Description of the condition
Bipolar disorder is a chronic illness which has a lifetime prevalence of 1% to 5% in adults (Akiskal 2000). It presents classically with depression, mania or (less frequently) mixed episodes.
An acute phase of bipolar disorder can present as:
1. an acute manic episode, where there is an onset of symptoms such as elated mood, disinhibition and rapidity of thought;
2. an acute depressive episode, where mood is lowered and there are a range of changes in thoughts, emotions, and bodily functions such as sleep and appetite; or
3. a mixed episode, where there are features of 1 and 2.
There is usually complete inter-episode recovery of affective symptoms, though cognition remains impaired. The prevalence rate in the older person in the community is low (up to 0.1%). However, in selected populations, i.e. in nursing homes and in hospital inpatients, there are significantly higher prevalence rates; 9.7% in nursing homes (Koenig 1992) and varying from 8% to 10% in hospital inpatients (Depp 2004). This is likely to relate in part to the association of late onset bipolar disorder with physical illness, especially dementia and cerebrovascular disease.
The category of 'older people' may include those over a certain age who have developed late onset bipolar disorder and those who have had bipolar disorder in their earlier adult life, and now are older.
Description of the intervention
Different medications are recommended for the various episodes of the illness (Goodwin 2009). A broad class of drugs used to treat the manic episode are anti-manic agents. These agents could be antipsychotic medications (such as olanzapine, quetiapine etc); or drugs traditionally used in the treatment of epilepsy such as sodium valproate or lamotrigine; or drugs used to help stabilise mood such as lithium. The anti-manic agents may be used in conjunction with drugs that help the person become sedated, e.g. lorazepam (given orally or by an injection). The depressive episode, on the other hand, is usually treated with a combination of antidepressants and an anti-manic agent. Lastly, a mixed episode is treated by a combination of medications.
Beside the acute phase, treatment is often needed for prevention of further relapse of the illness. Such treatments are relatively well established in the adult population as there is a good amount of research trial data available. Late life bipolar disorder, particularly late onset bipolar disorder, may be a distinct diagnostic entity compared to bipolar disorder in younger adults and hence might need different treatment strategies (Vasudev 2010). However, such treatments have not been tested well in older people. There are some data for the efficacy of lamotrigine (a medication used to treat epilepsy) in bipolar disorder in those greater than 55 years of age (Marcotte 2004). There is also some evidence for levetiracetam (Kyomen 2006), another agent used in the treatment of epilepsy.
How the intervention might work
Antipsychotic, anti-epileptic and antidepressant agents have a complex and incompletely understood mechanism of action in the treatment of bipolar disorder in the adult person. Added to this complexity, treatment of any disorder in the older person is difficult as there are various changes that occur in the body associated with aging. They also may have other physical problems as a result of old age, and hence they may be on other prescribed medications, which may sometimes interact with each other. Lithium remains the mood stabiliser in longest use for the treatment and prevention of relapses in bipolar disorder in the adult population. However, even in this population the response to lithium treatment may often be unsatisfactory. It has a narrow margin of safety and is frequently associated with adverse effects, often at therapeutic doses (Goodwin 2003). It is thought to precipitate episode recurrence on withdrawal, making its use in the non-compliant patient problematic. On the other hand, there may be a protective effect of lithium in preventing dementia (Kessing 2008). Further, in long-term usage it may sometimes be toxic to the kidneys (Bendz 1996). The role of other agents in late life bipolar disorder is even less clear, mostly due to lack of randomised controlled studies.
Why it is important to do this review
Bipolar disorder remains a complex illness to treat and this is evidenced by publication of a number of reviews (Vasudev 2010) and meta-analyses within The Cochrane Library itself (Cipriani 2009; Justo 2007; Macritchie 2009a; Macritchie 2009b; Morriss 2009; Rendell 2006; Vasudev 2006; Vasudev 2012). There are guidelines from the National Institute of Clinical Excellence (NICE), Canadian Network for Mood and Anxiety (CANMAT) and British Association for Psychopharmacology (BAP) for the management of various phases of bipolar disorder in the adult population. However, there are no recommendations for medications in the treatment of late life bipolar disorder. This may be because of a lack of randomised controlled studies of pharmacological treatment in this age group. There is limited evidence in late life bipolar disorder from open label studies using lamotrigine in an adjuvant trial (Kyomen 2006), naturalistic studies, case reports (Marcotte 2004) and clinical experience. Hence, different pharmacotherapeutic interventions, particularly for the older person, need to be evaluated for their role in bipolar disorder.
