Dementia with Lewy bodies (DLBs) is a common cause of dementia but has higher mortality than Alzheimer’s disease (AD). The reasons for this are unclear, but antidementia drugs (including acetylcholinesterase inhibitors [AChEIs] and memantine) symptomatically benefit people with DLB and might improve outcomes. We investigated whether AChEIs and/or memantine were associated with reduced hospital admissions and mortality.
Methods and findings
We performed a retrospective cohort study of those diagnosed with DLB between 1 January 2005 and 31 December 2019, using data from electronic clinical records of secondary care mental health services in Cambridgeshire and Peterborough NHS Foundation Trust (CPFT), United Kingdom (catchment area population approximately 0.86 million), as well as linked records from national Hospital Episode Statistics (HES) data. Eligible patients were those who started AChEIs or memantine within 3 months of their diagnosis (cases) and those who never used AChEIs or memantine (controls). Outcomes included admission, length of stay, and mortality. Cox proportional hazard and linear regression models were used.
Of 592 patients with DLB, 219 never took AChEIs or memantine, 100 took AChEIs only, and 273 took both AChEIs and memantine. The cohorts were followed up for an average of 896 days, 981 days, and 1,004 days, respectively. There were no significant differences in the cohorts’ baseline characteristics, except for socioeconomic status that was lower in patients who never took AChEIs or memantine (χ2 = 23.34, P = 0.003). After controlling for confounding by sociodemographic factors (age, sex, marital status, ethnicity, socioeconomic status), antipsychotic use, antidepressant use, cognitive status, physical comorbidity, anticholinergic burden, and global health performance, compared with patients who never took AChEIs or memantine, patients taking AChEIs only or taking both had a significantly lower risk of death (adjusted hazard ratio (HR) = 0.67, 95% CI = 0.48 to 0.93, p = 0.02; adjusted HR = 0.64, 95% CI = 0.50 to 0.83, P = 0.001, respectively). Those taking AChEIs or both AChEIs and memantine had significantly shorter periods of unplanned hospital admission for physical disorders (adjusted coefficient −13.48, 95% CI = [−26.87, −0.09], P = 0.049; adjusted coefficient −14.21, 95% CI = [−24.58, −3.85], P = 0.007, respectively), but no difference in length of stay for planned admissions for physical disorders, or for admissions for mental health disorders. No significant additional associations of memantine on admission, length of stay, and mortality were found (all P > 0.05). The main limitation was that this was a naturalistic study and possible confounds cannot be fully controlled, and there may be selection bias resulting from nonrandom prescription behaviour in clinical practice. However, we mimicked the intention-to-treat design of clinical trials, and the majority of baseline characters were balanced between cohorts. In addition, our series of sensitivity analyses confirmed the consistency of our results.
Why was this study done?
- Compared to Alzheimer’s disease (AD), dementia with Lewy bodies (DLBs) is associated with accelerated cognitive decline, lower quality of life, higher caregiver burden, shorter lifespan, and higher costs of care, as well as increased rates of admission to general hospitals and residential care and longer length of stay.
- Although there are no disease-modifying medications for DLB, there is evidence suggesting that acetylcholinesterase inhibitors (AChEIs) and glutamate N-methyl-D-aspartate receptor antagonists (specifically, memantine) are efficacious in DLB, though an impact on mortality is not clear.
- Distinct evidence on any survival benefit of AChEIs and memantine in DLB needs to be explored specifically, as DLB is pathologically and clinically different from AD and other forms of dementia.
What did the researchers do and find?
- We identified 592 patients with DLB, including 219 who never took AChEIs or memantine, 100 who took AChEIs only, and 273 who took both AChEIs and memantine, between 1 January 2005 and 31 December 2019, using data from electronic clinical records of secondary care mental health services in Cambridgeshire and Peterborough NHS Foundation Trust (CPFT), United Kingdom and linked records from national Hospital Episode Statistics (HES) data.
- We investigated the associations of antidementia drug use in DLB with hospital admissions, length of stay, and mortality.
- We found that taking AChEIs alone or with memantine was associated with a significantly reduced risk of death and shorter hospital stays after unplanned admissions for physical disorders. Taking AChEIs alone or with memantine was not associated with the probability of admission for mental disorders or physical disorders (planned or unplanned), or the length of stay after admission for mental disorders or planned admission for physical disorders, compared to those not taking AChEIs or memantine. We found no significant additional association of memantine with admission, length of stay, or mortality, compared to those taking AChEI therapy alone.
What do these findings mean?
- We observed that AChEIs (with or without memantine) may be associated with a shorter length of stay associated with unplanned hospital admissions for physical disorders and reduced risk of mortality, warranting further study.
- The findings in our study may help to address potential concerns that AChEIs could have the opposite effect (to increase mortality in DLB), thus giving both clinicians and people with DLB more evidence to support treatment decisions.
- The main limitation was that as a naturalistic study, possible confounds cannot be fully controlled and there may be selection bias resulting from nonrandom prescription behaviour in clinical practice. However, we mimicked the intention-to-treat design of clinical trials, and our series of sensitivity analyses confirmed the consistency of our results.
Citation: Chen S, Price AC, Cardinal RN, Moylett S, Kershenbaum AD, Fitzgerald J, et al. (2022) Association between antidementia medication use and mortality in people diagnosed with dementia with Lewy bodies in the UK: A retrospective cohort study. PLoS Med 19(12): e1004124. https://doi.org/10.1371/journal.pmed.1004124
Academic Editor: Raquel C. Gardner, University of California San Francisco, UNITED STATES
Received: March 22, 2022; Accepted: October 14, 2022; Published: December 6, 2022
Copyright: © 2022 Chen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Patient-level data is not publicly available, but can be accessed by researchers who meet the criteria for access to de-identified sensitive data under NHS Research Ethics terms. Full details of how to request access to CPFT data for research can be found at: https://www.cpft.nhs.uk/research-database/ or email firstname.lastname@example.org for further information.
