Description of the condition
The World Health Organization (WHO) defines cardiovascular diseases (CVDs) as a collection of conditions affecting the heart and blood vessels in the body. More individuals die from CVDs annually than from any other disease, making it the leading cause of mortality across the globe. It is no longer just a disease of developed nations, as over 80% of CVD mortality now occurs in developing countries (WHO 2013). Approximately 17.3 million people died from CVDs in 2008; this is predicted to rise to 23.3 million by 2030, with CVDs projected to remain the foremost cause of death globally (WHO 2013).
Cardiovascular events, such as myocardial infarctions and strokes, are primarily triggered by a blood clot in an artery narrowed by atherosclerosis, which restricts the flow of blood to the heart or the brain. They are usually acute events, but the risk of such an event is attributed to long-term health behaviours. In more than 90% of cases worldwide, the risk of a first myocardial infarction is related to nine independent, potentially modifiable, risk factors: an abnormal blood lipid profile, smoking, hypertension, diabetes mellitus, abdominal obesity, diet, alcohol, physical activity and psychosocial factors such as depression (Yusuf 2004).
The effect of modifying these risk factors is clearer for some behaviours than for others. For example, there is substantial evidence that smoking cessation (Novello 1999); treatment to reduce blood pressure, treatment to reduce blood lipids and adequately control diabetes (Graham 2007); adopting a Mediterranean diet (Estruch 2013; Rees 2013); and increasing levels of physical fitness (Lee 2011) reduce cardiovascular mortality. However, findings regarding intentional weight loss are less clear. Although a body mass index (BMI) above 25 kg/m
A 15-year prospective cohort study in 1160 men suggested that weight fluctuations result in a higher risk of all-cause mortality compared to steady weight maintenance, based on an adjusted hazard rate ratio of 1.86 (95% confidence interval (CI) 1.31 to 2.66). This study adjusted for pre-existing illness, but did not adjust for intentionality of weight loss (Rzehak 2007). Two further prospective studies, which adjusted for intentionality of weight loss, did not find a significantly higher all-cause mortality rate in those with a history of severe weight cycling than in those whose weight remained stable (Field 2009; Stevens 2012). A meta-analysis of 26 prospective studies by Harrington 2009 found that intentional weight loss had a small benefit in obese individuals with comorbidities (relative risk (RR) 0.84, 95% CI 0.73 to 0.97). However, weight loss appeared to be associated with slightly higher mortality in healthy individuals (RR 1.11, 95% CI 1.00 to 1.22) and those who were overweight but not obese (RR 1.09, 95% CI 1.02 to 1.17). The analysis showed no evidence that weight loss conferred either benefit or risk in healthy obese individuals. Therefore, the evidence supporting the benefit of intentional weight loss, particularly in individuals without comorbidities, or the risk of associated weight cycling in this population is unclear.
The problem of recurrent weight cycling and the difficulty in achieving and maintaining weight loss may lie in the complex aetiology of overweight and obesity. Although the concept of energy balance is simple - when an individual’s energy intake exceeds their energy expenditure, weight is gained (Söderlund 2009) - the factors that cause intake to exceed expenditure are varied and complex. They comprise interrelated personal and environmental influences that were described in the 2007 Foresight report on tackling obesities to cluster into seven themes (Foresight 2007). The first of these themes, food consumption, encompasses personal food preferences, food habits and portion control, which may be either self-determined or influenced by current social norms. Food production is a theme that encompasses factors such as the availability of cheap, palatable, energy-dense foods. Demand for these foods fuels food production, which leads to further societal acceptance and higher demand, and so the cycle continues. Another theme is physiological factors, which result from an individual's biological and genetic composition. These may, for example, partially determine an individual's regulatory mechanisms of hunger and satiety. However, the mechanisms may be overridden by external factors such as the availability of food. Individual physical activity is another theme and includes individual barriers to engaging in physical activity; for example, learned patterns of activity, sedentary occupations and commuting needs. Physical activity environment, also a theme, relates to barriers within the environment, such as the cost of formal exercise and the perceived safety of being out alone. Another key theme is individual psychology; this encompasses stress levels, confidence and self-esteem, and the associated emotive drive to eat, which again may override physiological mechanisms. The last theme is social psychology. This is related to the thoughts and opinions that permeate society and are prevalent in printed media, television and education. For example, the societal perception of obesity and pressure related to body image has psychological consequences of stress and low self-esteem, and overeating may develop as a coping strategy to deal with these issues (Foresight 2007).
