Intervenciones para el manejo de la obesidad en personas con trastorno bipolar
Resumen
Antecedentes
El trastorno bipolar es una de las enfermedades mentales graves más frecuentes con 60 millones de personas afectadas en todo el mundo. Se caracteriza por alteraciones extremas del estado de ánimo, la cognición y la conducta, y puede tener un impacto negativo significativo sobre el funcionamiento y la calidad de vida de la persona que lo sufre. En comparación con la población general, la prevalencia de la obesidad comórbida es significativamente mayor en el trastorno bipolar. Aproximadamente el 68% de los pacientes con trastorno bipolar que buscan tratamiento tienen sobrepeso u obesidad. Los médicos son conscientes de que la obesidad puede contribuir a otros problemas de salud física de las personas con trastorno bipolar, como la diabetes, la hipertensión, el síndrome metabólico, las enfermedades cardiovasculares y las enfermedades coronarias. Las enfermedades cardiovasculares son la principal causa de muerte prematura en el trastorno bipolar, ya que se producen una década o más antes que en la población general. Entre los factores contribuyentes, figuran los relacionados con la enfermedad (factores relacionados con el estado de ánimo, es decir, manía o depresión), factores relacionados con el tratamiento (implicaciones para el peso y otros efectos secundarios de los medicamentos) y factores relacionados con el estilo de vida (inactividad física, dieta deficiente, tabaquismo, abuso de sustancias). Los enfoques para tratar la obesidad en los pacientes con trastorno bipolar son diversos e incluyen intervenciones no farmacológicas (es decir, dietéticas, de ejercicio, conductuales o de múltiples componentes), intervenciones farmacológicas (es decir, medicamentos para la pérdida de peso o el cambio de medicación) y cirugía bariátrica.
Objetivos
Evaluar la efectividad de las intervenciones para el manejo de la obesidad en pacientes con trastorno bipolar.
Métodos de búsqueda
Se efectuaron búsquedas en el Registro de Ensayos Controlados de Cochrane de Trastornos Mentales Comunes (CCMDCTR) y el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register for Controlled Trials) (CENTRAL) hasta febrero de 2019. Se realizaron búsquedas adicionales a través de las bases de datos de Ovid, incluidas MEDLINE, Embase y PsycInfo hasta mayo de 2020. Se hicieron búsquedas en la Plataforma de registros internacionales de ensayos clínicos de la Organización Mundial de la Salud (OMS) (International Clinical Trials Registry Platform) (ICTRP) y en ClinicalTrials.gov. También se comprobaron las listas de referencia de todos los artículos llevados a la fase de texto completo y todas las revisiones sistemáticas pertinentes.
Criterios de selección
Fueron elegibles para su inclusión los ensayos controlados aleatorizados (ECA), a nivel individual o de grupo, y los diseños cruzados de intervenciones para el tratamiento de la obesidad, en los que al menos el 80% de los participantes del estudio tenían un diagnóstico clínico de trastorno bipolar y obesidad comórbida (índice de masa corporal (IMC) ≥ 30 kg/m²). No se siguieron criterios de exclusión según el tipo de trastorno bipolar, fase de la enfermedad, edad o género. Se incluyeron intervenciones no farmacológicas que comprendían intervenciones dietéticas, de ejercicio, conductuales y de múltiples componentes; intervenciones farmacológicas que consistían en medicamentos para la pérdida de peso e intervenciones de cambio de medicación; e intervenciones quirúrgicas como bypass gástrico, bandas gástricas, derivación biliopancreática y gastroplastia vertical con banda. Los comparadores incluyeron los siguientes enfoques: intervención dietética versus comparador inactivo; intervención con ejercicio versus comparador inactivo; intervención conductual versus comparador inactivo; intervención con múltiples componentes en el estilo de vida versus comparador inactivo; intervención con cambio de medicación versus comparador inactivo; intervención con medicación para la pérdida de peso versus comparador inactivo; e intervención quirúrgica versus comparador inactivo. Los principales desenlaces de interés fueron los cambios en la masa corporal, los eventos adversos informados por los pacientes y la calidad de vida.
Obtención y análisis de los datos
Cuatro autores de la revisión participaron en la selección de los estudios. Dos autores de la revisión examinaron de forma independiente los títulos y los resúmenes de los estudios identificados en la búsqueda. Los estudios llevados a la etapa de texto completo fueron luego revisados por otros dos autores de la revisión que trabajaron de manera independiente. Sin embargo, ninguno de los estudios con texto completo cumplió los criterios de inclusión. Si se hubiesen incluido estudios, se habría evaluado la calidad metodológica mediante los criterios recomendados en el Manual Cochrane para Revisiones Sistemáticas de Intervenciones (Cochrane Handbook for Systematic Reviews of Interventions). Se pretendían combinar los datos dicotómicos utilizando riesgos relativos (RR) y los datos continuos utilizando diferencias de medias (DM). Para cada uno de los desenlaces, se pretendió calcular un tamaño del efecto general con intervalos de confianza (IC) del 95%.
Resultados principales
Ninguno de los estudios cribados cumplió los criterios de inclusión.
