A pragmatic, multicentre, double-blind, placebo-controlled randomised trial to assess the safety, clinical and cost-effectiveness of mirtazapine and carbamazepine in people with Alzheimer's disease and agitated behaviours: the HTA-SYMBAD trial

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 13/115/76. The contractual start date was in December 2015. The draft report began editorial review in July 2021 and was accepted for publication in September 2022. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2023 Banerjee et al. This work was produced by Banerjee et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2023 Banerjee et al.

Scientific background

Dementia is one of the most common and serious public health issues of our time. 1 Over 46 million people have dementia worldwide, a figure set to double in the next 20 years. 2 The commonest cause of dementia is Alzheimer’s disease (AD), which causes irreversible and progressive decline in memory, reasoning, communication skills and the ability to carry out daily activities. Alongside this cognitive and functional decline, individuals may develop neuropsychiatric symptoms (NPS) such as agitation, sleep disturbance, depression and psychosis. 3 These are common, occurring in up to 90% of people with dementia, with agitation as one of the most persistent symptoms. 4 Agitation is defined as inappropriate verbal, vocal or motor activity that is not thought to be caused by unmet need; it encompasses physical and verbal aggression and is particularly problematic. 5 It affects nearly half of people with AD over a month6 and 80% of those with clinically significant symptoms will have them 6 months later. 7 Agitation is associated with deteriorating relationships with family and professional carers, care home admission, increased costs of care, carer burden and burnout and decreased quality of life. 5,7,8

Agitation in dementia has substantial economic consequences, accounting for between 12% and 44% of dementia care costs annually9,10 and imposing significant costs on unpaid carers. 11 Costs rise as the severity of agitation increases. 10,12 The annual excess health and social care cost of agitation in AD in the UK has been estimated at £2B. 9

Agitation in dementia is therefore a legitimate target for therapeutic intervention, but it is a symptom with a number of possible causes, including pain, physical or psychological distress, misperception of threat (e.g. when receiving personal care), and response to hallucinations or delusions. Using non-pharmacological interventions that investigate aetiology and provide a tailored response as a first-line treatment for agitation in dementia, such as the DICE approach (Describe the problem, Investigate the cause, Create a plan, Evaluate its effectiveness), is recommended as best practice. 1,13 However, given the clinical significance of agitation, there is a need for second-line treatments when no underlying causes are found or when correction of these has not resulted in improvement. The mainstay of drug treatment is the use of antipsychotic medication. These drugs, however, have low efficacy, with the American Psychiatric Association guideline group reporting that they ‘demonstrate minimal or no efficacy with strong placebo effects’14 and have been shown to cause particular harms in those with dementia, including excess dementia-specific mortality. In 2009, around 180,000 people with dementia were prescribed antipsychotic medication across the UK per year and this equated to an additional 1800 deaths and an additional 1620 cerebrovascular adverse events (AEs) attributable to the use of antipsychotics in dementia. 15 While their rate of prescription to people with dementia has decreased,16 they are still commonly used and such treatment is largely unlicensed. In most countries, few or no treatments have been given regulatory approval for such use. In the UK, the only drugs with a relevant license are risperidone and haloperidol and these are highly restrictive. Risperidone is indicated for the ‘short-term treatment (up to 6 weeks) of persistent aggression in patients with moderate to severe Alzheimer’s dementia unresponsive to non-pharmacological approaches and when there is a risk of harm to self or others’ and haloperidol for ‘persistent aggression and psychotic symptoms in moderate to severe Alzheimer’s dementia and vascular dementia [when non-pharmacological treatment is ineffective and there is a risk of harm to self or others]’.

Other drug treatments have been suggested for agitation in dementia but trials of antidementia medication, the acetylcholinesterase donepezil17 and the N-Methyl-D-Aspartate (NMDA) inhibitor memantine,18 have been tested in randomised controlled trials (RCTs) and not demonstrated efficacy. In a large multicentre trial, the anticonvulsant sodium valproate did not delay or prevent NPS in dementia. 19 Benzodiazepines are used short-term clinically but there are no trials and adverse effects such as falls are common and of concern. 20 Antidepressants have also been investigated as an alternative to antipsychotics. The CitAD trial of citalopram for agitated behaviours provided evidence that a target dose of citalopram 30 mg per day had a small positive effect on agitation in dementia21 in those who were less agitated and less cognitively impaired. 22 Adverse cardiac and cognitive effects identified in the trial limit its use in clinical practice. 21 Antidepressants are not mentioned as a potential treatment for agitation in the English National Institute for Health and Care Excellence (NICE) guideline on dementia assessment and management23 but they are increasingly used as a treatment of agitation in dementia. This substitution strategy to seek to avoid the prescription of antipsychotics has been reported in a large study of US nursing homes where the prescription rates of mood stabilisers such as sodium valproate, carbamazepine and particularly gabapentin increased as those for antipsychotics decreased. 24,25 Such prescribing of antidepressants is part of the common polypharmacy seen among people with dementia in the community. 26