As the factors that contribute to the development of obesity are many, multicomponent interventions that tackle several of these areas have been developed for the effective prevention and management of obesity, so as to reduce the burden of diseases such as CVD and diabetes. The National Institute for Health and Clinical Excellence (NICE 2006) recommend that interventions for obese individuals comprise components that address diet and physical activity, and incorporate behavioural change techniques. It recommends that individuals should strive to reach and maintain a realistic target weight loss of 5% to 10% of their original weight through a weekly weight loss of 0.5 kg to 1 kg (NICE 2006). Although many overweight or obese people are able to lose weight in this way, however, a large proportion of these people are unable to maintain these new health behaviours or this weight loss. Several meta-analyses of randomised trials (Curioni 2005; McTigue 2003; Turk 2009) have shown that about one-third of weight lost is regained within one year and almost all of the weight lost is regained within five years (Katzer 2008). Sometimes, this weight gain continues; a review of 31 long-term studies found that between one- and two-thirds of dieters gained more weight than they had lost originally (Mann 2007).
Lack of a sustained change in weight and, hence, a lack of a reduction in long-term chronic disease risk, may be due to an emphasis on food restriction. Food restriction can lead to hunger and feelings of deprivation or preoccupation with food, which in turn trigger overconsumption. Overconsumption and weight gain may lead to feelings of low self-esteem, depression and guilt. To manage feelings of guilt associated with overconsumption, further attempts are made to restrict eating, and a cycle of dieting and bingeing, weight loss and weight gain is perpetuated, with little long-term gain in risk reduction (Cole 2010; Hawley 2008; Neumark-Sztainer 2000; Rapoport 2000).
This leads us to question what is the best approach to reducing the chronic disease risk associated with obesity? Should we continue to focus on weight loss in all who are overweight or obese, or is there a possibility that some individuals would achieve a greater improvement in their cardiovascular risk profile through greater emphasis being placed on the maintenance of a healthy lifestyle and improved psychosocial well being, irrespective of weight loss?
Description of the intervention
Interventions that have been developed to focus on the health gains of dietary change, physical activity and psychosocial well being in those who are overweight or obese, rather than on weight loss only, we have termed 'health, not weight loss, focused' (HNWL) programmes. Among these is the Health At Every Size (HAES) programme. Randomised controlled trials (RCTs) have shown that HNWL programmes may have a greater effect on reducing cardiovascular risk factors, such as improving the blood lipid profile and blood pressure, and reducing depression, compared with conventional weight loss (CWL) programmes (Bacon 2002; Bacon 2005; Carroll 2007; Gagnon-Girouard 2010; Provencher 2007; Provencher 2009).
HNWL programmes promote healthy lifestyle change and foster a positive body image, irrespective of BMI or weight loss success. Behaviour change techniques help individuals adopt more healthy food choices and engage in more physical activity, with the enjoyment of food and exercise as an important focus. Techniques include cognitive restructuring, and attentive and intuitive eating. A holistic approach is taken that seeks to address the multifactorial aspects of achieving a healthy lifestyle, including emotional, spiritual and social needs. Promoting mental well being through encouraging self-acceptance and promoting feelings of self-worth are all part of this approach.
CWL programmes include behaviour change techniques that facilitate adopting a healthy diet and increasing physical activity for the purpose of achieving a weekly weight loss target (NICE 2006). The weight loss target is a priority of the interventions, unlike HNWL programmes, which focus on the benefits of a healthy lifestyle, improved mental well being and self-acceptance, regardless of the amount of weight loss achieved. Any weight loss that does occur through these changes is a bonus rather than a target.
How the intervention might work
In taking the focus away from weight loss and food restriction, the HNWL approach engages in a more holistic method of promoting healthy behavioural change. The combination of addressing psychological and spiritual, as well as physical, well being may lead to greater contentment and more-sustained lifestyle changes, resulting in a greater reduction in cardiovascular risk factors than CWL programmes. A key element of HNWL is teaching individuals to think positively about food, and also to recognise and respond to internal cues, such as hunger and satiety (Bacon 2005). Responding to internal cues to eat is called intuitive eating (IE). This helps people to eat in response to genuine hunger and not to consume food for other reasons (e.g. in response to emotion). This may work as feelings of satisfaction through eating a healthy diet replace feelings of deprivation. In addition, recognising the body's natural desires for food and working in tune with these may improve body satisfaction, self-image and mental well being (Katzer 2008; Miller 2001; Robison 2005).