Conclusiones de los autores
Ninguno de los estudios evaluados cumplió los criterios de inclusión para esta revisión. Por lo tanto, no se pudo determinar la efectividad de las intervenciones para el manejo de la obesidad en pacientes con trastorno bipolar. Dada la magnitud y el impacto del problema y la ausencia de evidencia, esta revisión recalca la necesidad de investigar en esta área. Se sugiere que los ECA se centren solamente en poblaciones con trastorno bipolar y obesidad comórbida. Se identificaron varios estudios en curso que pueden incluirse en la actualización de esta revisión.
PICO
Resumen en términos sencillos
Intervenciones para el manejo de la obesidad en personas con trastorno bipolar
¿Por qué es importante esta revisión?
El trastorno bipolar es una de las enfermedades mentales graves más frecuentes y tiene impacto sobre el estado de ánimo, el pensamiento, la conducta, el funcionamiento y la calidad de vida. Aproximadamente 60 millones de personas en todo el mundo están afectadas por este trastorno. La obesidad a menudo se asocia con esta enfermedad, y cuando esto sucede, pueden aparecer otras enfermedades de salud física como la diabetes y las enfermedades cardiovasculares, lo que conduce a una muerte prematura. Se utilizan muchos enfoques para manejar la obesidad en el trastorno bipolar, pero hasta ahora, no está claro que alguno o una combinación de ellos sea efectiva.
¿Quién puede estar interesado en esta revisión?
Psiquiatras y cualquier miembro de un equipo multidisciplinario que atienda a personas con trastorno bipolar y obesidad comórbida. Los hallazgos de esta revisión serán también de interés para investigadores y personas con trastorno bipolar y sus familias.
¿Qué pregunta pretende contestar esta revisión?
Esta revisión trató de evaluar la efectividad de las intervenciones utilizadas para abordar el problema de la obesidad en el trastorno bipolar.
¿Qué estudios se incluyeron en la revisión?
Se realizaron búsquedas en las bases de datos hasta febrero de 2019 de estudios de intervenciones utilizadas en el manejo de la obesidad en individuos con trastorno bipolar. Ninguno de los estudios revisados cumplió los criterios de inclusión.
¿Qué revela la evidencia de la revisión?
No hay evidencia de la efectividad de las intervenciones farmacológicas, no farmacológicas o quirúrgicas para el manejo de la obesidad en personas con trastorno bipolar.
¿Qué debe suceder a continuación?
Existe una necesidad urgente de realizar ensayos controlados aleatorizados para evaluar la efectividad de los enfoques no farmacológicos, farmacológicos y quirúrgicos para el tratamiento de la obesidad en poblaciones con trastorno bipolar. Los ensayos deben incluir solamente a personas con un diagnóstico clínico de trastorno bipolar y obesidad comórbida.
Conclusiones de los autores
Implicaciones para la práctica
Los hallazgos de esta revisión sugieren que no existe evidencia que apoye los enfoques no farmacológicos, farmacológicos o quirúrgicos para el manejo de la obesidad en el trastorno bipolar.
Implicaciones para la investigación
Se necesitan realizar ensayos controlados aleatorizados bien diseñados de intervenciones no farmacológicas, farmacológicas y/o quirúrgicas para abordar la obesidad en el trastorno bipolar. Los estudios deben centrarse únicamente en personas con trastorno bipolar, en lugar de en poblaciones mixtas con enfermedades mentales graves, y todos los participantes en el ensayo deben haber recibido un diagnóstico de obesidad comórbida. Se identificaron varios estudios en curso que pueden incluirse en la actualización de esta revisión.
Authors' conclusions
Background
Bipolar disorder is one of the most common serious mental illnesses (SMIs) characterised by mood instability, results in marked impairment in overall functioning and health‐related quality of life (De Hert 2011; Klienman 2003). Bipolar disorder is the sixth leading cause of disability worldwide in people aged 15 to 44 years (Klienman 2003), and it has a worldwide prevalence of 2.4% (Merikangas 2011). Globally, bipolar disorder affects approximately 60 million people (WHO 2015).
Weight gain and obesity have long been recognised in mental health practice as matters of significant concern (Baptista 1999; McIntyre 2001; McIntyre 2010). Individuals with bipolar disorder are more frequently overweight (body mass index (BMI) 25.0 to 29.9) or obese (BMI ≥ 30), or have a higher prevalence of central obesity (abdominal fat), or both, compared with the general population (McElroy 2004). Clinical research suggests that up to 68% of bipolar patients seeking treatment are overweight or obese (De Hert 2011). Analysis of data from the US National Comorbidity Survey Replication demonstrated that obesity is associated with a significant increase in a lifetime diagnosis of bipolar disorder (Simon 2006).
Clinical research suggests that approximately 68% of treatment‐seeking individuals with bipolar disorder are overweight or obese (McElroy 2004). Traditionally, the issue of weight gain in people with mental illness was perceived as less important than mental wellness (Fontaine 2001). However, clinicians today recognise that obesity has the potential to contribute to other physical health conditions in people with bipolar disorder, including diabetes, hypertension, metabolic syndrome (MetS), cardiovascular disease, and coronary heart disease (De Hert 2011). Cardiovascular disease is the leading cause of premature death in bipolar disorder, occurring a decade or more earlier, on average, than in the general population (Goldstein 2014).
