Pripravci valproata za agitaciju u demenciji
Abstract
Background
Agitation affects up to 70% of older people with dementia. Valproic acid derivatives have been used for the past 10 years to control agitation in dementia, but no systematic review of the effectiveness of this treatment has been published to date. A systematic review of 2004 examined three randomised, placebo‐controlled trials of the effect of valproate therapy on older people with dementia who were agitated. The review was updated (October 2008) to include two additional studies.
Objectives
To determine whether evidence supports the use of valproate preparations in the treatment of agitation of people with dementia.
Search methods
We searched ALOIS ‐ the Cochrane Dementia and Cognitive Improvement Group’s Specialized Register on 30 July 2010 using the terms: valproic OR valproate OR divalproex. ALOIS contains records from all major health care databases (The Cochrane Library, MEDLINE, EMBASE, PsycINFO, CINAHL, LILACS) as well as from many trials databases and grey literature sources.
Selection criteria
Randomized, placebo‐controlled trials with concealed allocation where agitation and dementia of participants were assessed
Data collection and analysis
1. Two reviewers extracted data from published trials
2. Odds ratios of average differences were calculated
3. Only "intention to treat" analyses were included
4. Analysis compared participants treated with valproic acid with controls
Main results
Meta‐analysis in 2004 of the pooled results was limited because of the following problems.
In Porsteinsson 2001, although the physicians having direct responsibility for patient care were blinded, a non‐blinded physician, who had no direct contact with these physicians, adjusted divalproex sodium dosage on the basis of reports from blinded raters and from confidential laboratory reports. Therefore, because the physician who controlled therapy knew which patients were receiving divalproex, the trial did not satisfy the criterion of concealed allocation.
In Tariot 2001, 54% of the treated patients dropped out compared with 29% of control patients. Of all treated patients, 22% dropped out because of adverse effects, and the study had to be discontinued prematurely.
The third trial (Sival 2002) had a cross‐over design. No results from the first phase of the study were available, and although the statistical section stated, "the t‐test for independent samples is used to analyse the two‐period cross‐over trial", because the samples were not independent ‐ they are the same patients in the treatment and placebo groups ‐ a question must be raised about the correctness of the analyses.
The valproate preparation used in the trials varied ‐ one used short‐acting sodium valproate, one long‐acting divalproex sodium, and the third early‐onset acting divalproex sodium. Average doses differed (480 mg/d ‐ 1000 mg/d), as did duration of therapy (3 weeks ‐ 6 weeks), and ways of evaluating patients and their response to therapy.
A limited meta‐analysis, pooling the results concerning adverse effects (Porsteinsson 2001, Tariot 2001) revealed the following: sedation occurred more frequently in patients treated with valproic acid than in controls. Urinary tract infection was more common among patients treated with valproic acid than controls.
An updated systematic review (October 2008) of two new studies (Tariot 2005, Herrmann 2007) applied meta‐analysis to the effect of valproate on agitation in demented patients and also combined these studies with the earlier reports to examine adverse effects among valproate treated patients. Because the study of Herrmann et al involved a cross‐over design, only those results from the first part of this study were included in the updated review.
The new meta‐analysis of pooled results showed no improvement of agitation among valproate treated patients, compared with controls, and showed an increase in adverse events (falls, infection, gastrointestinal disorders) among valproate treated patients.
Authors' conclusions
The updated review corroborates the earlier findings that valproate preparations are ineffective in treating agitation among demented patients, and that valproate therapy is associated with an unacceptable rate of adverse effects. More research on the use of valproate preparations for agitation of people with dementia is needed. On the basis of current evidence, valproate therapy cannot be recommended for management of agitation in dementia.
PICOs
Laički sažetak
Nema dokaza o djelotvornosti pripravaka valproata u liječenju agitacije u osoba s demencijom
Obnovljeni pregled (listopad 2008.) liječenja agitacije u dementnih bolesnika valproatom nije pokazao poboljšanje agitacije kod liječenih bolesnika u usporedbi s onima koji nisu bili liječeni, a također je pokazao veću stopu štetnih učinaka, kao što su padovi, infekcije i gastrointestinalni poremećaji (proljev, mučnina) kod onih koji su primali pripravke valproata. Iako su potrebna daljnja istraživanja koristi pripravaka valproata, trenutni dokazi ne podupiru primjenu ovog lijeka za kontrolu agitacije u osoba s demencijom.
Authors' conclusions
Background
Agitation has been reported in up to 70% of people with dementia. No formal definition of agitation has been published, although a widely accepted definition is: "inappropriate verbal, vocal, or motor activity that is not explained by needs or confusion per se" (Billig 1991; Cohen‐Mansfield 1989). The descriptors of agitation include wandering, crying out, aggressiveness, repetitive movements, and uncooperative behaviour. Agitation of demented patients has been shown to worsen caregiver burden, increase the risk of injury, and add to the need for institutionalizations.
Attempts to control agitation in dementia with behavioural modification, benzodiazepines, neuroleptics, antidepressants, and beta‐blocking agents have yielded varying results. During the past decade valproic acid, a drug employed initially to treat epilepsy (Trimble 1984), has been used to manage agitation of people with dementia (Herrmann 1998; Kasckow 1997; Porsteinsson 2001; Sival 2002; Tariot 2001). The mechanism of action of valproic acid is unclear; suggested mechanisms included enhancement of the intracerebral neurotransmitting agent, gamma butyric acid, anti‐manic action, and mood stabilizing effect (Lon 1995). Recently, a more readily tolerated pro‐drug for valproate, divalproex, has been used. This drug differs slightly from valproic acid in that peak blood levels occur later (3 to 6 hrs, compared with 3 hrs), but the dosage and half‐life of this drug are identical to those of valproic acid. Valproic acid has a record of low incidence of adverse effects. Despite some reports describing favourable effects of valproic acid, no clear consensus has emerged concerning the value of this drug. To date, no systematic reviews or meta‐analyses of the effect of valproic acid or valproate preparations on agitation in dementia have been published.
A recent systematic review of the effect of the neuroleptic, haloperidol, on agitation of people with dementia revealed problems in the methodology used by different authors (Lonergan 2002). Interpretation of study results was difficult because of variations in the description of agitation, in the dosage and duration of drug treatment, and in the assessment of response to therapy. These findings must therefore influence the interpretation of the results of all studies on treatment of agitation in dementia, including the present review.
With these considerations in mind the question "what is the effect of valproic acid on agitation in demented patients?" was explored from several points of view:
-
The effect of valproate preparations on overall response of agitation to treatment
-
The effect of valproate compounds on one or more specific manifestations of agitation (e.g. wandering, aggression, crying out)
-
Examination of adverse effects of valproate treatment, and how adverse effects contributed to drop‐outs among study populations.
The initial review, here updated to October 2008, was designed to provide guidance to clinicians and to researchers who are interested in the use of valproate derivatives for agitation in demented patients.
Objectives
Primary objective: To determine the effect of valproate preparations on agitation of people with dementia.