Mirtazapine for agitated behaviours in dementia

Mirtazapine, a noradrenergic and specific serotonergic antidepressant (NASSA), is widely used in older people; from 2009 to 2014, in a study of 4.8 million antidepressant initiations in Europe, it was the antidepressant that was most commonly prescribed for older people and also to those with dementia. 27 A centrally active presynaptic α₂-antagonist, it stimulates both noradrenergic and serotonergic systems mediated via 5-HT1 receptors, with 5-HT2 and 5-HT3 receptors blocked by mirtazapine. Histamine H1-antagonistic activity is thought to cause its sedative properties. It has little anticholinergic activity, unlike citalopram, and, at therapeutic doses, has few cardiovascular effects. Mirtazapine is a relatively potent antagonist/inverse agonist at key receptors likely to be pivotal in target symptoms including antagonism of α₂-adrenergic, 5-HT1A and histamine H1 receptors. The overall effects are to increase noradrenergic and serotonergic neurotransmission which may explain its use in depression while the H1 antagonism is associated with useful acute sedative benefits. The pharmacological profile of mirtazapine is such that at higher dosages, the sedative effect decreases due to noradrenaline stimulation. It is a well-established treatment for depression and is well tolerated by older people and so a popular choice by psychiatrists which will encourage recruitment. It is available generically at low cost in the NHS. Cost implications, were it found to be effective, would be therefore minimal.

In pre-specified secondary analyses of the HTA-SADD data set, reported in the HTA-SADD final report,28 we have found a positive effect of mirtazapine on decreasing Behavioural and Psychological Symptoms in Dementia (BPSD) [as measured by Neuropsychiatric Inventory (NPI) score] at 13 weeks. Taking the top 50% of raw NPI scores (i.e. those with appreciable BPSD), there was a 7.1-point difference in NPI score [95% confidence interval (CI) −0.50 to 14.68; p = 0.067] between mirtazapine and placebo and a 13.2-point difference between mirtazapine and sertraline (95% CI 4.47 to 21.95; p = 0.003). An additional surprising but encouraging positive finding was from the cost-effectiveness analyses. Over the course of the trial, the time spent by unpaid carers caring for participants in the mirtazapine group was almost half that for patients in the placebo group (6.74 vs. 12.27 hours per week) and sertraline group (6.74 vs. 12.32 hours per week). Informal care costs were £1510 (95% CI −3088 to −136) and £1522 (95% CI −3398 to −72) less for the mirtazapine-treated group when compared with placebo and sertraline respectively. In the secondary outcome evaluation, looking at quality-of-life gains and costs, treatment with mirtazapine had a high likelihood of cost-effectiveness compared to placebo or sertraline. 28,29 The improvements in quality of life for mirtazapine relative to the other treatments contributed to the cost-effectiveness result, and there is a plausibility that comes from the putative ability of mirtazapine to ameliorate sleep disturbances and anxiety. 30,31 Improvements in sleep could potentially improve life quality and therefore patient-reported EuroQol-5 Dimension (EQ-5D) scores; they could also release carer time directly and also ameliorate an important source of carer distress. 32 In this way, mirtazapine might have a general effect, beneficial for both the patient and the carer, even without exerting an antidepressant effect. Two small-scale open-label pilot studies give supportive evidence for the potential of a trial in this area [Cakir and Kulaksizoglu33 (those on mirtazapine did better); Reichman et al. 34 (NPI decreased by 5.8 points)]. This would be the first placebo-controlled RCT of mirtazapine for agitation in dementia. Given the paucity of alternatives and the priority of finding safe and effective treatments for BPSD, these data suggest that a placebo-controlled trial of mirtazapine would be of value.

