1. WHAT IS THE EVIDENCE ON EFFECTIVENESS OF ANTIPSYCHOTICS IN PERSONS WITH DEMENTIA? 1

2. Atypical Antipsychotic Doses Used in Dementia MedicationLow DoseNormal Dose Aripiprazole (Abilify)<2 mg/d2-15 mg/d Olanzapine (Zyprexa)<5 mg/d5-10 mg/d Quetiapine (Seroquel)<50 mg/d50-100 mg/d Risperidone (Risperdal)<1 mg/d1-2 mg/d

3. Effectiveness in Dementia is weak Meta-Analysis (JAMA 2011) Zyprexa, Risperdal, and Abilify- small but statistically significant effect (12 – 20%) compared to placebo Seroquel – no statistically significant effect Antipsychotics led to an average change on the NPI of: 35% from a patient’s baseline 3.41 point difference from placebo group 30% change or 4.0 difference = minimum clinically meaningful No conclusive evidence found on comparative effectiveness of different antipsychotics Source: JAMA 306:1359-69 2011; Meta-analysis 38 RCTs in dementia

4. Net Effectiveness 4 “For every 100 patients with dementia treated with an antipsychotic medication, only 9 to 25 will benefit” Drs Avorn, Choudhry & Fishcher Harvard Medical School Dr Scheurer Medical University of South Carolina Source: Independent Drug Information Service (IDIS) Restrained Use of antipsychotic medications: rational management of irrationality. 2012

5. Effectiveness with Low Dose Low dose Risperdal <1 mg/d): small positive effect increased risk of adverse events Low dose Zyprexa (5 mg/d): no positive effect increased risk of adverse events Low dose Abilify and Seroquel effectiveness unknown, but Seroquel at normal dose is ineffective 5 Source: Cochrane Review 2012; Meta-analysis 16 RCTs in dementia

6. Adverse outcomes Off-label use of antipsychotics in nursing facility residents are associated with an increase in: Death Hospitalization Falls & fractures Venothrombolic events Conventional antipsychotics are worse than atypical antipsychotics 6

7. Odds of having an adverse event after receiving Risperidone 1 mg/d compared to placebo 7 Source: Cochrane Review 2012; Meta-analysis 4 RCTs in dementia Adverse EventOdds Ratio 95% Confidence Interval Mortality1.250.73 to 2.16 Somnolence2.401.70 to 3.20 Falls0.840.63 to 1.14 Extrapyramidal disorder1.781.00 to 3.17 UTI1.400.92 to 2.13 Edema2.751.51 to 5.03 Abnormal Gait5.312.24 to 12.62 Urinary Incontinence13.61.81 to 101 CVA3.641.72 to 7.69 Drop out (had to stop meds) 1.431.01 to 2.03

8. Evidence for Discontinuing Meds RCTs comparing withdrawal of medication to continuing antipsychotics will show the medication: to be effective, if more people randomized to stop the medication get worse than those randomized to continue on the medication to be ineffective, if the same percentage of people randomized to stop the medication as continue the medication get worse or do not change to be harmful, if more people randomized to stop the medication get better compared to those who continue the medication

9. RCT to withdraw antipsychotics 2 100 w/ dementia on antipsychotics 54 continue med Outcomes assessed over 3 months Outcomes - 76% no change in behaviors - NPI total worse - Agitation worse - QOL worse - 9% stopped due to behaviors 46 stopped med Outcomes - 67% no change behaviors - NPI total worse - Agitation worse - QOL better - 13% stopped due to behaviors Meds stopped abruptly and given a placebo 2 Ballard C et al J Clin Psychiatry 2004: 65:114-119 Statistical Difference None

10. RCT to withdraw antipsychotics 3 165 w/ dementia on antipsychotics 83 continue med Outcomes assessed after 6 months Outcomes (N=51) - Cognitive Fxn worse - NPI total worse - Verbal fluency worse - ADLs worse - Agitation 32% 82 stopped med Outcomes (N=51) - Cognitive Fxn worse - NPI total worse - Verbal Fluency better - ADLs worse - Agitation 34% Meds stopped abruptly and given a placebo 3 Ballard C et al Plos Medicine 2008; 5:e76: 587-599 Statistical Difference None YES None

11. RCT to withdraw antipsychotics 4 110 w/ Dementia with psychosis who responded to antipsychotics 32 continue med Outcomes assessed @ 4 & 8 months Outcomes - 33% Relapse (n=14) - Adverse events worse - Completed trial (N=10) 40 stopped med Outcomes - 60% Relapse (n= 23) - Adverse events worse - Completed trial (n=10) Meds tapered over 1 week to placebo Statistical Difference YES None 4 Devandand DP et al NEJM 2012; 367:1497-1507 Third group not shown here: continued med for 4 moths then discontinued meds