Secondary objectives:
-
To examine the incidence and types of adverse effects of valproic acid derivatives, especially where they may confound the evaluation of response of agitation to treatment, and the effect of adverse effects on patients withdrawing from treatment
-
To examine the effect of valproate preparations on caregiver burden for those who provide support for demented patients with agitation
To determine if the response to valproate therapy is influenced by: ·
-
Dose of drug
-
Duration of treatment
-
Degree of dementia and type of dementia
-
Manifestations of agitation
-
Age of patients
-
Sex of patients
Methods
Criteria for considering studies for this review
Types of studies
Types of studies included only unconfounded, randomised, placebo‐controlled trials with concealed allocation of subjects. Trials had to have included pre‐ and post‐treatment assessment of agitation. Interrupted time series trials were excluded. Length of trial and number of measurements did not influence the selection of trials for study. Where indicated, individual patient data were requested for further examination.
Types of participants
The report included participants of either sex and of any age (although almost all were 65 years of age or older). Both inpatients and outpatients (with or without caregivers) were included. The presence and type of dementia, unclassified or diagnosed, were identified according to the classifications provided by DSM IV (APA 1994) and ICD‐10 (WHO 1991). Studies prior to 1994 used criteria for dementia found in DSM III (APA 1980) and DSM IIIR (APA 1987). In the absence of these criteria other acceptable evidence such as the Mini‐Mental State Examination (Folstein 1975), psychiatric evaluation, psychological evaluation, or a medical evaluation were accepted. Subset analysis by type of dementia was carried out where the number of patients was sufficient; otherwise, patients were combined into the category of "dementia" regardless of type.
No formal definition of agitation has been published. Therefore, either the description, "inappropriate verbal, vocal, or motor activity that is not explained by needs or confusion per se." (Billig 1991; Cohen‐Mansfield 1989), or other descriptive language (e.g., wandering, aggression, crying out, uncooperative behavior) provided by individual investigators to identify agitation was accepted. Because agitation is common in delirium, all studies had to have demonstrated that a clinical evaluation was undertaken to rule out delirium, and all studies had to have ruled out other treatable causes of agitation (e.g. pain, infection, drug effect, urinary or faecal retention) prior to entering patients into the trial.
Types of interventions
At least one week of treatment with valproate preparations, of any dosage given by mouth, compared with placebo. Patients receiving chronic therapy with other psychoactive medications were excluded from the review.
Types of outcome measures
For professionals and for non‐professionals who care for demented patients, agitation is one of the most difficult behavioural problems to manage. The outcomes focus on the benefits and hazards of valproate treatment of demented patients who manifest agitation, and therefore should be meaningful to experts and non‐experts alike. The most important of the outcomes were: overall effect of treatment on agitation; effect of treatment on one or more specific aspects of agitation; adverse effects of treatment; and withdrawal from treatment.
It would have been desirable to analyse response to treatment in terms of factors such as age and sex of patient, type of dementia, and type of agitation, but the number of patients available in the studies was not large enough to permit subgroup analysis. A more feasible approach was to consider the overall effect of treatment, particularly if it could be shown to modify specific manifestations of agitation, and to examine the effect of treatment on dropout rates.
For continuous or ordinal variables, such as psychometric test scores and clinical global impression scales, if the ordinal scale data appeared to approximate a normal distribution, or if analysis indicated that parametric tests were appropriate, outcome measures were treated as continuous data. A second approach, not excluding the first, was to dichotomise the data (e.g. improved versus not improved) and treat the variable as binary. For binary outcomes the Peto method of the typical odds ratio was used.
Where studies used a cross‐over design, examination of the results was limited to the first part of the study.
Search methods for identification of studies
Electronic searches
We searched ALOIS (www.medicine.ox.ac.uk/alois) ‐ the Cochrane Dementia and Cognitive Improvement Group’s Specialized Register on 14 May 2010. The search terms used were: valproic OR valproate OR divalproex.
ALOIS is maintained by the Trials Search Co‐ordinator and contains studies in the areas of dementia prevention, dementia treatment and cognitive enhancement in healthy. The studies are identified from:
-
Monthly searches of a number of major healthcare databases: Medline, Embase, Cinahl, Psycinfo and Lilacs
-
Monthly searches of a number of trial registers: ISRCTN; UMIN (Japan's Trial Register); the WHO portal (which covers ClinicalTrials.gov; ISRCTN; the Chinese Clinical Trials Register; the German Clinical Trials Register; the Iranian Registry of Clinical Trials and the Netherlands National Trials Register, plus others)
-
Quarterly search of The Cochrane Library’s Central Register of Controlled Trials (CENTRAL)
-
Six‐monthly searches of a number of grey literature sources: ISI Web of Knowledge Conference Proceedings; Index to Theses; Australasian Digital Theses
To view a list of all sources searched for ALOIS see About ALOIS on the ALOIS website.
Details of the search strategies used for the retrieval of reports of trials from the healthcare databases, CENTRAL and conference proceedings can be viewed in the ‘methods used in reviews’ section within the editorial information about the Dementia and Cognitive Improvement Group.
Additional searches were performed in many of the sources listed above to cover the timeframe from the last searches performed for ALOIS to ensure that the search for the review was as up‐to‐date and as comprehensive as possible. The search strategies used can be seen in Appendix 1.
The previous update search conducted on the 7 February 2008 retrieved two new studies (Tariot 2005; Herrmann 2007) for inclusion in the review.
The latest search (July 2010) retrieved a total of 304 results. After a first‐assess and a de‐duplication of these results the authors were left with 12 references to further assess. Of these references no new studies were identified for either inclusion or exclusion within the review.
Data collection and analysis
Searching and screening of the results were be performed independently by two reviewers (ETL, JL). The reviewers selected trials for relevance and against defined inclusion criteria. Trials that did not meet the criteria were excluded. Reviewers' selection of trials were compared and the final list of studies was agreed by the reviewers.
Inclusion criteria: trials were ranked using one of the Cochrane approaches (Mulrow 1997).
Two reviewers (ETL, JL) independently estimated the quality of included studies. The Cochrane approach to assessing adequacy of allocation concealment was used:
1. Category A (adequate concealment). Allocation is determined by: 1) a centralized randomised scheme in which subjects are enrolled at a registry which codes and randomises the participants and notifies the investigator by telephone about treatment and control group allocation; 2) a randomizations controlled by a pharmacy; 3) coded containers with identical appearing capsules which are administered sequentially to participants; 4) an on‐site or coded computer, given the allocations were in a locked, unreadable file that could be accessed only after inputting the characteristics of an enrolled subject; 5) the use of sequentially numbered and sealed, opaque envelopes; 6) combinations of the above.
2. Category B (unclear concealment). The report describes the allocation of treatment by: 1) the use of a list or table to allocate assignments; 2) the use of envelopes or of sealed envelopes to distribute medications; 3) a statement in the report to the effect that the report is randomised without further characterization.
3. Category C (inadequate). Allocation of treatment by: 1) alternation of subjects to treatment or control arms of the study; 2) use of "fixed data" such as birth dates, record numbers, days of the week; 3) allocation that is transparent such as an open list of random numbers of assignments. Unclear concealments have been shown to increase estimates of treatment effects compared with adequate methods of concealment.