Carbamazepine for agitated behaviours in dementia

Carbamazepine stabilises the inactivated state of voltage-gated sodium channels and potentiates GABA receptors. It is recommended in the BNF for epilepsy, prophylaxis of bipolar disorder and trigeminal neuralgia. It is generally safe within the proposed dose ranges; there are few data on people with AD, but it seems that there is no increase in mortality as in antipsychotics for AD. 35 Carbamazepine has been widely used in psychiatric disorders and AD, off licence, to treat symptoms including agitation, aggression, irritability and impulsivity. Open-label studies and case reports have indicated promise in agitation in AD. 36 Two small 6-week parallel-group RCTs of carbamazepine for BPSD have been published. 37,38 The first in 55 patients (modal dose 300 mg) showed significant symptom decrease. It was well tolerated with no decrease in cognition, function or increased side effects relative to placebo. The second (400 mg in 21 patients not responding to antipsychotics) showed a trend but not a significant advantage over placebo. Meta-analysis indicated significant benefit compared with placebo treatment on the Brief Psychiatric Rating Scale (mean difference −5.5 points, 95% CI −8.5 to −2.5 points) and on the Clinical Global Impression Scale [odds ratio (OR) 10.2, 95% CI 3.1 to 33.1]. 39 A third small trial of the similar compound oxcarbazepine (n = 103) indicated a trend towards benefit (p = 0.07) with active drug performing better than placebo in all analyses. 40

Why is the study needed?

Agitated behaviours drive poor quality of life in dementia and poor outcomes including hospitalisation and care home placement and high cost. They have profound negative effects on people with dementia themselves, their family carers, services and society. They are a major issue in care homes, in general hospitals and in people’s own homes. The non-drug treatments we have are not always successful and the antipsychotic drugs that we use are associated with unacceptable increases in mortality and morbidity and low clinical effectiveness. There is a pressing need nationally and internationally for safe alternative pharmacological treatments. Research into better treatments for agitated behaviours in dementia was identified as a top 10 research priority by the Alzheimer’s Society and the James Lind Alliance. 41 This involved extensive engagement with people with dementia, carers, health and social care practitioners and organisations that represent these groups. Over 4000 questions on prevention, diagnosis, treatment and care of dementia were considered and the top 10 identified, including: ‘What non-pharmacological and/or pharmacological (drug) interventions are most effective for managing challenging behaviour in people with dementia?’ The need for better research into pharmacological treatments for agitation and aggression is also articulated in the National Dementia Strategy,42 the outputs of the 2010 Ministerial Dementia Research Summit and 2011 NIHR Dementia Research Workshop summarised in the Ministerial Advisory Group for Dementia Research (MAGDR) final report ‘Priority Topics in Dementia Research’ published by the MRC in February 2011. 43 MAGDR concluded ‘further research into behavioural and psychological symptoms in order to provide more effective management of challenging behaviour and improved quality of life’ was one of the top six headline priorities for research.

In this study, we therefore aimed to establish the clinical and cost-effectiveness and safety profile of carbamazepine (discontinued when 40 people had been randomised into this arm due to slower than projected recruitment) or mirtazapine in reducing agitation in AD relative to placebo.

Study design

This is a pragmatic, multicentre, double-blind, placebo-controlled superiority RCT of safety, clinical and cost-effectiveness of mirtazapine (with usual care) at 6 and 12 weeks on agitated behaviours in dementia. We included a long-term follow-up period to allow limited assessment of longer-term outcomes at 26 and 52 weeks. An internal pilot phase assessed trial recruitment, with progression to a full trial dependent on the number of patients recruited within the pilot recruitment period.

Important changes to methods

SYMBAD was initially designed as a three-arm trial, comparing both carbamazepine and placebo and mirtazapine and placebo for a difference in change in Cohen-Mansfield Agitation Inventory (CMAI) score at 12 weeks as the primary objective. Challenges in recruitment in this population resulted in the funder requesting that the available data be reviewed to July 2018, with the aim of dropping one arm of the trial.

The SYMBAD independent Data Monitoring Committee (DMC) reviewed available data comparing the two active arms with placebo. They were asked to consider efficacy data (the primary end point, CMAI at 12 weeks), safety data [frequency of AEs and serious adverse events (SAEs) on an individual basis] and compliance with treatment (dropouts and adherence with the prescribed amount of treatment medication). This was done subgroup blind: the DMC knew which was the placebo group but not the identity of the two active groups. Based on each set of data, the DMC were asked which arm they would recommend stopping or if they felt unable to make a recommendation. Taking all three sets together, the DMC were again asked whether they would recommend stopping one arm or were unable to make a recommendation.