Funding: SC’s, AP’s, AK’s, SM’s, and JOB’s research was supported by the UK Alzheimer’s Society (grant AS-PG-16-006 to JOB) (https://www.alzheimers.org.uk/). RNC’s research was supported by the Medical Research Council (grants MC_PC_17213, MR/W014386/1) (https://www.ukri.org/councils/mrc/). The CPFT Research Database was supported by the UK National Institute of Health Research Cambridge Biomedical Research Unit in Dementia and the Biomedical Research Centre (https://cambridgebrc.nihr.ac.uk/), and Cambridgeshire and Peterborough NHS Foundation Trust (https://www.cpft.nhs.uk/). CM and RS were part-funded by the National Institute for Health Research (NIHR) Biomedical Research Centre at the South London and Maudsley NHS Foundation Trust and King’s College London (https://www.maudsleybrc.nihr.ac.uk/). RS is additionally part-funded by the National Institute for Health Research (NIHR) Applied Research Collaboration South London at King’s College Hospital NHS Foundation Trust (https://arc-sl.nihr.ac.uk/) and the DATAMIND HDR UK Mental Health Data Hub (MRC grant MR/W014386) (https://www.ukri.org/councils/mrc/). This research was supported in part by the NIHR Cambridge Biomedical Research Centre (BRC-1215-20014) (https://cambridgebrc.nihr.ac.uk/) and the Cambridge Centre for Parkinson’s Plus disorders (https://ccpp.cam.ac.uk/). The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the article.
Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: SC, ACP, SM, ADK, JF, and CM, declare no conflict of interest with this work. RNC consults for Campden Instruments Ltd and receives royalties from Cambridge University Press, Cambridge Enterprise, and Routledge. RS has received research support in the last 36 months from Janssen, GSK and Takeda. J.O.B. has acted as a consultant for TauRx, Eisai, Novo Nordisk, Biogen and GE Healthcare and received grant support from Avid/ Lilly, Merck and Alliance Medical.
Abbreviations: ACB, anticholinergic cognitive burden; ACE-III, Addenbrooke’s Cognitive Examination III; AChEIs, acetylcholinesterase inhibitors; AD, Alzheimer’s disease; CCI, Charlson comorbidity index; COVID-19, Coronavirus Disease 2019; CPFT, Cambridgeshire and Peterborough NHS Foundation Trust; DLB, dementia with Lewy body; HES, Hospital Episode Statistics; HoNOS, Health of the Nation Outcomes Scales; HR, hazard ratio; ICD, International Classification of Diseases; IMD, Index of Multiple Deprivation; LBD, Lewy body dementia; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment; NICE, National Institute for Health and Care Excellence; PDD, Parkinson’s disease dementia; RCT, randomised controlled trial; SD, standard deviation; WHO, World Health Organization
Dementia with Lewy bodies (DLBs) is the second most common form of neurodegenerative dementia after Alzheimer’s disease (AD) in people over 65 [1,2]. Together with Parkinson’s disease dementia (PDD), it constitutes Lewy body dementia (LBD). In a systematic review of 31 studies , the incidence of DLB was 3.8% of new dementia cases, and within the UK, DLB diagnostic rates were found to be 4.6% of all dementia cases .
Compared to AD, DLB is associated with accelerated cognitive decline, poorer quality of life, higher caregiver burden, higher costs of care, increased rates of hospital admission, longer hospital stays, and increased mortality (approximately twice that for AD) . One study  estimated that in the United Kingdom, there are 80,000 people living with DLB who collectively experience >27,000 more hospitalizations per year than the same number of patients with AD, and spend >300,000 days more in hospital, costing >£35 million per year.
Interventions that could improve the prognosis of DLB even modestly would provide a major social and economic benefit for patients, their families, and the healthcare system. However, there are no disease-modifying therapies for DLB, and the management of DLB focuses on symptomatic control, largely based on the evidence base from AD and PDD. Data across all dementia , from AD [6–8] and from PDD , suggests an effect of acetylcholinesterase inhibitors (AChEIs) to reduce mortality; however, there is no such evidence for DLB despite its relatively high prevalence and poor prognosis. There is evidence in DLB, and in the broader group of LBD, of possible (though inconsistent) benefits of memantine for cognition and neuropsychiatric symptoms [10–13], but similarly no study has focused on any survival benefit. Evidence speaking to the impact of AChEIs and memantine on mortality is important, as DLB is pathologically and clinically different from, and has higher mortality than, AD and other forms of dementia.
People with DLB have complex physical and psychiatric symptoms and are more likely to be admitted to hospital than those without DLB. In this study, our primary aim was to investigate whether the use of AChEIs and/or memantine was associated with an altered risk of death in people with DLB. Our second aim was to investigate the association of AChEIs and/or memantine with the risk of hospital admission (planned or unplanned) and corresponding duration of hospital stay, for mental and physical disorders, in people with DLB. We hypothesised that both AChEIs and memantine would be associated with reduced risk of admission, decreased length of stay, and reduced mortality in people with DLB.
Study design and participants
We performed a retrospective cohort study using data from de-identified electronic clinical records of the secondary care mental health services of Cambridgeshire and Peterborough NHS Foundation Trust (CPFT), which provides specialist mental health and community physical health services to a population of approximately 0.86 million people in the United Kingdom. Therefore, all patients reported in CPFT were under the care of a specialist. CPFT’s electronic records contain patient information recorded during routine treatment, such as sociodemographic information, diagnosis, prescription data, death status, and clinical notes in free text. Clinicians enter aspects of this information in a systematic and structured/standard way to ensure its accuracy . Data from patients with DLB were linked to national Hospital Episode Statistics (HES) data (from NHS Digital). Data were de-identified before researchers were given access [15,16].