Trials conforming to Categories A and B were accepted. Trials falling into category C were excluded from further study.
Data extraction
-
Intention‐to‐treat (ITT) analysis was applied to data obtained on every randomised patient, without exception. In the absence of ITT data, data for 'on‐treatment analysis' were extracted and indicated as such.
-
For continuous variables, or ordinal variables approximated to continuous variables, outcomes of interest were the assessment score at the time point considered, and the change from baseline (i.e. pre‐randomization or at randomization) at this point.
-
Where patients were not randomised and where treatment allocation was not concealed when titration periods were used, data from these titration periods were not extracted to assess safety or efficacy of valproic acid. For the same reason, data from follow‐up periods were not extracted.
-
Data on adverse effects and drop‐outs were recorded.
-
For some binary and ordinal outcomes (i.e. improved versus not improved) the end‐point itself was of clinical relevance; because all patients, by definition, had the same initial score.
-
Where present, numerical scores such as those provided by Cohen‐Mansfield (Cohen‐Mansfield 1989b) were used to measure response to treatment.
-
In some instances, because of variation in the way response to treatment was measured, it was necessary to ope rationalize outcomes as 'improved' versus 'not improved', regardless of the scale used by the authors.
-
Where a cross‐over design was present, only results from the first part of the study were examined.
Analysis of data
The null hypothesis tested was that for the outcomes examined, treatment with valproic acid derivatives had no effect compared with placebo.
For continuous or ordinal variables (e.g. psychometric test scores, clinical global impression scales (Guy 1976), Cohen‐Mansfield Agitation Scale (Cohen‐Mansfield 1989b), if analyses indicated that parametric tests were appropriate, outcome measures were treated as continuous data; the data were be treated as continuous if ordinal scale data appeared to approximate a normal distribution.
Not excluding the above approach, data were concatenated into the two categories that best represent the contrasting states of interest, and the variable was treated as binary. For binary outcomes the Peto method of typical odds ratio was used. A test for heterogeneity of the treatment effect between the trials was done using a standard chi‐square statistic. Where no heterogeneity was found, a fixed‐effect parametric approach was taken. Where the included studies use the same outcome measures the method of weighted mean difference was to be used for meta‐analysis. When different scales were used in the studies the method of standardized mean difference was used.
As indicated in the Background section, where different forms of valproic acid (e.g. divalproex or valproic acid) were used, subset analysis of the results was applied, and where different scales of assessment were used, again subset analysis was be applied to the results.
Results
Description of studies
See Additional Tables: Baseline characteristics (Table 1) and Outcomes, instruments and studies (Table 2).
| Name | Country | Population | Mean Age | % Female | Intervention | Diagnoses | Mean MMSE |
| Porsteinsson 2001 | United States | Multicentric, Institutionalised; Alzheimer's dementia, vascular dementia and other dementias | 85.0 | 61.0% | Valproic acid (N=28); placebo (N=28). divalproex sodium titrated to mean dose of 826mg/d; 6 wk course. | Dementia: MMSE; DSM IV; NICDS‐ADRDA. Agitation: CMAI; Aggression: CMAI sub scale. Global: CGI. | 6.8 |
| Sival 2002 | The Netherlands | Institutionalised, Alzheimer's dementia, vascular dementia, and other dementias. | 80.4 | 59.5% | Valproic acid (N=42); placebo (N=42); Sodium valproate 480mg/d; 3 wk course. | Dementia: MMSE; DSM IV; NINCDS‐ADRDA,Clinical Dementia Rating Scale. Agitation: BPRS subset. Aggression: Patel's method; SDAS‐9 sub scale; CGI; GIP. Global: CGI | |
| Tariot 2001 | United States | Multicentric, Institutionalised; Alzheimer's dementia, vascular dementia, and other dementias. | 83.3 | 64.0% | Valproic acid (N=87); placebo (N=85); Divalproex sodium (delayed release); patients titrated to target dose of 20mg/kg/d; median dose 1000mg/d; 6 wk course. | Dementia: MMSE; DSM IV. Agitation: CMAI. Aggression: CMAI sub scale. Global: CGI | 7.4 |
| Tariot 2005 | United States | Multicentric, Institutionalized, Alzeheimers' Dementia | 84.0 | 68.6% | Divalproex (N=48). placebo (N=78);patients titrated to target dose of 750mg/day; 6 wk course | Alzheimer's Dementia (probable or possible) diagnosed on basis of National Institute of Neurological and Communicative Disorders and Stroke ‐Alzheimr's Disease and Related Disorders Criteria (NINDCDS‐ADRDA). Agitation and hostility, uncooperativeness measured by Brief Psychiatric Rating Scale (BPRS). | 10.8 |
| Herrmann 2007 | Canada | Multicentric, institutionalized;Alzheimer's Dementia | 85.6 | 42.8% | Valproate (N=14); placebo (N=13); Valproate titrated to dose of 1500mg/day; 6 wk course | Alzheimer's Dementia diagnosed by NINDCDS‐ADRDA criteria;Agitation/aggression diagnosed by Cohen‐Mansfield Agitation Inventory (CMAI). | <15 |
| Outcomes | Instruments | Studies |
| Agitation | Cohen‐Mansfield Agitation Inventory (CMAI) | Porsteinsson 2001 |
| CMAI | Tariot 2001 | |
| CMAI | Tariot 2007 | |
| CMAI | Herrmann 2007 | |
| Brief Psychiatric Rating Scale Subset (BPRS) | Porsteinsson 2001 | |
| Aggression | CMAI | Porsteinsson 2001 |
| CMAI | Tariot 2001 | |
| Social Dysfunction and Aggression‐9 Scale (SDAS‐9)14 | Sival 2002 | |
| Clinical Global Impression Scale (CGI) | Sival 2002 | |
| Nurse Observation Scale | Sival 2002 | |
| Patel's Method | Sival 2002 | |
| Overt Aggression Scale | Porsteinsson 2002 | |
| Other Types of Disturbed Behavior | Behavior Scale for Intramural Psychogeriatric Patients (GIP) | Sival 2002 |
| Global Impression | CGI | Porsteinsson 2001 |
| CGI | Sival 2002 | |
| CGI | Tariot 2001 | |
| CGI | Tariot 2005 | |
| Cognitive Function | Mini‐mental State Examination (MMSE) | Porsteinsson 2001 |
| MMSE | Sival 2002 | |
| MMSE | Tariot 2001 | |
| MMSE | Tariot 2005 | |
| MMSE | Herrmann 2007 | |
| Side Effects | Number of Adverse Reactions (checklist) | Porsteinsson 2001 |
| Number of Adverse Reactions (checklist) | Sival 2002 | |
| Coding Symbols for Thesaurus of Adverse Reaction Terms (COSTART 1989) | Tariot 2001 | |
| Number of Adverse Reactions (checklist) | Tariot 2005 | |
| Number of Adverse Reactions (checklist) | Herrmann 2007 |
At the time of the original review, three trials met the inclusion criteria:
1. One study was conducted in Europe (Sival 2002). Two studies were multi‐site and from the United States (Porsteinsson 2001; Tariot 2001). All studies involved institutionalised patients.