The DMC could provide no recommendation on the basis of treatment compliance but recommended on the basis of efficacy and safety data the discontinuation of the carbamazepine arm. This recommendation did not provide, in any way, any indication that sufficient evidence had accrued to deem either drug to be effective or non-effective.

In August 2018, the protocol was submitted for a substantial amendment to change to a two-arm trial design, comparing mirtazapine with placebo as the primary objective. Protocol version 2.0 shows the new trial design. Up to the date of approval of this substantial amendment, 40 patients had been randomised to receive carbamazepine. These data have been analysed in the same way as the mirtazapine data. Chapters 2–4 reflect the amended protocol and refer only to the mirtazapine/placebo comparisons.

See Appendix 1 for summary of protocol changes.

Aim

The overall trial aim was to assess the safety, clinical and cost-effectiveness of mirtazapine in the treatment of agitation in dementia.

The null hypothesis is that there is no difference in CMAI scores between patients treated with placebo and mirtazapine at 12 weeks.

The primary objective is to determine if mirtazapine is more clinically effective in reducing agitated behaviours in dementia than placebo, measured by CMAI score 12 weeks post randomisation.

The secondary objectives are:

1. to determine if mirtazapine is more cost-effective than placebo at 12 weeks post randomisation

2. to determine if mirtazapine is more clinically and cost-effective than placebo in reducing CMAI score at 6 weeks post randomisation

3. to determine differences in effectiveness between mirtazapine and placebo on carers

4. to determine whether there are differences between the groups in AEs and adherence

5. to determine long-term differences between those randomised to placebo and mirtazapine in a head-to-head comparison of agitation (measured by CMAI score), institutionalisation, death and clinical management at 26 and 52 weeks post randomisation.

Participants

Setting

Participants were drawn from existing patients and new patient referrals to old age psychiatric services, memory clinics, specific Participant Identification Centres (PICs), primary care centres and those in care homes in 26 UK sites.

Interventions

Initially there were three groups: (1) mirtazapine, (2) carbamazepine and (3) placebo. As noted above, the carbamazepine arm was dropped during the course of the study in August 2018. The target dose was 45 mg of mirtazapine or 300 mg of carbamazepine per day. Drugs and their placebos were identically presented with participants aiming to take three capsules orally once a day.

Randomisation

Once a patient’s screening CMAI score has been assessed as being ≥ 45, the research worker discussed the case with the site PI who was permitted to prescribe Investigational Medicinal Product (IMP). The PI confirmed or not the patient’s eligibility to join the study, and on confirmation, the research worker used an online randomisation system to randomise the patient for the trial. This system required confirmation of eligibility criteria. Details of the randomisation were confirmed by e-mail to the research worker, site PI, Chief investigator and co-ordinating team at Norwich Clinical Trials Unit (NCTU). A semi-blinded randomisation e-mail detailing IMP allocation was sent to site pharmacy contact/s only. The PI provided a signed prescription for the patient’s trial medication. The research worker then collected this prescription from the central pharmacy and delivered it to the patient at a scheduled ‘Week 0’ IMP delivery visit. Local policies for treating patients outside of their registered NHS Trust were followed as appropriate. Random allocation was block stratified by centre and type of residence (care home vs. own household) with random block lengths of three or six before the discontinuation of the carbamazepine arm and thereafter of two or four. The trial was double-blind, with drug and placebo identically encapsulated. Referring clinicians, participants, the trial management team and the research workers who did baseline and follow-up assessments were masked to group allocation.

Primary outcome

Cohen-Mansfield Agitation Inventory score (Long Form) at 12 weeks.

Secondary outcomes

1. Costs derived from Client Service Receipt Inventory (CSRI), and quality-adjusted life-years (QALYs) from cost data alongside supplemented information from Dementia-Specific Quality of Life (DEMQOL) and EuroQol-5 Dimensions, five-level version (EQ-5D-5L) interviews 12 weeks post randomisation.

2. CMAI score and cost at 6 weeks post randomisation.

3. Patient and carer quality of life, and carer outcomes at 6 and 12 weeks post randomisation.

4. AEs from week 0 to week 16 and adherence at 6 and 12 weeks post randomisation.

5. CMAI score, AEs and adherence at 6 and 12 weeks, conditional on evidence of effectiveness of IMP over placebo.

6. Longer-term follow-up: CMAI score, institutionalisation, death and clinical management at 26 and 52 weeks post randomisation.

