A European Academy of Neurology guideline on medical management issues in dementia

A European Academy of Neurology guideline on medical management issues in dementia K. S. Frederiksen a , C. Cooper b , G. B. Frisoni c ,L.Fr € olich d , J. Georges e , M. G. Kramberger f , C. Nilsson g,h , P. Passmore i , L. Mantoan Ritter j , D. Religa k,l , R. Schmidt m , E. Stefanova n , A. Verdelho o , M. Vandenbulcke p , B. Winblad q,l and G. Waldemar a a Department of Neurology, Danish Dementia Research Center, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark; b Department of Clinical Educational and Health Psychology, University College London, London, UK; c Memory Clinic, University Hospital of Geneva, University of Geneva, Geneva, Switzerland; d Department of Geriatric Psychiatry, Zentralinstitut f€ur Seelische Gesundheit Mannheim, Medical Faculty Mannheim, University of Heidelberg, Mannheim, Germany; e Alzheimer Europe, Luxemburg, Luxembourg; f Department of Neurology, Center for Cognitive Impairments, University Medical Center Ljubljana, and Medical faculty, University of Ljubljana, Ljubljana, Slovenia; g Department of Neurology and Rehabilitation Medicine, Sk ane University Hospital, Lund, Sweden; h Clinical Memory Research Unit, Department of Clinical Sciences, Lund University, Malm€ o, Sweden; i Queens University, Belfast, UK; j Epilepsy Centre, King’s College NHS Foundation Trust, King´s College London, London, UK; k Division of Clinical Geriatrics, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Neo, Stockholm, Sweden; l Karolinska University Hospital, Theme Aging, Huddinge, Sweden; m University Clinic for Neurology, Medical University of Graz, Graz, Austria; n Medical Faculty, Neurology clinic CCS, University of Belgrade, Belgrade, Serbia; o Department of Neurosciences and Mental Health, CHLNorte-Hospital de Santa Maria, IMM, ISAMB, Faculty of Medicine, University of Lisbon, Lisbon, Portugal; p Department of Neurosciences, KU Leuven, Geriatric Psychiatry, University Hospitals Leuven, Leuven, Belgium; and q Division of Neurogeriatrics, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, BioClinicum, Solna, Sweden Keywords: antipsychotics, dementia, epilepsy, follow-up, guideline, medical management, pain, vascular Received 6 May 2020 revision requested 12 June 2020 Accepted 13 June 2020 European Journal of Neurology 2020, 0: 1–16 doi:10.1111/ene.14412 Background and purpose: Dementia is one of the most common disorders and is associated with increased morbidity, mortality and decreased quality of life. The present guideline addresses important medical management issues including systematic medical follow-up, vascular risk factors in dementia, pain in dementia, use of antipsychotics in dementia and epilepsy in dementia. Methods: A systematic review of the literature was carried out. Based on the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) framework, we developed a guideline. Where recommendations based on GRADE were not possible, a good practice statement was formulated. Results: Systematic management of vascular risk factors should be performed in patients with mild to moderate dementia as prevention of cerebrovascular pathology may impact on the progression of dementia (Good Practice statement). Individuals with dementia (without previous stroke) and atrial fibrillation should be treated with anticoagulants (weak recommendation). Discontinuation of opioids should be considered in certain individuals with dementia (e.g. for whom there are no signs or symptoms of pain or no clear indication, or suspicion of side effects; Good Practice statement). Behavioral symptoms in persons with dementia should not be treated with mild analgesics (weak recommendation). In all patients with dementia treated with opioids, assessment of the individual risk–benefit ratio should be performed at regular intervals. Regular, preplanned medical follow-up should be offered to all patients with dementia. The setting will depend on the organization of local health services and should, as a minimum, include general practitioners with easy access to dementia specialists (Good Practice statement). Individuals Correspondence: K. S. Frederiksen, Department of Neurology, Danish Dementia Research Center, Rigshospitalet, University of Copenhagen, Blegdamsvej 9, 2100 Copenhagen, Denmark (Tel.: +45 35456911; e-mail: Kristian.steen.fr[email protected]). ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology 1 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. GUIDELINES EUROPEAN JOURNAL OF NEUROLOGY with dementia and agitation and/or aggression should be treated with atypical antipsychotics only after all non-pharmacological measures have been proven to be without benefit or in the case of severe self-harm or harm to others (weak recommendation). Antipsychotics should be discontinued after cessation of behavioral disturbances and in patients in whom there are side effects (Good Practice statement). For treatment of epilepsy in individuals with dementia, newer anticonvulsants should be considered as first-line therapy (Good Practice statement). Conclusion: This GRADE-based guideline offers recommendations on several important medical issues in patients with dementia, and thus adds important guidance for clinicians. For some issues, very little or no evidence was identified, highlighting the importance of further studies within these areas. Introduction An estimated 10.5 million people in Europe live with dementia, with the number expected to increase to 