International Psychogeriatrics: page 1 of 17 C International Psychogeriatric Association 2014 doi:10.1017/S1041610214001720 REVIEW development for agitation and aggression in Alzheimer disease: review and discussion of recent randomized clinical trial design

...... Maria Soto,1 Sandrine Andrieu,1,2 Fati Nourhashemi,1 Pierre Jean Ousset,1 Clive Ballard,3 Philippe Robert,4 Bruno Vellas,1 Constantine G. Lyketsos5 and Paul B. Rosenberg5 1Gerontopôle, INSERM U 1027, Alzheimer’s Disease Research and Clinical Center, Toulouse University Hospital, France 2Department of Epidemiology, Toulouse University Hospital Toulouse, France 3Wolfson Centre for Age-Related Diseases, King’s College, London, UK 4EA CoBTeK / ICMRR University of Nice Sophia Antipolis -CHU, France 5Department of Psychiatry, The Johns Hopkins Bayview Medical Center, Baltimore, MD, USA

ABSTRACT

Background: The management of disruptive neuropsychiatric symptom (NPS) such as agitation and aggression (A/A) is a major priority in caring for people with Alzheimer’s disease (AD). Few effective pharmacological or non-pharmacological options are available. Results of randomized clinical trials (RCTs) of drugs for A/A have been disappointing. This may result from the absence of biological efficacy for tested in treating A/A. It may also be related to methodological issues such as the choice of outcomes. The aim of this review was to highlight key methodological issues pertaining to RCTs of current and emerging medications for the treatment of A/A in AD. Methods: We searched PubMed/Medline, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov for RCTs comparing medications with either placebo or other drugs in the treatment of A/A in AD, between January 2008 and December 2013. Results: We identified a total of 18 RCTs; of these, 11 were completed and 7 ongoing. Of the ongoing RCTs, only one is in Phase III. Seven of 10 completed RCTs with reported results did not report greater benefit from drug than placebo. Each of the completed RCTs used a different definition of “clinically significant A/A.” There was considerable heterogeneity in study design. The primary endpoints were largely proxy-based but a variety of scales were used. The definition of caregiver and scales used to assess caregiver outcomes were similarly heterogeneous. Placebo response was notable in all trials. Conclusions: This review highlights a great heterogeneity in RCTs design of drugs for A/A in AD and some key methodological issues such as definition of A/A, choice of outcome measures and caregiver participation that could be addressed by an expert consensus to optimize future trials design.

Key words: behavior, agitation, aggression, Alzheimer’s, measurement, therapeutics, clinical trial

Introduction Alzheimer’s disease (AD); 98% of people with AD (PwAD) develop at least one neuropsychiatric The number of people living with dementia symptom (NPS) over the course of the disease worldwide is estimated at 35.6 million and expected (Steinberg et al., 2008; Gonfrier et al., 2012). At to increase to 115.4 million by 2050 (Prince et al., least 20% of outpatients (Lyketsos et al., 2000) 2013). Seventy percent of dementia is due to and 40% of long-term care residents (Selbæk et al., 2013) exhibit disrupted NPS such as agitation and

Correspondence should be addressed to: Maria Soto, Alzheimer’s Disease Research aggression (A/A) encompassing a range of affective, and Clinical Center, Toulouse University Hospital, 170 avenue de Casselardit, verbal, and motor disturbances such as restlessness, 31059 Toulouse, Cedex 9, France. Phone: +33561779665;Fax:+33 5 cursing, aggression, hyperactivity, combativeness, 61 49 71 09. E-mail: [email protected]. Received 24 Feb 2014; revision requested 23 Apr 2014; revised version received 21 Jul 2014; accepted wandering, repetitive calling out, irritability, and 22 Jul 2014. disinhibition (Cohen-Mansfield et al., 1995). A/A 2 M. Soto et al. tends to co-occur with sleep disorders, delusions, the Brief Psychiatric Rating Scale (BPRS), the hallucinations, anxiety or dysphoria (Canevelli Behavior Pathology in Alzheimer’s Disease Rating et al., 2013). A/A is associated with greater Scale (BEHAVE-AD), the Neuropsychiatric caregiver burden (Okura and Langa, 2011), earlier Inventory (NPI), the Cohen-Mansfield Agitation institutionalization and death (Okura et al., 2011), Inventory (CMAI) and subscales (proxy-based poorer functioning (Okura et al., 2010), greater more common than direct observation), and cost of care (Murman et al., 2002), and more global assessments (Salzman et al., 2008). A acute hospitalizations (Soto et al., 2012). Thus, the non-pharmacologic intervention before enrolling management of A/A is a major priority in caring for a patient in a clinical trial and a placebo run-in PwAD. period were not common. Repeated measurement Consensus guidelines and expert opinion analyses were not performed in most trials. statements recommend non-pharmacological Atypicals APs, mainly risperidone, have the approaches to be first line (Benoit et al., 2006; best evidence for short-term efficacy (6–12 Lyketsos et al., 2006; Rabins et al., 2007; weeks), although meta-analyses have not indicated Gauthier et al., 2010;Kaleset al., 2014) but there significant benefit for non-aggressive symptoms of are limited options. Examples include caregiver agitation (Ballard and Howard, 2006; Schneider education, training in problem solving, and targeted et al., 2006). Efficacy is modest and AP use is interventions to causes for specific behaviors (Gitlin associated with serious adverse effects including et al., 2012). Patients in both community dwelling cerebrovascular accidents and mortality (Schneider (CD) (Brodaty et al., 2012) and nursing home et al., 2006;Kaleset al., 2012; Langballe (NH) settings benefit (Deudon et al., 2009; Husebo et al., 2013). In the European Union, risperidone et al., 2011; Ritcher et al., 2012). is indicated for the short-term treatment of Pharmacological treatment for A/A is recommen- severe aggression. In Australia the regulatory ded when non-pharmacological interventions fail authority, the Pharmaceutical Benefits Advisory or when A/A is linked to dangerousness to others Committee (PBAC), indicates risperidone for the or marked distress. The most studied medication treatment of psychotic symptoms and aggression class is antipsychotics (APs), both conventional and with unsuccessful non-pharmacological methods. atypical. Between 1999 and 2008, several RCTs as- The Food and Drug Administration (FDA) has sessed APs for treating A/A in PwAD. Eleven RCTs published a black box warning for the use of atypical used conventional APs, which mostly involved small APs in PwAD. In North America there are no sample sizes and with durations of 4 and 12 weeks approved drugs for treatment of NPS in AD. As a (Schneider et al., 1990;DeDeynet al., 1999; result, most agents are used off-label (Maher et al., Teri et al., 2000; Lonergan et al., 2002; Ballard 2011). Thus, management of severe, persistent or et al., 2009). Outcome was defined as a 30% im- recurrent A/A unresponsive to non-pharmacologic provement on standardized behavioral rating scales, intervention is a real challenge for clinicians. as per convention. A high placebo response was Emerging neurobiological research about patho- found in these RCTs. Since 1995, 18 RCTs have genesis has led to investigation of repositioned examined the efficacy of atypical APs in patients and novel therapeutics for A/A in PwAD, as with AD, mainly with durations of 6–12 weeks (only an alternative to APs. However, the limited three trials of 6–12 months) (Ballard and Howard, benefits reported so far may result from limited 2006; Schneider et al., 2006). Small scale trials of understanding of pathogenesis but also from key treatment with drugs other than APs (antidepress- methodological issues. We hypothesized that a ants and anti-convulsants mood stabilizers) have great heterogeneity in the design of recent RCTs produced equivocal results (Ballard et al., 2009). of drugs for A/A in AD would be found and The available data are limited by small numbers of that specific key methodological issues could be subjects or shortcomings in study design such as the identified. Thus, the objective on this paper was to (non-random) statistical distribution of behavior review methodological aspects from recent RCTs test scores and lack of consideration of effect size. of drugs for A/A in AD since 2008; the date of the In a general description all the previous most recent consensus statement on clinical trials studies since 1990 were placebo-controlled methodology of treatments for A/A in dementia and were parallel-group, fixed-dose range, or (Salzman et al., 2008). adjustable/titrated-dose trials, in the majority involving nursing home patients with a mean age over 80 years of age. Among subjects studied, there Methods was a wide degree of variation in type and severity of symptomatology. The clinical trials endpoints Reports of RCTs of medications for the treatment of were based on behavior rating scales, including A/A in AD published in the English language were Clinical trials of drugs for agitation in AD 3 identified by searching PubMed between January Completed RCTs do not report superiority of 2008 and December 2013, using terms (“De- any drug over placebo, or over active comparator, mentia”[Mesh] OR “Alzheimer”[Mesh]) AND but rather improvement in both groups (Table 1), (“Clinical Trial”[Mesh] OR “therapeutics”[MeSH with three exceptions. The prazosin pilot study Terms]) AND (“Agitation”[Mesh] OR “Aggres- (Wang et al., 2009) reported superiority of drug sion”[Mesh]) OR “Behavioral symptoms”[Mesh]). to placebo; the drug is being further studied in an Free text was used to identify articles on ongoing RCT. Improvements over placebo were “neuropsychiatric symptoms,” “treatment for reported in a trial of intramuscular aripiprazole neuropsychiatric behaviors,” and “behavioral and (Rappaport et al., 2009). In the very recent CitAD psychological symptoms of dementia.” This search trial, citalopram showed significant improvement was supplemented by hand searching of reference compared to placebo on both primary outcome lists of selected articles, meta-analyses, and review measures (Porsteinsson et al., 2014). articles. Google Scholar was searched for additional Two trials report no effect on primary articles, especially of ongoing RCTs and new drugs. outcomes but improvement in secondary outcomes. In addition, we searched the Cochrane Central Mibampator (Trzepacz et al., 2013a) led to better Register of Controlled Trials and ClinicalTrials.gov outcomes on the Frontal Systems Behaviors Scale, between January 2008 and December 2013. We and -NH on NPI total score (Fox et al., included reports (1) whose publication appeared 2012). between 2008 and 2013 and (2) registered in The next section highlights key methodological a clinical trial registry (including RCTs with no aspects of the completed and on-going RCTs. publications or posted results). We included studies where A/A was the primary or co-primary outcome. Population studied Studies, where A/A was a secondary outcome, were A GE not included. Studies focused only on psychosis, Most trials included patients aged between 50 and depression or apathy in PwAD were excluded. 90 years; with the youngest of 40 in THC and the Only randomized, parallel-group, controlled trials oldest of 95 in versus risperidone trial. comparing medication to placebo or to another There was no age limit in both prazosin RCTs. medication were included.