Instruments used in the study – range and scoring

1. CMAI (agitation): score ranges from 29 (no agitated behaviour) to 203 (very agitated behaviour).

2. Mini Mental State Examination (MMSE) (cognition): score ranges from 0 (severe impairment) to 30 (normal).

3. General Health Questionnaire (GHQ-12) (mental health): score ranges from 0 (least severe problems, good) to 36 (most severe).

4. Zarit Carer Burden: range from 0 (no burden) to 88 (most severe).

5. DEMQOL (quality of life): DEMQOL range from 28 (poor) to 112 (good); DEMQOL-Proxy: range from 31 (poor) to 124 (good).

6. NPI (neuropsychiatric symptoms) total: total score range from 0 (none) to 144 (severe); NPI agitation/aggression: range from 0 (none) to 12 (severe); NPI depression/anxiety/irritation: range from 0 (none) to 36 (severe); and NPI carer distress: range 0 (none) to 60 (severe).

Change in outcomes over the time of the trial

There were no changes to the primary outcomes of the trial following the registration of the trial.

Sample size

An initial calculated sample size of 400 (randomised 1 : 1 : 1) provided 90% power using two-sided 5% significance tests to detect a drug versus placebo mean difference in CMAI score at 12 weeks of 6 points. This equated to an effect size of d = 0.4 [assuming a common standard deviation (SD) of 15] or a clinically significant 30% decrease in CMAI from placebo to active drug. With a realistic 15% attrition, a sample of 471 (157 per arm) was aimed for. Due to slower than projected recruitment, an arm of the study (carbamazepine) was dropped and sample size was amended in consultation with the funder, DMC and TMC (see Important changes to methods). Based on the same parameters, an amended sample size target of 222 was calculated (randomised 1 : 1) allowing for a 10% attrition (111 per arm).

The primary outcome measure in this trial was the CMAI. Active drug treatment, compared with placebo, may be associated with changes in the CMAI that are much > 6 points, but SYMBAD was powered to detect the smallest difference in the CMAI that could be considered clinically meaningful. This estimation was based on the changes and SD of change score seen in the CALM trial which included a similar patient population treated with donepezil where 6 CMAI points was 35% of the SD.

Blinding and unblinding

Data management

Monitoring and site visits

For each site, a site initiation visit (SIV) was arranged by the trial manager with the core study team. A remote visit (teleconference or video conference call) was considered if they met predefined criteria of experience.

The minimum attendance for the SIV (in person or remotely) was the PI, lead research nurse, lead data manager (if applicable), pharmacy lead and research worker, any sub-Investigators were also encouraged to attend. All sites received a standardised copy of site initiation slides to aid the training of new staff working on the trial.

Each site was provided with a paper site file containing all the documents required to be held at site, generated at NCTU to comply with ICH GCP guidelines. A confirmation receipt was sent with the file for completion at site to confirm all documents had been received. Prior to a site being activated, the receipt was required to be received and returned to the TM or delegate at NCTU.

Before COVID-19, all sites recruiting at least five participants received one routine monitoring visit during the course of the trial. After lockdown, three sites (South West London, Central and North West London and Sheffield) that recruited more than five and were not visited, checks for these were made online. Site monitoring included the following checks:

• ensuing that key eligibility variables match source data

• blood test results and electrocardiogram (ECG) printouts

• SAE [and serious adverse reaction (SAR)/suspected unexpected serious adverse reaction (SUSAR) where applicable] reports were verified against clinical notes where possible

• clinic notes checked for unreported notable or serious events, where possible

• data from participants experiencing study drug discontinuations/dose lowering

• 50% Source Data Verification from patient packs transcribed to Electronic Case Report Forms (eCRFs) will be checked for at least 20% of patients recruited at the site at the time of the visit; if time allows, more patient packs will be checked. The CMAI questionnaire/score will always be checked for the randomly selected 20% of patients; a selection of the other questionnaires up to a minimum of 50% of the data will be checked

• sites will not be warned in advance which patient packs will be checked; the monitor will take a list with them and select the listed patient numbers from all of the available packs

• any major or critical findings should prompt the monitoring team to increase the level of monitoring to cover as many participant files as time allows

• completeness of trial drug dispensing, accountability and drug supply inventories

• documentation and procedures will be checked for protocol deviations and serious breaches

• consent forms

• delegation logs

• confirmation that safety checks had been completed and reviewed by the PI prior to randomisation

• accuracy of site file (and pharmacy file checks where relevant).