1. To determine the efficacy of any pharmacotherapeutic intervention compared to placebo or an alternative drug:
1.1 in alleviating acute manic episodes of bipolar disorder in the older person;
1.2 in alleviating depressive symptoms in acute episodes of bipolar disorder in the older person;
1.3 in alleviating mixed affective symptoms in acute episodes of bipolar disorder in the older person.
2. To review the acceptability of treatment with medications to patients, measured by numbers and reasons for dropping out of trials, compliance, and by reference to patients' expressed views regarding treatment.
3. To investigate the adverse effects of pharmacological treatments including the general prevalence of adverse effects. All adverse effects documented in the included studies will be extracted where possible.
4. To determine the overall mortality rates on various pharmacological treatments.
5. To prevent relapse of the illness in the medium or long term.
Criteria for considering studies for this review
Types of studies
All randomised controlled studies comparing various pharmacotherapeutic interventions in the treatment of any phase of bipolar disorder in older people.
We will exclude quasi-randomised studies, such as those allocating by using alternate days of the week.
We will include cross-over studies where participants receive different treatments sequentially.
We will include cluster randomised controlled trials but will be cautious of unit-of-analysis error issues (Higgins 2009b).
Types of participants
We will include males and females over the age of 55 years with a diagnosis of bipolar disorder according to internationally accepted criteria such as those laid out by the World Health Organization in their International Classification of Diseases (ICD). The age of 55 years will be used as a threshold for inclusion for a person who is "old". This is because there is evidence that aging has a variable influence on bipolar disorder for the older person as early as 50 years of age (Young 2004) rather than the traditional age cut off of 65 years, a threshold for provision of services for geriatric population in most parts of the world.
In its tenth edition, ICD defines a range of conditions relevant to this review and we refer the reader to the book for further details of the classification (ICD-10 Code F31* (WHO 1992)). The American Psychiatric Association has similarly brought out the Diagnostic and Statistical Manual (DSM) and in its current version (DSM IV) bipolar disorder's classification starts with the number 296* (APA 1994).
Bipolar disorder presents with a number of subtypes as per these criteria. We plan to include all (rapid cycling-suffering from four or more affective episodes per year, types I and II and other); however, cyclothymia will be excluded.
It is recognised that some trials may involve heterogeneous groups of participants: in particular, they may include schizo-affective disorder and unipolar depressive disorder (diagnoses approximating to ICD-10 F25 and DSM 295.70, and ICD-10 F33 and DSM IV 296.3, respectively). Where possible, data from these studies will be separated into diagnostic groups. If necessary, the authors of the studies will be requested to provide original data. Where such separation is impossible, the studies will be included but a sensitivity analysis will be conducted to examine the effect of their inclusion. Furthermore, if we are unable to obtain a separation of such heterogeneous data, we will extract information from these studies only if at least 80% of randomised patients have a bipolar diagnosis.
Examples of diagnostic codes, for e.g. participants with acute affective episodes which will include:
1. participants with depressive episodes, with or without psychotic symptoms, approximating to ICD-10 Codes F31.3-31.5* and/or DSM IV 296.5x;
2. participants with a diagnosis of mixed affective disorder, with or without psychotic symptoms, approximating to ICD-10 Code F31.6* and/or DSM IV Code 296.6x;
3. participants with a diagnosis of hypomania or mania with or without psychotic symptoms approximating to ICD-10 codes F30.0 and F31.0-31.2* and/or DSM Code 296.40 or 296.4x.
*Trials with ICD-9 and/or DSM III/IIIR diagnoses approximating to these codes will also be included together with any prior versions to DSM manuals/ICD classifications.