D EMENTIA DIAGNOSIS Diagnosis of AD was based on DSM IV and/or NINDS-ADRDA criteria (McKhann et al., 1984). Results The ongoing scyllo-inositol trial used recent AD We identified 18 RCTs evaluating efficacy of criteria of the National Institute of Aging-AD medications for treatment of A/A in AD. Of these, Association (McKhann et al., 2011). The ongoing 11 were completed RCTs and seven ongoing. THC trial includes patients with vascular or mixed These trials were characterized by a great deal of dementia. methodological heterogeneity. The therapeutic agents divide into: (1) D EMENTIA SEVERITY repurposed drugs marketed for other indications Dementia severity was assessed on Mini-Mental (e.g., citalopram, , delta-9- State Examination (MMSE) (Folstein, 1975)or tetrahydrocannabinol (THC) or prazosin) or Clinical Dementia Rating scale (CDR) (Rosen, (2) new chemical entities not approved for any 1984). The divalproex trial included patients with indication (e.g., mibampator or scyllo-inositol). Of moderate to severe dementia, while memantine these drugs only two were APs. and oxcarbazepine trials studied severe dementia. Table 1 lists RCTs completed between 2008 Citalopram, mibampator, and aripiprazole trials and 2013. Nine compared drug to placebo; two included a wide range of severity. Among used an active drug comparator (risperidone vs. ongoing trials, Scyllo-inositol include moderate to escitalopram or risperidone vs. topiramate). Five moderately-severe AD (MMSE 10–20), while the RCTs had < 50 subjects. THC, dextromethorphan-quinidine, and brexpi- Table 2 lists seven ongoing RCTs, three prazole trials include patients at all severities. assessing new chemical entities, and four repur- posing drugs marketed for other indications. Only S ETTING brexpiprazole is in phase III. Most are U.S. trials Of the 11 completed RCTs, 5 were NH studies, 3 with some in Europe, Canada, and other regions. CD studies, while 1 (risperidone vs. escitalopram) The mix of outcomes and industry sponsorship is assessed hospitalized patients (see Table 1). Only similar to completed trials. two RCTs (CitAD and prazosin) included patients 4 M. Soto et al. NBRS & mADCS- CGIC for total NPI outcomes outcomes only or FrSBe Positive for Null for all Null, positive RESULTS ...... a N ↓ 169 48 Null 40 None 450 ↓ 369 150 ↓ 132 2009 2010 2010 2011 Jan 2013 186 Jan 2010 153 Null, positive END DATE May Sept June ...... 2009 2007 2008 2003 2009 START DATE ...... USA 8 July UK Sept Israel 1 April Canada 23 Dec USA 16 129 Positive for all STUDY LOCATION USA 16 Feb & SITES ∗ ADCS- ADL, MMSE, NPI- distress, cumulative lorazepam dose 12wk, NPI, CGI-C, MMSE, SIB discontinu- ation drug plus, ADCS- ADL CGI-I NPI-10, Cornell, FrSBe, Adas-cog, CGI-S-A/A, NPI- distress NPI, CMAI, CMAI at CMAI, CIBIC SECONDARY OUTCOMES CMAI, ...... p from baseline to 6wk 0.31 in mean change from baseline to end point 0.41 in mean change from baseline to end point NBRS:3–5 p CGIC: 20% Difference of 6 Not reported MMSE Iran 1 Jan 2008 Mar Effect size of DESIRED CLINICAL DIFFERENCE Effect size of ...... NBRS CMAI ranks test analyses of the changes from baseline to Week 24 in NPI and SIB total scores based, MMRM DEFINITION OF PRIMARY COMPARISON ITT LME model of LME model of Wilcoxon signed & ANALYSIS ITT. Likehliood- (max b -A/A modified- ADCS- CGIC score 48) ...... OUTCOME PRIMARY MEASURES NPI-4 domain NBRS CMAI at 6 wk ITT Total NPI CMAI CGI PANNS-EC ACES, CGI-S, 4 Total NPI SIB ANCOVA > dementia depression memantine usage EXCLUSION CRITERIA Delirium Vascular Major Hachinski Previous Delirium Total NPI Not reported Time to 19 ...... < 13 ࣙ 10 45 y ADRDA DSM MMSE 6–26 NPI-10 ࣙ ADRDA MMSE 5–28 ADRDA & DSM-IV NPI psy wards ࣙ MMSE and mixed 60 50 y CD ࣙ NINCDS- INCLUSION CRITERIA CD & NH NINCDS- CD ࣙ MMSE 5–15 NINDS- NH & Acute CD 55–95 y DSM-IV MMSE10–15 Hospitalized 55–95 y DSM-IV MMSE 5–16 NH AD, vascular 1 4 > 4 ࣙ 2 31 ࣙ ࣙ 45 2, s < 1for ࣙ 1on1 10 > ࣙ ࣙ > ...... on 1 domain with s D/H, AMB agitation domain H, Irritability on 1 domain 14 & NPI-4-A/A Agitation Aggression NPI PRIMARY AIM DEFINITION Agitation NPI A/A NPS: A/A, NPI f Agitation NPI Agitation CMAI NPS: A/A, D, NPI Acute agitation PANNS-EC > TT doses doses doses doses doses doses DESIGN Ph 2 12 wk fixed Ph 3 9wkfixed 6wk Ph 3 24 wk fixed Ph 4 12 wk fixed PERIOD 8wkfixed 2IMfixed 24 hours ...... , , et al. , 2014 2013 2009 , , , et al. et al. 2013a et al. et al. 2012 2010 et al. Trzepacz Portsteinsson None Ph 4 Herrmann Fox Mowla Rappaport ...... () potentiator serotonin reuptake inhibitor (SSRI) atypical an- tipsychotic agonist agonist modulator vs. atypical anti- psychotic partial agonist MECHANISM OF ACTION REFERENCE SSRI vs. NMDA NMDA AMPA/GABA Completed randomized controlled trials (RCTs) of drugs for agitation and aggression in Alzheimer’s disease c CD vs. Risperidone NH Risperidone Mibampator AMPA Memantine- Table 1. MOLECULE Citalopram Selective Escitalopram Memantine- Topiramate vs. Aripiprazole Dopamine D2 ...... Clinical trials of drugs for agitation in AD 5 outcomes outcomes outcomes, significant adverse events outcomes RESULTS ...... a National 22 Positive for all 103 Null for all N 313 Null for all = Aberrant Motor = 2007 2006 2009 Oct Mar END DATE linear mixed effects model; ...... = 2005 2005 START DATE ty/emotional lability; and disinhibition. Clinical Global Impression of Change; CGI-I ...... USA 1 Jan 2005 Sep STUDY LOCATION Norway 35 Sept USA 46 Nov & SITES = delusions/hallucinations; AMB Diagnostic and Statistical Manual of Mental intention to treat; NINDS-ADRDA = = Clinician’s Interview-Based Impression of Change-plus intramuscular; LME model = caregiver distress, = = BPRS SE Blood pressure BARS, NPI- SECONDARY OUTCOMES Total NPI Total CMAI ADCS-CGIC Qol-AD MMSE ADAS-cog ADCS-ADL CDR ...... Positive and Negative Syndrome Scale – Excited Component; = points; IN between treatments of 1.2 p in incidence rate = difference 33% difference Not reported USA 1 43 Null for all DESIRED CLINICAL DIFFERENCE nursing home; ITT = agitation/aggression; D/H Alzheimer’s Disease Cooperative Study-Activities of Daily Living Inventory; ...... = = test (CGIC) measures model with an autoregressive covariance structure endpoint hazard ratio. variances DEFINITION OF PRIMARY COMPARISON LME model (NPI,BPRS) Mann-Whitney Time to Cox proportional Total sleep time. Analysis of & ANALYSIS Frontal Systems Behaviors Scale; CGI-C = 3 ࣙ Quality of Life-AD; p = CGIC on 1 or more NPI items 2wks actigraphy CMAI & ABRS Severe impairment Battery; CIBIC-plus ...... OUTCOME PRIMARY MEASURES Time to NPI-NH A/A repeated- A/A, D, H for Sleep: Agitation: = community-dwelling; NH = Psycho- tropic drugs hypotension Total NPI EXCLUSION CRITERIA Changes in neuropsychiatric symptoms; A/A = ...... 1in 54 y ࣙ < ADRDA ADRDA A/A, D, H since AD onset AD; ICD-10 ADRDA 55 y 60 y MMSE10–20 NINDS- NPI CD & NH No age limit NINDS- INCLUSION CRITERIA CD, NH ࣙ Vascular or MMSE 5–16 NH ࣙ NINDS- Severe AD week; NPS Clinical Global Impression Severity of Illness Score; PANSS-EC 6 = ࣙ 4inat National Institute of Aging; QoL-AD = /wk ࣙ Brief Psychiatric Rating Scale; SIB Cohen-Mansfield Agitation Inventory; ADCS-ADL ࣙ = ...... Aggression 2 least 1 item psychosis more NPI items 2wks Aggression Sleep 3on1or = = Agitation & BPRS Agitation & ࣙ A/A, D, H for Agitation & NPI A/A PRIMARY AIM DEFINITION Agitation & years; wk = mixed-effects model repeated measures analysis. TT American Academy of Neurology; CD = = doses doses doses doses 10 days fixed DESIGN Pilot 8 wk flexible 24 month fixed Ph 3 8wkfixed PERIOD ...... , , , , Treatment; y et al. et al. = et al. et al. 2009 2009 2011 2009 frequency; AAN Gerrman Wang Tariot Sommer = phase; TT ...... = Neuropsychiatric Inventory; CMAI Agitation Behavior Rating Scale; NIA Alzheimer’s Disease Cooperative Study-Clinical Global Impression of Change; FrSBe = receptor agonist receptor antagonist GABAergic function receptor potentiator = = 1-adreno severity; f analysis of covariance; MMRM α MECHANISM OF ACTION REFERENCE = Continued. = Agitation-Calmness Evaluation Scale; BPRS = Clinical Global Impression Improvement Score; CGI-S Completed RCT registered in a clinical trial registry between 2008 and 2013 and/or published between 2008 and 2013. The NPI-4 A/A is a 4-domain subscale chosen which combines from the NPI-10 four domains: agitation/aggression; aberrant motor behavior; irritabili Randomized, completed. This study was prematurely terminated due to recruitment problems. Melatonin Melatonin Prazosin Abbreviations: Ph Table 1. MOLECULE Valproate Enhanced ∗ a b c Oxcarbazepine GABA Behavior; s Institute Of Neurological and CommunicativeDisorders; Disorders NPI and Stroke and the Alzheimer’s Disease and Related Disorders Association, DSM ADCS-CGIG = version; ABRS ACES ANCOVA ...... 6 M. Soto et al. ) TARGET N ( 22 230 150 400 120 560 ...... a 2013 2016 2014 2013 2015 2017 END DATE May Feb Dec Dec July June 2013 2012 2011 2012 2010 2013 ...... START DATE Sept Nov July SITES Europe Europe STUDY LOCATION & USA, Canada, Netherlands 2 June USA Canada Europe 100 USA 1 March USA, Canada, ...... NH, CMAI subscales, NPI individual items, NPI-D/H, caregiver distress, CGI-I, CGI-E, QoL-AD PACSLAC-D, Caregiver- CGIC, Qol-AD, Barthel NPI, NPI-C domains, CMAI, ADCS-ADL, MMSE completed NH, CMAI subscales, NPI individual items, NPI-D/H, caregiver distress, CGI-I, CGI-E, QoL-AD ...... SECONDARY OUTCOMES ADCS-CGIC, Number of days BPRS Total BPRS combined A/A scores & total NPI OUTCOME PRIMARY MEASURES Total CMAI CGI-S, total NPI Total NPI CMAI, Total NPI CMAI, Zarit. Netherlands 2 Sept NPI-C Total CMAI CGI-S, total NPI ...... psychiatric disorder psychiatric disorder depression EXCLUSION CRITERIA Axis I (DSM 4) Axis II Major Major Major Psychosis Hypotension ADCS-CGIC Axis I (DSM 4) Axis II ...... 40 y 18 y ADRDA MMSE 5–22 AD/VaD ࣙ CDR 1–3 Persistent pain AD/VaD ࣙ CDR 1–3 Persistent pain ADRDA MMSE 5–22 INCLUSION CRITERIA NH 55–90 y NINDS- CD VaD, AD or CD VaD, AD or CD 50–88 y NIA-AA MMSE 10–21 CD No age limit NH 55–90 y NINDS- 1 1 2) ࣙ ࣙ 4 ࣙ 4 ࣙ ...... 4 ࣙ ࣙ 10 and 10 and 2&f A/A ࣙ ࣙ ࣙ at least 1 domain A/A or AMB at least 1 domain A/A or AMB Aggression (s agitated behavior at least twice/wk PRIMARY AIM DEFINITION NPS NPI NPS NPI Agitation & NPI A/A Disrupted Agitation NPI-A/A Agitation CGI-S b ...... dose dose doses doses label follow-up Doses follow-up 2 wk fixed 12 wk open 3 fixed Doses 30 days 30 days Ph2 3 wk fixed Pilot Ph2 Ph 2 12 wk fixed Ph 2 12 wk fixed Ph3 12 wk Ph3 12 wk flexible ...... agonist agonist metabolism & phosphoinositol signaling antagonist receptor partial agonist receptor partial agonist 1-adrenoreceptor CB-1 receptor MECHANISM OF ACTION DESIGN TT PERIOD CB-1 receptor α Myo-inositol Anti-amyloid Dopamine D2 Dopamine D2 ...... Ongoing randomized controlled trials (RCT) of drugs for agitation and aggression in Alzheimer’s disease hydrocannabinol hydrocannabinol (ELND005) (OPC-34712) (OPC-34712) Delta-9-tetra Delta-9-tetra Table 2. MOLECULE Scyllo-inositol Prazosin Brexpiprazole Brexpiprazole ...... Clinical trials of drugs for agitation in AD 7 ) = from both NH and CD. Regarding ongoing RCTs, four are only CD, two NH only (brexpiprazole), and