After the visit, the PI and site team were provided with a report summarising the documents that had been reviewed and the corrective actions that were required by the site team. A response was required to be provided to the TT. The Trial Manager reviewed responses and compiled them alongside the on-site monitoring findings. The final report was signed off by the TT member performing the visit and the CI (reviewer). Additional monitoring visits were conducted on a ‘for cause’ approach. Monitoring of data quality, recruitment rates, pharmacy, Investigator Site File documents, consent and safety also occurred centrally.

Assessment by time point

For an overview of assessments over time, please see Table 1.

Safety assessments

Definitions of harm of the EU Directive 2001/20/EC Article 2 based on the principles of ICH GCP applied to this trial: any unfavourable and intended sign, symptom or illness that developed or worsened during the period of the study was classified as an AE, whether or not it was considered to be related to the study treatment. AEs included unwanted side effects, sensitivity reactions, abnormal laboratory results, injury or intercurrent illnesses, and may be expected or unexpected. These were recorded on the CRF.

The period for SAE reporting was from the time of randomisation until 4 weeks post final trial medication administration. The participants were followed up by a telephone interview 4 weeks after the last dose of trial medication (the week 16 call). All events were followed until resolution, including if that meant beyond 4 weeks’ post final trial medication implementation.

Statistical methods

Sensitivity analyses

Sensitivity analyses explored the impact on results of varying key assumptions made in the base case for primary and secondary analyses: including accommodation of participants in domestic as well as residential care in total health and social care costs; examining total (EQ-5D-5L) QALY and costs for the dyad (person with dementia and unpaid carer); and using an alternative valuation and definition of unpaid carer time. Accommodation costs of domestic residence were sourced from UK Household expenditure statistics (Office for National Statistics 2019) and ‘sheltered’ domestic housing. 54 Unpaid carer time was valued at replacement cost, using the hourly cost of a home care worker. This valuation was also used to calculate unpaid care time defined as the hours of the day that the person with dementia could not be left alone by the carer.

In addition, we explored the impact on results of varying the modelling approach in the primary cost-effectiveness analyses. First, we included a covariate for gender in the MLM to adjust for a baseline imbalance between groups. Second, as an alternative approach to the MLM and to address skewness typical of cost data, we applied seemingly unrelated regressions67 (where cost and outcome equations were the same as in the MLM) to 4000 replicates generated by a two-stage bootstrapping procedure suitable for clustered data. 68 This analysis was conducted in R (The R Foundation for Statistical Computing, Vienna, Austria). 69

Ethics and regulatory approvals

The study was approved by the Hampshire A South Central Research Ethics Committee (15/SC/0606), and the MHRA. It received local NHS Trust approvals and consent or assent (with legal representative consent) was obtained from all participants (see trial protocol for more details). This protocol was submitted to the UK national competent authority (MHRA).

This is a clinical trial of an IMP as defined by the EU Directive 2001/20/EC. The progress of the trial, safety issues and reports, including expedited reporting of SUSARs, was reported to the competent authority (MHRA).

Patient and public involvement

Ensuring the involvement of people living with dementia and their carers has been integral to the SYMBAD trial. This section provides further information on their important role in the trial.

Patient flow

We recruited participants between January 2017 and February 2020 and completed week 12 follow-up interviews by May 2020 (See Figure 1). See Appendix 2 for recruitment by site and month.

FIGURE 1.

CONSORT flow diagram of recruitment and testing for mirtazapine and placebo groups.

Baseline characteristics

Table 4 shows baseline demographic and clinical characteristics of participants and carers. Groups were similar at baseline except for sex with more females randomised to mirtazapine (n = 77, 75%) than placebo (n = 59, 58%). In light of this difference, sex was included in an additional model as a sensitivity analysis. By week 12, similar numbers remained in the mirtazapine (80/102, 78%) and the placebo group (89/102, 87%).

Primary outcome measures

Severity of agitation decreased in both groups at 6 weeks by around 10 points and continued to be lower than baseline scores at 12 weeks (see Figure 2); this change between baseline and 6- and 12-week outcomes is illustrated by the separation in 95% confidence limits. At no point was the unadjusted or adjusted CMAI difference between the groups statistically significant (see Table 5). Table 5 presents the results from the general linear mixed modelling for the primary outcome. There was no evidence that mirtazapine improved agitation relative to placebo. The estimated adjusted effect on the CMAI was −1.74 (95% CI −7.17 to 3.69; p = 0.530). This changed little with the addition of sex into the model.