Types of interventions
An intervention, for the purpose of this protocol, will be a pharmacological treatment used either to alleviate symptoms of an existing acute episode, or as a maintenance therapy for chronic symptoms and to prevent relapse. Intervention parameters will include pharmacotherapeutic interventions of various kinds in the treatment of acute manic, mixed affective or depressive episodes in the context of bipolar disorder which could include as below:
- anti-manic agents including lithium and anticonvulsants. We recognise that the phrase "mood-stabiliser" is frequently used in this context - we acknowledge the use of such a phrase and would categorise such agents as anti-manic for the purpose of this review;
- combination anti-manic agents;
- antidepressants of any sort (which are medications used to treat low mood);
- antipsychotics of any sort (which are medications used to help with symptoms such as false beliefs, hearing voices etc);
- combinations of anti-manic agents and antipsychotics;
- any other pharmacological therapy.
The above intervention(s) could be compared with another pharmacological treatment or a control arm which may include a placebo or no treatment.
There will be no restrictions on dose, frequency, intensity or duration of treatment of the illness.
Please see Appendix 1 for further details of such medications which will be considered for this review.
Types of outcome measures
1. Efficacy measures
We will use validated depression or mania ratings scales to judge the outcome in two ways:
(1) according to changes in these scores (i.e. continuous outcomes); and
(2) outcomes where improvement on a rating scale has been dichotomised (such as clinical improvement as defined by a fixed percentage reduction in scores from baseline, e.g. 50%).
This could be any standardised scale including but not restricted to the Young Mania Rating Scale (YMRS) (Young 1978) and the Beck Depression Inventory, including current version (Beck 1996) and original version (Beck 1961).
2. Adverse effects
We will take note of subjects experiencing adverse effects of any nature.
Specific adverse effects and their prevalence will be estimated.
1. For each type of affective episode the following outcome measures will be estimated.
For acute manic episodes, efficacy of treatment will be measured by:
- requirement for hospital admission;
- length of hospital admission;
- time to cessation of additional treatment for manic symptoms;
- scores in manic symptom rating scales;
- scores in psychotic symptom scales.
For depressive episodes, efficacy of treatment will be measured by:
- requirement for hospital admission;
- length of hospital admission;
- time to cessation of additional treatment for depressive symptoms;
- scores in depression rating scales.
For mixed affective episodes, efficacy of treatment will be measured by:
- requirement for hospital admission;
- length of hospital admission;
- time to cessation of additional medication;
- scores in mania and/or depression rating scales.
If we come across studies where an investigational agent is used for maintenance and/or relapse prevention, they will be measured by:
- requirement of hospital admission;
- length of hospital admission.
2. Psychological social and occupational functioning as measured by:
- global impression of the clinician;
- global impression of the subject, family or significant others.
3. Acceptability of the pharmacological treatment, as measured by:
- subjects dropping out of the treatment during the study period;
- subjects' reports of satisfaction or otherwise with treatment.
4. Overall mortality rate:
- mortality rate on investigational agent versus others.
Search methods for identification of studies
CCDAN's Specialised Register (CCDANCTR)
The Cochrane Depression, Anxiety and Neurosis Group (CCDAN) maintains two clinical trials registers at their editorial base in Bristol, UK, a references register and a studies based register. The CCDANCTR-References Register contains over 30,000 reports of trials in depression, anxiety and neurosis. Approximately 65% of these references have been tagged to individual, coded trials. The coded trials are held in the CCDANCTR-Studies Register and records are linked between the two registers through the use of unique Study ID tags. Coding of trials is based on the EU-Psi coding manual. Please contact the CCDAN Trials Search Coordinator for further details. Reports of trials for inclusion in the Group's registers are collated from routine (weekly), generic searches of MEDLINE (1950-), EMBASE (1974-) and PsycINFO (1967-); quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL) and review specific searches of additional databases. Reports of trials are also sourced from international trials registers via the World Health Organization's trials portal (ICTRP), ClinicalTrials.gov, drug companies, the handsearching of key journals, conference proceedings and other (non-Cochrane) systematic reviews and meta-analyses. Details of CCDAN's generic search strategies can be found on the Group's website.
The CCDANCTR registers will be searched using terms for condition and age group only. Results will be screened for reports of trials using pharmacotherapeutic interventions in the acute phase of bipolar disorder. See Appendix 2 for details of search terms.
