Article Text

Download PDFPDF

Cohort study
Use of psychotropic medication in Alzheimer's disease is associated with more rapid cognitive and functional decline
  1. Edward C Lauterbach
  1. Departments of Psychiatry and Internal Medicine (Neurology Section), Mercer University School of Medicine, Macon, Georgia, USA
  1. Correspondence to : Dr Edward C Lauterbach
    Departments of Psychiatry and Internal Medicine (Neurology Section), Mercer University School of Medicine, 655 First Street, Macon, GA 31201, USA; eclbgnp{at}

Statistics from

Commentary on OpenUrlCrossRefPubMed


Psychotropic medicines are commonly prescribed in Alzheimer's disease (AD), yet their long-term effects are unknown. Rosenberg and colleagues determined AD clinical outcome trajectories after psychotropic exposure in the Cache County Dementia Progression Study. The authors hypothesised a more rapid clinical decline with antipsychotics and benzodiazepines and an improved status with antidepressants.


The authors studied 230 participants diagnosed with AD using the Diagnostic and Statistical Manual of Mental Disorders 1987 (DSM-III-R) and the NINCDS-ADRDA Alzheimer's Criteria. Participants were followed up for an average of 3.7±2.49 years (median follow-up of one visit). Mean changes in adjusted Mini-Mental State Examination (MMSE), Clinical Dementia Rating sum of six domains (CDR-Sum) and 10-item Neuropsychiatric Inventory—Clinician Rating total score (NPI-Total) ratings since AD diagnosis were analysed by the persistency index (PI) for each of the six psychotropic classes in a mixed effects model regression analysis adjusting for non-linear time effects and risk covariates. PIs (years of psychotropic use/years of observation) ranged from zero to one. The psychotropic classes examined were antidepressants, selective serotonin reuptake inhibitors (SSRIs), antipsychotics, atypical antipsychotics, typical antipsychotics and benzodiazepines.


At baseline, psychotropic use correlated with greater dementia severity and poorer medical status, conditions that are associated with greater neuropsychiatric symptoms and medications that can impair clinical ratings. Of the 230 predominantly later-onset (average age mid-80s), early-stage AD participants, persistent psychotropic use (PI≥0.50) was observed in 69 (30%) receiving antidepressants and 59 (25.7%) receiving SSRIs, but only 15 (6.5%) prescribed antipsychotics, 8 (3.5%) atypicals, 6 (2.6%) typicals and 16 (7%) benzodiazepines, and presumably even fewer individuals at follow-up. The participants received several classes of psychotropic drugs during the observation period, with only six participants exclusively taking antidepressants, none taking antipsychotics and eight with benzodiazepines at baseline. Indications for psychotropic administration, specific NPI domain scores and psychotropic present consumption status at evaluation were not ascertained, important because recent drug initiation or withdrawal could worsen performance on each outcome measure. Significantly more rapid decline with greater PIs was obtained for MMSE across all the psychotropic categories and for CDR-Sum and NPI-Total scores for some psychotropic classes. Correlates of persistent use (ie, PI≥0.50) were found for most of the psychotropic classes. The authors conclude that psychotropic administration identifies a group with a poor clinical prognosis, but they could not exclude that psychotropics may hasten disease progression.


This is a well-diagnosed population-based sample with incident AD and reliable ascertainment of psychotropic use. The limitations include a sample with unusual longevity, varying observation periods, limited data points (averaging two per individual), unequal group sizes, unknown psychotropic indications and efficacy, small sample sizes for antipsychotics and benzodiazepines, multiple psychotropics in most of the participants, unknown medication status at evaluation, non-significance of most of the NPI-Total and some CDR-Sum findings after Bonferroni correction and uncertain attribution of findings.

The findings indicate that AD patients tend to be prescribed multiple psychotropics simultaneously, have greater dementia severity and functional impairment, deteriorate more rapidly and may represent an AD subset with poorer prognosis with trajectories resembling later-stage AD. Indeed, psychotropic administration correlated with AD progression risk factors (younger age with SSRIs, apolipoprotein E ɛ4 (APOE-ɛ4) positivity with antipsychotics). AD recognition may be mildly delayed in antidepressant candidates, especially women with poorer health. Although the study design precludes conclusions that psychotropics cause deterioration, the authors recommend considering ‘non-pharmacological treatments first, unless the risk of harm or distress is high’. This is a good advice in light of important psychotropic adverse events in this population.

The findings accord with previous investigations cited in Rosenberg et al., as well as the Clinical Antipsychotic Trials of Intervention Effectiveness-Alzheimer's Disease (CATIE-AD)1 and benzodiazepine dementia risk studies.2 These data indicate that the clinical state may mask disease-modifying effects, necessitating methodological enhancements such as ≥24 months continuous psychotropic exposure analysed as a continuous variable, excluding multiple psychotropics and matching or controlling for AD duration, cognition, functional status, sex, APOE-ɛ4, pre-exposure and postexposure trajectories (and for SSRIs, age and medical health), requiring huge populations. Preclinical data reveal both prodegenerative and neuroprotective effects of psychotropics, as well as varying effects between and within drug classes.3 They indicate lithium, valproate, paroxetine, fluoxetine, quetiapine and perhaps other antipsychotics as potentially promising AD neuroprotectants,3 ,4 further supported by a recent lithium trial.5 Prospective randomised controlled trials incorporating these features (eg, the recent negative valproate trial)6 would be informative.


View Abstract


  • Competing interests None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.