Data were collected from 1 January 2005 to 31 December 2019, to exclude the influence of the Coronavirus Disease 2019 (COVID-19) pandemic. Eligible patients were those diagnosed with DLB and who had at least 3 months’ follow-up (see below). DLB was established based on coded World Health Organization (WHO) International Classification of Diseases (ICD-10) diagnoses, as well as by searching de-identified free text for diagnostic terms. In detail, all patients with a clinician-recorded diagnostic code G31.8 or the presence of “Lewy,” “LBD,” and “DLB” in their assessment/diagnostic documents, progress notes, or clinical correspondence from the examining clinicians, were included as potential DLB cases. Then, all potential DLB cases were manually checked by 2 experienced dementia clinicians based on the DLB diagnosis guidelines . The confirmed DLB cases thus selected were used for mortality studies in previous publications [14,18]. The procedure is summarised in Fig 1.
DLB, dementia with Lewy bodies; AChEIs, acetylcholinesterase inhibitors; LBD, Lewy body disease; DLB, dementia with Lewy body; CPFT, Cambridgeshire and Peterborough NHS Foundation Trust; ICD-10, World Health Organization International Statistical Classification of Diseases and Related Health Problems, 10th revision.
The index date was the date of diagnosis (for AChEIs and memantine non-users) or the first date of AChEIs or memantine prescription (for AChEIs or memantine users). Follow-up was until the patients’ final CPFT record date, the date of their death, or the study end date of 31 December 2019, whichever occurred first.
Outcomes and measures
Admission information was acquired from the CPFT and HES databases, which also indicated whether admissions were planned or unplanned (for emergencies), were for mental disorders or for physical disorders, and gave the time of admission and discharge.
Length of stay was calculated as the sum of duration from admission to discharge for each admission during the follow-up. As the admission is a recurrent event, the longer a person was followed up, the higher probability of a longer total length of stay because of multi-admissions. To eliminate discrepancies due to differences in follow-up duration, the length of stay was standardised as the average annual length of stay per person per follow-up year, by calculating the total length of stay (days) during follow-up divided by the length of follow-up (days) and then multiplied by 365 (days).
Death status was ascertained by weekly linkage to national NHS Spine mortality data.
Medications. AChEIs were defined as donepezil hydrochloride, rivastigmine tartrate, and galantamine hydrobromide. Patients were defined as having been initiated on AChEIs or memantine if these were prescribed within 3 months after the DLB diagnosis. Patients who started AChEIs/memantine beyond this period after diagnosis, or used AChEIs/memantine before diagnosis, were excluded from analyses (Fig 1). This conservative definition was used in previous studies  and is intended to mimic the intention-to-treat design of clinical trials and to avoid reverse causation (in which possible adverse effects would cause changes in medication status). According to the UK National Institute for Health and Care Excellence (NICE) guideline, both AChEIs and memantine are licensed for the treatment of dementia; however, AChEIs are the first choice, so in clinical practice, it is uncommon to find patients who only receive memantine (21 cases). Also because our primary focus was on the AChEIs and few patients received only memantine, we excluded those who only received memantine (Fig 1). Thus, in this study, we categorised medication use as AChEIs only, a combination of AChEIs and memantine (AChEI+memantine), or neither/none.
We examined the following sociodemographic variables: age at baseline (years), sex (male versus female), marital status (married, cohabiting or civil partnership versus single, divorced or widowed), ethnicity (white versus other), and socioeconomic status (measured by the 2015 Index of Multiple Deprivation [IMD]). The IMD is the official measure of relative deprivation in England for a geographical area of residence and incorporates 7 domains: income, health and disability, barriers to housing and services, education, employment, crime, and living environment and education . We also investigated physical co-morbidity (measured by the Charlson comorbidity index [CCI] , with co-morbid disorders identified via ICD codes in electronic medical records), use of antipsychotics, use of antidepressants, cognitive status (measured by a relative score of the Mini-Mental State Examination [MMSE], Montreal Cognitive Assessment [MoCA], or Addenbrooke’s Cognitive Examination III [ACE-III], closest to their index follow-up time), anticholinergic burden (measured by the anticholinergic cognitive burden [ACB] scale), and global health performance (measured by the Health of the Nation Outcomes Scales [HoNOS]). The CCI contains 19 categories of comorbidity and has been shown to be able predict 10-year mortality for patients who have a range of comorbid conditions . The formula and ICD-10 codes used for identifying comorbidities can be found elsewhere [20,21]. Antipsychotics and antidepressants were selected according to UK NICE guidelines (Table A in S1 Appendix). Scores from 3 validated cognitive scales, the MMSE, MoCA, and ACE-III [22–24] were used, varying across different people according to clinical practice; we rescaled the scores to a consistent maximum of 100. The ACB scale was developed based on a systematic literature review of medicines with known anticholinergic activity . In its newest (2012) version, the ACB scale included 115 medicines and scores each medicine 1, 2, or 3, based on their anticholinergic properties. The HoNOS is a clinician-rated instrument measuring the health and social functioning of patients . It comprises 12 scales covering behaviour, impairment, symptoms, and social functioning, for which values are recorded ranging from 0 (no problem) to 4 (severe to very severe problem).
For categorical variables, we have reported them as numbers (percentages) and continuous variables as mean (standard deviation, SD). Baseline differences between groups were assessed using ANOVAs (continuous variables) and chi-square tests (categorical variables).
Estimated projections of the cumulative hazard of admission and survival probability of mortality were obtained by Kaplan–Meier analysis. The log-rank test was used to determine whether survival curves differed statistically between individuals taking different medications.