TARGET one (dextromethorphan) CD and NH. N ( 200 = ......

2014 D IAGNOSIS OF CLINICALLY SIGNIFICANT END DATE March A / A There is not a “gold standard” definition of A/A Caregiver Strain Index; nursing home; NIA-AA 2012 ...... enrollment; A/A =

= in AD (Salzman et al., 2008). In its absence two START DATE = approaches have been used: (1) judgment by ex- perienced clinicians that medication is appropriate and/or (2) severity rating above a scale cut-off

SITES indicative of at least moderate A/A. CitAD and STUDY LOCATION & USA 25 June ...... dextromethorphan-quinidine trials combined both.

Cohen-Mansfield Agitation Inventory; A/A encompass a range of recognizable features

= (such as hitting, pacing, or disinhibition) but these

community-dwelling; NH are shared with other behavioral disturbances of = AD. It is not surprising that inclusion criteria vary neuropsychiatric symptoms; N Qol-AD, ADCS-ADL, CSI between RCTs. Several trials have used cutoffs on = ...... SECONDARY OUTCOMES total NPI (mibamptor, and THC). The THC trials additionally require high scores on two domains of the NPI: A/A or motor disturbance.

week; NPS Although NPI is widely used to define A/A across = OUTCOME PRIMARY MEASURES Total NPI ADCS-CGIC, RCTs, the specific definitions have varied (Tables 1

...... and 2). Some trials (CitAD, brexpiprazole, and Clinical Global Impression Severity of Illness Score; CSI Alzheimer’s Disease Cooperative Study-Clinical Global Impression of Change; CGI-C

delusions/hallucinations; CD scyllo-inositol) used the NPI A/A domain, while the Neuropsychiatric Inventory; CMAI = years; wk = = = mibampator trial required the presence of at least =

pain assessment checklist for seniors with limited ability to communicate Dutch version. one of four NPI domains. gravis = EXCLUSION CRITERIA Myasthenia Thresholds for agitation severity similarly

...... vary: in memantine-NH trial agitation was high

Treatment; y (CMAI > 45) (Fox et al., 2012), while in the frequency; D/H

= memantine CD trial it was low (NPI A/A ࣙ 1) = (Herrmann et al., 2013). Recent RCTs (CitAD INCLUSION CRITERIA NH & CD 50–90 y DSM IV MMSE 6–24 and mibampator), and the ongoing scyllo-inositol ࣙ

phase; TT and brexpiprazole trials use a threshold of 4on severity; F ...... 4 =

Clinical Dementia Rating; NPI NPI A/A; CiTAD adds the requirement that A/A ࣙ = A/A = occur at least several times/week (“frequently”) PRIMARY AIM DEFINITION Agitation CGI-S

quality of life-AD; PACSLAC-D to improve relevance and applicability to real-life

