How Long Should We Prescribe Medications for Agitation in the Nursing Home?

abstract & commentary

Source: Cohen-Mansfield J, et al. Withdrawal of haloperidol, thioridazine, and lorazepam in the nursing home: A controlled, double-blind study. Arch Intern Med 1999;159:1733-1740.

This double-blind crossover study assessed outcome when medication (haloperidol, lorazepam, and thioridazine) initiated for agitation was tapered in nursing home patients. Haloperidol, thioridazine, and lorazepam are commonly administered for agitation in nursing homes, but have significant side effects. Haloperidol (Haldol) is associated with stiffness and other movement disorders and anticholinergic side effects, which include urinary retention and constipation. Thioridazine (Mellaril) is also anticholinergic, and can potentially worsen cognition. Lorazepam (Ativan) is associated with sedation, confusion, and sometimes paradoxical agitation.

Although medications are often useful to control agitation in the elderly, concern about excessive use prompted the Nursing Home Reform Amendments of the Omnibus Budget Reconciliation Act (OBRA-87), which requires that psychotropic medication be prescribed only when a specific diagnosis has been made, and that dose reductions and behavioral interventions be attempted. One study noted a two-week increase in agitation after medication withdrawal, followed by an overall reduction in behavioral difficulties eight weeks later.3 An educational program targeted to physicians, nurses, and aides has also been shown to reduce the use of psychotropic drugs without adversely affecting overall behavior.7 The current study used a 550-bed, skilled, nonprofit nursing facility whose population reflects national norms in terms of age, sex, and functional impairment. Participants were on haloperidol, thioridazine, and lorazepam for at least four weeks and were older than 70. Exclusion criteria were concomitant administration of other psychotropic drugs, life expectancy of less than three months, diagnosis of a mood disorder or schizophrenia, current infection or uncontrolled diabetes, or expected nursing home stay of less than three months. Participants were randomly assigned to two groups by level of cognitive functioning (using the Brief Cognitive Rating Scale) and by psychotropic medication. In the first phase, group I (50%) had haloperidol, thioridazine, or lorazepam tapered over three weeks and then took placebo for seven additional weeks; group II continued without a medication taper. In the second phase, the groups were switched; group II tapered and took placebo, while medication was reinstituted for group I. Daily assessments were done by nurses regarding symptoms and medication side effects. Formal assessments were done at baseline, week 3 (after the first taper), week 10 (after the 7-week trial of placebo by group I), week 13 (after second taper and/or reinstitution), and week 20 (after the 7-week trial of placebo by group II). Assessments included the Brief Psychiatric Rating Scale, the Cohen-Mansfield Agitation Inventory, the Mini-Mental Status Examination, sleep and activity ratings, two screening questions on having a positive mood, weight, the Abnormal Involuntary Movement Scale, Clinical Global Impressions Scale, Brody’s Physical Self-Maintenance Scale (for activities of daily living), and adverse events. Paired t-tests were used to compare the groups and a two-way, repeated-measures analysis of variance was conducted to assess for within-subject changes.

Of 362 residents considered for the study, 251 were excluded (29% had prohibited medication use, 18% had medication discontinued before the study started, 19% had a mood disorder or schizophrenia, and 12% died before the study started). Nonparticipants had similar demographics to participants but statistically used more psychotropic medication. Participants were taking medication for a median of nine months and an average of 16.5 months before the study. Only 23 residents discontinued the study and 12 of those did so before the taper in the first phase; they did not statistically differ from ongoing participants in terms of severity of agitation. Medication adherence by participants was 100%. The average doses of haloperidol, thioridazine, and lorazepam were 1.03 mg, 25.75 mg, and 0.79 mg, respectively. No differences in agitation and side effects were found between groups I and II during either phase. Interestingly, by informal assessment, nurses attributed worsening of agitation to medication discontinuation even with the double-blind design. A chart review was done of the nonparticipants who had had medication discontinued outside of the study (the 18% mentioned above; n = 45). Over 22 weeks, 29% remained off the medication, 20% had drug reinstated, and 67% received another drug for the treatment of agitation (mean time to receiving it was 21 days). Cohen-Mansfield and colleagues pondered why no effect was shown. Low doses of the medications were used, but these doses were consistent with significant clinical efficacy in other studies, at least initially. Longer-term efficacy data are notably lacking.

Comment by Donald M. Hilty, MD

This was a well-designed study with a reasonable sample size given the difficult population. The lack of significant results could imply that haloperidol, thioridazine, and lorazepam do not have long-term clinical efficacy—this has not been formally studied but is believed to be true by many clinicians working in these settings. Another problem is the heterogeneity of the conditions that lead to a final, common pathway of "agitation." In terms of recommendations for practice, ongoing medical and psychiatric evaluation of patients with agitation is recommended to avoid missing treatable causes. Though more research is needed, gradual tapering of these medications with re-evaluation appears to be indicated because the risk of side effects may well outweigh benefit. Whenever possible, other psychotropic agents with more favorable central nervous system side effect profiles should be considered for long-term use (e.g., atypical antipsychotics, buspirone, divalproex, propanolol, SSRIs, or trazodone).6


1. Beardsley RS, et al. Prescribing of psychotropics in elderly nursing home patients. J Am Geriatr Soc 1989; 37:327-330.

2. Buck JA. Psychotropic drug practice in nursing homes. J Am Geriatr Soc 1988;36:409-418.

3. Fitz D, Mallya A. Discontinuation of a psychogeriatric program for nursing home residents: Psychotropic medication changes and behavioral reactions. J Appl Gerontol 1992;11:50-63.

4. Auer SR, et al. The treatment of behavioral symptoms in dementia: Haloperidol, thioridazine, and fluoxetine: A double-blind, placebo-controlled, eight-month study [abstract]. Neurobiol Aging 1996;17(suppl 4):652.

5. Schneider LS, et al. A metaanalysis of controlled trials of neuroleptic treatment in dementia. J Am Geriatr Soc 1990;38:553-563.

6. Hales RE, et al. Aggression and agitation. In: Rundell JR, Wise MG, eds. Textbook of Consultation-Liaison Psychiatry. 2nd ed. Washington, DC: American Psychiatric Press; in press.

7. Avorn J, et al. A randomized trial of a program to reduce the use of psychoactive drugs in nursing homes. N Engl J Med 1992;327:168-173.