The Cox proportional hazard model was used to estimate the hazard ratios (HRs) for medicine usage pattern in relation to admission or mortality. Results were reported as both unadjusted and adjusted hazard ratios (HRs). For adjusted HRs, the controlled covariates included sociodemographic factors (age, sex, marital status, ethnicity, socioeconomic status), antipsychotic use, antidepressants use, cognitive status, physical comorbidity, anticholinergic burden, and global health performance. As admission is a recurrent measure, the Cox model for admission was estimated by additionally using “person” as a cluster variable . The proportional hazard assumption was tested as the correlation between the Schoenfeld residuals and survival time, with a significance level of p < 0.05 indicating nonproportionality. We also plotted the Schoenfeld residuals to test graphically the proportional hazard assumption (Fig A in S1 Appendix).
To analyse length of stay, we used linear regression, not Cox proportional hazard models. Linear regression was used to estimate both unadjusted and adjusted coefficients through which medicine pattern predicted length of stay. Controlled covariates were the same as that in the above Cox model.
To estimate the add-on association of memantine compared to those taking AChEIs only, we repeated the analyses above but taking “AChEIs only” as the reference.
Data were complete for outcomes and predictors except for cognitive scores. For this, we used multiple imputations with chained equations based on age, sex, marital status, ethnicity, CCI, IMD, antipsychotic use, and antidepressant use.
We conducted 4 sensitivity analyses. First, we repeated our primary analysis by propensity score weighting following a tutorial on propensity score estimation for multiple treatments . In the propensity score estimation, our outcome was the medicine pattern and the predictors were the covariates controlled in the Cox model or linear regression. Second, we re-constructed the cohorts with a longer window, defining AChEI/memantine use as being within 6 months after the DLB diagnosis. Third, we repeated the analysis by setting the index date as the date of DLB diagnosis throughout in consideration of the start date being later for AChEI/memantine users, which may convey an automatic survival advantage. Fourth, we repeated the analysis by excluding the cognitive status in consideration of the missing and imputation of cognitive scores.
We used R (version 3.5.0) for all analyses and defined statistical significance as P < 0.05.
Planning of analyses
No document analysis plan was created prior to the start of the study, although the analysis plan was discussed and agreed upon among co-authors. No data-driven changes to the analysis plan were made. One additional sensitivity analysis was included in response to peer review, by repeating our primary analysis using propensity score weighting.
For identifiable records linkage, approval was obtained from the UK Confidentiality Advisory Group (CAG reference 18/CAG/0015) under section 251 of the UK National Health Service Act 2006. Data were de-identified before researchers were given access and analysed under NHS Research Ethics approvals (reference 18/EE/0029). Patients and carers have been involved throughout the process of cohort identification and data linkage that has allowed this analysis to take place. Patients and carers consent to the use of de-identified data for research purposes.
We identified 592 patients with the diagnosis of DLB, of whom 219 took neither AChEIs nor memantine, 100 took AChEIs only, and 273 took both AChEIs and memantine (Fig 1). The cohorts were followed up for an average of 896 days (range 30 to 3,736), 981 days (range 49 to 4,598), and 1,004 days (range 32 to 3,714), respectively (Table 1). There were no cohort differences in the baseline characteristics (Table 1), including age, sex, marital status, ethnicity, cognitive status, physical comorbidity, antipsychotic use, and antidepressant use (p > 0.05), except socioeconomic status (p = 0.003).
Fig 2 present the survival curves. In unadjusted analyses, mortality rates were lower in people taking AChEIs or AChEI+memantine than those taking neither (p = 0.02, Fig 2D), and the AChEI and AChEI+memantine groups had higher rates of admission for mental health disorders than those taking neither (p = 0.008, Fig 2A). There were no significant differences in the rates of planned admissions for physical disorders (p = 0.38, Fig 2B), or unplanned admissions for physical disorders (p = 0.70, Fig 2C).
Panels A, B and C present the recurrent events. P values are calculated from the log-rank test. AChEIs, acetylcholinesterase inhibitors.
However, after controlling for confounding by sociodemographic factors (age, sex, marital status, ethnicity, socioeconomic status), antipsychotic use, antidepressants use, cognitive status, physical comorbidity, anticholinergic burden, and global health performance, patients taking AChEIs only had a significantly lower risk of death than patients taking neither drug class (adjusted HR = 0.67, 95% CI = 0.48 to 0.93; p = 0.02, Fig 3G), and had shorter total hospital stays for unplanned admission for physical disorders (adjusted coefficient −13.48, 95% CI = [−26.87, −0.09]; p = 0.049, Fig 3F). Associations between AChEIs use and other outcomes were not significant, including the risk of admission for mental disorders (adjusted HR = 0.59 95% CI = 0.16 to 2.25; p = 0.44, Fig 3A), length of stay because of admission for mental disorders (adjusted coefficient −9.70, 95% CI = [−74.11, 54.72]; p = 0.76, Fig 3B), risk of planned admission for physical disorders (adjusted HR = 0.76 95% CI = 0.44 to 1.31; p = 0.33, Fig 3C), length of stay because of planned admission for physical disorders (adjusted coefficient −0.09, 95% CI = [−16.04, 15.85]; p = 0.99, Fig 3D), and risk of unplanned admission for physical disorders (adjusted HR = 0.95, 95% CI = 0.77 to 1.17; p = 0.63, Fig 3E). Similarly, patients taking both AChEIs and memantine had a significant lower risk of death (adjusted HR = 0.64, 95% CI = 0.50 to 0.84; p = 0.001, Fig 3G) and shorter stay because of unplanned admission for physical disorders (adjusted coefficient −14.21, 95% CI = [−24.58, −3.84]; p = 0.007, Fig 3F), but associations for other outcomes were not significant, including risk of admission for mental disorders (adjusted HR = 1.31, 95% CI = 0.45 to 3.74; p = 0.62, Fig 3A), length of stay because of admission for mental disorders (adjusted coefficient −15.68, 95% CI = [−66.08, 34.73]; p = 0.53, Fig 3B), risk of planned admission for physical disorders (adjusted HR = 0.94, 95% CI = 0.66 to 1.35; p = 0.75, Fig 3C), length of stay because of planned admission for physical disorders (adjusted coefficient −6.46, 95% CI = [−18.67, 5.75]; p = 0.30, Fig 3D), and risk of unplanned admission for physical disorders (adjusted HR = 0.90, 95% CI = 0.76 to 1.06; p = 0.20, Fig 3E). The proportional hazard assumption was valid for medicine usage pattern across outcomes (all p > 0.05, Fig A in S1 Appendix).