FIGURE 2.

Unadjusted mean CMAI scores (95% CI) by treatment group.

Secondary outcome measures

Table 5 shows the effect of mirtazapine compared with placebo on secondary outcomes in participants and Table 6 in carers. Again, there was no evidence of difference between the groups, apart from: a single statistically significant difference in the Zarit Carer Burden Inventory at 12 weeks which indicated higher carer burden in the mirtazapine group (adjusted difference 5.01 points, 95% CI 0.80 to 9.23; p = 0.020); weaker evidence at 6 weeks (3.76, −0.03 to 7.83); p = 0.069) in the same variable; and a weak association between higher proxy-rated ED-5D quality of life in the placebo group at 6 weeks (−0.07, −0.13 to 0.00, p = 0.061) that was not maintained at 12 weeks (−0.01, −0.08 to 0.07, p = 0.822).

Adverse events and severe AEs were ascertained to 16 or 4 weeks after last dose of IMP; deaths were recorded up to 16 weeks after randomisation. Examining AEs by week 16, there were 192 in 102 participants in the placebo group, of whom 65 (64%) individuals had at least one AE, compared with 225 events in 102 participants in the mirtazapine group of whom 67 (66%) had at least one. There were 35 SAEs in 18 individuals in the placebo group, compared with 13 in 8 individuals in the mirtazapine group. Mortality differed between groups with a potentially higher rate in the mirtazapine group (seven deaths in the mirtazapine and one in the placebo group by 16-week safety follow-up). Post hoc statistical analysis suggested weak evidence of a mortality difference between groups (Fisher’s exact test p = 0.065). Causes of death coded with MedDRA (Medical Dictionary for Regulatory Activities) terms showed no consistent pattern with the one death in the placebo group attributed to dementia, and the seven in the mirtazapine group to: (i) dementia; (ii) pneumonia, aspiration; (iii) emphysema, dementia, pneumonia, aspiration; (iv) dementia Alzheimer’s type; (v) cardiac failure; (vi) pelvic fracture, osteoporosis, vascular dementia; and (vii) chronic kidney disease, dementia, congestive cardiac failure. See Table 7. See Appendix 3 for a summary of AEs and severe AEs by randomisation group.

Cost-effectiveness analyses

Primary analysis

Mean raw outcome scores and costs of the complete cases samples showed no differences between groups (see Table 19). Adjusting for baseline measure and living arrangement, the estimate for the difference between groups in both CMAI and costs had wide CIs crossing zero. The point estimate for the ICER on CMAI was negative because costs were slightly lower and outcome slightly better in the mirtazapine group compared to the placebo group. The NMB line (see Figure 3) shows that net benefit is positive at all willingness-to-pay thresholds from £0 to £30,000: there is monetary benefit once the cost of the intervention has been deducted. However, the CIs of the line do not cross zero, illustrating that 95% confidence limits of the ICER could not be defined and therefore neither mirtazapine nor placebo can be judged to be the more cost-effective strategy with a high level of confidence. The CEAC (see Figure 4) illustrates that probability of cost-effectiveness was 81% at a willingness to pay of £3000 and 80% at £20,000; also that a 10% CI for the ICER can be defined between willingness to pay of approximately £0 and £3000 per QALY, giving a low degree of certainty that mirtazapine is cost-effective (see Glick et al. 63 2014; Gray et al. 59).

TABLE 19 - Primary outcome/costs: CMAI score and health and social care costs over 12-week study follow-up, raw and adjusted difference between groups and incremental cost-effectiveness ratio

Reversed so that a higher score indicates a better outcome.

Outcomes and costs	Mirtazapine	Placebo	Mirtazapine-placebo difference			ICER
	Mean (SE)	Mean (SE)	Mean (95% CI)	Adjusted (95% CI)	p	Cost per 6-point differencea
Health and social care
Observations	N = 72	N = 79
CMAI score	61.847 (2.659)	60.848 (2.49)	0.999 (−6.193 to 8.191)	−2.446 (−8.243 to 3.352)	0.408	−273/0.408 = −670
Total costs	5752 (513)	5877 (591)	−125 (−1686 to 1435)	−273 (−1754 to 1208)	0.718

FIGURE 3.