Why it is important to do this review
As described above, we are aware of several RCTs comparing HNWL with CWL programmes. While two narrative review papers have been published (Bacon 2011; King 2007), these do not use robust quantitative methodology to combine and analyse the whole body of evidence in this area. To do this, a systematic review and meta-analysis is needed. A scoping review of the literature showed that to date no systematic review or meta-analysis has been conducted in this area and, as such, no clinical recommendation for or against the effectiveness of HNWL approaches can be made. Undertaking a systematic review with meta-analysis will inform clinical practice regarding whether HNWL approaches show the potential to improve cardiovascular risk factors compared with CWL programmes.
The aim of this review is to compare the effects of HNWL programmes with those of CWL programmes on cardiovascular risk factors in adults with a BMI greater than 25 kg/m
Criteria for considering studies for this review
Types of studies
We will include RCTs comparing HNWL programmes with CWL programmes. We will also include cluster randomised trials and take these different study designs into account in the analysis.
Types of participants
We will include adults, 18 years old and over, with a BMI greater than 25 kg/m
Types of interventions
We will include trials comparing HNWL programmes with CWL programmes.
Any programme that promotes an increase in physical activity and improved healthy eating without a primary focus on weight loss, whether the intervention is branded HAES or not. The focus will instead be on improving physical and mental health through addressing a variety of factors, which may include lifestyle, emotional, social and spiritual factors.
Any diet, exercise or behavioural programme, or a combination of these, focusing on achieving weight loss of between 0.5 kg/week and 1 kg/week.
We have devised a checklist from the descriptions of interventions in studies already known to us in order to identify suitable programmes and differentiate between programme types. We have paid particular attention to the behaviour change techniques that are unique to each programme ( Table 1).
Types of outcome measures
Our outcome measures are based on seven of the nine main risk factors described previously (Yusuf 2004). Trials need to include at least one of these outcome measures. The seven risk factors chosen are those most typically measured as part of weight management interventions and exclude the assessment of changes in the prevalences of smoking and diabetes. Included studies must have a minimum follow-up period of two months after the end of treatment.
Our primary outcomes are the following physiological markers of cardiovascular risk:
- blood lipids;
- blood pressure;
- body weight.
In trials using more than one measure for any outcome, we will prefer the measure with the strictest criteria (e.g. measured over self-reported weight).
Our secondary outcomes are those risk factors that may mediate the primary outcomes, but that are also known independent cardiovascular risk factors. They may include a variety of dimensions and may be measured in a variety of ways. We will pool together measures of the same dimension and include only those studies that have measured these outcomes with validated tools:
- physical activity;
- alcohol intake;
- psychosocial well being;
- adverse events.
Search methods for identification of studies
Our search for trials will begin in 1970, the time when the concept of HNWL was first developed.
We will identify trials through systematic searches of the following bibliographic databases:
- Cochrane Central Register of Controlled Trials (CENTRAL) on The Cochrane Library;
- MEDLINE (Ovid);
- EMBASE (Ovid);
- PsycINFO (Ovid);
- CINAHL (Ebsco);
- ASSIA (Applied Social Sciences Index and Abstracts).
We will adapt the preliminary search strategy for MEDLINE (Ovid) (Appendix 1) for use with the other databases. We will apply the Cochrane sensitivity-maximising RCT filter (Lefebvre 2011) to MEDLINE (Ovid) and adaptations of it to the other databases, with the exception of CENTRAL. We will search all databases from 1970 to the present, and we will impose no restriction on language of publication.
Searching other resources
We will identify any relevant ongoing trials through searching trial registers and trial result registers, which are the best sources for trials that are either ongoing or unpublished:
- ClinicalTrials.gov (www.ClinicalTrials.gov);
- WHO International Clinical Trials Registry Platform Search Portal (http://apps.who.int/trialsearch/);
- National Institute for Health Research search portal (https://portal.nihr.ac.uk/Pages/NIHRResearchInfoStatement.aspx);
- UK Clinical Trials Gateway - Current Controlled Trials (www.controlled-trials.com/ukctr/);
- National Research Register Archive (portal.nihr.ac.uk/Pages/NRRArchive.aspx);
- University Hospital Medical Information Network Clinical Trials Registry (www.umin.ac.jp/ctr/).
