Light levels of sedation and DSM-5 criteria for delirium

Abstract

Dear Editor, In a recent opinion essay, Devlin and colleagues argue that sedation confounds reliable detection of delirium when using current assessment tools [1]. To support this, they state that a low Richmond Agitation and Sedation Score (RASS) or bispectral index (BIS) value increase the likelihood to screen delirium positive [2], and that sedation-induced delirium has a better prognosis than delirium that is unrelated to sedation. The issues raised by Devlin et al. are fundamental, as these refer to the question what delirium ‘really’ means. In the end, this question considers the definition that is used to diagnose delirium, also debated by Brummel and Ely [3]. Recently, the American Psychiatric Association adjusted the criteria for delirium in its fifth Diagnostic and Statistical Manual (DSM) of mental disorders (DSM-5) [4]. Compared to the fourth, revised edition (DSM-IV-R) [5], less emphasis has been put on a disturbance of consciousness, Table 1. Still, a patient at a light level of sedation may fulfil DSM-5 criteria for delirium. Low RASS and BIS values are not characteristic of sedation, and can also be observed in patients with hypoactive delirium who did not receive sedatives. Besides the unpublished study that is cited by the authors, other unpublished material suggests the opposite: the prognosis of sedation-induced delirium was found to be as bad as that of delirium due to other causes (T. Girard, personal communication, Annual Congress European Delirium Association, 20–21 September, 2013). Moreover, because of the complex, multifactorial nature of delirium, it may be questionable whether these two categories can really be distinguished. Delirium in critically ill patients usually results from a multitude of pathogenic factors including inflammation, multi-organ failure, metabolic disturbances, and medication effects. With current assessment tools light levels of sedation may be clinically indistinguishable from hypoactive delirium in which no sedation was used. There is a need for an objective delirium assessment tool that can also identify sedation effects. Electroencephalography with a limited number of electrodes and automatic processing is a promising