13.4 million in 2030 [1]. Many patients with dementia are elderly persons with comorbidities, such as cardiovascular conditions and diabetes, and also have an increased risk of developing epilepsy and behavioral symptoms. Opioids, anti-epileptic drugs (AEDs) and antipsychotics can worsen cognitive function in persons with dementia. Moreover, treatment with antipsychotics in this population is associated with an increased risk of cardiovascular accidents and mortality [2]. Inequality in access to physicians means that not all patients with dementia are offered systematic medical follow-up, and therefore treatable conditions may be overlooked or mismanaged, leading to accelerated functional decline, hospitalizations and increased mortality, and thus increased healthcare costs. Polypharmacy with adverse interactions and an unfavorable risk–benefit ratio is another frequent occurrence in this population, which requires systematic medical follow-up [3,4]. Moreover, for patients with dementia, treatment practices for associated medical conditions may vary due to uncertainty with regard to the balance between benefits and adverse effects, and due to a lack of scientific certainty. The scope of the Guideline is to address pressing medical issues in patients with dementia typically managed by a medical doctor in both primary and secondary care sectors. Despite increasing focus on the underlying etiology of dementia and biomarkerbased diagnosis [5,6], we have chosen to focus on the syndrome of dementia since many of the management issues are shared across disorders. Moreover, in a substantial number of patients, an etiological diagnosis may remain elusive. A number of European Academy of Neurology (EAN) and European Federation of Neurological Societies guidelines exist on different aspects of diagnosis and treatment of dementia disorders [7–10], none of which overlap with the present Guideline. The present Guideline is based on systematic reviews of the evidence and the use of Grading of Recommendations, Assessment, Development and Evaluations (GRADE) framework [11] for the development of guidelines. Both are standard operating procedures for the EAN [12] regarding guideline development. Methods The Guideline was proposed at a meeting of the EAN Scientific Panel on Dementia and Cognitive Disorders in 2016. From this group a Task Force was convened, based on previous clinical experience in dementia management and research and geographical representation of the EAN member states. The EAN epilepsy panel was invited to appoint a representative (L.M.R.; following the decision to include treatment of epilepsy in dementia in the Guideline). Furthermore, related organizations were invited to appoint representatives for the task force: the European Alzheimer Disease Consortium [dementia experts within geriatrics (B.W.) and old age psychiatry (L.F.)], the European Geriatric Medicine Society (P.P.) and the European Association of Geriatric Psychiatry (M.V.). The patient organization, Alzheimer Europe, was invited to appoint a representative (J.G.). Within the scope of medical management issues, the Task Force decided to address and prioritize those issues which were deemed most pressing. This included issues where there were significant variations in treatment across centers, regions and countries, unmet needs in relation to the management issues, additional costs associated with mismanagement, and ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology 2K.S. FREDERIKSEN ET AL. where mismanagement was associated with negative health consequences. The process by which the issues were selected were as follows: first a brain-storming session was carried out at a meeting of the entire EAN Scientific panel on Dementia and Cognitive Disorders. The brain-storming session resulted in the production of a list of possible issues to be addressed in the Guideline. This was followed by initial scoping searches, and further discussion and consensus-seeking via e-mail correspondence between members of the Task Force. Full consensus was reached for all recommendations. A list of five issues was agreed upon: systematic medical follow-up in dementia; management of vascular risk factors in dementia; management of pain in dementia; treatment of agitation/aggressive behavior with antipsychotics in dementia; and management of epilepsy in dementia. The Guideline was developed according to the Standard Operating Procedures of the EAN as described in the paper by Leone et al. [12], which included the development of an a priori protocol approved by the EAN (Appendix S1), and use of the GRADE framework [11] for development of guidelines, followed by final approval of the Guideline by the EAN Scientific Committee (Date of approval 14 April 2020). All co-authors were familiar with the GRADE framework. Please refer to other papers for details of the GRADE framework (e.g. Guyatt et al. [13]). The Task Force met at two in-person meetings. At the first meeting and according to the GRADE methodology a number of PICO (Population, Intervention, Comparator, Outcome) questions were formulated. At the final face-to-face meeting where all but two of the co-authors were present, the level of the quality of the body of the evidence (high, moderate, low, very low), the direction of the recommendation (for or against) and the strength of recommendation (’weak’ or ’strong’) were decided by consensus. This process also applied to those research questions where a Good Practice statement was made. The two