= situations.

years old; S Moreover, the CitAD, brexpiprazole, and scyllo- = inositol trials require a history of poor response to ...... dose National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association, non-pharmacological interventions for inclusion. Clinical Global Impression Improvement Score; CGI-S = Vascular Dementia; Ph The FDA appears to be increasingly accepting = = A/A as a target indication, examples being the on- going dextromethorphan and scyllo-inositol trials. side effects; QoL-AD Ph2 10 wk fixed = ...... C AREGIVER PARTICIPATION Most trials require a reliable caregiver as informant Aberrant Motor Behavior; y

= for patient symptoms and caregiver burden. In CitAD the requirement was for a primary caregiver antagonist and sigma-1 receptor agonist Alzheimer’s disease; VaD MECHANISM OF ACTION DESIGN TT PERIOD NMDA receptor who spends at least several hours a week with = ...... the patient, supervises his/her care and attends placebo, delta-9-tetra hydrocannabinol and acetaminophen. Alzheimer’s Disease Cooperative Study-Activities of Daily Living Inventory; ADCS-CGIG = = clinic visits with the patient. Mibampator required

Continued. “frequent or daily contact with the patient.” Diagnostic and Statistical Manual of Mental Disorders; CDR National Institute of Aging; SE Dextromethorphan requires a caregiver who is with = = the patient a minimum of three times per week on Clinical Global Impression of Change; CGI-I Quinidine (AVP-923) Three arms Active trial but with a closed recruitment with a 6-month open label safety extension until March 2014. Table 2. MOLECULE Dextromethorphan/ a b Abbreviations: AD National Institute of Aging-AD; NINDS-ADRDA DSM ADCS-ADL agitation/aggression; AMB = NIA ...... three separate days, THC requires a caregiver who 8 M. Soto et al. is in touch with patient at least twice weekly, while 2010) by providing “enhanced usual care” to the ongoing prazosin trial requires a caregiver who patients and caregivers regardless of treatment spends 10 hours per week with the patient. assignment (Drye et al., 2012).

C ONCOMITANT PSYCHOTROPIC Outcome measures MEDICATION All trials allow AD specific medications (cholin- NPS measures may be narrow spectrum, (assessing esterase inhibitors and memantine) with doses depression or agitation, for example), or broad stable for at least one month prior to enrollment. spectrum (Steinberg and Lyketsos, 2008), including Criteria for concomitant psychotropics vary con- the NPI. The NPI originally included 10 domains siderably. Mibamptor and CiTAD excluded most (Cummings et al., 1994) but expanded to 12 by the concomitant psychotropics (except for “rescue” addition of sleep and appetite domains (Cummings medications). Both memantine trials allowed et al., 1997). Domains of NPI are commonly psychotropic medications at baseline (including used in studies with varied degrees of validation. antidepressants, antipsychotics, or anxiolytics), Dennehy et al. (Dennehy et al., 2012) evaluated while the ongoing prazosin trial allows such a cluster of items from the NPI in the hope of medication if the patient is a partial responder. The validating A/A assessment. Based on epidemiology ongoing THC trial allows psychotropic medications and consultation with clinicians, they selected the except tricyclics, fluoxetine or carbamazepine. individual NPI-10 domains of agitation, irritability, disinhibition, and aberrant motor behavior as a Study design four-item measure of agitation and aggression (NPI- 4-A/A). The 4-A/A subscale of NPI-Q was validated R ANDOMIZATION in the Alzheimer’s Disease Neuroimaging Initiative Stratification of randomization varies with some and National Alzheimer’s Coordinating Center trials stratifying by clinical site and others adding cohorts (Trzepacz et al., 2013b). presence of antidepressant treatment (scyllo- One limitation of NPI is that assessment relies inositol), or severity (scyllo-inositol, mibamptor). solely on subjective caregiver input with resultant bias. Recently, the Neuropsychiatric Inventory P HARMACOLOGICAL INTERVENTION Clinician Rating (NPI-C) was developed in part These are fixed-dose or adjustable/titrated-dose to address this issue (de Medeiros et al., 2010). trials. All studies used oral formulations except Aberrant vocalization was added as a new domain, one trial of intramuscular aripiprazole (Rappaport whereas the A/A domain of the NPI was split to et al., 2009). In completed RCTs, the duration of arrive at a total of 14 domains. Unlike NPI, each treatment was 8–12 weeks except for melatonin domain and each item within a domain can be rated (ten days), memantine (24 weeks), or divalproex on the NPI-C. Trained clinicians use input from (24 months). Among ongoing RCTs, the duration caregiver and patient to rate frequency, severity, and of treatment is 12 weeks except for dextrometh- distress of each item. NPI-C was field tested in an orphan (10 weeks), and 3 weeks for THC. international validation study and compared with focused scales to determine convergent validity (de A LLOWED RESCUE MEDICATION Medeiros et al., 2010). It was found to have greater All RCTs allowing “rescue” medication for severe reliability than the conventional NPI. A/A used lorazepam with doses ranging from 0.5 The CMAI (Cohen-Mansfield, 1996) is widely mg to 4 mg daily. Trazodone was used at low doses used particularly in more advanced dementia and (50 mg per day) for sleep disorders (CitAD) and at NH settings. Items are rated on a seven-point higher doses (50–150 mg per day) for severe A/A in scale assessing the frequency of agitated behaviors the memantine trial (Fox et al., 2012). (ranging from “never” to “several times an hour”). was used in the oxcarbazepine trial. Items are presented in four subscales: aggressive behavior; physical non-aggressive behavior; verbally N ON- PHARMACOLOGICAL INTERVENTION agitated behavior; hiding and hoarding. There are In CitAD, mibampator and divalproex trials, versions for both settings, NH (29 items) and a different interventions, at inclusion and regularly non-validated CD (36 items). CMAI provides the over treatment period, have been used but vary richest description of A/A but has the limitations of a in concept, content, and intensity. In CitAD the relatively long administration time and assessments psychosocial intervention during the study was based solely on subjective caregiver input. more intense than mibampator, for example. The The Alzheimer’s Disease Cooperative Study- goal of the CiTAD intervention is to systematize Clinical Global Impression of Change (ADCS- education and support in RCTs (Rosenberg et al., CGIC) is a clinician-rated judgment of change Clinical trials of drugs for agitation in AD 9 from baseline (Schneider et al., 1997) and is Most of trials reported AE and SAV based on based on the Clinical Global Impression of Change caregivers and physicians. AE could be recorded as (CGI-C) (Guy, 1976). The modified ADCS-CGIC spontaneous but also a checklist could be used like (mADCS-CGIC) version targets global functioning in CitAD and mibampator trials. in specific NPS domains is particularly suited to blinded assessments in RCTs. The CitAD and Analytic strategies scyllo-inositol trials have chosen a modified version All statistical analysis plans for these RCTs were of the ADCS-CGIC assessing specific A/A related intention-to-treat (ITT). Primary comparisons domains. were made using: (1) Analysis of covariance The divalproex trial (Tariot et al., 2011) uniquely (ANCOVA) with “last observation carried forward” measured incident A/A as the primary outcome. Of (LOCF) to estimate change from baseline to 18 RCTs, NPI was the primary outcome in ten trials endpoint in the memantine CD trial, or (2) mixed and a secondary outcome in eight. The subscale models (CitAD, mibampator, memantine in NH, NPI-4-A/A domain was primary outcome in the and prazosin). Recently, the general superiority of mibampator trial (Trzepacz et al., 2013a). The mixed models over LOCF was established (Siddiqui ongoing scyllo-inositol trial has chosen as primary et al., 2009), especially for RCTs with longer outcome the NPI-C (combined scores on agitation treatment periods. + aggression domains) and other NPI-C domains Among mixed models, linear mixed models or as secondary outcomes. Three completed and two mixed models of repeated measures (MMRM) are ongoing trials, such as the brexpiprazole trials, used options. The latter treat time as a continuous and are using CMAI as primary outcome. In eight or categorical variable, estimating mean change RCTs CMAI was a secondary outcome. All RCTs from baseline, adjusting for baseline performance. used a global impression of change measure as MMRMs are attractive because they make no secondary outcome except for both prazosin trials assumption about the shape of the outcome’s and the ongoing CitAD where they were co-primary trajectory over time. A recent paper compared outcomes. MMRM to linear models using data from Other outcomes include: Alzheimer’s Disease several trials (Donohue and Aisen, 2012). Neither Cooperative Study Activities of Daily Living approach was more robust to missing data, an (ADCS-ADL) (Galasko et al., 1997)infiveRCTs important issue in AD trials. CitAD and prazosin and seven quality of life (Qol-AD, Logsdon et al., used linear effects models with random intercept 2002) in five RCTs. MMSE is the most widely used and slope in the primary comparison, while cognitive (secondary) outcome. Caregiver burden mibampator used MMRM. was assessed in four completed and three ongoing trials, using the NPI-caregiver distress item (total or specific items); one THC trial uses the Zarit Burden Choice of drug Inventory (Zarit et al., 1980). NPS in AD are thought to reflect one or more types of CNS dysfunction: (1) synaptic or circuit disconnections in specific neuronal S AFETY OUTCOMES networks, (i.e., frontal-subcortical and cortico- Most RCTs of A/A include a typical safety cortical networks); (2) dysfunction in ascend- evaluation: adverse events (AE), serious adverse ing monoaminergic systems involving serotonin, events (SAV), and AEs leading to withdrawal. norepinephrine, or dopamine neurons primarily Specific safety outcomes related to psychotropic located in the brain stem and diffusely projecting medication effects have included cognition (based via long axons to virtually all parts of the on MMSE), falls, sedation, weight change, and QT brain; and (3) glutamate-mediated excitatory prolongation. After the FDA issued an advisory . These CNS alterations are not regarding dose-dependent risk of QT prolongation mutually exclusive and likely synergize to mediate with citalopram, CiTAD increased surveillance NPS. of QT intervals and reported notable QT Scyllo-inositol appears to improve synaptic prolongation with citalopram at 30 mg. Balance, activity in networks underlying NPS via a dual gait, and mobility are targets of assessments for mechanism of action: (1) regulation of brain myo- CitAD, brexpiprazole, and THC. Brexpiprazole inositol metabolism and phosphoinositol signaling, also assesses specific AEs related to AP use such and (2) protection from oligomer-induced as extra-pyramidal symptoms (Simpson-Angus due to amyloid anti-aggregation effects (Townsend Scale), tardive dyskinesia (Abnormal Involuntary et al., 2006). Loss of serotonin in the inferior Movement Scale), and akathisia (Barnes Akathisia frontal cortex was reported to be limited to Rating Scale). AD patients with prominent aggressive behavior 10 M. Soto et al.

