THE AUTHOR REPLIES

I read Dr. Spentzas’ interesting letter, where he reminds us that delirium is multifactorial, with a complex underlying pathophysiology and a final common pathway that likely involves alteration in neurotransmitter activity(1). He makes several important points in his letter, but I would like to focus on his observations regarding the difficulty distinguishing between delirium and dementia in adults. I agree with the comparison he draws between geriatric patients with dementia and the children in our pediatric intensive care units (ICUs) with neurocognitive disorders. These children are notoriously hard to assess for delirium. In addition to establishing altered cognition and awareness, one must also confirm that these behaviors are acute and fluctuating (i.e.: not reflective of underlying static encephalopathy). Although it would be convenient to exclude these hard to assess children from delirium research, I fear it would be unethical. Carefully designed prospective research studies have used the gold-standard psychiatric assessment – with an in-depth interview and exam to establish alteration from neurocognitive baseline – to diagnose delirium in children with neurocognitive disorders. Interrater reliability of delirium diagnosis has been established in this complex group of children(2). Results indicate that when compared to children with normal development, cognitively impaired children are at heightened risk for delirium. Similar to the geriatric patients referenced above, these children likely have decreased cognitive reserve, rendering them more susceptible to development of delirium when exposed to the stress of critical illness(3). In fact, the odds ratio for development of delirium in children with underlying cognitive impairment has been shown to be 3.5 (95% CI 1.5–7.8, p=0.003), after carefully controlling for the other epidemiologic predictors of delirium in children(4).

In children with typical development, a positive delirium screen (using either the Cornell Assessment for Pediatric Delirium (CAPD) or the Pediatric Confusion Assessment Method for the ICU (pCAM-ICU) has been used as a valid and reliable surrogate for delirium diagnosis in research studies, where it has been shown to track with outcomes(5,6). However, in a developmentally delayed child, we cannot automatically assume that a positive bedside screen is diagnostic of delirium. In addition to the positive screen, it is also necessary for the intensivist to determine that the observed behavior is an acute and fluctuating change from pre-hospital baseline. Despite this added layer of complexity, it would be shortsighted to omit this high-risk group from future delirium research, as they are likely to benefit from the knowledge gained regarding modifiable risk factors, opportunities for delirium prevention, and therapeutic approaches. I applaud Dr. Spentzas’ insightful conclusion that “looking for delirium vigilantly in critical care pediatrics is a good practice with immediate benefits for our patients. We should be looking for its expression in all our patients and not only the ones we can easily screen”(1).