Hazard ratios (HR), 95% confidence intervals (CI), and p values were estimated from Cox proportional hazards models. Coefficients, 95% confidence intervals (CI), and p values were estimated from linear regression. Adjusted HRs/coefficients were adjusted for age, sex, marital status, ethnicity, socio-economic status (Index of Multiple Deprivation), antipsychotic use, antidepressant use, cognitive status/score, physical comorbidity, anticholinergic burden, and global health performance. The blue lines show the unadjusted HRs/coefficients, the red lines show the adjusted ones, and the grey dotted line shows the cut-off for negative or positive associations (0 for coefficients and 1 for hazard ratios). AChEIs, acetylcholinesterase inhibitors.
After controlling for confounding by sociodemographic factors (age, sex, marital status, ethnicity, socio-economic status), antipsychotic use, antidepressants use, cognitive status, physical comorbidity, anticholinergic burden, and global health performance, the differences between the AChEI+memantine group and the AChEI-only group were not significant for any of these outcomes (p > 0.05, Fig B in S1 Appendix).
Sensitivity analyses using propensity score weighting (Fig C in S1 Appendix), a longer initiation window for AChEIs/memantine (Fig D in S1 Appendix), setting the index date as the date of DLB diagnosis throughout (Fig E in S1 Appendix), and excluding cognitive status as a predictor (Fig F in S1 Appendix) all confirmed our primary results, that use of AChEIs with or without memantine in DLB was associated with decreased mortality and shorter total hospital stays for unplanned admission for physical disorders.
Using a retrospective cohort study based on a large and comprehensive electronic clinical records database, we investigated the associations of antidementia drug use in DLB with hospital admissions, length of stay, and mortality. Groups at baseline were similar in terms of age, sex, and cognitive test score. After controlling for sociodemographic factors, antipsychotic use, antidepressant use, cognitive status, and physical comorbidity, we found that patients who took AChEIs only, or took both AChEIs and memantine, had a significantly lower risk of death than those taking neither. Although no significant associations were found between use of AChEIs with or without memantine and the number of hospital admissions, the length of stay in patients taking AChEIs with or without memantine was significantly shorter for unplanned admissions for physical disorders. No significant additional associations of memantine over AChEIs were found.
Our findings supported the hypothesis that taking AChEIs is associated with a reduced risk of death in DLB. This association is consistent with the survival benefits of AChEIs identified in those with unspecified dementia , AD [6–8], and PDD . Patients with AD, PDD, and DLB have severe deficits in brain levels of acetylcholine and its synthetic enzyme, choline acetyltransferase, and AChEIs are beneficial for improving related clinical symptoms by lowering degradation of acetylcholine once released into the synapse [30,31]. In addition, AChEI use in DLB can cause improvements in cognition, including attention, and improvements in neuropsychiatric features including visual hallucinations, a hallmark of the disorder. These benefits have led to AChEIs being recommended for use in DLB by national and international expert groups including the International DLB Consortium, the UK NICE, and the British Association for Psychopharmacology [17,32,33].
No significant associations were found between AChEIs usage and the risk of admissions (for mental disorders, or for planned or unplanned physical disorders). This inconsistency was unexpected given the significant association between AChEI use and lower risk of mortality identified in this study, the established evidence that AChEIs have been associated with improvements in global cognitive function, behavioural disturbances, activities of daily living, and overall function in people with LBD or DLB [11,13,34], and the considerable overlap between the primary death reason of DLB  and primary admission reason of DLB  or LBD . Therefore, one of the possible mechanisms is that the symptomatic benefit of AChEIs could be translated into better physical health and then lower mortality by reducing events such as falls, but the absence of a significant association between AChEIs and admission (especially unplanned admission) might result from unmet needs or underutilised admission. For example, fall-related injuries and behavioural/neuropsychiatric symptoms of dementia accounted for about 13% of all admissions of DLB in South London, UK, far lower than the proportion (about 64%) observed for LBD in Florida, United States of America [4,36]. Another possible mechanism is that AChEIs have been directly associated with reductions in physical conditions. Growing studies support the beneficial role of AChEIs in the cardiovascular system and suggest that AChEIs have anti-inflammatory properties and reduce markers of endothelial and platelet activation [37–40]. Evidence from AD and all dementia spectrum also indicated that AChEIs is associated with reduced risk of stroke  and myocardial infarction . Although an effect of AChEIs to improve cognitive function is well established, the possibility that AChEIs reduce mortality by improving the course of dementia itself should be treated cautiously, as the association between AChEI use and the onset of severe AD has not been found to be significant .
Taking AChEIs or AChEIs+memantine was associated with reduced unplanned admission days for physical disorders. This finding also supports AChEI usage in this group, given the substantial amount of time that was spent in hospital by people with DLB (Table 1).
We found no additional associations with respect to memantine. In relation to hospital admissions, this is to some extent consistent with another study reporting no significant association between memantine and nursing home admission in mild dementia (including DLB and AD) . Meta-analysis and randomised controlled trials (RCTs) have found inconsistent but primarily negative findings that memantine improved  or did not improve [11–13] neuropsychiatric symptoms, improved quality of life , had no effect on general cognition , no effect on the clinical global impression of change , and no effect on motor symptoms  in DLB, versus placebo. There is also evidence from 1 meta-analysis  and an RCT study  indicating that AChEIs, but not memantine, significantly improved cognitive function, behavioural symptoms, and activities of daily living in DLB.