Weight gain is a commonly reported side effect of medications used in the treatment of bipolar disorder and is associated with lower quality of life in this population. Factors contributing to obesity in the bipolar disorder population are diverse but generally stem from illness‐related factors (mood‐related factors, i.e. mania or depression), treatment‐related factors (weight implications and other side effects of medications), or lifestyle factors (physical inactivity, poor diet, smoking, and substance abuse) (De Hert 2011; Firth 2019; Newcomer 2007; Stahl 2009; Ussher 2011), or a combination of some or all of these factors. Obesity and related physical illnesses are associated with aggravation of depression, morbid course of illness, non‐adherence with treatment regimens, poor treatment outcomes, and an increased prevalence of suicide in people with bipolar disorder (Fagiolini 2003).
Bipolar disorder leads to a significant economic burden on both the individual and society as a whole. Indirect physical health costs account for most of this burden, which is due largely to lost work productivity of people with bipolar disorder and their carers. Hospitalisation and emergency department services, psychiatric visits, and costs of medication are the main contributors to direct costs (Guvstavsson 2011).
Description of the condition
Bipolar disorder is a recurrent and sometimes chronic mental illness. The term 'bipolar disorder' refers to a group of affective or mood disorders, typically characterised by episodes of depression and either mania (elated or irritable mood or both), manifested as increased energy and reduced need for sleep, or hypomania, symptoms of which are less severe or less protracted than those of mania. The Fifth Edition of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM‐V) identifies four subtypes of bipolar disorder (APA 2013).
-
Bipolar disorder type I is defined as episodes of depression and at least one episode of mania.
-
Bipolar disorder type II is defined as at least one episode of major depression and at least one hypomanic episode but no manic episodes.
-
Cyclothymic disorder refers to a number of episodes of hypomanic and depressive symptoms in which the depressive symptoms do not meet the criteria for depression.
-
Bipolar disorder not otherwise specified (NOS) refers to depressive and hypomanic episodes that may change rapidly, yet do not meet the full diagnostic criteria for any other illnesses
The World Health Organization International Classification of Diseases (ICD‐10) defines bipolar disorder as characterised by two or more episodes in which the individual's mood and activity levels are very disturbed; bipolar disorder may involve an elevation of mood with increased energy and activity (hypomania or mania), and at other times lowering of mood and decreased energy and activity (depression) (WHO 1992). Repeated episodes of hypomania or mania only are classified as bipolar disorder.
Description of the intervention
A diversity of interventions are available for the treatment of obesity. The main interventions can be classified into non‐pharmacological, pharmacological, and surgical.
Non‐pharmacological interventions
Non‐pharmacological interventions include four approaches.
Dietary interventions
This approach focuses on lifestyle modification, generally encompassing diet changes, whereby there is an attempt to enhance dietary restraint by providing adaptive dietary strategies and by discouraging maladaptive dietary practices (Shaw 2005). This approach also includes self‐monitoring. The aim is to achieve weight loss by reducing daily intake of food (Harvey 2004; Shaw 2005).
Exercise interventions
This approach aims to increase daily expenditure of energy via increased physical activity (Harvey 2004;Shaw 2005). Physical activity approaches include all types of aerobic and anaerobic exercise approaches.
Behaviour change interventions
The aim of behaviour change strategies is to elicit behaviour change by enabling individuals to explore and resolve mixed feelings they may have about change. Interventions such as motivational interviewing, which is grounded in a client‐centred approach, help individuals move to greater readiness to change behaviour (Rollnick 1995). Another approach is cognitive‐behavioural therapy, which teaches individuals behavioural and cognitive strategies, for example, breaking negative behaviour cycles, with focus on achieving and maintaining lifestyle changes (Beck 1979). This can be achieved by stimulus control, goal‐setting, and self‐monitoring. All of these interventions can be delivered at an individual or group level.
Multi‐component interventions
Multi‐component interventions may contain one or more components of any of the three interventions described above.
Pharmacological interventions
Pharmacological interventions may involve the following approaches.
Weight loss medication interventions
This approach involves taking drugs that inhibit appetite or food absorption, or both, or that act centrally. Obesity guidelines currently recommend that drug therapy be considered for patients with a BMI of 30 kg/m² or more (Lau 2007; Padwal 2007). Five anti‐obesity drugs are commonly used, namely, orlistat, lorcaserin, phentermine/topiramate, naloxone/bupropion, and liraglutide (Patel 2015).
Medication switching interventions
Weight gain is a well‐recognised side effect of some of the medications used in treatment for bipolar disorder (Narasimhan 2007). For example, second‐generation atypical anti‐psychotic drugs such as olanzapine have the propensity to induce weight gain (Lieberman 2005; McIntyre 2001). The mood‐stabilising drug lithium is also associated with weight gain (Shrivastava 2010). Consequently, the second pharmacological approach for managing weight gain in bipolar disorder involves switching anti‐psychotic or mood‐stabilising medications to alternative anti‐psychotic or mood‐stabilising medications with less potential for weight gain (Weiden 2007; White 2013).
Adjunctive metformin
The British Association for Psychopharmacology recommends initiation of metformin, the glucose‐lowering drug for the treatment of type 2 diabetes, for people with a history of psychosis and comorbid obesity for prevention of diabetes and treatment of anti‐psychotic‐induced weight gain (Cooper 2016).