(Lai et al., 2003). Citalopram, a selective Discussion serotonin reuptake inhibitor, improves functional serotonorgic neurotransmission. Aggressive beha- We reviewed recent 11 completed and 7 ongoing vior has been associated with up regulation of RCTs of current and emerging medications for α1-adrenoceptors in AD prefrontal cortex and the treatment of A/A in AD. The major result excessive noradrenergic reactivity is associated with of this review is the great heterogeneity found anxiety and agitation in dementia (Szot et al., in the design of these RCTs. Only three (pilot 2007). Prazosin antagonizes norepinephrine at prazosin, intramuscular aripiprazole, and CitAD) postsynaptic α1-adrenoreceptors. report results favoring drug treatment of A/A, Brexpiprazole is chemically similar to aripi- emphasizing the lack of current evidence for a prazole and has broad activity across multiple definitive treatment strategy. Most trials report monoamine systems with reduced partial agonism improvement on placebo likely due to intangible for D2, 5HT1A receptors, and enhanced antagon- benefits of being in RCTs, variability in outcomes, ism for 5-HT2A, and α1-adrenoreceptors. There and variability in the course of A/A (Garre-Olmo is growing interest in modulators of glutamate et al., 2010), with spontaneous remissions or neurotransmission, implicated in many neuro- improvement without treatment. This high rate of psychiatric diseases including schizophrenia and response on placebo decreases the statistical power AD (Blanchard et al., 2004). Dextromethorphan of trials to detect drug effects. Additional challenges hydrobromide (DM) (the active main molecule) include: (1) disentangling treatment effects from modulates glutamate signaling in two ways: (1) competing effects on the outcome in the presence pre-synaptic inhibition of glutamate release (by of psychosocial interventions; (2) varied approaches sigma-1 receptor agonism), and (2) post-synaptic to use of rescue and concomitant psychotropic glutamate response modulation (by weak blockage medications; (3) heterogeneity of target symptoms of NMDA receptor and modulation of NMDA and overlap of NPS (Lyketsos et al., 2011); (4) response to glutamate by the sigma-1 receptor). variable inclusion criteria; and (5) the use of global In the compound, quinidine sulfate increases the AD measures rather than measures specific to bioavailability of DM. DM-quinidine is FDA- A/A or individual A/A symptoms (Drye et al., approved for treatment of pseudo-bulbar affect 2012). whose symptoms overlap partially with A/A in The field does not yet have a tight consensus on AD. Mibampator is an AMPA primary outcome measures, on what constitutes a receptor potentiator (O’Neill, 2004). Activation clinically significant effect size (Mohlar et al., 2009), of AMPA receptors strengthens synapses and or how to translate effect size between different changes in glutamatergic synaptic transmission A/A scales. There are two overall approaches to that contributes to neural plasticity in the central assess treatment response both of which should nervous system (Yamada, 2000). Memantine’s be used: (1) outcomes based on the judgment of mechanism of action is related to glutamatergic experienced clinicians (such as the mADCS-CGIC) synaptic transmission. and/or (2) outcomes measuring the severity of A/A Delta-9- tetrahydrocannabinol (THC) is the symptoms. most biologically active isomer of THC, a The heterogeneous approach toward defining psychoactive compound extracted from the resin caregiver role and assessing caregiver outcomes of Cannabis sativa (marijuana). The ongoing THC is another issue in RCTs of A/A. The field trial is testing a very similar molecule to dronabinol, needs consensus on who is a caregiver (family, synthetic delta-9-THC, indicated for severe formal/professional caregiver, time spent with the nausea and vomiting caused by chemotherapy, patient, etc.). Caregiver opinion is clearly necessary or for anorexia with weight loss in patients in the evaluation of A/A symptoms, which by their with acquired immunodeficiency syndrome. THC nature vary over time, and need to be assessed activates cannabinoid receptors (mainly type 1), by history-taking rather than merely observation in repressing neurotransmitter release in the brain. the clinic. But caregiver opinion is by its nature The rational for its use in NPS is based on its subjective and, therefore, vulnerable to reporting psychoactive effects and association with reduced bias; there is a role for training the caregiver in the pain sensation. Although earlier retrospective data assessment in order to obtain more standardized reported benefits of dronabinol for A/A in severely reports. One important innovation is the NPI- demented patients (Woodward et al., 2014), the C that provides anchors for experienced clinician exact mechanism of action is not known. This could judgments that can overcome caregiver biases and be explained by preventing aggregation of amyloid- allows for input from all available information beta with consequent microglial activation (Ramirez including the clinical record, caregiver input, and et al., 2005). direct observation. Clinical trials of drugs for agitation in AD 11

Table 3. Recommendations for future RCTs targeting A/A in patients with AD

METHODOLOGICAL ASPECT RECOMMENDATIONS ...... Population studied Age • No limit Dementia severity • Mild to severe based on CDR rating of 1–3; stratification Settings • Different RCTs for NH or CD preferred; or stratification Clinically significant • A/A needs consensus criteria A/A • “Clinically significant” = medication is needed based on judgment of experienced clinician combined with severity rating above a cut-off on a A/A scale Concomitant • “AD treatments” allowed on stable doses for 30–60 days medications • APs not allowed; or allowed stable doses for 30–60 days • Antidepressants, mood stabilizers, anticonvulsants: allowed on stable doses for 30–60 days Caregiver participation • Caregiver needs a consensus definition • Standardized training in recognizing NPS and in rating behavior scales • Use of a caregiver diary for real time observations Study design Pharmacological • Run-in-period before randomization (2–4 weeks) intervention • 8–12-week treatment period • Consolidation response: to assess time to relapse within responders in each group during a 6–12-month period Non-pharmacological • Psychosocial intervention during the run-in and the treatment periods in both groups. intervention • Etiologic, non-pharmacologic, person-centered approach during run-in and treatment periods in both groups Allowed rescue • Defined allowable dosing, monitored use medication Outcome measures Primary • Global measure of A/A as primary • Validated scales assessing A/A, co-primary or secondary • Rated by clinicians with patient and caregiver input Secondary • Consider actigraphy • Agitation symptoms • Aggression symptoms • Other NPS: irritability, anxiety, depression, psychosis • Cognition, functional ability, quality of life • Caregiver distress, other caregiver measures • Allowed rescue medication cumulative dose Analytic strategies • Intention to treat analysis • Mixed models: LMM or MMRM

Abbreviations: AD = Alzheimer’s disease; NPS = neuropsychiatric symptoms; A/A = agitation/aggression; CD = community dwelling; NH = nursing home; CDR = clinical dementia rating; MMRM: mixed model of repeated measures; linear mixed models.