co-authors not present were consulted following the meeting via email on whether they were in agreement with the decisions reached at the meeting. In instances where there was no evidence, the Task Force gave a Good Practice statement. ’Weak’ and ’strong’ have specific meanings which relate to the interpretation of a given recommendation within GRADE. Work on the Guideline was divided amongst all authors, so that two persons were responsible for the work on each topic, with K.F. and G.W. coordinating the work. Remaining members of the Task Force were regularly updated and involved in all aspects of the work throughout the process by the Task Force chairperson via email, and input was requested when necessary. Search strategy, selection of studies, data extraction, risk-of-bias assessment and data synthesis Six databases (including MEDLINE and Embase) were searched using relevant search terms including terms from controlled vocabularies and thesauruses. A single search string was developed for each management issue (See Appendix S1 for full search strategy). Each search strategy was reviewed using the Peer Review of Electronic Search Strategies (PRESS) checklist [14]. Eligible studies were randomized controlled trials (RCTs) which reported on relevant interventions and outcomes. For pain and vascular risk factors, observational studies were also included, since it was the opinion of the Task Force that, for these two issues, observational studies might contribute important evidence. Two of the co-authors (follow-up: D.R., K.F.; epilepsy: C.N., K.F.; vascular: A.V., K.F.; pain: M.K., K.F.; antipsychotics: E.S., K.F.) independently screened the identified references using an online software tool and piloted questionnaires first on title and abstract and secondly on full-text level. Data extraction was carried out by one author (K.F.) in a piloted Excel spreadsheet, and subsequently checked by the second co-author as with the screening. Financial disclosure statements and conflict of interest statements were extracted for all included studies and are reported in Table S1. Risk of bias for RCTs was assessed using the Cochrane´s risk-of-bias assessment tool, version 1 [15] and, for observational studies, the Newcastle–Ottawa Quality Assessment Scale [16] was used. Data synthesis could be either as meta-analysis or narrative synthesis depending on the number of studies identified and which type of synthesis heterogeneity across studies permitted (See Appendix S1 for further details on the methods). Due to space restraints, results will largely be presented in supplementary tables and figures referred to throughout the text. Section 1: Systematic medical follow-up in dementia Background for systematic medical follow-up in dementia Many patients with dementia are not followed by a medical doctor after initial diagnosis or may only be followed regularly for less than a year. Although no disease-modifying treatment is available for any of the neurodegenerative dementias, adequate management of associated medical conditions may reduce the burden of disease and functional decline [17–19]. In patients with dementia, some guidelines recommend ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology EAN GUIDELINE ON MEDICAL MANAGEMENT ISSUES IN DEMENTIA 3 that patients should be followed by a specialist (e.g. neurologist or psychiatrist or geriatrician) [20]. However, many patients with dementia may never see a specialist [21]. Lack of insight and autonomy, an early and prominent symptom in many patients with dementia [22], may diminish the ability to make decisions about treatment [23] and is also associated with caregiver burden [24]. Cognitive and non-cognitive symptoms of dementia may constitute some of the barriers which impair access to medical care for patients with dementia [25]. Research question 1.1 Should home-living (non-institutionalized) patients with dementia be offered systematic medical follow-up in a memory clinic setting? PICO question Population: Home living (non-institutionalized) patients with dementia. Intervention: Planned structured follow-up in the form of consultations offered in a medical dementia specialist team. Comparator: Usual care. Outcome: 1. Institutionalization (Important); 2. Caregiver burden (Important); 3. Acute hospital admissions (Important); 4. Activities of daily living (ADL; Critical). Summary of evidence We identified 400 references (after deduplication) for screening. No studies were eligible for inclusion (Figure S1 and Table S3). Level of evidence Not applicable. Good Practice statement The authors conclude that patients with dementia should be offered regular, preplanned medical followup. The setting will depend on the organization of local health services and should as a minimum include general practitioners with easy access to dementia specialists. Supplemental considerations for systematic medical follow-up in dementia We believe that regular, prescheduled follow-up in patients with dementia is important in order to address psychosocial as well as medical issues during the course of the condition. For optimal medical follow-up, visits in a multiprofessional setting (e.g. a memory clinic), where possible, are important. Interaction between healthcare providers and patients with dementia is challenging due to lack of insight and self-awareness and other cognitive and non-cognitive symptoms of dementia leading to impaired autonomy [22,23], and may thus be better handled in a multiprofessional setting. Furthermore, we