Surgical interventions
Bariatric surgery (weight loss surgery) is considered almost exclusively for patients with a BMI of 40 kg/m² or more. It may also be considered for patients with a BMI of 35 kg/m² in the presence of an associated serious physical illness such as diabetes (NICE 2006).
How the intervention might work
Non‐pharmacological interventions
Common to all non‐pharmacological interventions for weight management is the concept of lifestyle modification of diet, physical exercise, or behaviour, or a combination of these approaches (McElroy 2009; Patel 2015). A central aim of lifestyle change approaches is to achieve and maintain weight loss by reducing caloric intake and increasing physical activity. The behavioural component seeks to inspire behaviour change toward food intake and physical activity. Weight loss, however modest, can have positive effects on overall well‐being and is considered a successful outcome, especially if maintained over time. Weight loss of between 5% and 15% has been shown to improve lipid levels and reduce low‐density lipoprotein (LDL), total cholesterol, blood pressure, and risk of cardiovascular disease (Aucott 2005; Wing 2011). Weight loss of 10 kg has been shown to decrease systolic blood pressure by 5 mmHg and diastolic blood pressure by 6 mmHg (Aucott 2005). Weight loss has also been reported to improve psychological well‐being and quality of life and to facilitate a more active lifestyle, which can help to maintain or increase weight loss.
Pharmacological interventions
The five most commonly used weight loss drugs are orlistat, lorcaserin, phentermine/topiramate, naloxone/bupropion, and liraglutide (Patel 2015). Orlistat acts peripherally by preventing intestinal fat absorption by inhibiting the pancreatic lipase enzyme (Heck 2000). Lorcaserin is a selective agonist of 5‐HT2C receptors (Hurren 2011). Phentermine/topiramate is a combination treatment that consists of phentermine, which increases central noradrenaline levels, and topiramate, which is a central modulator of the inhibitory neurotransmitter GABA (Khorassani 2015). Another combination therapy comprises the mu‐opioid receptor antagonist naloxone and bupropion (Yanovski 2015). Finally, liraglutide is a glucagon‐like peptide 1 (GLP‐1) receptor agonist that was originally developed to treat type 2 diabetes mellitus but has now been approved for long‐term weight management. Drug therapies that are no longer used focused on centrally acting agents that targeted the catecholaminergic or serotonergic system, or both (such as sibutramine, mazindol, diethylpropion, benzphetamine, and phendimetrazine), or the cannabinoid receptor antagonist rimonabant. Due to safety concerns of previous therapeutic approaches, great care is advised in treatment for weight reduction with current drugs (American College of Cardiology 2014).
Medication switching involves changing from prescribing drugs associated with weight gain to prescribing drugs associated with less weight gain. A Cochrane Review evaluating the effects of anti‐psychotic switching for people with schizophrenia who have neuroleptic‐induced weight gain concluded that patients who were switched to aripiprazole or quetiapine from olanzapine lost weight, had a reduced BMI, and had improved profiles of fasting glucose and lipids (Mukundan 2010).
Metformin may contribute to weight loss by reducing insulin resistance and suppressing appetite. A systematic review and meta‐analysis demonstrated that metformin is effective in treating anti‐psychotic‐induced weight gain in patients with schizophrenia or schizoaffective disorder (De Silva 2016).
Surgical interventions
Bariatric surgery for obesity includes a variety of approaches. Weight loss is achieved by reducing the size of the stomach, so that food intake is restricted and weight loss is induced (Colquitt 2014). Surgical approaches include gastric bypass, gastric bands, biliopancreatic diversion, and vertical banded gastroplasty.
Why it is important to do this review
A previous systematic review evaluated interventions targeting physical health comorbidities in people with SMI (Cabassa 2010). However, the term 'serious mental illness' is an umbrella term incorporating schizophrenia, major depressive disorder, and bipolar disorder. Our proposed systematic review seeks to focus only on people with bipolar disorder given potential differences in patterning of health risks between individuals with bipolar disorder and those with other conditions (e.g. schizophrenia) within the broader category of serious mental illness (Hayes 2017). Given the prevalence and devastating effects of obesity in this population, as well as the enormous economic burden to society, any intervention that is effective in addressing the problem of obesity in bipolar disorder would have a major impact. Psychological and health‐related quality of life improvements have been shown in people with bipolar disorder following weight loss. Moreover, weight loss allows for improved psychological functioning, a more active lifestyle, and increased physical activity, which in turn may induce further weight loss, weight maintenance, or both.
Objectives
To assess the effectiveness of interventions for the management of obesity in people with bipolar disorder.
Methods
Criteria for considering studies for this review
Types of studies
We considered randomised controlled trials (RCTs), randomised at the level of the individual or cluster. We also considered studies employing a cross‐over design, using data from the first active treatment only, that is, before the first cross‐over. Non‐randomised controlled studies in which assignment to treatment group was decided through non‐random methods were not eligible for inclusion. We intended to include studies regardless of their publication status.
Types of participants
Participant characteristics
All participants with a clinical diagnosis of bipolar disorder and comorbid obesity were the focus of this review. No exclusion was based on type of bipolar disorder, stage of illness, age, or gender.