2. The average age of participants ranged between 80.4 years and 85.0 years. Women made up 59.5 % to 69.6% of the patients studied.
3. All studies used one or more standard methods: (DSM IV (APA 1994); NINCDS‐ADRDA (McKhann 1984)) to diagnose dementia, including Alzheimer's disease. All studies included patients with Alzheimer's disease and vascular dementia.
4. All three studies examined the tolerability, adverse effects, and safety of valproic acid preparations. Other goals included the effect of divalproex treatment on agitation, aggression, mania, and overall function of demented patients.
5. One study (Sival 2002) used the Clinical Dementia Rating Scale (Hughes 1982), in which 2/42 patients were classified as "light", 24/42 patients as "moderate", and 14/42 patients as "severe". This same study used the Mini‐Mental State Examination (Folstein 1975), but 14 patients could not be scored because of low level of function. The second study (Tariot 2001) used the MMSE and found an average score of 7.4 among their patients (severe dementia). The third study (Porsteinsson 2001) used the MMSE and found an average score of 6.8 among study subjects.
6. Two studies (Porsteinsson 2001; Tariot 2001) used the Cohen‐Mansfield Agitation Inventory (CMAI) scale (Cohen‐Mansfield 1989b) for agitation and aggression. One study (Sival 2002) used the Social Dysfunction and Aggression‐9 Scale (SDAS‐9) (Wistedt 1990), the Clinical Global Impression Scale (CGI) (Guy 1976) and the Nurse Observation scales (Colenda 1991), and Patel's method (Patel 1993) to measure aggressiveness. This same study used the Behavior Observation Scale for Intramural Psychogeriatric Patients (GIP) (Verstraten 1988) to measure other types of disturbed behaviour. One study (Porsteinsson 2001) used the Overt Aggression Scale (Yudofsky 1986) to measure aggression. This study used the Brief Psychiatric Rating Scale subset for agitation (BPRS) (Overall 1988) to measure other forms of disturbed behaviour.
7. In one study (Tariot 2001) patients were treated with divalproex sodium delayed‐release tablets or placebo. Dosage was begun at 125 mg bid and titrated to a dosage of 20 mg/kg/d ‐ 30 mg/kg/d, to be reduced if intolerable adverse effects appeared. The median dose for treated patients at the end of 6 weeks was 1000 mg/d. In another study (Sival 2002) participants received sodium valproate 240 mg bid for 3 weeks. In the third study (Porsteinsson 2001) participants received rapid‐acting divalproex sodium at an initial dose of 375 mg/day which was titrated to an average daily dose of 826 mg. In this trial a non‐blinded supervising physician who had no contact with the blinded reviewers altered drug dosage by 125 mg per day, depending on written reports by the reviewers of response and adverse effects. Length of treatment was 6 weeks.
8. All studies permitted short‐term use of short‐acting psychotropics. In Porsteinsson 2001 chloral hydrate was used; in Tariot 2001 short‐term use of lorazepam, oxazepam, or chloral hydrate were permitted as needed. In Sival 2002 oxazepam was allowed for severe anxiety or insomnia.
9. In an update review (October 2008) two new studies were examined (Tariot 2005; Herrmann 2007). Each of these used the CMAI to estimate effect of valproate preparations on agitation in dementia, compared with placebo controls, and each used a checklist to measure the incidence of adverse events among treated versus placebo patients. One study (Herrmann 2007) involved two institutions and used a cross‐over design, and a two week washout period. The other study (Tariot 2005) was a multi‐institution study. The methodology of both studies included intention‐to treat analysis.
Outcome measures
General
-
In one study (Tariot 2001) the primary goal was to investigate the safety and tolerability of divalproex sodium in the treatment of mania among demented patients. The design of this study, however, included an examination of agitation as measured by total agitation and verbal agitation scores (Cohen‐Mansfield Agitation Inventory), to compare the response of placebo treated patients with patients who received divalproex sodium. In a second study (Sival 2002) the response of aggressive behavior, a manifestation of agitation, was examined by comparing sodium valproate treated patients with control patients. In the third study (Porsteinsson 2001) the response of patients with agitation was measured, comparing divalproex sodium treated patients with controls. This study was described as a preliminary examination of efficacy, tolerability and safety of divalproex sodium and was to be followed by a more extensive investigation of the value of this drug in treating demented patients who should aggression/agitation.
-
One study (Sival 2002) used a cross‐over design with each participant serving as his or her own control. Intention‐to‐treat analysis was not used in one study (Sival 2002). In this study 4 patients withdrew, none of whom was receiving valproic acid at the time of withdrawal. Two studies (Porsteinsson 2001; Tariot 2001) used an intention‐to‐treat" analysis.
-
In the update of 2008 one study (Herrmann 2007) used a cross‐over design with a two week washout interval between study periods. Both studies (Tariot 2005; Herrmann 2007) use the CMAI to measure response of agitation to valproate preparation treatment, compared with controls.
Agitation and aggression
-
The CMAI (Cohen‐Mansfield 1989b) examines 29 types of agitated behaviour, including pacing, verbal or physical aggression, screaming, and restlessness. The frequency of these behaviours is measured on a 7‐point scale from 1 (never occurs) to 7 (occurs several times an hour) and scores for physical and verbal aggression and overall aggression may be aggregated.
-
The SDAS‐9 (Wistedt 1990) measures several aspects of behaviour having to do with patient interaction with other persons, and physical and verbal aggression. The scale is a 9‐point observation scale covering outward aggressive behaviour, with total sores ranging from 0 ‐ 36).
-
Patel's (Patel 1993) description of aggression includes any overt act of noxious stimuli affecting, though not necessarily aimed at, an object, organism or self which is not accidental. Intention is not included in the definition.
-
The CGI (Guy 1976) is a 7‐point scale with scores ranging from 1 (no aggressive behaviour) to 7 (severely aggressive behaviour). It is also uses to measure overall response to treatment (see below)
-
The "Nurse Observation" scale (Colenda 1991) assesses the incidence of aggressive behaviour at the moment the behaviour occurs.
-
The Behavior Observation Scale of Intramural Psychogeriatric Patients (Verstraten 1988) consists of 14 observational scales to describe agitated and aggressive behaviour.
-
The Brief Psychiatric Rating Scale (Overall 1962; Overall 1988) measures physical and verbal aggression, hallucinatory behaviour, and abnormal thought content.
-
The Overt Aggression Scale (Yudofsky 1986) quantifies aggressive verbal and physical behaviours and includes the number, specific nature, and intervention response.
Overall response
-
Porsteinsson 2001, Tariot 2001 and Sival 2002 used the CGI (Guy 1976). This is a 7‐point scale (1 ‐ very much improved; 4 ‐ no change; 7 ‐ very much worse) which rates all aspects of the individual's condition, comparing the current state with baseline, and using direct observation as well as information from significant others.
-
Tariot 2005 used the ADCS‐CGIC (Schneider 1997).
Measurement of adverse effects
-
Two studies (Porsteinsson 2001; Sival 2002) used check‐lists of adverse effects (e.g. drowsiness, nausea, vomiting, diarrhoea, confusion, disturbance in speech, disturbance of coordination, tremor, seizures, oedema, fever thrombocytopenia) which were reviewed at regular intervals by interviewing patients and nursing staff; and by reviewing chart entries.