believe that multiprofessional memory clinics may be better suited for the management of cognitive, neuropsychiatric and other specific symptoms related to dementia disorders. Management of complex comorbidities, providing access for patients to participate in research and for meeting the needs of the caregiver are other important reasons. However, regular and preplanned medical follow-up may also be carried out in other settings, depending on the local organization of healthcare, for example, at the general practitioner´s office. Easy access to a dementia specialist, such as a neurologist, psychiatrist or geriatrician, is desirable. Certain patients (e.g. patients with frontotemporal dementia, patients with Lewy-body dementia, patients with diagnostic uncertainty, patients with severe psychiatric and behavioral symptoms, and patients with familial disorders) should be given high priority for follow-up in a specialist setting, for example, a memory clinic setting. Section 2: Management of vascular risk factors in dementia Background for management of vascular risk factors in dementia Vascular risk factors are associated with an increased risk of cognitive decline and dementia [26] and with the rate of progression of dementia and the occurrence of brain pathology such as small vessel disease [27,28], lacunar infarcts [29,30] and microbleeds [31]. Hence, optimal management of vascular risk factors may potentially modify the disease course in dementia. On the other hand, pharmacological treatment of vascular risk factors is associated with unwanted side effects such as an increased risk of intracerebral hemorrhage with anticoagulants [32] and dizziness and falls with anti-hypertensives. However, differentiating treatment based on cognitive status and the aforementioned side effects may not be warranted. It is likely that in many patients with dementia, vascular conditions and risk factors are identified at the time of the dementia diagnosis. Usually medical doctors would treat such conditions. However, the medical doctor may be in doubt as to whether there is any additional benefit for the prognosis of the dementia condition, or additional harm in patients with dementia, associated with treatment of the vascular condition. For ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology 4K.S. FREDERIKSEN ET AL. example, treatment with anticoagulation may be withheld in patients with dementia in instances where treatment would be initiated in non-demented patients due to potentially differential risk–benefit profiles. A number of medical societies have addressed the issue of anticoagulation in patients with dementia, either implicitly or explicitly. The American Academy of Neurology guidelines for stroke prevention in atrial fibrillation conclude that data in patients with atrial fibrillation who have ’advanced’ dementia or frequent falls are insufficient to determine whether anticoagulants are safe and effective [33]. European Stroke Organization guidelines do not address the use of anticoagulants in dementia patients, but do not recommend them in patients with comorbid conditions (e.g. falls, uncontrolled epilepsy, gastrointestinal bleeding or poor compliance) [34]. The European Society of Cardiology guidelines state that anticoagulation should only be withheld ’in selected patients with dementia where compliance and adherence cannot be ensured by a caregiver’ [35]. Research question 2.1 Should patients with atrial fibrillation (without previous stroke, but where there is indication for anticoagulants), and dementia be treated with anticoagulants? PICO question Population: Patients with dementia and atrial fibrillation and indication for treatment with anticoagulants and no previous stroke or transitory ischemic attack. Intervention: Treatment with new oral anticoagulants or warfarin. Comparator: No treatment with new oral anticoagulants or warfarin. Outcome: 1. Major hemorrhagic events (Critical); 2. Global cognitive function (Important); 3. Mortality (Important); 4. Ischemic cerebrovascular event (Critical). Summary of evidence We identified 7135 (after deduplication) references (See Figure S2 and Table S3). We found one study that was eligible for inclusion [36]. In a retrospective cohort study in patients with dementia and atrial fibrillation (but no previous stroke), results from 3724 patients followed for a median (interquartile range) of 669 (805) days, were reported. Patients were either not treated (for unknown reasons) or treated with warfarin. Treatment was not associated with an increased risk of major hemorrhagic events [hazard ratio 1.07, 95% confidence interval (CI) 0.82–1.39], a decreased risk of mortality (hazard ratio 0.88, 95% CI 0.77–1.01) and ischemic stroke (hazard ratio 0.74, 95% CI 0.54–1.03). The study did not report on global cognitive function (Tables S4 and S5). A financial disclosure statement and a conflict of interest statement were reported for the study (Table S1). Level of evidence The level of evidence was graded as very low (Tables S5 and S6). Recommendation The authors conclude that there should be a weak recommendation for treatment with anticoagulants in patients with dementia (without previous stroke) and atrial fibrillation. Justification for recommendations It is the opinion of the authors that the recommendation may be extended to include non-vitamin K oral anticoagulants, as these offer a better protection against ischemic stroke and a comparable safety profile [37]. There is a trend in the directions of