Searching other resources
Specialist journals and conference proceedings will be handsearched. These will include, but not be restricted to, the annual conference proceedings of the American Psychiatric Association (2001 onwards), Collegium Internationale Neuro-Psychopharmacologicum (CINP) (2001 onwards), British Association of Psychopharmacology (2001 onwards).
Handsearching of textbooks on affective disorders including Essential Psychopharmacology of Depression and Bipolar Disorder, Oxford Textbook of Psychiatry and Textbook of Psychopharmacology will also be carried out. We will check reference lists of papers or books. We plan to contact authors, experts in the field of treatment of late life bipolar disorder, and representatives of pharmaceutical companies to request information on published or unpublished trials.
Data collection and analysis
Selection of studies
Studies generated by the search strategies will be checked by two review authors (AGN and AV) to ensure they meet the inclusion criteria. Any disagreements will be resolved by consensus discussions with a third member of the review team.
The abstracts of the individual studies will be read to check if they fulfil the preliminary criteria for inclusion:
a) participants with bipolar disorder;
b) age over 55 years;
c) comparison of a pharmacotherapeutic intervention with a comparator.
Care will be taken to avoid inclusion of studies with duplicated results.
Data extraction and management
Two review authors (AGN and AV) will independently extract data concerning participant characteristics, age, bipolar diagnosis, intervention details and outcome measures from the included studies. Any disagreements will be resolved by consensus discussions with a third member of the review team.
Non-concurrence in selection and quality assessment will be reported.
1. Pharmacological treatment (mood stabiliser, antidepressants, antipsychotics, others) versus placebo.
2. Pharmacological treatment versus pharmacological treatment.
3. Combined mood stabilisers/antipsychotics versus placebo.
4. Combined mood stabilisers/antipsychotics versus other pharmacological treatment.
Assessment of risk of bias in included studies
Risk of bias will be assessed for each included study using The Cochrane Collaboration 'risk of bias' tool (Higgins 2009b). The following six domains will be considered:
- Sequence generation: was the allocation sequence adequately generated?
- Allocation concealment: was allocation adequately concealed?
- Blinding of participants, personnel and outcome assessors for each main outcome or class of outcomes: was knowledge of the allocated treatment adequately prevented during the study?
- Incomplete outcome data for each main outcome or class of outcomes: were incomplete outcome data adequately addressed?
- Selective outcome reporting: are reports of the study free of suggestion of selective outcome reporting?
- Other sources of bias: was the study apparently free of other problems that could put it at a high risk of bias?
A description of what was reported to have happened in each study will be provided, and a judgement on the risk of bias will be made for each domain within and across studies, based on the following three categories:
A. low risk of bias;
B. unclear risk of bias;
C. high risk of bias.
Two independent review authors will assess the risk of bias in selected studies (AGN and AV). Any disagreement will be discussed with a third review author. Where necessary, we will contact the authors of the studies for further information. All risk of bias data will be presented graphically and described in the text. We will use allocation concealment as a marker of trial quality for the purposes of undertaking sensitivity analyses.
Measures of treatment effect
For dichotomous outcomes the Odds Ratio (OR) will be calculated with 95% confidence limits. For continuously distributed outcome data the weighted mean difference (WMD) with 95% confidence limits will be used if all the trials used the same assessment scale. However, if the trials use different assessment scales, the standardised mean difference (SMD) with 95% confidence limits will be calculated. If effects on binary outcomes reached statistical significance, the Number Needed to Treat for an Additional Beneficial Outcome (NNTB) for effectiveness outcomes and the Number Needed to Treat for an Additional Harmful Outcome (NNTH) for side effects will be calculated. A P value of 0.05 and below will be chosen to indicate statistical significance of effects.
When it appears that data are skewed, they will be reported descriptively using recommended methods for identifying skewness from summary data (Deeks 2008). A negative outcome will be assumed if subjects drop out of the study for any other reason than response to treatment. Non-quantitative data will be presented descriptively. Outcomes concerning relapse or recurrence of bipolar disorder will be analysed excluding data from studies of discontinuation design.
Unit of analysis issues
We envisage that for most studies we shall be able to extract data from the end point data. However, if the study design is of cross-over, multiple arm or cluster randomised we shall address unit of analysis issues as below.
