November 2005
Reviewer: Jeff C. Huffman, MD
Association between psychoactive medications and delirium in hospitalized patients: a critical review
Gaudreau J, Gagnon P, Roy M, et al
Psychosomatics 2005; 46(4):302-316
Summary: A thorough, well-designed review of the association between medications and the development of delirium found little consensus between studies and no clear evidence to strongly link any class of medication to the development of delirium.
Background/Method: Delirium is an exceedingly common and important problem in psychosomatic medicine. Delirium increases length of stay and independently predicts mortality in the hospital and after discharge. Medications are frequently implicated in the causation of or contribution to delirium, with several classes of medications (benzodiazepines, opioids, and anticholinergics) commonly identified as ‘deliriogenic.' Despite this, there has been no systematic meta-analysis or other critical review of the literature to confirm the clinical sense that these (and other) medication classes are associated with delirium. The authors comprehensively reviewed literature from 1975-2003 to identify articles that studied hospitalized patients using standardized assessments of delirium to compare rates of delirium in users and nonusers of specific psychoactive medications; they eventually found 22 studies that met their criteria. They then performed a systematic review of these articles (data could not be pooled because of significant heterogeneity between studies).
Results: A variety of medications have been studied; 5 of 11 studies examining grouped psychoactive medications found association with delirium. Of the most-studied individual classes of medication, 4 of 11 studies examining benzodiazepines, 4 of 12 studies examining opioids, 0 of 7 studies examining anticholinergics, and 2 of 5 studies examining antipsychotics found associations with the development of delirium. The studies were all widely heterogeneous in methods, time course, and data analysis, and no clear association between any class of agent and delirium could be supported.
Comment: This appears to be an exceptionally careful and well-designed review. Unfortunately, a review can only be as good as the original studies it assimilates, and thus far studies of delirium and medication use have been small (and often inadequately powered) and with multiple limitations. The surprising results of this review—that no class of medications has been associated with delirium in more than 50% of studies and that antipsychotics are more associated with delirium than anticholinergics—must be carefully examined.
One major limitation of this area of research is that it is often difficult to separate medication use from all other variables that may represent or cause delirium: for example, antipsychotics may be associated with delirium because they are actually being used to treat symptoms of (diagnosed or undiagnosed) delirium and not because they are causing delirium, benzodiazepines may be prescribed for patients with sleep disturbance that represents early stages of delirium and then are seen to “cause” the delirium, and opioids are often prescribed post-operatively, when multifactorial delirium is very common, and then are seen to “cause” the delirium. The results of this review seem to collide with clinical experience and published guidelines concerning delirium (i.e., that anticholinergics and benzodiazepines are among the greatest offenders with regard to delirium, and that antipsychotics effectively reduce core symptoms of delirium).
Given the frequency of delirium, the profound effects of delirium on patient outcomes, and the lack of clarity from the existing literature, it seems that a continued push for delirium research is required to allow practitioners of psychosomatic medicine to best understand and treat this serious condition. This well-done review clearly brings this point home.