the point estimates towards a beneficial profile, which importantly is in line with many other studies clearly demonstrating a beneficial effect of anticoagulation in patients with atrial fibrillation but no dementia [38]. In the opinion of the authors there is no reason to believe that some patients with dementia would not have a similar benefit from anticoagulation. Atrial fibrillation remains the commonest cause of ischemic stroke in the older population [39], and a study showed that elderly persons would accept a higher risk of a hemorrhagic event for a smaller reduction in the risk of an ischemic stroke [40]. There is no reason to believe that patients with dementia differ in this regard. Research question 2.2 Does systematic management of vascular risk factors in patients with dementia slow the progression of dementia? PICO question Population: Patients with dementia and hypertension, hypercholesterolemia and type 2diabetes mellitus. Intervention: Systematic management of vascular risk factors (concomitant hypertension, hypercholesterolemia, type 2 diabetes mellitus). Comparator: Usual care. Outcome: 1. Institutionalization (Important), 2. Global cognitive function (Critical), 3. Mortality (Important), 4. ADL (Important). ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology EAN GUIDELINE ON MEDICAL MANAGEMENT ISSUES IN DEMENTIA 5 Summary of evidence One single-center retrospective cohort study was identified which reported on dementia patients with all three vascular risk factors [41] (Figure S2 and Tables S3 and S7). Twenty patients were followed for a mean (SD) of 30 (6) months. Treatment of one or more vascular risk factors was associated with less decline on Mini Mental State Examination [mean difference 2 (95 % CI 3.1 to 7.2)] compared to no treatment at all (Table S8). The study did not report data on the three remaining outcomes. No grading of evidence, using the GRADE framework, was carried out since it was the opinion of the authors that the evidence was insufficient to make a recommendation. Financial disclosures and a statement on conflict of interest were reported for the study (Table S1). Level of evidence Not applicable. Good Practice statement The authors conclude that systematic management of vascular risk factors should be performed in patients with mild to moderate dementia since prevention of cerebrovascular pathology may impact on the progression of dementia. Justification for statement We believe that systematic management of vascular risk factors is as important in patients with mild to moderate dementia as in patients without dementia as these factors are associated with adverse health outcomes and because an effect is unlikely to differ between patients with and without dementia [26–31]. It is the opinion of the authors that the value and attitudes of patients with dementia would not go against treatment, at least for a majority of patients. Since treatment may be associated with side effects, we believe that the recommendations should not extend to patients with severe dementia as the risk–benefit ratio becomes less clear, and because a short life expectancy may exclude any benefit from treatment. Supplemental considerations for management of vascular risk factors in dementia In patients with dementia, vascular risk factors may be neglected, and undertreated [42]. We believe that there is no reason to assume that patients with dementia benefit less from management of vascular risk factors than other patient groups. Furthermore, prevention of vascular pathology may provide an important added value in patients with vascular risk factors and dementia [43]. We believe that treatment should be individualized, issues of compliance should be considered, and measures to ensure compliance and safety should be instituted where necessary. Review of existing medication should be carried out for drugs which may increase the risk of vascular pathology. Whether or not to institute pharmacological treatment should, in the opinion of the authors, be carefully considered in patients with falls, low compliance and advanced dementia and when life expectancy may be short because of safety concerns and because there may be uncertainty regarding the balance between harm and benefit. Section 3: Management of pain in dementia Background for management of pain in dementia Painful conditions, such as arthritis are as frequent in patients with dementia as in patients without dementia [44]. Moreover, patients with advanced dementia may not be able to report pain adequately because of impairment of memory or language, and pain may be difficult to assess [45]. Further, pain has been found to be associated with behavioral symptoms in patients with dementia [46]. Therefore, identification and treatment of pain is important but poses several challenges, which may both lead to overtreatment and undertreatment. For example, pain associated with malignancies and treatment of pain in an end-of-life situation may warrant use of opioids. However, use of opioids for the treatment of agitation when it is suspected to be attributable to pain may lead to overuse, which can be detrimental and constitutes a potential safety issue [47]. Conversely, use of milder analgesics more routinely in the presence of behavioral changes may be a rational intervention [48]. Research question 3.1 In patients with dementia, should opioids be discontinued? PICO question Population: Patients with dementia who are treated with opioids. Intervention: Discontinuation of opioid treatment. Comparator: Continuation of opioid treatment. Outcome: 1. Psychotropic treatment (Important); 2. Global cognitive function (Critical); 3. Mortality (Important); 4. Pain (Critical); 5. Neuropsychiatric symptoms (Important). Summary of evidence We identified 1834 references (after deduplication) for screening. No references were eligible for inclusion (Figure S3 and Table S9). ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology 6K.S. FREDERIKSEN ET AL. Level of evidence Not applicable. Good Practice statement The authors conclude that discontinuation of opioids should be considered in patients for whom there are no complaints of pain and no clear indication, where mild analgesics have not been tried and in patients in whom there is suspicion of side effects, such as rapid cognitive decline, sedation, falls, respiratory problems, constipation, nausea or reduced appetite. Research question 3.2 Should behavioral symptoms in patients with dementia be treated with mild analgesics? PICO question Population: Patients with dementia and behavioral symptoms. Intervention: Treatment with mild analgesics (paracetamol). Comparator: No treatment with analgesics. Outcome: 1. Psychotropic treatment (Important); 2. Global cognitive function (Important); 3. Agitation/ aggression (Critical); 4. Neuropsychiatric symptoms (Important). Summary of evidence After screening, we included a single RCT [49] (Figure S3 and Table S9). In a placebo-controlled crossover study, 25 patients were randomized to either paracetamol or placebo. There was no difference in change in agitation over 8 weeks (as measured by the CohenMansfield Agitation inventory). Use of paracetamol was found to be associated with less use of as-needed antipsychotics [mean difference 0.28 (95% CI –1to0.45); Tables S10 and S11]. The study did not report on neuropsychiatric symptoms or global cognitive function. Financial disclosure statement and statement on conflict of interest was reported for the study (Table S1). Level of evidence The level of evidence was graded as very low (Tables S11 and S12). Recommendation The authors conclude that there should be a weak recommendation against treatment of behavioral symptoms in persons with dementia with mild analgesics. Justification for recommendations The study identified for the present PICO did not demonstrate an effect of treatment with paracetamol on behavioral symptoms in patients with dementia. Behavioral symptoms may be caused by many underlying causes, of which pain is one, and the rationale for the use of analgesics against behavioral symptoms would be to alleviate pain. Therefore, routine use of analgesics as a treatment for behavioral symptoms would neglect the identification of the true underlying cause and instead assume that in most cases pain is the causative factor. Moreover, the recommendation extends to all mild analgesics, as there is no reason to suspect efficacy will differ. Supplemental considerations for management of pain in dementia Assessment of pain in patients with dementia is important, but complicated by impaired communication, particularly in patients with advanced dementia [45]. We believe that behavioral signs must be included in the assessment of pain in patients who are impaired in verbal communication. Conversely, pain should also always be considered as a possible cause of behavioral signs. Conditions other than pain which may cause altered behavior must be ruled out before starting analgesics, and, if relevant, non-pharmacological treatment instituted. In patients receiving analgesics, always review treatment at regular intervals. We believe that when initiating treatment with analgesics, a stepwise approach starting with mild analgesics should be used. For safety reasons, long-acting opioids should not be introduced before short-acting opioids. Follow-up is important after initiation of analgesics (particularly opioids) to assess sedation, nausea and cognitive deterioration and after discontinuation to assess re-emergence of pain. Section 4: Treatment of agitation/aggressive behavior with antipsychotics in dementia Background for treatment of agitation/aggressive behavior with antipsychotics in dementia Aggression or agitation may occur during the course of the condition without obvious reasons or when patients with dementia experience stress (e.g. changes in the environment during hospitalization), as a result of a physical condition, or as side effects of drug treatment. A number of meta-analyses have found that the efficacy of antipsychotics for treating agitation/aggression in patients with dementia, is modest and confers increased mortality and risk of cerebrovascular accidents [50–52]. In 2007 European regulatory agencies and the US Food and Drug Administration (FDA) issued warnings regarding the use of atypical antipsychotics in patients with ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology EAN GUIDELINE ON MEDICAL MANAGEMENT ISSUES IN DEMENTIA 7 dementia. This was extended by the European Medicines Agency in 2008 and the FDA in 2009 to include a warning regarding the use of all antipsychotics, partly based on a number of large observational studies and meta-analyses that showed an increased mortality associated with treatment [53–59], which has been confirmed in subsequent studies [2,60]. However, antipsychotics may be used as first-line pharmacological treatment in agitation/aggression under certain preconditions (after all non-pharmacological measures have been proven to be without