Methodological enhancements of recent features of AD. Therefore, disease severity should RCTs: suggestions for improvement be based on CDR. We propose in Table 3 methodological considera- In recent trials there is a tendency to include tions for future trials for A/A in PwAD. patients at all levels of disease severity and, thus, patients at milder stages. This is important, given that the pathological processes leading to AD begins Population studied decades before a clinical diagnosis, and drugs D ISEASE SEVERITY targeting NPS may need to be assessed earlier in The THC trial is the first trial to use the CDR the disease process. to measure dementia severity for inclusion. The MMSE’s major limitation is that the estimate S ETTING of disease severity is based solely on cognitive The majority of trials have included patients from performance, and does not capture, as CDR does, CD or from NH settings and only two (CitAD cognitive and physical impairments which are key and dextromethorphan trials) have patients from 12 M. Soto et al. both. It is important for generalizability to include to be short due to the acuity of symptoms (see participants from both settings, either in separate Table 3). trials or in a single trial; the latter strategy might be best implemented by stratifying randomization P HARMACOLOGICAL INTERVENTION by care setting (as is being done in CiTAD, for Most completed and current RCTs administer drug example). Additionally, it is important to validate treatment for 9 to 12 weeks, adequate for assessing methods of caregiver data collection since most acute response, but not stability of response over caregivers in CD are family members and most in longer periods of time. Some design strategies partly NH are paid staff who may have different agendas address this issue. The prazosin trial has included and different reporting bias. It is important not to a 12-week open label observation period after the exclude NH participants since the majority of A/A strict RCT period. There is an ongoing safety in advanced dementia is in the NH setting. extension RCT to evaluate persistence of the effects of scyllo-inositol on A/A, beyond the treatment C ONCOMITANT PSYCHOTROPIC period covered by the trial. All patients will receive MEDICATIONS scyllo-inositol, but masking from the first trial will Most patients with A/A will already be taking be maintained. Another strategy could be a two- psychotropic medications, and the approach toward week washout after the RCT period to assess if there these needs to find a middle ground between is a rebound in symptoms. maximal rigor (no concomitant medications) Open-label observation under treatment drug and maximal generalizality (allow concomitant it could be interesting if the period was long medications and assess intervention as an “add- enough to assess “consolidation” of response, on” therapy). Many RCTs have chosen the former preferably 6 to 12 months for maximizing clinical approach to assess efficacy of an intervention as generalizability. However, having no comparator, monotherapy, with the trade offs of rendering such a methodology will not inform efficacy. recruitment more challenging, possibly inadvert- In order to address consolidation efficacy the ently excluding the most severe A/A, and limiting observation period should be long enough (six6 generalizability. The alternative approach could be months to one year) and the analysis should proposed for these reasons. The ongoing prazosin measure time. It will also be interesting to assess trial follows this approach, allowing all stable time to remission or relapse, as is done in evaluation concomitant psychotropic medications at inclusion, of anticonvulsants, for example. if the patient is a partial responder. However, doses of concomitant psychotropic should remain N ON- PHARMACOLOGIC INTERVENTION stable over the treatment period. Regarding APs The citalopram and mibampator trials include medication, since they are the only one having a psychosocial intervention for all participants proved efficacy, even modest, their concomitant use regardless of treatment assignment, and this it is questionable and should be address by an expert design feature is important to provide ethically panel (see Table 3). “enhanced usual care” to all participants, to improve recruitment and retention, (Drye et al., Study design 2012) and, to reduce variance in outcome reporting by caregivers via education about the features of the P LACEBO RUN- IN disease (Tariot et al., 2011). One design feature to diminish the effect of placebo response is to have a placebo run- in period, which is becoming the norm in A LLOWED RESCUE MEDICATION other fields of psychopharmacology (Iovieno and A/A can truly be a crisis for family or institutional Papakostas, 2012), accounting for the benefit of caregivers, and rescue medication is often needed to non-pharmacologic interventions, and improving have viable retention. The CiTAD trial is a model, statistical power (Frost et al., 2008). None of monitoring the use and dose of rescue medication these trials included a placebo run-in period before by treatment group (Drye et al., 2012). randomization. Two trials in related fields used two to four week placebo run-ins and reported relatively Outcome measures low on-placebo response rates in the randomized The choice of the optimal primary outcome is treatment period (Howard et al., 2007, Cummings probably the most important decision in study et al., 2013). We suggest that for RCTs of A/A design in this field. As this review shows, the choices the period be kept to two weeks because this is a have been very heterogeneous in completed and reasonable period of time for relying solely on non- ongoing RCTs due to the lack of a gold standard pharmacologic interventions, the duration needs outcome measure. However, the field appears to be Clinical trials of drugs for agitation in AD 13 reaching a consensus in using both agitation-specific and thus with a high quality, differences in the quantitative measures (e.g., relevant domains of methodological quality were not deeply analyzed. NPI-C or CMAI) plus a global rating of change Second, comparisons between RCTs’ reports inter- for agitation outcomes (mADCS-CGIC). This rater reliability were not available. Third, only combined approach has been successfully used in publications in English were included. the CitAD trial. Although the pimavanserin trial is Finally, in comparison to pre-2008 RCTs, it not included in this review (targeting psychosis in is notable that other potential pharmacological Parkinsons’ disease) the assessment process is worth alternatives to atypical APs, have been and are noting: central raters assessed the primary outcome, being tested in recent RCTs. However, despite site-based raters assessed CGI, and caregivers considerable efforts in crafting appropriate designs assessed the Zarit burden scale (Cummings et al., for RCTs of promising therapeutics agents for A/A, 2013). Assessing different outcomes with different it is urgent to gain more clarity regarding the raters may provide robust support for convergent underlying neural regions and circuitry involved clinical benefits. Actigraphy offers the possibility of in NPS, to thereby shed light on symptom objective measurement of activity associated with pathogenesis. However, so far, there has been agitation and has reasonable validation versus NPI very little developmental work based on this better in two studies (Mahlberg et al., 2007; Kirste et al., understanding that could offer an opportunity to 2014). develop more new targeted drug treatments (Ballard et al., 2013;Gedaet al., 2013). Analytical strategies An issue to take into account in statistical analyses is measurement variance that may be exaggerated Conclusion in some trials especially multinational studies Despite the urgent need to identify effective where translations and cultural interpretations pharmacological treatments for A/A in PwAD, are required. In most recent RCTs (citalopram, progress has been slow. In the past six years a mibampator, memantine (Fox et al., 2012)trials) small number of RCTs of drugs for treatment mixed models have been estimated. Currently, of A/A in AD have been conducted with mixed models seem to be more appropriated than disappointing results. These trials are characterized ANCOVA and LOCF in RCTs of drugs in AD. by methodological heterogeneity. Several issues However, up to date there is no available data to have been encountered: the need for stronger better propose MMRM or linear mixed models as consensus on the syndromal definition of A/A as a statistical strategy in RCTs of drugs for A/A in AD. target, choice of primary efficacy outcome measure, There are several important methodologic the content and timing of the non-pharmacological obstacles to drug development in this area that intervention in placebo and drug arms before might be best amenable to expert consensus on and/or during the trial, concomitant psychotropic the: (1) definition of A/A for inclusion; (2) choice medication, definition of caregivers and their of primary outcome; (3) role of adaptive designs participation. Consensus is necessary to enhance to minimize exposure to a drug or placebo that is the design of future trials. The fact that placebo “not working” (Mugno et al., 2004; Kairalla et al., effects are substantial and consistently observed 2012); (4) standardization of non-pharmacological suggests that non-pharmacological approaches are person-centered interventions (Gitlin et al., 2012) currently the standard of care that we are still in both arms; and (5) placebo run-in-period prior waiting for clear and consistent evidence on drug to randomization. Once eligibility is confirmed at efficacy. baseline, only patients with persisting A/A or at higher levels of A/A severity would participate in the treatment period. This etiologic non- pharmacological centered person approach, in Conflict of interest addition to the psychosocial intervention, should be prolonged over the treatment period. This approach Dr Soto has served as consultant for Ethypharm. will reflect real-life clinical practice, reduce placebo Dr. Rosenberg has received research support from response rate, and address the real-life issue which National Institute on Aging (1K08AG029157- is whether a drug can demonstrate superior efficacy 01; A15R01AG039384-03), American Federation to placebo in the presence of non-pharmacological for Aging Research, Lilly, Functional Neur- intervention for all participants. omodulation, and Pfizer, and has served as a Our review has limitations. First, even all consultant/advisor to Janssen, Pfizer, Lundbeck, studies were randomized and controlled trials and Abbvie. 14 M. Soto et al.