The findings in our study help to address potential concerns that the AChEIs could have the opposite effect and increase mortality in DLB by, for example, their propensity to cause bradycardia [44,45]. There are often concerns about the use of AChEI in DLB, including that long-term use may be associated with syncope, bradycardia, autonomic dysfunction, postural hypotension, and cardiac problems. It is therefore reassuring that we found mortality was lower in patients taking an AChEI, which do not support the association between AChEI use and negative long-term outcomes, thus giving both clinicians and people with DLB more evidence to support treatment decisions.
Strengths of our study included the use of an anonymised electronic records database derived from routinely collected clinical records enabling a larger sample size than many previously published examinations of DLB cohorts, also reducing sampling bias in relation to diagnosed cases . Because we used routine NHS e-record data for a mental health trust covering a substantial geographical area and population, we can be confident our findings reflect antidementia drug use and outcomes for DLB within routine NHS services. We also adopted a comprehensive approach to identifying DLB cases, not just relying on ICD codes (which can be used inconsistently) but also identifying DLB cases by text searches of diagnostic terms followed by clinical validation. In addition, by using linked HES data, we were able to study the associations of drug use with admission and length of stay and to examine planned/unplanned admissions as well as admissions for mental/physical disorders.
Limitations include that as a naturalistic study, possible confounds cannot be fully controlled; as this is not a RCT, use of AChEIs and memantine is not random between cohorts. In practice, medicines may be more likely to be prescribed to healthy patients, and the use of medicines with anticholinergic effects may also affect the prescription of AChEIs. In addition, unequal presence of adverse events of AChEIs (like syncope, bradycardia, autonomic dysfunction, postural hypotension, and cardiac problems), which may also function as contra-indications for starting AChEIs, may also have resulted in relevant selection bias. However, we sought to mimic the intention-to-treat design of clinical trials, excluding those who used AChEIs before the index date, and the majority of baseline characters (including global health performance and anticholinergic burden) were balanced between cohorts. Despite this, we conducted a sensitivity analysis with propensity score weighting, and the results remained consistent.
A second limitation is the conservative design in which we defined medication exposure as being within the 3 months after DLB diagnosis. Patients who were defined as taking AChEIs only, or AChEIs+memantine, might have stopped taking 1 or both medications (e.g., due to adverse effects), but would still be analysed within their starting cohort. On one hand, this limitation could attenuate the association in question if it takes prolonged exposure for AChEIs and/or memantine to work. On the other hand, this limitation could also lead to overestimation of this association if treatment of behavioural problems improved users’ awareness of the specific drug benefits, and such people receive more or are willing to take more antidementia medication. However, there is no evidence supporting such a possible cumulative effect. In addition, the sensitivity analysis with a longer drug initiation window confirmed our primary results.
Thirdly, although at baseline there were no significant differences in the use of antipsychotics and antidepressants among our cohorts, over time, there may have been a different pattern in groups of antidementia medication users. The possibility is that there may have less use of antipsychotics and antidepressants in those using antidementia medications, as these groups already have more polypharmacy. This may have an influence on the associations in which we were interested, when our models only controlled for antipsychotics and antidepressants at baseline as confounders. However, we reported both unadjusted and adjusted estimations, and both estimations were consistent.
Fourthly, we used scaled scores from multiple cognitive tests and imputed some missing cognitive data, which might have affected (underestimated or overestimated) associations of interest. However, results from Cox regressions with and without cognitive status were similar and robust.
Finally, because this was a sample based on clinical diagnosis, we were unable to determine how much the association with mortality was driven by those with mixed AD/DLB pathology, compared to more “pure” AD cases, since concurrent AD pathology has been shown to be associated with poorer outcome in some studies on DLB [46–48].
One unanswered question from this work is the dose–response relationship between AChEIs and the outcomes we examined. Previous evidence indicates that higher doses of AChEIs appear to lead to better survival benefits in patients with AD  and better improvement on neuropsychiatric symptoms in patients with DLB . Additionally, we did not examine specific AChEIs. There are differences between donepezil, rivastigmine, and galantamine in the magnitude of the association with mortality in patients with AD  and in associations with neuropsychiatric symptoms and general cognition in patients with DLB . One study suggested a possible cutoff dose for these 3 AChEIs to reduce the risk of death in patients with AD . However, we were unable to check these 2 questions, because our real-world data did not provide sufficient DLB cases with single-AChEI therapy, we lacked precise data on dosage, and patients’ medication dosages may have changed frequently. These unanswered questions require further study.
In patients with DLB, taking AChEIs (with or without memantine) was associated with a significantly reduced risk of death, and reduced length of hospital stay due to unplanned admissions for physical health problems. Our findings provided new evidence for the possible benefits of AChEI treatment for those with DLB and suggest corresponding RCT studies. In addition, further studies should investigate possible mechanisms for the benefit of increased survival.
S1 Appendix. Additional information on study methods and findings.
We thank Jonathan Lewis for providing us with data management assistance.
The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care.