Diagnosis
We considered in this review only participants with a clinical diagnosis of bipolar disorder and comorbid obesity. We considered bipolar disorder diagnosed according to the criteria laid out by the American Psychiatric Association in DSM‐IV or DSM‐V (APA 1994; APA 2013), or criteria specified by the WHO in ICD‐10 (WHO 1992). A clinical diagnosis of obesity was defined in the context of this review as BMI of 30 kg/m² or more. A priori, we anticipated that there would be few studies in which all participants had a clinical diagnosis of bipolar disorder and comorbid obesity. Therefore, we would have included all studies in which 80% or more of the participants had both bipolar disorder and comorbid obesity.
Comorbidities
Comorbid obesity, but participants would have been included regardless of any other diagnosed comorbidity.
Setting
We applied no restriction on the treatment setting.
Types of interventions
Experimental interventions
We planned to include the following interventions.
1. Non‐pharmacological interventions (i.e. lifestyle modifications).
-
Diet.
-
Exercise.
-
Behaviour.
-
Multi‐component lifestyle interventions, which may include one or more of the components detailed in 1, 2, and 3 above.
We placed no restriction on who delivered the intervention or where the intervention was delivered, nor on frequency, intensity, or duration of the intervention.
2. Pharmacological interventions.
-
Weight loss drugs.
-
Medication switching.
-
Adjunctive metformin.
Had we included studies, there would have been no restriction on type of drug, dose, frequency of delivery, route of delivery, or length of exposure.
3. Surgical interventions.
-
Gastric bypass.
-
Gastric banding.
-
Biliopancreatic diversion.
-
Vertical banded gastroplasty.
We placed no restriction on type of surgical procedure nor on length of follow‐up.
Comparator interventions
We planned to include the following comparators.
1. Inactive comparator.
-
No treatment.
-
Treatment as usual (TAU), also called standard care or usual care.
-
Placebo (inactive/dummy), defined as a control condition that is regarded by researchers as inactive but is regarded by participants as active.
2. Active comparator.
-
A dietary intervention versus a different dietary intervention.
-
An exercise intervention versus a different exercise intervention.
-
A behavioural intervention versus a different behavioural intervention.
-
A medication switching intervention versus a different medication switching intervention.
-
A weight loss medication intervention versus a different weight loss medication intervention.
-
A surgical intervention versus a different surgical intervention.
Main comparisons
Planned main comparisons to be reported in 'Summary of findings' tables were as follows.
-
Dietary intervention versus inactive comparator.
-
Exercise intervention versus inactive comparator.
-
Behavioural intervention versus inactive comparator.
-
Multi‐component lifestyle intervention versus inactive comparator.
-
Medication switching interventions versus inactive comparator.
-
Weight loss medication interventions versus inactive comparator.
-
Surgical interventions versus inactive comparator.
Types of outcome measures
Planned primary and secondary outcomes of interest included the following:
Primary outcomes
-
Changes in body mass, measured as change in BMI.
-
Patient‐reported adverse events (e.g. pain, distress).
-
Quality of life measured by the following validated quality of life measurement scales: the Quality of Life in Bipolar Disorder Scale (QOL‐BD) (Michalak 2010), as well as the suite of Short Form (SF) Health Surveys originating from the SF‐36 tool (Ware 1993).
Secondary outcomes
-
Mood measured by validated measurement scales, for example, the Hamilton Depression Rating Scale (HDRS) may be used to indicate depression and severity of depression (Hamilton 1960). The Young Mania Rating Scale (YMRS) may be used to assess the severity of manic episodes (Young 1978).
-
Global functioning measured by validated measurement scales, for example, the Global Assessment of Functioning (GAF) and the Clinical Global Impression Bipolar Scale (CGI‐BP) (Jones 1995; Spearing 1997).
-
Clinician‐reported adverse events (e.g. infection, drug toxicity).
-
Blood pressure.
-
Total cholesterol.
-
LDLs.
-
Blood glucose levels.
Timing of outcome assessment
We anticipated that study authors would report response rates at various time points during and post intervention; we therefore planned to subdivide the timing of outcome assessments as follows.
-
Short‐term effects, measured up to 12 months after completion of the intervention, which was the primary time point to be reported in the 'Summary of findings' table.
-
Sustained effects, measured at least 12 months after completion of the intervention.
If included studies provided data at more than one time point, we would have included one set of data in the short‐term effects meta‐analysis, choosing the time point nearest 12 months, and one set of data in the sustained‐effects meta‐analysis, choosing the longest data collection period possible.
Hierarchy of outcome measures
We planned to measure data on the primary outcome ‐ change in body mass ‐ and include them in the meta‐analysis using BMI only.
If data on change in quality of life were available from studies that were sufficiently similar to allow a meta‐analysis, we planned to perform one meta‐analysis using data measured by QOL‐BD, and a separate meta‐analysis using data from the SF suite of scales; we planned to report both of these in our 'Summary of findings' table.
For our secondary outcomes, we planned to include data from HDRS and the Beck Depression Inventory (BDI) in our meta‐analysis for depressed mood, and data from the YMRS in the meta‐analysis of mania. We planned to conduct separate meta‐analysis for overall global assessment using data from GAF and CGI‐BP. We did not plan to report secondary outcomes in our 'Summary of findings' table.
If a study reported an outcome in more than one way, we planned to include relevant data in each meta‐analysis.