If any included trial has a cross-over design, only results from the first randomisation period will be considered.
Studies with multiple treatment groups
If the trial has three (or more) arms, for dichotomous outcomes, data from relevant active intervention arms will be collapsed into a single arm for comparison or data from relevant active intervention arms will be split equally between comparator arms. For continuous outcomes, means, SDs and number of participants for each active treatment group will be pooled across treatment arms as a function of the number of participants in each arm to be compared against the control group.
We shall use the generic inverse variance technique and the intraclass correlation coefficients to adjust for cluster effects using Review Manager 5 software (RevMan 2008).
Dealing with missing data
We will use intention-to-treat (ITT) data where available. For binary efficacy outcomes and continuously distributed outcomes, ITT analysis will be performed using available case analysis and if not available, using imputation. The results arising from these two methods will be compared during sensitivity analyses. Where ITT analysis is not possible, end-point data for trial completers will be used.
Assessment of heterogeneity
Inconsistency across studies will be quantified with the I
Assessment of reporting biases
The risk of publication bias will be graphically illustrated with the funnel plot method (Light 1984) and quantified with a linear regression test (Egger 1997), determining the linear regression coefficient between the log Odds Ratio (OR) and its standard error.
Data will be entered into Review Manager 5 software (RevMan 2008) by one reviewer (AV).
For synthesizing aggregate outcome measures, a random-effects model (Der Simonian 1986) will be chosen according to the recommendation of Brockwell 2001, with study effects being weighted with the Mantel-Haenszel approach (Mantel 1959). For outcomes with low effect heterogeneity (I
When it appears that data are skewed, using recommended methods for identifying skewness from summary data (Deeks 2008), they will be reported descriptively. A negative outcome will be assumed if subjects drop out of the study for any other reason than response to treatment. Non-quantitative data will be presented descriptively. Outcomes concerning relapse/recurrence of affective disorder will be analysed excluding data from studies of discontinuation design. Data from these studies will be analysed separately, to assess the effects of drug discontinuation.
Subgroup analysis and investigation of heterogeneity
We envisage that the number of studies which shall achieve the inclusion criteria of this meta-analysis will be few. Hence, to avoid Type I errors we plan to have only a few subgroup analyses which we anticipate will be possible and appropriate for the following:
- Presence of rapid cycling versus not. Rapid cycling tends to be more difficult to treat clinically; the ability of a particular pharmacological agent to treat this needs to be assessed.
- Variation in dose (low dosages versus adequate dosage range used in adult populations). Clinically it is observed that the older people with bipolar disorder need lower dosages of most psychotropics. It would be important to see if in a clinical trial setting low dosages have a beneficial response or not.
- Age of onset of bipolar disorder. Recent epidemiological data suggest that it is useful to differentiate between Early Onset Bipolar (EOB) disorder and Late Onset Bipolar (LOB) disorder as they have different etiopathogenetic profiles and response to medication (Vasudev 2010).
Heterogeneity between studies will be investigated. Potential sources include those listed above. Any other potential source of heterogeneity which is apparent on examining the studies will also be included as a post hoc analysis.
The following sensitivity analyses for primary outcomes are planned a priori.
- We shall exclude trials with unclear concealment of random allocation or unclear double blinding, or both.
- We shall exclude trials with more than 50% non-completers.
- We shall exclude trials that are sponsored by the pharmaceutical industry versus those that were investigator-driven and of non-profit funding.
- We shall exclude trials which include patients with unipolar disorder or schizoaffective disorder, or both.
- We shall perform the worst-case scenario ITT: that all patients in the experimental group experience the negative outcome and all those in the comparison group experience the positive outcome.
- We shall perform the best-case scenario ITT: that all patients in the experimental group experienced the positive outcome and all those in the comparison group experienced the negative outcome.
CRG Funding Acknowledgement:
The National Institute for Health Research (NIHR) is the largest single funder of the Cochrane Depression, Anxiety and Neurosis Group.
The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR, NHS or the Department of Health.