We will handsearch the reference list of all included studies, background articles and the narrative review articles described above, to identify any additional relevant studies.
We will search commercial and non-profit organisation websites for additional information that may be relevant to us:
- HAES UK (http://www.healthateverysize.org.uk/);
- Association for Size Diversity and Health (https://www.sizediversityandhealth.org/);
- National Association to Advance Fat Acceptance (NAAFA) (http://www.naafaonline.com/dev2/);
- the resource list for the HAES curriculum (http://haescurriculum.files.wordpress.com/2013/07/haes-curriculum-resource-list.pdf).
We will keep and maintain a hard or electronic copy of findings and study details, with specific reference to the date when we accessed the websites.
We will contact the authors of relevant papers where trials are ongoing or data are missing.
Data collection and analysis
Selection of studies
NK and DL will check the titles and abstracts of studies generated by the search. We will obtain full-text copies of papers reporting relevant trials and have them translated if necessary. NK and DL will independently review the trials, either accepting or rejecting them in accordance with the eligibility criteria ( Table 2). We will resolve any disagreement regarding study inclusion through discussion with a third author (GF). In the case of different reports of the same study, we will use the data only once.
Data extraction and management
For each trial, NK will extract the data and DL will check them. We will use a data collection form to record the study methods, participant characteristics and outcomes. We will present this information in a 'Characteristics of included studies' table. The information collected will include the following.
- Design of the trial, for example RCT or cluster RCT
- Method of randomisation
- Method of randomisation concealment
- How participants were selected
- Details of ethical approval and when consent forms and information sheets were given to participants
- The location and setting of the study
- The date of the study
- Total duration of the study
- Time elapsed before following up with participants or any withdrawal of participants
- Details of any 'run in' period used with HNWL or weight loss programmes. For example, individuals may be required to avoid weight loss attempts in the months preceding the trial
- The number of participants randomly assigned to each intervention group
- Baseline demographic information relating to participants (age, gender, ethnicity)
- Baseline BMI of participants
- Presence of comorbidities
- Whether having any treatment or taking any medication that may influence the cardiovascular risk factors being measured
- Inclusion and exclusion criteria used in the study
- Description and duration of any form of HNWL programme used
- Aherence to the intervention
- Description and duration of any form of CWL programme used
- Supplementary interventions (e.g. low calorie diets or extreme exercise) that might result in weight loss of more than 1 kg/week
- Whether any other forms of intervention were used and further details if relevant
- Details of any dietary, physical activity, psychological or behavioural support provided to participants
- Behaviour change techniques, using the taxonomy of behaviour change (Michie 2011), used in the control and intervention groups
- Details of any HNWL philosophy discussed
- Details of who delivered the intervention
- Outcomes measured
- Tools and methods used to record outcomes and whether these are validated
- Timepoints when outcomes were measured
- Whether intention-to-treat (ITT) analysis was used/how dropout was dealt with in the analysis
- Dropout rates
- Reasons for withdrawal
- Details of any adverse events in each group
- Additional outcome results
- Amount of missing data in both groups
Assessment of risk of bias in included studies
We will assess risk of bias using the guidelines given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We will include assessments of the domains of random sequence generation, allocation concealment and incomplete outcome data. NK will assess the risk of bias in each study. DL will check this, and GF will, if necessary, assist in resolving any disputes that arise during this process. We will grade each potential source as having a high, low or unclear risk of bias, and provide a quote from the study report together with a justification for our judgement in the 'Risk of bias' table ( Table 3). We will summarise our 'Risk of bias' judgements across different studies for each of the domains listed. Where information on risk of bias relates to unpublished data or to correspondence with a trialist, we will note this in the 'Risk of bias' table. We will also note any other type of bias (e.g. where there are financial conflicts of interest).
When considering treatment effects, we will consider whether the contributing studies have a high, low or unclear risk of bias.
Measures of treatment effect
Most of our outcomes (e.g. weight, blood pressure, lipids) will be continuous data. We will analyse continuous data as mean differences with 95% confidence intervals, in order to compare the change in outcome between the intervention and control arms. We will enter data presented as a scale with a consistent direction of effect. We will narratively describe skewed data reported as medians and interquartile ranges. Some of our outcomes may use ordinal scales (e.g. measures of psychosocial well being, levels of physical activity). Where possible (i.e. with scales of five or more ordinal categories) we will treat these as continuous data. If there is variation in the scales used to measure the same outcome, we will analyse these outcomes using standardised mean differences and 95% confidence intervals.