Description of authors’ roles Gornbein, J. (1994). The neuropsychiatric inventory: comprehensive assessment of psychopathology in dementia. M. Soto, B. Vellas, C. Lyketsos, and P. Rosenberg Neurology, 44, 2308–2314. designed the manuscript. M. Soto wrote the Cummings, J. L. et al. (2013). Pimavanserin for patients manuscript. C. Lyketsos and P. Rosenberg assisted with Parkinson’s disease psychosis: a randomised, with writing the article. B. Vellas, C. Ballard, S. placebo-controlled phase 3 trial. Lancet, 383, 533–540. Andrieu, F. Nourhashemi, P. Robert, and PJ Ousset de Deyn, P. P. et al. (1999). A randomized trial of revised and corrected the manuscript. risperidone, placebo, and haloperidol for behavioral symptoms of dementia. Neurology, 53, 946–955. de Medeiros, K. et al. (2010). The neuropsychiatric inventory-clinician rating scale (NPI-C): reliability and Acknowledgment validity of a revised assessment of neuropsychiatric This work was supported by NIH grant symptoms in dementia. International Psychogeriatrics, 22, P50AG005146 to the Johns Hopkins Alzheimer’s 984–994. Dennehy, E. B., Kahle-Wrobleski, K., Sarsour, K. and Disease Research. Milton, D. R. (2012). Derivation of a brief measure of agitation and aggression in Alzheimer’s disease. International Journal of Geriatric Psychiatry, 28, 182–189. References Deudon, A. et al. (2009). Non-pharmacological management of behavioural symptoms in nursing homes. International Ballard, C., Francis, P. and Corbett, A. (2013). Journal of Geriatric Psychiatry, 24, 1386–1395. Randomized controlled trial of mibampator for behavioral Donohue, M. C. and Ainsen, P. S. (2012). Mixed model of and psychological symptoms of dementia: comments on the repeated measures versus slope models in Alzheimer’s trial and thoughts for future studies. International disease clinical trials. Journal of Nutrition Health and Aging, Psychogeriatrics, 25, 687–689. 16, 360–364. Ballard, C. and Howard, R. (2006). Neuroleptic drugs in Drye, L. T. et al. (2012). CitAD Research Group. dementia: benefits and harm. Nature Reviews Neuroscience, Citalopram for agitation in Alzheimer’s disease: design and 7, 492–500. methods. Alzheimer’s and Dementia, 8, 121–130. Ballard, C. et al. (2009). Management of agitation and Folstein, M. F., Folstein, S. E. and McHugh, P. R. aggression associated with Alzheimer’s disease. Nature (1975). “Mini-mental state”: a practical method for Reviews Neurology, 2, 245–255. grading the cognitive state of patients for the clinician. Benoit, M. et al. (2006). Professional consensus on the Journal of Psychiatry Research, 12, 196–198. treatment of agitation, aggressive behaviour, oppositional Fox, C. et al. (2012). Efficacy of memantine for agitation in behaviour and psychotic disturbances in dementia. Journal Alzheimer’s dementia: a randomised double-blind placebo of Nutrition Health and Aging, 10, 410–415. controlled trial. PLoS One, 7, e35185. Blanchard, B. J., Chen, A., Rozeboom, L. M., Stafford, Frost, C., Michael, G., Kenward, I. and Fox, C. (2008). K. A., Weigele, P. and Ingram, V. M. (2004). Efficient Optimizing the design of clinical trials where the outcome is reversal of Alzheimer’s disease fibril formation and a rate. Can estimating a baseline rate in a run-in period elimination of neurotoxicity by a small molecule. increase efficiency? Statistical Medicine, 27, 3717–3731. Proceedings of the National Academy of Sciences USA, 101, Galasko, D. et al. (1997). An inventory to assess activities of 14326–14332. daily living for clinical trials in Alzheimer’s disease. The Brodaty, H. and Arasaratnam, C. (2012). Meta-analysis of Alzheimer’s disease cooperative study. Alzheimer Disease nonpharmacological interventions for neuropsychiatric and Associated Disorders, 11, S33–S39. symptoms of dementia. American Journal of Psychiatry, 169, Garre-Olmo, J., López-Pousa, S., Vilalta-Franch, J., de 946–953. Gracia Blanco, M. and Vilarrasa, A. B. (2010). Canevelli, M. et al. (2013). Behavioral and psychological Grouping and trajectories of neuropsychiatric symptoms in subsyndromes in Alzheimer’s disease using the patients with Alzheimer’s disease. Part II: two-year patient neuropsychiatric inventory. International Journal of Geriatric trajectories. Journal of Alzheimers Disease, 22, 1169–1180. Psychiatry, 28, 795–803. Gauthier, S. et al. (2010). Management of behavioral Cohen-Mansfield, J. (1996). Conceptualization of agitation: problems in Alzheimer’s disease. International results based on the Cohen-Mansfield agitation inventory Psychogeriatrics, 22, 346–372. and the agitation behavior mapping instrument. Geda, Y. E. et al. (2013). Neuropsychiatric syndromes International Psychogeriatrics, 8, 309–315. professional interest area of ISTAART. Neuropsychiatric Cohen-Mansfield, J., Werner, P., Watson, V. and Pasis, symptoms in Alzheimer’s disease: past progress and S. (1995). Agitation among elderly persons at adult anticipation of the future. Alzheimer’s and Dementia,9, day-care centers: the experiences of relatives and staff 602–608. members. International Psychogeriatrics, 7, 447–458. Gehrman, P. R., Connor, D. J., Martin, J. L., Shochat, Cummings, J. L. (1997). The Neuropsychiatric Inventory: T., Corey-Bloom, J. and Ancoli-Israel, S. (2009). assessing psychopathology in dementia patients. Neurology, Melatonin fails to improve sleep or agitation in 48, S10–S16. double-blind randomized placebo-controlled trial of Cummings, J. L., Mega, M., Gray, K., institutionalized patients with Alzheimer disease. American Rosenberg-Thompson, S., Carusi, D. A. and Journal of Geriatric Psychiatry, 17, 166–169. Clinical trials of drugs for agitation in AD 15

Gitlin,L.N.,Kales,C.K.andLyketsos,C.G.(2012). Lonergan, E., Luxenberg, J., Colford, J. M. and Birks, J. Nonpharmacologic management of behavioral symptoms in (2002). Haloperidol for agitation in dementia. Cochrane dementia. Journal of the American Medical Association, 308, Database of Systematic Reviews, Issue 2, CD002852. 2020–2029. Lyketsos, C. G., Steinberg, M., Tschanz, J. T., Norton, Gonfrier, S., Andrieu, S., Renaud, D., Vellas, B. and M. C., Steffens, D. C. and Breitner, J. C. (2000). Robert, P. H. (2012). Course of neuropsychiatric Mental and behavioral disturbances in dementia: findings symptoms during a 4-year follow up in the REAL-FR from the Cache County Study on in Aging. cohort. Journal of Nutrition Health and Aging, 16, American Journal of Geriatric Psychiatry, 157, 708–714. 134–137. Lyketsos, C. G. et al. (2006). Task Force of American Guy, W. (1976). ECDEU Assessment Manual for Association for Geriatric Psychiatry. Position statement of Psychopharmacology, Revised. Bethesda, MD: US the American Association for Geriatric Psychiatry regarding Department of Health, Education, and Welfare. principles of care for patients with dementia resulting from Herrmann, N., Gauthier, S., Boneva, N. and Lemming, Alzheimer disease. American Journal of Geriatric Psychiatry, O. M. (2013). A randomized, double-blind, 14, 561–572. placebo-controlled trial of memantine in a behaviorally Lyketsos, C. G. et al. (2011). Neuropsychiatric symptoms in enriched sample of patients with moderate-to-severe Alzheimer’s disease. Alzheimer’s and Dementia, 7, 532–539. Alzheimer’s disease. International Psychogeriatrics, 25, Maher, A. R. et al. (2011). Efficacy and comparative 919–927. effectiveness of atypical antipsychotic medications for Howard, R. J. et al. (2007). Donepezil for the treatment of off-label uses in adults: a systematic review and agitation in Alzheimer’s disease. New England Journal of meta-analysis. JAMA, 28, 306, 1359–1369. Medicine, 357, 1382–1392. Mahlberg, R., Walther, S., Eichmann, U., Tracik, F. Husebo, B. S., Ballard, C., Sandvik, R., Nilsen, O. B. and Kunz, D. (2007). Effects of rivastigmine on and Aarsland, D. (2011). Efficacy of treating pain to actigraphically monitored motor activity in severe agitation reduce behavioural disturbances in residents of nursing related to Alzheimer’s disease: a placebo-controlled pilot homes with dementia: cluster randomised clinical trial. study. Archives of Gerontolology Geriatric, 45, 19–26. BMJ, 343, d4065. McKhann, G. et al. (1984). Clinical diagnosis of Alzheimer’s Iovieno, N. and Papakostas, G. I. (2012). Does the disease: report of the NINCDS-ADRDA Work Group presence of an open-label antidepressant treatment period under the auspices of Department of Health and Human influence study outcome in clinical trials examining Services Task Force on Alzheimer’s disease. Neurology, 34, augmentation/combination strategies in treatment partial 939–944. responders/nonresponders with major depressive disorder? McKhann, G. M. et al. (2011). The diagnosis of dementia Journal of Clinical Psychiatry, 73, 676–683. due to Alzheimer’s disease: recommendations from the Kairalla, J. A., Coffey, C. S., Thomann, M. A. and National Institute on Aging Alzheimer’s Association Muller, K. E. (2012). Adaptive trial designs: a review of workgroups on diagnostic guidelines for Alzheimer’s barriers and opportunities. Trials, 13, 145. disease. Alzheimer’s and Dementia, 7, 263–269. Kales,H.C.,Gitlin,L.N.andLyketsos,C.(2014) (For Molnar, F., Man-Son-Hing, M. and Dean Fergusson, D. the Detroit Expert Panel on the Assessment and (2009). Systematic review of measures of clinical Management of Neuropsychiatric Symptoms of Dementia). significance employed in randomized controlled trials of Management of neuropsychiatric symptoms of dementia in drugs for dementia. Journal of the American Geriatrics clinical settings: recommendations from a multidisciplinary Society, 57, 536–546. expert panel. Journal of the American Geriatrics Society, 62, Mowla, A. and Pani, A. (2010). Comparison of topiramate 762–769. and risperidone for the treatment of behavioral Kales, H. C. et al. (2012). Risk of mortality among individual disturbances of patients with Alzheimer disease. A antipsychotics in patients with dementia. American Journal double-blind, randomized clinical trial. Journal of Clinical of Psychiatry, 169, 71–79. Psychopharmacology, 30, 40–43. Kirste, T. et al. (2014). Detecting the effect of Alzheimer’s Mugno, R., Zhus, W. and Rosenberger, W. F. (2004). disease on everyday motion behavior. Journal of Alzheimers Adaptive urn designs for estimating several percentiles of a Disease, 38, 121–132. dose–response curve. Statistical Medicine, 23, 2137– Lai, M. K. et al. (2003). Reduced serotonin 5-HT1A receptor 2150. binding in the temporal cortex correlates with aggressive Murman, D. L., Chen, Q., Powell, M. C., Kuo, S. B., behavior in Alzheimer disease. Brain Research, 974, Bradley, C. J. and Colenda, C. C. (2002). The 82–87. incremental direct costs associated with behavioral Langballe, E. M., Engdahl, B., Nordeng, H., Ballard, C., symptoms in AD. Neurology, 59, 1721–1729. Aarsland, D. and Selbæk, G. (2013). Short- and Okura, T. and Langa, K. M. (2011). Caregiver burden and long-term mortality risk associated with the use of neuropsychiatric symptoms in older adults with cognitive antipsychotics among 26,940 dementia outpatients: a impairment: the Aging, Demographics, and Memory Study population-based study. American Journal of Geriatric (ADAMS). Alzheimer Disease and Associated Disorders, 25, Psychiatry, 22, 321–331. 116–121. Logsdon, R. G., Gibbons, L. E., McCurry, S. M. and Okura, T., Plassman, B. L., Steffens, D. C., Llewellyn, Teri, L. (2002). Assessing quality of life in older adults D. J., Potter, G. G. and Langa, K. M. (2010). with cognitive impairment. Psychosomatic Medicine, 64, Prevalence of neuropsychiatric symptoms and their 510–519. association with functional limitations in older adults in the 16 M. Soto et al.