Appendix 1. Drug classes and medications
Anti-manic agent: lithium
Anticonvulsants: valproic acid and its congeners, carbamazepine, oxcarbazepine, gabapentin, topiramate, tiagabine, lamotrigine, levetiracetam and other anticonvulsants
a) Typical antipsychotics including haloperidol, chlorpromazine and other such drugs
b) Atypical antipsychotics including olanzapine, quetiapine, ziprasidone, aripiprazole, clozapine, asenapine and other such drugs
a) Tricyclic antidepressants: amitriptyline, imipramine, trimipramine, doxepin, desipramine, protriptyline, nortriptyline, clomipramine, dothiepin, lofepramine
b) SSRIs: zimelidine, fluvoxamine, fluoxetine, paroxetine, sertraline, citalopram, escitalopram
c) SNRIs: venlafaxine, milnacipram, duloxetine
d) NASSAs: mirtazapine
e) MAOIs: Irreversible: phenelzine, tranylcipromine, izocarboxazid; reversible: brofaramine, moclobemide, tyrima.
Other antidepressants: NARIs: reboxetine, atomoxetine; NDRIs: amineptine, buproprion; SARIs: trazodone; unclassified: agomelatine, vilazodone; other heterocyclic antidepressants: mianserin, amoxapine, maprotiline.
We will only combine drugs in analyses where they are from the same class. The 'other antidepressants' will be presented together, but again only drugs of the same class will be combined to produce a pooled effect, e.g. duloxetine and venlafaxine.
SSRI: selective serotonin reuptake inhibitor
MAOI: monoamine oxidase inhibitor
NARI: noradrenaline reuptake inhibitor
NDRI: norepinephrine-dopamine reuptake inhibitor
SNRI: serotonin–norepinephrine reuptake inhibitor
NASSA: noradrenergic and specific serotonergic antidepressant
SARI: serotonin antagonist and reuptake inhibitor
Other pharmacological therapy: including, but not restricted to omega 3 fatty acids, inositol, pramipexole.
Appendix 2. Search terms
The CCDANCTR-Studies Register will be searched using the following terms:
Condition = (bipolar or mania or manic or hypomania or ("affective disorder*" and (psychos* or psychotic)) or schizoaffective)
Age Group = Aged
The CCDANCTR-References register will be searched using a more sensitive set of terms to identify additional untagged/uncoded references. This search will also identify references tagged to adult studies where participants were young to old, i.e. 55 to 65 years.
Free-Text = (bipolar or mania or manic or hypomani* or (affective and (psychos* or psychotic)) or ((*rapid or ultradian) and cycling) or schizoaffective)
Keywords = (Aged) or Free-Text=(elder* or frail or geriatric* or seniors or retired or retirement or “late life” or “later life” or “old age” or “old people” or “older people” or "old person*" or "older person*" or "old adult*" or “older adult*” or “old men” or “older men” or “old women” or “older women” or “old male*” or “older male*” or “old female*” or “older female*” or “old patient*” or “older patient*” or “old old” or old-old or “very old” or "senior citizen*" or sedentary or "care home*" or "nursing home*" or "middle age*" or middle-age* or midlife or mid-life or young-old or "young old")
or Abstract = ("50 years" or "55 years" or "60 years" or "64 years" or "65 years" or "70 years" or "75 years" or "79 years" or "80 years" or "85 years" or "90 years" or "95 years" or "older than 50" or "older than 55" or "older than 60" or "older than 65" or "older than 70" or "older than 75" or "older than 80" or "older than 85" or "older than 90" or "older than 95")
Contributions of authors
AV wrote the protocol.
AT and HG reviewed the protocol and offered critical comments.
Declarations of interest
AV has in the last five years received educational grants from Pfizer Inc. He has also received honoraria for presentations from Eli-Lilly.
AT none known.
Within the last five years HG has received grants/ research support from the Medical Research Council (UK), the NHS National Institute for Health Research, Pfizer Inc.. He has received consulting fees from AstraZeneca, Bial, BMS, Cephalon, Merck, Organon, Gedeon Richter, Sepracor and UBC. He has also received honoraria for presentations for AstraZeneca, BMS, Desitin, Eli-Lilly, Janssen-Cilag, Merck, Pfizer Inc., Sanofi-Aventis, Servier and UBC.
Sources of support
- Newcastle University, UK.
- University of Western Ontario, Canada.
- No sources of support supplied