Primary outcome: NMB plot.

FIGURE 4.

Primary outcome: cost-effectiveness acceptability curve.

Participant outcomes

Carer outcomes

In cases with complete health and support and QALY data, unadjusted and adjusted between-group differences in QALYs were not significant (see Table 21). The NMB of mirtazapine at £20,000 was negative but the CI crossed zero. The ICER on this measure was unbounded. Results for carers’ societal costs and QALY were similar. Probability of cost-effectiveness from the health and social care perspective did not exceed 45% over WTP-per-QALY thresholds ranging between £0 and £50,000; probability of cost-effectiveness from the societal perspective did not exceed 19% over the same range.

Introduction

As discussed above, this trial was designed to include a carbamazepine arm as well as mirtazapine and placebo arms. The carbamazepine arm was discontinued after 40 randomisations due to slower than projected recruitment. In this chapter, we present the rationale for this arm and mirrored analyses of the clinical effectiveness of carbamazepine versus placebo in the treatment of agitation in dementia using the same methodology that was used for mirtazapine versus placebo.

Carbamazepine for agitated behaviours in dementia

Carbamazepine stabilises the inactivated state of voltage-gated sodium channels and potentiates GABA receptors. It is recommended in the BNF for epilepsy, prophylaxis of bipolar disorder and trigeminal neuralgia. It is generally safe within the proposed dose ranges; there are few data on people with AD, but it seems that there is no increase in mortality as in antipsychotics for AD. 35 Carbamazepine has been widely used in psychiatric disorders and AD, off licence, to treat symptoms including agitation, aggression, irritability and impulsivity. Open-label studies and case reports have indicated promise in agitation in AD. 36 Two small 6-week parallel-group RCTs of carbamazepine for BPSD have been published. 37,38 The first in 55 patients (modal dose 300 mg) showed significant symptom decrease. It was well tolerated with no decrease in cognition, function or increased side effects relative to placebo. The second (400 mg in 21 patients not responding to antipsychotics) showed a trend but not a significant advantage over placebo. Meta-analysis indicated significant benefit compared with placebo treatment on the Brief Psychiatric Rating Scale (mean difference −5.5 points, 95% CI −8.5 to −2.5 points) and on the Clinical Global Impression Scale (OR 10.2, 95% CI 3.1 to 33.1). 39 A third small trial of the similar compound oxcarbazepine (n = 103) indicated a trend towards benefit (p = 0.07) with active drug performing better than placebo in all analyses. 40

Aim

To determine if carbamazepine is more clinically effective in reducing agitated behaviours in dementia than placebo measured by change in CMAI score 12 weeks post randomisation.

Methods

Please see Chapter 3.

Results

Patient flow

We recruited participants between January 2017 and August 2018 to the three-arm trial and recruitment continued until February 2020 for mirtazapine and placebo arms. Follow-up interviews were completed accordingly. See Figure 5.

FIGURE 5.

CONSORT flow diagram of recruitment and testing for carbamazepine and placebo groups.

Primary outcome measures

Severity of agitation decreased in the placebo group by 9.6 points on the CMAI and by 4.4 points in the carbamazepine group. At 12 weeks, the placebo group was 60.8 and the carbamazepine group 63.9 (see Figure 6). At no point was the unadjusted or adjusted CMAI difference between the groups statistically significant (see Tables 27 and 28). Table 27 presents the results from the general linear mixed modelling for the primary outcome. There was no evidence that carbamazepine improved agitation relative to placebo. The estimated adjusted effect on the CMAI was 2.46 (95% CI −5.01 to 9.93; p = 0.518). This changed little with the addition of sex into the model.

FIGURE 6.

Unadjusted mean CMAI scores (95% CI) by treatment group, carbamazepine vs. placebo.

TABLE 27 - Primary outcomes of research worker rated CMAI score at 12 weeks

Adjusted for pre-specified factors: baseline CMAI, household status and centre.

Adjusted for sex and pre-specified factors: baseline CMAI, household status and centre.

Carbamazepine (n = 40)	Placebo (n = 102)	Difference	(95% CI)	Adj diffa	(95% CI)	p-value	Adj diffb	(95% CI)	p-value
n = 32	n = 87
63.9 (sd 24.7)	60.8 (sd 21.8)	3.1	(−6.17 to 12.37)	2.46	(−5.01 to 9.93)	0.518	2.67	(−5.05 to 10.40)	0.498