benefit or in the case of severe self-harm or harm to others) [61]. This includes the use of risperidone, haloperidol and aripiprazole. Antipsychotics confer modest benefits for short-term treatment of aggression/agitation in dementia [50–52], but these benefits have to be balanced against the risk of serious adverse events including increased mortality. The benefits are less clear-cut with long-term prescribing, but the mortality risk remains significantly elevated [53–59]. Therefore, minimizing and individualizing (with regular assessment of the individual risk–benefit ratio) the use of antipsychotics and ensuring that treatment with antipsychotics is only instituted when non-pharmacological treatment proves to be ineffective and when treatable causes of agitation/aggression have been ruled out, is necessary and will improve quality of life and function. A prescription or dose optimization of cholinesterase inhibitors or memantine should be carried out in parallel to other measures, since the effects of this may only be evident with a delay of 8 to 12 weeks. Because of the risk of severe hypersensitivity to antipsychotics in patients with Lewy-body dementia and Parkinson´s disease dementia, the recommendations for Research questions 4.1 and 4.2 do not extend to these patients. Research question 4.1 Should patients with dementia and agitation/aggressive behavior be treated with modern (atypical) antipsychotics compared to no pharmacological treatment? PICO question Population: Patients with dementia and agitation/aggressive behavior. Intervention: Treatment with aripiprazole, zotepine, olanzapine, quetiapine, risperidone or clozapine. Comparator: No pharmacological treatment. Outcome: 1. Mortality (Important); 2. Agitation/ Aggression (Critical); 3. Global cognitive function (Important); 4. Serious adverse events (Critical); 5. Caregiver burden (Important). Summary of evidence A total of 4058 references were screened and seven references were included [62–68] (Figure S4 and Table S13 and S14). Two references reported on different outcomes from the same study [65,66]. All studies included were RCTs. Studies were on risperidone (0.25 to 2 mg), quetiapine (25 to 200 mg) or olanzapine (2.5 to 15 mg). Meta-analysis was not performed due to heterogeneity in populations, length of interventions and outcome measures. There were no differences in mortality between interventions and across studies [62,64,65,67,68] (relative risk point estimates 0.85–2.08; not significant) or serious adverse events [62,63,66,67] (relative risk point estimates 0.89–1.9; not significant in all but one study [68]). However, there were, in general, very few events for the two outcomes, possibly because of the relatively low number of patients and short length of included trials (6– 36 weeks; four studies ≤12 weeks). For both global cognitive function and agitation/aggression scores, there was no difference in change score between treatment and placebo, except for a single study [67]. None of the studies reported on caregiver burden (Table S15). All studies reported financial statements. All but one study [67] reported receiving some or all of the funding for the study from pharmaceutical companies. One study did not include a statement on conflict of interest [62] (Table S1). Level of evidence The level of evidence was graded as low (Tables S15 and S16). Recommendation The authors conclude that there should be a weak recommendation against treatment of patients with dementia and agitation/aggressive behavior with modern (atypical) antipsychotics compared to no pharmacological treatment. Justification for recommendations It is the opinion of the authors that treatment of agitation and aggression (in the absence of severe selfharm or harm to others) should principally be directed at underlying causes, which in many instances are identifiable and to some extent treatable or manageable. Further, the effect of antipsychotics will (in the absence of psychotic symptoms) in many instances rely on a sedating effect, which in general is unwanted. Finally, use of antipsychotics in elderly and patients with dementia is associated with an increased risk of severe morbidity and mortality [2,53–60]. We believe that in light of this, and the fact that the included studies did not demonstrate an effect ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology 8K.S. FREDERIKSEN ET AL. of treatment, a majority of patients will prefer no pharmacological treatment. Research question 4.2 Should patients with dementia and agitation/aggressive behavior be treated with modern (atypical) antipsychotics compared to haloperidol? PICO question Population: Patients with dementia and agitation/aggressive behavior. Intervention: Treatment with aripiprazole, zotepine, olanzapine, quetiapine, risperidone or clozapine. Comparator: Haloperidol. Outcome: 1. Mortality (Important); 2. Agitation/ Aggression (Critical); 3. Global cognitive function (Important); 4. Serious adverse events (Critical); 5. Caregiver burden (Important). Summary of evidence Four references reporting on the effect of atypical antipsychotics versus haloperidol were identified (Figure S4 and Tables S13 and S17). Three studies were RCTs [69–71] and one was an RCT with a crossover design [72]. Three studies compared risperidone (0.25 to 2 mg) to haloperidol (0.25 to 3 mg), and one study compared it with olanzapine (2.5 to 7.5 mg). There were no events in the studies reporting on mortality [69,70,72] or serious adverse events [71,72]. With regard