-
Another study (Tariot 2001) measured adverse events based on the Coding Symbols for Thesaurus of Adverse Reaction Terms (COSTART 1989).
-
In the update of 2008 both of the newly examined studies (Tariot 2005; Herrmann 2007) measured adverse events using a check list.
See Additional table: Outcomes, instruments and studies (Table 2).
Risk of bias in included studies
Each of the studies reviewed here conformed to category B (unclear concealment).
In Porsteinsson 2001, although the physicians having direct responsibility for patient care were blinded, a non‐blinded physician, who had no direct contact with these physicians, adjusted divalproex sodium dosage based on reports from the blinded raters. Therefore, because one physician who controlled therapy knew which patients were receiving divalproex, and adjusted that therapy based on the reports that were submitted, the trial did not satisfy the criterion of concealed allocation. Another methodological problem is that in the statistical analysis section it is stated that the t‐test was used to examine within subject changes, but is unclear in describing treatment effects between groups. The results section does state that between‐group treatment differences were examined.
In Tariot 2001 54% of the treated patients dropped out, compared with 29% of control patients; 22% of all treated patients dropped out because of adverse effects, and the study had to be discontinued prematurely. Further, since patients had been on therapy for varying periods of time when the study was terminated, interpretation of the effects of treatment is difficult.
Sival 2002 had a cross‐over design. No results from the first phase of the study were available, and although the statistical section stated, "the t‐test for independent samples is used to analyse the two‐period cross‐over trial", because the samples were not independent ‐ they are the same patients in the treatment and placebo groups ‐ a question must be raised about the correctness of the analyses.
The trials varied in the type of valproic acid used (one trial, short‐acting sodium valproate; one trial, long‐acting divalproex sodium; one trial early‐onset acting divalproex sodium), average dosage of drug (480 mg/d ‐ 1000 mg/d), duration of therapy (3 weeks ‐ 6 weeks), and ways of evaluating patients and their response to therapy.
In the updated review of 2008, Herrmann 2007 had a cross‐over design and a small group of patients (N = 14) who served as their own controls in the study. Although only the results from the first part of the study were examined in this review, interpretation of the results must be tempered by the observation that the number of patients in this report was small.
Effects of interventions
The three included studies varied in the type of valproic acid preparation (Sival 2002, short‐acting sodium valproate; Porsteinsson 2001; Tariot 2001, long‐acting divalproex sodium), dosage (Sival 2002 daily dose 480 mg; Tariot 2001, median daily dose 1000 mg; Porsteinsson 2001, average daily dose 834 mg, Tariot 2005 initial divalproex sodium (sprinkle formulation) dose of 125 mg twice daily for 3 days, escalated in 125‐mg
increments every 3 days until the target dose of 750 mg/day was reached, Herrmann 2007, valproate initiated at 125 mg p.o., twice a day, with a forced titration to 500 mg p.o., twice a day, over the first 2 weeks. Thereafter, the dose of valproate was increased to a maximum of 1,500 mg per day or decreased based on efficacy and tolerability as determined by a blinded study physician), and duration of therapy (Sival 2002, 3 weeks; Tariot 2001; Porsteinsson 2001, 6 weeks; Tariot 2005, 6 weeks; Herrmann 2007, 6 weeks). In some instances, methods of evaluating the patients varied among the reports as well, with different scales being used to assess agitation and aggression, and response to therapy.
Agitation
Because of methodological problems, the effect of valproic acid derivatives on agitation in demented patients as reported in the studies of Tariot 2001and Sival 2002 cannot be interpreted, and therefore these studies do not support the use of valproic acid derivatives to control agitation in demented patients. The third study (Porsteinsson 2001) showed no significant difference in the response of agitation, between control and treated groups (P = 0.08). The authors report that when covariants were included, using analysis of covariance, a significant response of agitation to treatment (P = 0.05) was noted, comparing treated with placebo patients.
Aggression
No study demonstrated overall improvement of aggression in treated compared with placebo patients (P > 0.05).
Overall response
Because of methodological limitations, no interpretation of the results comparing treated and placebo patients could be derived from the studies of Tariot 2001and Sival 2002. The third study (Porsteinsson 2001) showed no difference comparing treated and control groups in overall response, measured by the CGI (P = 0.07).
Adverse effects and drop‐outs
In one study (Sival 2002) the mean incidence of adverse effects was low during three weeks of observation in both treated (0.17) and placebo (0.02) groups. The study provides no description of the kinds of adverse reactions that were identified.
In the second study (Tariot 2001) 47 (54%) of treated patients and 25 (29%) placebo patients withdrew from the study. Nineteen treated patients (22%) and 3 placebo patients (4%) withdrew from the study because of adverse effects, predominately somnolence (P < 0.01), and including hyponatraemia, accidents, and weight loss, and the study was prematurely terminated. The only adverse effect other than somnolence that reached statistical significance in the study of Tariot 2001 was thrombocytopenia: 6 participants (7%) in the treated group developed thrombocytopenia, compared with none in the placebo group (P = 0.01).
In the third study (Porsteinsson 2001) adverse effects occurred in 68% of the divalproex group compared with 33% of the placebo group (P = 0.035), but these were described as "mild" and were not responsible for any withdrawals. Two of Porsteinsson's treated patients developed a significant decrease in platelet count (baseline > 300,000/cubic mm to 261,000/cubic mm and 159,000/cubic mm), but not to the level of thrombocytopenia. Sival 2002 monitored blood values in his patients, but reported no instances of a drop in platelet counts. There were a total of five decreases in divalproex sodium in four subjects, presumably due to adverse effects. Individual adverse effects included:
-
sedation (placebo, 3/27; treatment, 11/28),
-
nausea, vomiting or diarrhoea (placebo, 2/27; treatment, 7/28);
-
urinary tract infection (placebo, 3/27; treatment 5/28);
-
ataxia (placebo, 3/27; treatment, 4/28);
-
skin problems (placebo, 3/27; treatment 4/28);
-
other gastrointestinal problem (placebo, 2/27; treatment 4/28);
-
fall with injury (placebo, 3/27 treatment, 3/28);
-
fall without injury (placebo, 2/27; treatment, 4/28);
-
trauma, other than falls (placebo, 2/27; treatment, 2'28);
-
postural instability (placebo, (placebo, 1/27; treatment 4/28);
-
respiratory problems (placebo, 0/27; treatment, 5/28);
-
joint problems (placebo, 3/27; treatment, 1/28); weakness (placebo, 0/27; treatment, 4/28);
-
cardiovascular problems (placebo, 1/27; treatment, 2/28);
-
oedema (placebo, 0/27; treatment, 3/28);
-
fever (placebo, 0/27; treatment, 3/28);
-
other infection (placebo, 1/27; treatment, 1/28);
-
hallucinations (placebo, 1/27; treatment, 1/28).