- 1. Vann Jones SA, Brien JT. The prevalence and incidence of dementia with Lewy bodies: a systematic review of population and clinical studies. Psychol Med. 2014;44(4):673–683. pmid:23521899
- 2. Mueller C, Ballard C, Corbett A, Aarsland D. The prognosis of dementia with Lewy bodies. Lancet Neurol. 2017;16(5):390–398. pmid:28342649
- 3. Kane JPM, Surendranathan A, Bentley A, Barker SAH, Taylor JP, Thomas AJ, et al. Clinical prevalence of Lewy body dementia. Alzheimers Res Ther. 2018;10(1):19. pmid:29448953
- 4. Mueller C, Perera G, Rajkumar AP, Bhattarai M, Price A, O’Brien JT, et al. Hospitalization in people with dementia with Lewy bodies: Frequency, duration, and cost implications. Alzheimers Dement (Amst). 2018;10:143–152. pmid:29780862
- 5. Tan ECK, Johnell K, Garcia-Ptacek S, Haaksma ML, Fastbom J, Bell JS, et al. Acetylcholinesterase inhibitors and risk of stroke and death in people with dementia. Alzheimers Dement. 2018;14(7):944–951. pmid:29706487
- 6. Xu H, Garcia-Ptacek S, Jonsson L, Wimo A, Nordstrom P, Eriksdotter M. Long-term Effects of Cholinesterase Inhibitors on Cognitive Decline and Mortality. Neurology. 2021;96(17):e2220–e2230. pmid:33741639
- 7. Mueller C, Perera G, Hayes RD, Shetty H, Stewart R. Associations of acetylcholinesterase inhibitor treatment with reduced mortality in Alzheimer’s disease: a retrospective survival analysis. Age Ageing. 2018;47(1):88–94. pmid:28655175
- 8. Nordstrom P, Religa D, Wimo A, Winblad B, Eriksdotter M. The use of cholinesterase inhibitors and the risk of myocardial infarction and death: a nationwide cohort study in subjects with Alzheimer’s disease. Eur Heart J. 2013;34(33):2585–2591. pmid:23735859
- 9. Emre M, Aarsland D, Albanese A, Byrne EJ, Deuschl G, De Deyn PP, et al. Rivastigmine for dementia associated with Parkinson’s disease. N Engl J Med. 2004;351(24):2509–2518. pmid:15590953
- 10. Stinton C, McKeith I, Taylor JP, Lafortune L, Mioshi E, Mak E, et al. Pharmacological Management of Lewy Body Dementia: A Systematic Review and Meta-Analysis. Am J Psychiatry. 2015;172(8):731–742. pmid:26085043
- 11. Chu CS, Yang FC, Tseng PT, Stubbs B, Dag A, Carvalho AF, et al. Treatment Efficacy and Acceptabilityof Pharmacotherapies for Dementia with Lewy Bodies: A Systematic Review and Network Meta-Analysis. Arch Gerontol Geriatr. 2021;96:104474. pmid:34256210
- 12. Meng YH, Wang PP, Song YX, Wang JH. Cholinesterase inhibitors and memantine for Parkinson’s disease dementia and Lewy body dementia: A meta-analysis. Exp Ther Med. 2019;17(3):1611–1624. pmid:30783428
- 13. Wang HF, Yu JT, Tang SW, Jiang T, Tan CC, Meng XF, et al. Efficacy and safety of cholinesterase inhibitors and memantine in cognitive impairment in Parkinson’s disease, Parkinson’s disease dementia, and dementia with Lewy bodies: systematic review with meta-analysis and trial sequential analysis. J Neurol Neurosurg Psychiatry. 2015;86(2):135–143. pmid:24828899
- 14. Price A, Farooq R, Yuan JM, Menon VB, Cardinal RN, O’Brien JT. Mortality in dementia with Lewy bodies compared with Alzheimer’s dementia: a retrospective naturalistic cohort study. BMJ Open. 2017;7(11):e017504. pmid:29101136
- 15. Herbert A, Wijlaars L, Zylbersztejn A, Cromwell D, Hardelid P. Data Resource Profile: Hospital Episode Statistics Admitted Patient Care (HES APC). Int J Epidemiol. 2017;46(4):1093–i. pmid:28338941
- 16. Cardinal RN. Clinical records anonymisation and text extraction (CRATE): an open-source software system. BMC Med Inform Decis Mak. 2017;17(1):50. pmid:28441940
- 17. McKeith IG, Boeve BF, Dickson DW, Halliday G, Taylor JP, Weintraub D, et al. Diagnosis and management of dementia with Lewy bodies: Fourth consensus report of the DLB Consortium. Neurology. 2017;89(1):88–100. pmid:28592453
- 18. Moylett S, Price A, Cardinal RN, Aarsland D, Mueller C, Stewart R, et al. Clinical Presentation, Diagnostic Features, and Mortality in Dementia with Lewy Bodies. J Alzheimers Dis. 2019;67(3):995–1005. pmid:30776008
- 19. Ministry of Housing CLG. English indices of deprivation 2015. 2015. https://www.gov.uk/government/statistics/english-indices-of-deprivation-2015.
- 20. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–383. pmid:3558716
- 21. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130–1139. pmid:16224307
- 22. Horton AM Jr, Alana S. Validation of the Mini-Mental State Examination. Int J Neurosci. 1990;53(2–4):209–212. pmid:2265941
- 23. Nasreddine ZS, Phillips NA, Bedirian V, Charbonneau S, Whitehead V, Collin I, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53(4):695–699. pmid:15817019
- 24. Hsieh S, Schubert S, Hoon C, Mioshi E, Hodges JR. Validation of the Addenbrooke’s Cognitive Examination III in frontotemporal dementia and Alzheimer’s disease. Dement Geriatr Cogn Disord. 2013;36(3–4):242–250. pmid:23949210
- 25. Campbell NL, Maidment I, Fox C, Khan B, Boustani M. The 2012 update to the anticholinergic cognitive burden scale. J Am Geriatr Soc. 2013;61(S1):S142–S143.
- 26. Wing JK, Curtis RH, Beevor AS. HoNOS: Health of the Nation Outcome Scales: Report on Research and Development July 1993-December 1995. London: Royal College of Psychiatrists; 1996.
- 27. Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. New York, NY: Springer; 2000.