Search methods for identification of studies
Cochrane Common Mental Disorders Group Specialised Register (CCMDCTR)
The Cochrane Common Mental Disorders Group (CCMD) maintains two archived clinical trials registers at its editorial base in York, UK: a references register and a studies‐based register. The CCMDCTR‐References Register contains over 40,000 reports of RCTs on depression, anxiety, and neurosis. Approximately 50% of these references have been tagged to individual, coded trials. The coded trials are held in the CCMDCTR‐Studies Register, and records are linked between the two registers through the use of unique Study ID tags. Coding of trials is based on a bespoke manual used in the creation of a database of mental health/psychiatry trials (PsiTri), as part of an EU funded project (please contact the CCMD Information Specialist for further details). Reports of trials for inclusion in the Group Register were collated from routine (weekly), generic searches of MEDLINE (1950 to 2019), Embase (1974 to 2019), and PsycINFO (1967 to 2019); quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL); and review‐specific searches of additional databases. Reports of trials were sourced from international trials registers via the World Health Organization trials portal (the International Clinical Trials Registry Platform (ICTRP)) and pharmaceutical companies, and by handsearching of key journals, conference proceedings, and other (non‐Cochrane) systematic reviews and meta‐analyses. Details of the core search strategies of CCMD (used to identify RCTs) can be found on the Group's website, and an example of the core MEDLINE search is displayed in Appendix 1.
Electronic searches
An information specialist with the Cochrane Common Mental Disorders Group searched the biomedical databases listed below, using relevant keywords, subject headings (controlled vocabularies), and search syntax, appropriate to each resource (to 19 February 2019) (Appendix 2),
An update search was performed on 22 May 2020, which employed a more sensitive set of terms for anti‐psychotic‐induced weight gain (Appendix 3).
-
Cochrane Common Mental Disorders Controlled Trials Register (CCMDCTR) (all years to 14 June 2016).
-
Cochrane Central Register of Controlled Trials, in the Cochrane Library (2020; Issue 5 of 12).
-
Ovid MEDLINE (1946 to 22 May 2020).
-
Ovid Embase (1974 to 2020 Week 21).
-
Ovid PsycINFO (all years to May Week 3 2020).
We searched international trials registries via the World Health Organization trials portal (ICTRP) and ClinicalTrials.gov to identify unpublished or ongoing studies.
We did not apply to the searches any restrictions on date, language, or publication status.
Searching other resources
Grey literature
We searched the following sources of grey literature: theses via PsycINFO (Appendix 2); the British Library E‐theses Online Service (EthOS); and the DART‐Europe E‐theses Portal. We also checked the reference lists of all papers brought to full‐text stage and all relevant systematic reviews.
Data collection and analysis
Selection of studies
Two review authors (AT and FJ) independently screened all titles and abstracts identified through the literature searches to identify those that met the inclusion criteria. We retrieved the full text of studies identified as potentially relevant by at least one review author. Two other review authors (SS and YC) independently screened the full‐text articles for inclusion or exclusion. Disagreements were resolved by discussion, or, when necessary, we consulted a third review author (FJ or AT). All studies excluded at the full‐text stage are listed as excluded studies, and reasons for their exclusion are presented in the Characteristics of excluded studies table. The screening and selection processes are presented in the adapted PRISMA flowchart (Liberati 2009).
Data extraction and management
We designed an electronic data extraction form, and we planned that two review authors (SS and FJ) would independently extract data from eligible studies. Any disagreements would have been resolved by discussion, or, if necessary, a third review author (DD) would have been consulted. We planned for one review author (FJ) to enter extracted data into Review Manager Software 5.3 (RevMan 2014), and for a second review author (AT) to check for accuracy and consistency against the data extraction sheets. We planned to extract the following data.
-
Methods: study design, date, total duration of and length of time each participant was part of the study, details of any 'run‐in' period, number of study centres and locations, study setting, date of the study.
-
Participants: inclusion criteria and exclusion criteria, method of recruitment, number of participants eligible and number randomised, reasons for not including eligible participants, baseline imbalances, and withdrawals and numbers lost to follow‐up in each arm. Participant characteristics: age, sex, duration and severity of condition, race, diagnostic criteria, occupational status.
-
Intervention(s): details of intervention (name; intervention components, including any materials used by study personnel or given to participants; dose, location, timing, and mode of administration; duration of intervention; providers; scope for and details of modifications during the study).
-
Comparison (including definition of usual care when appropriate): concomitant medications, excluded medications.
-
Outcomes: unit of analysis, primary and secondary outcomes specified and collected, time points reported, scales used to measure outcomes, person/method of recording outcomes, baseline and end of intervention data for outcomes of interest. Data to assess risk of bias of each study, as required by the Cochrane 'Risk of bias' tool.
-
Notes: funding for the study and notable conflicts of interest of study authors; any other study‐specific information of importance not already captured above.
Assessment of risk of bias in included studies
We planned to assess risk of bias of the included studies using Cochrane's tool for assessing risk of bias, as outlined in the Cochrane Handbook for Systematic Reviews of Interventions, and contained in RevMan 5 (Higgins 2017; RevMan 2014).
Two review authors (AT and FJ) planned to assess risk of bias independently for each study. We planned to resolve any disagreements through consultation with a third review author (DD).