Serious adverse effects:
-
seizure (placebo, 0/27; treatment 2/28);
-
chronic renal failure (placebo, 1/27; treatment, 0/28);
-
cerebrovascular accident (placebo, 0/27; treatment, 1/28);
-
pneumonia (placebo, 0/27; treatment, 1/28);
-
small bowel obstruction (placebo, 0/27; treatment, 1/28).
Analysis of individual adverse effects reported by Porsteinsson 2001 states no significant difference between treated and control groups. On repeating the analysis this was true for all adverse effects except sedation, weakness and respiratory infections. Sedation occurred more frequently in divalproex treated participants (odds ratio: 5.18, 95% confidence limits, 1.25 ‐ 21.41; p = 0.02). Weakness occurred more frequently in divalproex treated participants (odds ratio: 8.00, 95% confidence limits, 1.06 ‐ 60.16; p = 0.04). Respiratory tract infections were also more common among divalproex treated patients, compared with controls (odds ratio: 8.34; 95% confidence limits, 1.35 ‐ 51.60; p = 0.02).
One study (Sival 2002) lost 3 patients (two placebo, one during washout period) due to complications that were not ascribed to the study drug (fever, stroke, fractured hip) and one patient due to protocol violation (total 4, or 9.5%). No patient was receiving valproic acid at the time of withdrawal from the study. These patients were omitted from the statistical analysis of response to treatment. In another study (Tariot 2001) 47 treated patients (64%) dropped out, compared with 15 (30%) placebo treated patients, and the study was terminated prematurely. The major reason for drop‐outs was somnolence among the treated patients. In the third study (Porsteinsson 2001) four patients were dropped from the placebo arm, all with increased agitation; two patients left the divalproex arm, one because of small bowel obstruction, the other because of infection, delirium, and seizures.
Meta‐analysis of two pooled trials (Porsteinsson 2001, Tariot 2001) showed sedation to be more frequent among treated patients compared with controls (odds ratio 2.64, 95% confidence interval 1.43 ‐ 4.88; p = 0.002). Urinary tract infections were also more frequent among treated than control patients (odds ratio 3.02; 95% confidence interval 1.04 ‐ 8.80; p = 0.04). Overall the number of adverse events was more frequent in treated than control patients (odds ratio 3.22; 95% confidence interval 1.77 to 5.86, p = 0.0001).
Update of 2008
Meta‐analysis of three studies (Porsteinsson 2001; Tariot 2005; Herrmann 2007) showed, as measured by the CMAI, no improvement in agitation among valproate treated patients, compared with controls. Meta‐analysis of three pooled studies (Tariot 2001; Tariot 2005; Porsteinsson 2001) showed an overall increase in adverse effects among valproate treated patients, compared with controls. An examination of individual studies showed adverse effects such as falls, gastrointestinal disorders, infections and decreased platelet counts to be significantly more frequent among valproate patients, compared with controls. Because for the outcome CMAI in measuring response to agitation in the Herrmann 2007 study there was heterogeneity, sensitivity analysis was done and showed that removing the Herrmann results from the analysis resulted in no change from the original finding of lack of effect of valproate preparations on agitation among treated patients, compared with controls.
Plasma levels of valproate
With an average dose of divalproex of 826 mg/d the mean plasma level in Porsteinsson's study (Porsteinsson 2001) was 45.45 µg/ml. In Tariot 2001 the median divalproex dose was 100 mg/day and the mean weekly plasma level was 55.3 to 68.9 µg/ml. In Sival 2002 the average daily dose of sodium divalproate was 480 mg and the mean plasma level was 40.9 µg/ml, comparable to Porsteinsson's value. In Tariot 2005, the mean dose was 800 mg/day, and the mean serum valproate level at endpoint was 52.8 µg/ml. In Herrmann 2007, the average daily dose of valproate at the end of the valproate treatment phase was 1,134.6 ± 400.1 mg, resulting in average serum valproate levels of 48.8 ± 16.8 µg/ml.
Discussion
It must be appreciated that the primary goal in one of the studies (Tariot 2001) was to examine the safety and tolerability of divalproex sodium in the treatment of mania among demented patients. This study clearly showed that high dose divalproex sodium (median dose, 1000 mg per day) was associated with an unacceptable frequency of adverse effects, predominantly somnolence, among treated patients and the study was terminated prematurely. Although not a primary goal of the study, it is the case that the protocol of this investigation measured the effect of divalproex sodium on agitation in demented patients, as shown by total and verbal agitation scores of the Cohen‐Mansfield Agitation Inventory. Because of high drop rates among treated patients interpretation of the results of treatment on agitation in demented patients cannot be relied upon for clinical guidance.
Flaws in methodology and lack of standardization in the ways of treating and assessing patients made it inappropriate to apply meta‐analysis to these preliminary trials of the effect of valproic acid on people with dementia patients who were agitated. Variations in method, type of medication, dosage, duration of treatment, and ways of assessing participants in these studies also made it inappropriate to apply meta‐analysis.
For example, although Porsteinsson's trial (Porsteinsson 2001) described a randomised, double‐blind, placebo‐controlled design, the presence of a non‐blinded physician who varied dosage of medication based on reports from blinded raters, and who was informed about laboratory results among these patients, did not satisfy the requirement of concealed allocation and could have introduced an uncontrolled variable into the outcomes of this investigation.
Further, the premature termination of one study (Tariot 2001), in which 47 treated patients (54%) dropped out before the protocol could be completed, made questionable the conclusion that divalproex sodium improved agitation of people with dementia, and did not justify including this report in a pooled analysis of the results of these studies. In Sival's report (Sival 2002) the description of statistical analysis section describes "the t‐test for independent samples is used to analyse the two‐period cross‐over trial," but since they are the same patients in the treatment and placebo groups, and hence not independent samples, it is not possible to accept the analyses in this study. Finally, it would be difficult to interpret a meta‐analysis based on pooled results from studies that employed short‐acting sodium valproate (Sival 2002) or longer‐acting divalproex sodium (Porsteinsson 2001; Tariot 2001), and in which the dosage varied more than two‐fold (Sival 2002, mean dose 480 mg/day; Tariot 2001, median dose 1000 mg/day; Porsteinsson 2001, mean dose 875 mg/day).
A limited study of two pooled trials (Porsteinsson 2001; Tariot 2001), and applying meta‐analysis, showed that sedation occurred more frequently among patients who received the higher dose of valproic acid/derivative compared with controls (odds ratio 2.64, 95% confidence interval 1.43 to 4.88; p = 0.002). In the third, low‐dose study, (Sival 2002) adverse effects were described as rare and no mention of sedation was made.
A meta‐analysis of the pooled data (Porsteinsson 2001; Tariot 2001) concerning urinary tract infection showed a higher frequency of this adverse effect among participants taking valproic acid, compared with controls (odds ratio 3.02; 95% confidence interval 1.04 ‐ 8.88; p = 0.04). No explanation for this adverse effect was given in either of the reports, and no attempt to link sedation to urinary tract infections was made. Reanalysis of Porsteinsson's data also showed a higher rate of respiratory infections among divalproex treated patients, compared with controls (odds ratio 8.34; 95% confidence interval 1.35 ‐ 51.60; p = 0.02). It may be conjectured that because respiratory tract and urinary tract infections, in general, occur more frequently among immobilized patients, is possible that the high frequency of sedation found in these studies may have lessened mobility among treated patients and thus contributed to the increased incidence of infection.