to agitation/aggression, two studies on risperidone reported that treatment with atypical antipsychotics improved scores [71,72], and two studies, one on risperidone, the other on olanzapine, reported the converse [69,70]. However, none of the studies reported measures of variance (Tables S18 and S19). None of the studies reported on caregiver burden. One study did not include a financial statement [69], and two studies reported receiving funding for the study from pharmaceutical companies [71,72]. None of the studies included a conflicts of interest statement. Level of evidence The level of evidence was graded as very low (Tables S18 and S19). Recommendation The authors conclude that there should be a weak recommendation for treatment of patients with dementia and agitation/aggressive behavior with modern (atypical) antipsychotics compared to haloperidol when pharmacological treatment of agitation/aggressive behavior is necessary. Among modern (atypical) antipsychotics, risperidone may be considered as firstline treatment when pharmacological treatment of agitation/aggressive behavior is necessary. Justification for recommendations Atypical antipsychotics is superior compared to placebo in reducing behavioral symptoms in patients with dementia, whereas haloperidol is not [73] and in elderly patients with various conditions, risperidone has been found to be as efficacious as haloperidol [74]. Moreover, several very large observational studies have demonstrated that haloperidol is associated with higher mortality, risk of pneumonia and cardiovascular disease compared to atypical antipsychotics (including risperidone) [53,54]. Research question 4.3 Should treatment with antipsychotics routinely be discontinued? PICO question Population: Patients dementia who are currently being treated with antipsychotics. Intervention: Discontinuation of antipsychotics. Comparator: Continuation of treatment. Outcome: 1. Mortality (Important); 2. Neuropsychiatric symptoms (Critical); 3. Global cognitive function (Important); 4. Serious adverse events (Critical); 5. Caregiver burden (Important). Summary of evidence Eight references [75–82] reporting on the effect of discontinuation of antipsychotics were identified (Figure S4 and Table S13 and S20) from a total of six studies. Three studies were single-center and three were multi-center RCTs. A number of references reported on the same two studies. In detail, Devanand et al. [7879] reported on the same study as was the case with Ballard et al. [8182]. In both instances, the studies reported on the same cohort, but outcomes relevant for the PICO were reported in two different studies (e.g. mortality in one study and neuropsychiatric symptoms in another). A recent Cochrane metaanalysis [83] on discontinuation of antipsychotics assessed three studies not included in the present Guideline. One study did not report on any of the outcomes for the PICO [84] and two did not test a relevant intervention [85,86] (i.e. did not test the intervention which was defined in the PICO; Table S13). One additional study which was evaluated as satisfying the inclusion criteria and none of the exclusion criteria for the present Guideline, was included in the present Guideline but had not been included in the ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology EAN GUIDELINE ON MEDICAL MANAGEMENT ISSUES IN DEMENTIA 9 and epilepsy: A systematic review and meta-analysis. Epilepsia 2017; 58: 962–972. 90. Beagle AJ, Darwish SM, Ranasinghe KG, La AL, Karageorgiou E, Vossel KA. Relative incidence of seizures and myoclonus in Alzheimer’s disease, dementia with lewy bodies, and frontotemporal dementia. J Alzheimer’s Dis 2017; 60: 211–223. 91. Helmstaedter C, Witt JA. Clinical neuropsychology in epilepsy: Theoretical and practical issues. Handb Clin Neurol 2012; 107: 437–459. 92. Vossel KA, Ranasinghe KG, Beagle AJ, et al. Incidence and impact of subclinical epileptiform activity in Alzheimer’s disease. Ann Neurol 2016; 80: 858–870. 93. Vu LC, Piccenna L, Kwan P, O’Brien TJ. New-onset epilepsy in the elderly. Br J Clin Pharmacol 2018; 84: 2208–2217. 94. Kaur U, Chauhan I, Gambhir IS, Chakrabarti SS. Antiepileptic drug therapy in the elderly: a clinical pharmacological review. Acta Neurol Belg 2019; 119: 163–173. 95. de Campos MS, Ayres LR, Morelo MRS, Carizio FAM, Pereira LRL. Comparative efficacy of antiepileptic drugs for patients with generalized epileptic seizures: systematic review and network meta-analyses. Int J Clin Pharm 2018; 40: 589–598. 96. Zaccara G, Giovannelli F, Giorgi FS, Franco V, Gasparini S, Benedetto U. Tolerability of new antiepileptic drugs: a network meta-analysis. Eur J Clin Pharmacol 2017; 73: 811–817. 97. Fu J, Peng L, Li J, Tao T, Chen Y. Effects of secondgeneration antiepileptic drugs compared to first-generation antiepileptic drugs on bone metabolism in patients with epilepsy: a meta-analysis. Horm Metab Res 2019; 51: 511–521. 98. Beyenburg S, Back C, Diederich N, Lewis M, Reuber M. Is valproate encephalopathy under-recognised in older people? A case series. Age Ageing 2007; 36: 344–346. 99. Norris SL, Bero L. GRADE Methods for Guideline Development: Time to Evolve? Ann Intern Med 2016; 165: 810. 100. Rafiq M, Boccia S. Application of the GRADE Approach in the Development of Guidelines and Recommendations in Genomic Medicine. Genomics Insights 2018; 11: 1–8. ©2020 The Authors. European Journal of Neurology published by John Wiley & Sons Ltd on behalf of European Academy of Neurology 16 K.S. FREDERIKSEN ET AL.