For randomised studies, we planned to base our assessment on the following domains.
-
Random sequence generation.
-
Allocation concealment.
-
Blinding of participants and personnel.
-
Blinding of outcome assessment.
-
Incomplete outcome data.
-
Selective outcome reporting.
-
Other bias.
We planned to judge each potential source of bias as having high, low, or unclear risk, and to provide a supporting quotation from the study report together with a justification for our judgement in the 'Risk of bias' table. We planned to summarise risk of bias judgements across different studies for each of the domains listed. If information on risk of bias related to unpublished data or correspondence with a study author, we would have noted this in the 'Risk of bias' table.
For cluster‐randomised studies, we planned to assess these additional sources of bias, as detailed in Section 16.3.2 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
-
Recruitment bias.
-
Baseline imbalance.
We planned to judge each additional source of bias as having high, low, or unclear risk, and to provide a supporting quotation from the study report together with a justification for our judgement in the 'Risk of bias' table.
Dichotomous data
We planned to express results for dichotomous outcome measures using risk ratios (RRs) and associated 95% confidence intervals (CIs) to reflect uncertainty of the point estimate of effects.
Continuous data
For continuous outcome measures, we planned to calculate mean differences (MDs) and standard deviations (SDs) with corresponding 95% CIs. We planned to use standardised mean differences (SMDs) with 95% CIs to combine outcomes from trials that measure the same outcome using different scales (Higgins 2011).
Unit of analysis issues
We envisaged that for most studies, we would have been able to extract data from baseline and endpoint details. However, if the study was a cluster‐randomised, cross‐over, or multiple‐arm study, we planned to address unit of analysis issues as detailed below.
Cluster‐randomised studies
We planned on including cluster‐randomised studies in the analyses, along with individually randomised studies. We would have made an adjustment to the sample size in these studies for each intervention based on the method described in Deeks 2017, using an estimate of the intracluster correlation coefficient (ICC) derived from the study (when available) or from a similar study, or from a study of a similar population. If we used ICCs from other studies, we planned to conduct sensitivity analyses to explore the effect of variation in ICC values. We planned to include studies with data from more than one time point, selecting data from one clinically important time point for inclusion in a meta‐analysis if appropriate.
Cross‐over studies
We planned to consider only results from the first randomisation period, that is, before the first cross‐over.
Studies with multiple treatment groups
If studies had three (or more) arms testing relevant active interventions versus an inactive control, for continuous outcomes, we planned to pool means, SDs, and number of participants for each active treatment group across treatment arms, or we planned to divide the number of participants in the control group between the treatment arms. For dichotomous outcomes, we planned to pool data from relevant active intervention arms into a single arm for comparison, or we would have divided data from the comparator arm equally between the treatment arms.
Dealing with missing data
For studies with missing data, we planned to contact the corresponding study authors to try to obtain additional information. We planned to record missing and unclear data for each included study. We also aimed to perform all analyses using an intention‐to‐treat approach, that is, we planned to analyse all participants and their outcomes within the groups to which they were allocated, regardless of whether or not they received the intervention.
Assessment of heterogeneity
We planned to evaluate heterogeneity by visually inspecting point effect estimates and confidence intervals in forest plots, and by using Tau², the Chi² test, and the I² statistic (Higgins 2003), as outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017). We planned to interpret the I² statistic as follows: 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity; 75% to 100% represents considerable heterogeneity. We acknowledge an understanding that the importance of I² depends on (i) magnitude and direction of effect and (ii) strength of evidence for heterogeneity, for example, P value from the Chi² test, or a confidence interval for I².
Assessment of reporting biases
If 10 or more studies were included in the review, we planned to produce a funnel plot to investigate publication bias through visual inspection of asymmetry (Higgins 2011). If asymmetry was evident, we planned to perform a statistical test for funnel plot asymmetry as proposed by Egger and Rücker (Sterne 2017).
Data synthesis
We planned to use RevMan 5 to conduct statistical analysis (RevMan 2014).
When reasonable to assume that studies were homogeneous, that is, examining the same intervention in similar populations using the same methods, we planned to use a fixed‐effect meta‐analysis to combine data. If there was clinical heterogeneity (due to variations in participants, interventions, or outcomes), we planned to use random‐effects meta‐analysis to produce an overall summary if it were reasonable to assume that an average treatment effect across the included studies was clinically meaningful. If we did not perform a meta‐analysis, instead we had planned to present a narrative synthesis.
Subgroup analysis and investigation of heterogeneity
If we identified substantial heterogeneity, we planned to explore this using subgroup analyses. We planned to perform the following subgroup analyses on our primary outcomes.
-
Bipolar disorder type I versus bipolar disorder type II.
-
Duration of treatment (up to 12 months and longer than 12 months).
-
Setting (community versus hospital).
Sensitivity analysis
We planned to repeat the analyses including high‐quality studies only. For the purpose of this review, we would have classified studies judged as ‘low risk of bias’ for sequence generation and allocation concealment as high‐quality studies.
