptor effects, atypical antipsychotics may be more useful in mixed or hypoactive delirium and may lessen the cognitive side effects seen with conventional antipsychotics.

All antipsychotics carry the risk of QT prolongation, which can lead to torsade de pointes and ultimately sudden death (18, 23, 24). Children with medical illness may have electrolyte abnormalities, infection, hypoxia, or exposure to other cardiotoxic drugs, all of which can increase the likelihood of cardiac arrhythmias with antipsychotic treatment. These factors must be assessed and addressed before antipsychotics are introduced. While most children in PICU settings are already receiving continuous cardiac monitoring, the additional monitoring of QTc, electrolyte levels, and drug interactions during the introduction and course of antipsychotic treatment is prudent. Cessation of antipsychotic treatment must be considered if there is significant QTc prolongation, a QTc >500 msec, new T wave abnormalities, marked bradycardia, or a Brugada pheno-type. Risks may increase with the addition of metabolic inhibitors or other drugs that prolong QTc. Intravenous haloperidol is associated with particular risk for cardiac arrhythmia and should be avoided in high-risk patients whenever possible (18). There are no established clinical guidelines for the use of antipsychotics for delirium in children with preexisting cardiac disease. Each case must be evaluated individually, perhaps with the guidance of pediatric cardiologists, to weigh the risk of arrhythmia against the risks of untreated delirium, such as respiratory compromise and brain anoxia, as in the case presented here.

If cardiac risk is low and the use of antipsychotics is desired, multiple factors may bear on specific drug choice. Intravenous haloperidol has many benefits in cases of reduced gastrointestinal absorption or acute psychomotor agitation threatening the safety of catheters and ventila-tory support because it can be rapidly administered and absorbed. It also has minimal effects on blood pressure, respiration, and heart rate; a low sedative effect; and few anticholinergic side effects. In less acute situations and when oral medication is possible, atypical antipsychotics are preferred (5, 18). The choice of atypical agent can be based on side effect profile. The most serious side effects to consider are those of cardiac toxicity. Current limited data indicate that cardiac risks are similar across all atypical antipsychotics except that there may be less QTc prolongation with aripiprazole. Once the decision has been made to use an atypical agent, one can consider need for sedation, appetite stimulation/weight gain, nausea control, or glycemic control to identify the best drug. Table 2 describes drug profiles in detail.

The treatment of infants and young children with delirium is a new frontier in pediatrics and psychosomatic medicine, with limited information about assessment tools, best practices, and potential side effects of treatment. Clinical practitioners should be aware, however, that there are potential developmental consequences to untreated delirium in this population and consider intervention when appropriate.