- 28. McCaffrey DF, Griffin BA, Almirall D, Slaughter ME, Ramchand R, Burgette LF. A tutorial on propensity score estimation for multiple treatments using generalized boosted models. Stat Med. 2013;32(19):3388–3414. pmid:23508673
- 29. STROBE. STROBE checklist for cohort studies. 2007. https://www.strobe-statement.org/fileadmin/Strobe/uploads/checklists/STROBE_checklist_v4_cohort.pdf.
- 30. Perry EK, Haroutunian V, Davis KL, Levy R, Lantos P, Eagger S, et al. Neocortical cholinergic activities differentiate Lewy body dementia from classical Alzheimer’s disease. Neuroreport. 1994;5(7):747–749. pmid:8018843
- 31. Francis PT. The interplay of neurotransmitters in Alzheimer’s disease. CNS Spectr. 2005;10(11 Suppl 18):6–9. pmid:16273023
- 32. O’Brien JT, Holmes C, Jones M, Jones R, Livingston G, McKeith I, et al. Clinical practice with anti-dementia drugs: A revised (third) consensus statement from the British Association for Psychopharmacology. J Psychopharmacol. 2017;31(2):147–168. pmid:28103749
- 33. National Institute for Health and Clinical Excellence. Donepezil, galantamine, rivastigmine and memantine for the treatment of Alzheimer’s disease. 2011. https://www.nice.org.uk/guidance/ta217/resources/donepezil-galantamine-rivastigmine-and-memantine-for-the-treatment-of-alzheimers-disease-pdf-82600254699973.
- 34. Matsunaga S, Kishi T, Yasue I, Iwata N. Cholinesterase Inhibitors for Lewy Body Disorders: A Meta-Analysis. Int J Neuropsychopharmacol. 2015;19(2). pmid:26221005
- 35. Armstrong MJ, Alliance S, Corsentino P, DeKosky ST, Taylor A. Cause of Death and End-of-Life Experiences in Individuals with Dementia with Lewy Bodies. J Am Geriatr Soc. 2019;67(1):67–73. pmid:30291740
- 36. Spears CC, Besharat A, Monari EH, Martinez-Ramirez D, Almeida L, Armstrong MJ. Causes and outcomes of hospitalization in Lewy body dementia: A retrospective cohort study. Parkinsonism Relat Disord. 2019;64:106–111. pmid:30930058
- 37. Handa T, Katare RG, Kakinuma Y, Arikawa M, Ando M, Sasaguri S, et al. Anti-Alzheimer’s drug, donepezil, markedly improves long-term survival after chronic heart failure in mice. J Card Fail. 2009;15(9):805–811. pmid:19879468
- 38. Reale M, Iarlori C, Gambi F, Feliciani C, Salone A, Toma L, et al. Treatment with an acetylcholinesterase inhibitor in Alzheimer patients modulates the expression and production of the pro-inflammatory and anti-inflammatory cytokines. J Neuroimmunol. 2004;148:162–71. pmid:14975597
- 39. Borroni B, Agosti C, Martini G, Volpi R, Brambilla C, Caimi L, et al. Cholinesterase inhibitors exert a protective effect on endothelial damage in Alzheimer disease patients. J Neurol Sci. 2005;229–230:211–213. pmid:15760641
- 40. Isik AT, Soysal P, Stubbs B, Solmi M, Basso C, Maggi S, et al. Cardiovascular Outcomes of Cholinesterase Inhibitors in Individuals with Dementia: A Meta-Analysis and Systematic Review. J Am Geriatr Soc. 2018;66(9):1805–1811. pmid:29851022
- 41. Rongve A, Vossius C, Nore S, Testad I, Aarsland D. Time until nursing home admission in people with mild dementia: comparison of dementia with Lewy bodies and Alzheimer’s dementia. Int J Geriatr Psychiatry. 2014;29(4):392–398. pmid:23943275
- 42. Larsson V, Engedal K, Aarsland D, Wattmo C, Minthon L, Londos E. Quality of life and the effect of memantine in dementia with lewy bodies and Parkinson’s disease dementia. Dement Geriatr Cogn Disord. 2011;32(4):227–234. pmid:22122992
- 43. Choi SH, Park KW, Na DL, Han HJ, Kim EJ, Shim YS, et al. Tolerability and efficacy of memantine add-on therapy to rivastigmine transdermal patches in mild to moderate Alzheimer’s disease: a multicenter, randomized, open-label, parallel-group study. Curr Med Res Opin. 2011;27(7):1375–1383. pmid:21561398
- 44. Ballard C, Lane R, Barone P, Ferrara R, Tekin S. Cardiac safety of rivastigmine in Lewy body and Parkinson’s disease dementias. Int J Clin Pract. 2006;60(6):639–645. pmid:16805745
- 45. Rosenbloom MH, Finley R, Scheinman MM, Feldman MD, Miller BL, Rabinovici GD. Donepezil-associated bradyarrhythmia in a patient with dementia with Lewy bodies (DLB). Alzheimer Dis Assoc Disord. 2010;24(2):209–211. pmid:20505440
- 46. Lemstra AW, de Beer MH, Teunissen CE, Schreuder C, Scheltens P, van der Flier WM, et al. Concomitant AD pathology affects clinical manifestation and survival in dementia with Lewy bodies. J Neurol Neurosurg Psychiatry. 2017;88(2):113–118. pmid:27794030
- 47. Kraybill ML, Larson EB, Tsuang DW, Teri L, McCormick WC, Bowen JD, et al. Cognitive differences in dementia patients with autopsy-verified AD, Lewy body pathology, or both. Neurology. 2005;64(12):2069–2073. pmid:15985574
- 48. Howlett DR, Whitfield D, Johnson M, Attems J, O’Brien JT, Aarsland D, et al. Regional Multiple Pathology Scores Are Associated with Cognitive Decline in Lewy Body Dementias. Brain Pathol. 2015;25(4):401–408. pmid:25103200