'Summary of findings' table
For each of the main comparators (detailed under the 'Main comparators' heading), we would have prepared a ‘Summary of findings’ table using GRADEpro GDT 2015 software. We planned to include the three primary outcomes (i.e. changes in body mass as measured by change in BMI, patient‐reported adverse events, and quality of life). Short‐term effects, measured up to 12 months after completion of the intervention, would have been the primary time point reported in the 'Summary of findings' table. We would have graded the quality of evidence according to the GRADE approach (Guyatt 2008). We would have assigned one of four levels of quality ‐ high, moderate, low, or very low ‐ based on overall risk of bias of the included studies, directness of the evidence, inconsistency of results, precision of the estimates, and risk of publication bias.
Ensuring relevance to decisions in health care
The principle of assessing healthcare interventions using outcomes that matter to people making choices in health care underpinned our approach to defining the outcomes for this review. We consulted with a consumer group of people living with bipolar disorder on what outcomes matter to this population, and the primary and secondary outcomes of this review reflect their input. We continued to consult with this consumer group throughout the process of carrying out this systematic review.
Results
Description of studies
Results of the search
Searches of the CCMDCTR and other databases to 19 February 2019 retrieved 3541 records to screen, and we identified a further 38 duplicates within this set. A total of 3503 records were screened (on title/abstract), 3478 were judged irrelevant, and 25 were carried forward to full‐text review (including three ongoing Clinical Trials.gov protocols). None of the studies reviewed in 2019 met the inclusion criteria.
An update search (22 May 2020) retrieved a further 2684 records (after de‐duplication), and after screening titles and abstracts, we excluded a further 2652 records. We brought 32 studies forward to full‐text review (a majority of these records were ClinicalTrials.gov protocols). Thirteen studies were carried forward, six are additional ongoing studies, and six are studies that have been completed but remain unpublished. A further study was excluded once the drug company report had been retrieved. A cumulative summary of the results from both searches is displayed in the PRISMA flow diagram (Figure 1).
Study flow diagram.
Included studies
There are no included studies in this review.
Excluded studies
To date, we have not identified any completed and published studies that meet our eligibility criteria. We have formally excluded 23 studies (Alvarez‐JimÈnez 2006; Baptista 2007; Bartels 2015; Bobo 2011; Dauphinais 2011; Deberdt 2005; Deberdt 2008; Elmslie 2006; Erickson 2016; Evans 2005; Frank 2015; Gillhoff 2010; Goldberg 2013; Graham 2005; Kim 2014; Mc Elroy 2007; Milano 2007; Mostafavi 2017; NCT00044187; NCT00303602; NCT00472641; NCT00845507;Rado 2016). All excluded studies and reasons for exclusion are presented in the Characteristics of excluded studies table.
Ongoing studies
We identified a total of nine ongoing studies: three studies from the initial search in 2019 (NCT02515773;NCT02815813;NCT03158805), and six from the 2020 update (Daumit 2019;NCT03382782;NCT03541031;NCT03695289;NCT03980743;NCT04272541. Details are provided in the Characteristics of ongoing studies table.
Studies awaiting classification
The update search in May 2020 identified an earlier conference abstract (Wirshing 2009), together with five completed but unpublished trial protocols (ChiCTR‐IPR‐17013122; NCT00203450; NCT01828931; NCT02130596; NCT03743844). Details are provided in the Characteristics of studies awaiting classification table.
Risk of bias in included studies
It was not possible to assess the risk of bias due to the absence of eligible studies for inclusion.
Effects of interventions
Due to lack of published and unpublished studies eligible for inclusion, it was not possible to examine the effectiveness of interventions for the management of obesity in people with bipolar disorder.
Discussion
Summary of main results
The objective of this review was to evaluate the effectiveness of interventions for the management of obesity in people with bipolar disorder, but we found no published or unpublished randomised controlled trials (RCTs) addressing the objective of this review. However, we did find several ongoing studies that could potentially meet the inclusion criteria for this review and may be included in the updated review. For now, we cannot draw any conclusions regarding management strategies for obesity in bipolar disorder. It is evident that there is a need to undertake high‐quality RCTs to examine different approaches. The focus must be on a bipolar population only, with a clinical diagnosis of bipolar disorder and comorbid obesity. Interventions and outcomes should be clearly defined.
Overall completeness and applicability of evidence
We were unable to assess completeness and applicability of studies, as we found no studies that meet the inclusion criteria of this review.
Quality of the evidence
It was not possible to assess methodological quality or quality of evidence in the absence of studies eligible for inclusion.
Potential biases in the review process
We found no studies relevant for inclusion in this review. By establishing inclusion and exclusion criteria from the outset and performing an extensive literature search, we reduced the risk of bias.
Agreements and disagreements with other studies or reviews
Establishing agreement or disagreement with other evidence sources is not possible or beneficial in this instance because, to our knowledge, and indeed from our experience of undertaking this review, other reviews or studies in this area of research or evidence synthesis have not focused only on evaluating the effectiveness of interventions for the management of obesity in people with bipolar disorder. Most of the studies in this review were excluded because of "wrong population". Much to our disappointment, none of the studies that we reviewed focused exclusively on a bipolar population with a clinical diagnosis of obesity. Populations were mixed in terms of types of serious mental illness (SMI), and there was no differentiation between overweight and obesity in many of the studies we reviewed. Given our recommendations for research stemming from the findings of this review, we are hopeful that future updates of this review will allow us to make meaningful comparisons across studies and indeed other reviews.
Study flow diagram.