United States: the aging, demographics, and memory Schneider, L. S. et al. (1997). Validity and reliability of the study. Journal of the American Geriatrics Society, 58, 330– Alzheimer’s Disease Cooperative Study-clinical global 337. impression of change. The Alzheimer’s Disease Okura, T., Plassman, B. L., Steffens, D. C., Llewellyn, Cooperative Study. Alzheimer Disease and Associated D. J., Potter, G. G. and Langa, K. M. (2011). Disorders, 11, S22–S32. Neuropsychiatric symptoms and the risk of Selbæk, G., Høgset, L. D., Söderhamn, U. and institutionalization and death: the aging, demographics, Kirkevold, O. (2013). The course of neuropsychiatric and memory study. Journal of the American Geriatrics symptoms in nursing-home patients with dementia over a Society, 59, 473–481. 53-month follow-up period. International Psychogeriatrics, O’Neill, M. J. (2004). SMi 4th Annual Conference on 20, 1–9. Neurodegenerative Disorders: a focus on Alzheimer’s and Siddiqui, O., hung, M. J. and O’neill, R. (2009). MMRM Parkinson’s disease. Expert Opinion on Investigational Drugs, versus LOCF: a comprehensive comparison based on 13, 1369–1373. simulation study and 25 NDA datasets. Journal of Porsteinsson, A. P. et al. (2014). Effect of citalopram on Biopharmaceutical Statistics, 19, 227–246. agitation in Alzheimer’s disease–the CitAD randomized Sommer, O. H., Aga, O., Cvancarova, M., Olsen, I. C., controlled trial. Journal of the American Medical Association, Selbaek, G. and Engedal, K. (2009). Effect of 311, 682–691. oxcarbazepine in the treatment of agitation and aggression Prince, M., Bryce, R., Albanese, E., Wimo, A., Ribeiro, in severe dementia. Dementia and Geriatric Cognitive W. and Cleusa, P. (2013). The global prevalence of Disorders, 27, 155–163. dementia: a systematic review and meta-analysis. Soto, M. E. et al. (2012). Improving care of older adults with Alzheimer’s and Dementia, 9, 63–75. dementia: description of 6299 hospitalizations over 11 Rabins, P. V. et al. (2007). APA work group on Alzheimer’s years in a special acute care unit. Journal of the American disease and other dementias; steering committee on Medical Directors Association, 13, 486,.e1–6. practice guidelines. American Psychiatric Association Steinberg, M. and Lyketsos, C. G. (2008). practice guideline for the treatment of patients with Neuropsychiatric measures for cognitive disorders. In Alzheimer’s disease and other dementias: second edition. A. J. Rush, M. B. First and D. Blacker (eds.), Handbook of American Journal of Psychiatry, 164(S), 5–56. Psychiatric Measures (p. 864). Arlington, VA: American Ramírez, B. G. et al. (2005). Prevention of Alzheimer’s Psychiatric Publishing. disease pathology by cannabinoids: neuro-protection Steinberg, M. et al. (Cache County Investigators). (2008). mediated by blockade of microglial activation. Journal of Point and 5-year period prevalence of neuropsychiatric Neurosciences, 25, 1904–1913. symptoms in dementia: the Cache County Study. Rappaport, S. A., Marcus, R. N., Manos, G., McQuade, International Journal of Geriatric Psychiatry, 23, 170– R. D. and Oren, D. A. (2009). A randomized, 177. double-blind, placebo-controlled tolerability study of Szot, P., White, S. S., Greenup, J. L., Leverenz, J. B., intramuscular aripiprazole in acutely agitated patients with Peskind, E. R. and Raskind, M. A. (2007). Changes in Alzheimer’s, vascular, or mixed dementia. Journal of the adrenoreceptors in the prefrontal cortex of subjects with American Medical Directors Association, 10, 21–27. dementia: evidence of compensatory changes. Neuroscience, Richter, T. et al. (2012). Psychosocial interventions for 146, 471–480. reducing antipsychotic medication in care home residents. Tariot, P. et al. (2011). Chronic divalproex sodium to The Cochrane Database of Systemetic Reviews, 12, attenuate agitation and clinical progression of Alzheimer CD008634. disease. Archives of General Psychiatry, 68, 853–886. Rosen, W. G., Mohs, R. C. and Davis, K. L. (1984). A Teri, L. et al. (2000). Treatment of agitation in AD: a new rating scale for Alzheimer’s disease. American Journal randomized, placebo-controlled clinical trial. Neurology 55, of Psychiatry, 141: 1356–1364. 1271–1278. Rosenberg, P. B. et al. (2010). Sertraline for the treatment Townsend, M. et al. (2006). Orally available compound of depression in Alzheimer disease: DIADS-2 Research prevents deficits in memory caused by the Alzheimer Group. American Journal of Geriatric Psychiatry, 18, amyloid-beta oligomers. Annals of Neurolology, 60, 136–145. 668–676. Salzman, C. et al. (2008). Elderly patients with Trzepacz,P.T.et al. (2013a). Mibampator (LY451395) dementia-related symptoms of severe agitation and randomized clinical trial for agitation/aggression in aggression: consensus statement on treatment options, Alzheimer’s disease. International Psychogeriatrics, 25, clinical trials methodology, and policy. Journal of Clinical 707–719. Psychiatry, 69, 889–898. Trzepacz,P.T.et al. (2013b) Alzheimer’s disease Schneider, L. S., Dagerman, K. and Insel, P. (2006). neuroimaging initiative. Subscale validation of the Efficacy and adverse effects of atypical antipsychotics for neuropsychiatric inventory questionnaire: comparison of dementia: meta-analysis of randomized, placebo-controlled Alzheimer’s disease neuroimaging initiative and national trials. American Journal of Geriatric Psychiatry, 14, 191– Alzheimer’s coordinating center cohorts. American Journal 210. of Geriatric Psychiatry, 21, 607–622. Schneider, L. S., Pollock, V. E. and Lyness, S. A. (1990). Wang, L. Y. et al. (2009). Prazosin for the treatment of A meta-analysis of controlled trials of neuroleptic treatment behavioral symptoms in patients with Alzheimer disease in dementia. Journal of the American Geriatrics Society, 38, with agitation and aggression. American Journal of Geriatric 553–563. Psychiatry, 17, 744–751. Clinical trials of drugs for agitation in AD 17

Woodward, M. R., Harper, D. G., Stolyar, A., Yamada,K.A.(2000). Therapeutic potential of positive Forester, B. P. and Ellison, J. M. (2014). AMPA receptor modulators in the treatment of Dronabinol for the treatment of agitation and aggressive neurological disease. Expert Opinion on Investigational behavior in acutely hospitalized severely demented Drugs, 9, 765–778. patients with noncognitive behavioral symptoms. Zarit, S. H., Reever, K. E. and Bach-Peterson, J. (1980). American Journal of Geriatric Psychiatry, 22, Relatives of the impaired elderly: correlates of feelings of 415–419. burden. Gerontologist, 20, 649–655.