Sival 2002 found a low frequency of adverse effects in participants who received sodium valproate, 480 mg/day, but no improvement in agitation. Sival 2002 suggested that higher doses of valproate, and higher plasma levels, were associated with the increased frequency of serious adverse effects reported in Tariot 2001, with attendant high drop‐out rates, but also may have explained the improved CMAI (Cohen‐Mansfield 1989b) total scores (P < 0.05) and verbal agitation scores in this study. Neither Sival 2002 nor Porsteinsson 2001 demonstrated a significant response to treatment of agitation or aggression, although covariate analysis in Porsteinsson's study showed an improvement in the Brief Psychiatric Rating Scale (Overall 1988) among treated patients. There were four drop‐outs in the placebo arm of Porsteinsson's study, all associated with increased agitation. Given that there were also two drop‐outs among the treated patients in this study, it is unlikely that the placebo drop‐outs had any significant effect on the overall assessment of the results of this study. Sival 2002 concluded that higher doses of valproate were necessary to achieve therapeutic effect, and Tariot 2001 observed that lower doses than those used in his study were necessary to avoid adverse effects and high drop‐out rates.
In a new controlled study of the effect of divalproex sodium on agitation in demented patients, reported in abstract form,Tariot and associates (Tariot 2004a) were unable to demonstrate any significant difference in agitation among treated patients (target dose of divalproex, 750 mg per day), compared with placebo controls. This study will be reviewed in greater detail when the published article becomes available.
Update October 2008: Meta‐analysis of pooled studies showed no improvement in agitation among demented patient treated with valproate preparations, as compared with controls. Because of heterogeneity in the outcome for agitation as measured by the CMAI in one study (Herrmann 2007), sensitivity analysis was performed and no change in the result (no response of agitation to valproate treatment) was found after the Herrmann study of 2007 was subtracted from the pooled results. Heterogeneity in the results of Herrmann may have been due to the small sample size (N = 14).
Further meta‐analysis of pooled results showed both an increase in overall adverse events among valproate patients compared with placebo patients, and in subset analysis an increase in specific adverse events such as falls, infection, gastrointestinal disorders, and decreased platelet counts. In one study (Tariot 2005) there was no difference in total adverse events, comparing valproate and placebo patients, but in this report specific adverse effects (falls, gastrointestinal symptoms, decreased platelet counts) were significantly more frequent among treated patients, compared with controls.
Because of the limited number of subjects it was not possible to analyse secondary objectives such as the effect of valproate therapy on individual manifestations of agitations (e.g., crying out, wandering) or the influence of age or type of dementia on the response to therapy.
Comparison 1 Divalproex vs placebo, Outcome 1 Brief Psychiatric rating scale Total score. Change from baseline at 6 weeks (ITT).
Comparison 1 Divalproex vs placebo, Outcome 2 BPRS agitation factor. Change from baseline at 6 weeks (ITT).
Comparison 1 Divalproex vs placebo, Outcome 3 BPRS hostility factor. Change from baseline at 6 weeks (ITT).
Comparison 1 Divalproex vs placebo, Outcome 4 Overt aggression scale total score. Change from baseline at 6 weeks (ITT).
Comparison 1 Divalproex vs placebo, Outcome 5 Cohen‐Mansfield Agitation Index. Total Score. Change from baseline at 6 weeks (ITT).
Comparison 2 Divalproex vs placebo, Outcome 1 Total number of patients who suffered from sedation at 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 2 Total number of patients who suffered from nausea, vomiting, or diarrhea at 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 3 Total number of patients who suffered from an urinary tract infection by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 4 Total number of patients who suffered falls without injury by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 5 Total number of patients who suffered falls with injury by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 6 Total number of patients who suffered postural instability by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 7 Total number of patients who suffered weakness by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 8 Total number of patients who suffered cardiovascular problems by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 9 Total number of patients who suffered edema by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 10 Total number of patients who suffered a fever by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 11 Total number of patients who suffered a respiratory problem by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 12 Total number of patients who suffered any adverse effect by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 13 Total number of patients who suffered from ataxia at 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 14 Total number of patients who suffered from a skin problem at 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 15 Total number of patients who suffered other gastrointestinal problem by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 16 Total number of patients who suffered trauma (other than falls) by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 17 Total number of patients who suffered joint problems by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 18 Total number of patients who suffered other infection by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 19 Total number of patients who suffered hallucinations by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 20 Total number of patients who suffered accidental injury by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 21 Total number of patients who suffered anorexia by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 22 Total number of patients who suffered weight loss by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 23 Total number of patients who suffered dehydration by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 24 Total number of patients who suffered thrombocytopenia by 6 weeks.
Comparison 2 Divalproex vs placebo, Outcome 25 Total number of patients who suffered serious adverse events by 6 weeks.
| Name | Country | Population | Mean Age | % Female | Intervention | Diagnoses | Mean MMSE |
| Porsteinsson 2001 | United States | Multicentric, Institutionalised; Alzheimer's dementia, vascular dementia and other dementias | 85.0 | 61.0% | Valproic acid (N=28); placebo (N=28). divalproex sodium titrated to mean dose of 826mg/d; 6 wk course. | Dementia: MMSE; DSM IV; NICDS‐ADRDA. Agitation: CMAI; Aggression: CMAI sub scale. Global: CGI. | 6.8 |
| Sival 2002 | The Netherlands | Institutionalised, Alzheimer's dementia, vascular dementia, and other dementias. | 80.4 | 59.5% | Valproic acid (N=42); placebo (N=42); Sodium valproate 480mg/d; 3 wk course. | Dementia: MMSE; DSM IV; NINCDS‐ADRDA,Clinical Dementia Rating Scale. Agitation: BPRS subset. Aggression: Patel's method; SDAS‐9 sub scale; CGI; GIP. Global: CGI | |
| Tariot 2001 | United States | Multicentric, Institutionalised; Alzheimer's dementia, vascular dementia, and other dementias. | 83.3 | 64.0% | Valproic acid (N=87); placebo (N=85); Divalproex sodium (delayed release); patients titrated to target dose of 20mg/kg/d; median dose 1000mg/d; 6 wk course. | Dementia: MMSE; DSM IV. Agitation: CMAI. Aggression: CMAI sub scale. Global: CGI | 7.4 |
| Tariot 2005 | United States | Multicentric, Institutionalized, Alzeheimers' Dementia | 84.0 | 68.6% | Divalproex (N=48). placebo (N=78);patients titrated to target dose of 750mg/day; 6 wk course | Alzheimer's Dementia (probable or possible) diagnosed on basis of National Institute of Neurological and Communicative Disorders and Stroke ‐Alzheimr's Disease and Related Disorders Criteria (NINDCDS‐ADRDA). Agitation and hostility, uncooperativeness measured by Brief Psychiatric Rating Scale (BPRS). | 10.8 |
| Herrmann 2007 | Canada | Multicentric, institutionalized;Alzheimer's Dementia | 85.6 | 42.8% | Valproate (N=14); placebo (N=13); Valproate titrated to dose of 1500mg/day; 6 wk course | Alzheimer's Dementia diagnosed by NINDCDS‐ADRDA criteria;Agitation/aggression diagnosed by Cohen‐Mansfield Agitation Inventory (CMAI). | <15 |
| Outcomes | Instruments | Studies |
| Agitation | Cohen‐Mansfield Agitation Inventory (CMAI) | Porsteinsson 2001 |
| CMAI | Tariot 2001 | |
| CMAI | Tariot 2007 | |
| CMAI | Herrmann 2007 | |
| Brief Psychiatric Rating Scale Subset (BPRS) | Porsteinsson 2001 | |
| Aggression | CMAI | Porsteinsson 2001 |
| CMAI | Tariot 2001 | |
| Social Dysfunction and Aggression‐9 Scale (SDAS‐9)14 | Sival 2002 | |
| Clinical Global Impression Scale (CGI) | Sival 2002 | |
| Nurse Observation Scale | Sival 2002 | |
| Patel's Method | Sival 2002 | |
| Overt Aggression Scale | Porsteinsson 2002 | |
| Other Types of Disturbed Behavior | Behavior Scale for Intramural Psychogeriatric Patients (GIP) | Sival 2002 |
| Global Impression | CGI | Porsteinsson 2001 |
| CGI | Sival 2002 | |
| CGI | Tariot 2001 | |
| CGI | Tariot 2005 | |
| Cognitive Function | Mini‐mental State Examination (MMSE) | Porsteinsson 2001 |
| MMSE | Sival 2002 | |
| MMSE | Tariot 2001 | |
| MMSE | Tariot 2005 | |
| MMSE | Herrmann 2007 | |
| Side Effects | Number of Adverse Reactions (checklist) | Porsteinsson 2001 |
| Number of Adverse Reactions (checklist) | Sival 2002 | |
| Coding Symbols for Thesaurus of Adverse Reaction Terms (COSTART 1989) | Tariot 2001 | |
| Number of Adverse Reactions (checklist) | Tariot 2005 | |
| Number of Adverse Reactions (checklist) | Herrmann 2007 |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Brief Psychiatric rating scale Total score. Change from baseline at 6 weeks (ITT) Show forest plot | 2 | 202 | Mean Difference (IV, Fixed, 95% CI) | 0.23 [‐2.14, 2.60] |
| 2 BPRS agitation factor. Change from baseline at 6 weeks (ITT) Show forest plot | 1 | 55 | Mean Difference (IV, Fixed, 95% CI) | ‐1.30 [‐2.73, 0.13] |
| 3 BPRS hostility factor. Change from baseline at 6 weeks (ITT) Show forest plot | 1 | 55 | Mean Difference (IV, Fixed, 95% CI) | 0.10 [‐1.12, 1.32] |
| 4 Overt aggression scale total score. Change from baseline at 6 weeks (ITT) Show forest plot | 1 | 55 | Mean Difference (IV, Fixed, 95% CI) | 0.10 [‐3.42, 3.62] |
| 5 Cohen‐Mansfield Agitation Index. Total Score. Change from baseline at 6 weeks (ITT) Show forest plot | 3 | 216 | Mean Difference (IV, Fixed, 95% CI) | ‐2.20 [‐6.38, 1.99] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Total number of patients who suffered from sedation at 6 weeks Show forest plot | 3 | 241 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.48 [1.37, 4.47] |
| 2 Total number of patients who suffered from nausea, vomiting, or diarrhea at 6 weeks Show forest plot | 2 | 208 | Odds Ratio (M‐H, Fixed, 95% CI) | 7.09 [1.73, 29.02] |
| 3 Total number of patients who suffered from an urinary tract infection by 6 weeks Show forest plot | 2 | 227 | Odds Ratio (M‐H, Fixed, 95% CI) | 3.02 [1.04, 8.80] |
| 4 Total number of patients who suffered falls without injury by 6 weeks Show forest plot | 3 | 222 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.70 [0.84, 3.44] |
| 5 Total number of patients who suffered falls with injury by 6 weeks Show forest plot | 2 | 82 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.57 [0.38, 6.41] |
| 6 Total number of patients who suffered postural instability by 6 weeks Show forest plot | 1 | 55 | Odds Ratio (M‐H, Fixed, 95% CI) | 4.33 [0.45, 41.55] |
| 7 Total number of patients who suffered weakness by 6 weeks Show forest plot | 1 | 55 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 8.00 [1.06, 60.16] |
| 8 Total number of patients who suffered cardiovascular problems by 6 weeks Show forest plot | 1 | 55 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.0 [0.17, 23.44] |
| 9 Total number of patients who suffered edema by 6 weeks Show forest plot | 1 | 55 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 7.69 [0.77, 77.21] |
| 10 Total number of patients who suffered a fever by 6 weeks Show forest plot | 1 | 55 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 7.69 [0.77, 77.21] |
| 11 Total number of patients who suffered a respiratory problem by 6 weeks Show forest plot | 1 | 55 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 8.34 [1.35, 51.60] |
| 12 Total number of patients who suffered any adverse effect by 6 weeks Show forest plot | 4 | 394 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.99 [1.29, 3.08] |
| 13 Total number of patients who suffered from ataxia at 6 weeks Show forest plot | 1 | 55 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.33 [0.27, 6.61] |
| 14 Total number of patients who suffered from a skin problem at 6 weeks Show forest plot | 2 | 227 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.27 [0.45, 3.58] |
| 15 Total number of patients who suffered other gastrointestinal problem by 6 weeks Show forest plot | 1 | 55 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.08 [0.35, 12.45] |
| 16 Total number of patients who suffered trauma (other than falls) by 6 weeks Show forest plot | 1 | 55 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.62 [0.09, 4.01] |
| 17 Total number of patients who suffered joint problems by 6 weeks Show forest plot | 1 | 55 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.30 [0.03, 3.04] |
| 18 Total number of patients who suffered other infection by 6 weeks Show forest plot | 3 | 380 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.30 [0.70, 2.45] |
| 19 Total number of patients who suffered hallucinations by 6 weeks Show forest plot | 1 | 55 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 7.13 [0.14, 359.55] |
| 20 Total number of patients who suffered accidental injury by 6 weeks Show forest plot | 1 | 172 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 1.52 [0.79, 2.92] |
| 21 Total number of patients who suffered anorexia by 6 weeks Show forest plot | 1 | 172 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 2.40 [0.95, 6.07] |
| 22 Total number of patients who suffered weight loss by 6 weeks Show forest plot | 1 | 172 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 2.56 [0.76, 8.66] |
| 23 Total number of patients who suffered dehydration by 6 weeks Show forest plot | 1 | 172 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 3.13 [0.82, 11.94] |
| 24 Total number of patients who suffered thrombocytopenia by 6 weeks Show forest plot | 2 | 186 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 7.91 [1.92, 32.57] |
| 25 Total number of patients who suffered serious adverse events by 6 weeks Show forest plot | 3 | 235 | Peto Odds Ratio (Peto, Fixed, 95% CI) | 1.34 [0.74, 2.45] |