Unit of analysis issues
In the case of cluster RCTs we will either reduce the size of the trial to its effective sample size by taking into account the intracluster correlation coefficient or we will calculate an inflated standard error, whichever is most appropriate to the trial.
Dealing with missing data
We will contact investigators or study sponsors in order to verify key study characteristics and obtain missing numerical outcome data where possible. Where this is not possible, and the missing data are thought to introduce serious bias, we will explore the impact of including such studies on the overall assessment of the results using sensitivity analyses. It may be necessary to analyse separately those studies that used ITT analyses from those that carried out completer analyses.
Assessment of heterogeneity
We will identify and analyse inconsistencies across study results through an examination of the forest plots. We will observe the overlapping confidence intervals and will use the I² statistic to measure the heterogeneity among trials in each analysis. Significant heterogeneity will be defined as a P value of less than 0.05 for the Chi
Regardless of the model used, where there is significant heterogeneity we will explore possible causes through prespecified subgroup analyses, as described below. We will pay particular attention to those studies that have a poor overlap of confidence intervals.
Assessment of reporting biases
This is a controversial area of treatment for those with obesity and some relevant studies may not have been published as a result of this controversy. We will assess the risk of selective outcome reporting across the studies using a funnel plot, if sufficient studies are available.
We will use the Mantel-Haneszel fixed-effect model for pooling results provided there is no significant heterogeneity that would prevent us doing so. We will deal with anticipated sources of heterogeneity by the appropriate use of standardised mean differences and subgroup analyses.
Subgroup analysis and investigation of heterogeneity
We will stratify our findings according to outcomes and length of follow up. We may also need to stratify our analyses to analyse trial data separately, according to whether the intervention had an element of enhanced care, which may influence the effect size (e.g. weight loss medication), or where populations have specific characteristics (e.g. eating disorders, female gender or a specific ethnicity).
We plan to carry out sensitivity analyses to review the influence of studies with a low risk of bias on our findings. We may need to analyse separately those studies that used ITT analyses from those that carried out completer analysis. We may also need to analyse separately those trials that were cluster randomised.
Appendix 1. Preliminary search strategy - MEDLINE (Ovid)
1."health at every size".tw.
3.(weight adj4 (manag* or accept* or centred* or centered*)).tw.
4.(health* adj6 (size or weigh*)).tw
8.(Wholistic or holistic) adj4 weight.tw.
9.((Intuitiv* or attentiv* or mindful*) adj4 eat*).tw.
25.body mass index/
26.exp Abdominal Fat/
30.(weight adj2 (gain* or chang* or los* or maint*)).tw.
31.(body mass adj (index or indexes or indices)).tw.
34.((reduc* or increas* or decreas* or los* or gain*) adj2 weight).tw.
41.exp food and beverages/
42.exp drinking behavior/
43.physical adj activit*.tw.
45.exp Affective symptoms/
47.exp stress, psychological/
54.Psychological adj $stress.tw.
55.Body adj (image or satisfaction or dissatisfaction).tw.
57. randomized controlled trial.pt.
58. controlled clinical trial.pt.
61. drug therapy.fs.
65. 57 or 58 or 59 or 60 or 61 or 62 or 63 or 64
66. exp animals/ not humans.sh.
67. 65 not 66
68. 10 and 56
69. 67 and 68
Contributions of authors
NK, supervised by DL, drafted the protocol. ACF, APT, NL-H and GF reviewed and revised the draft protocol.
Declarations of interest
Amanda Farley is employed by the University of Birmingham as a Lecturer in epidemiology and carries out research in the area of smoking cessation, weight control and lung cancer. She had also done consultancy work for the Department of Health and has been paid as a guest speaker at various training events.
Nicola Lindson-Hawley is funded by a research grant awarded by the NIHR HTA programme (09/110/01), investigating the use of pre-quit nicotine patches for smoking cessation. The excess treatment provided for this research are nicotine patches supplied by GlaxoSmithKline.
Deborah Lycett, Andy Turner, Nazanin Khasteganan, and Gill Furze have no known conflict of interest.
Sources of support
- Dr Deborah Lycett, UK.Deborah Lycett's Salary is paid by Coventry University
- Nazanin Khasteganan, UK.Nazanin Khasteganan is funded by a PhD scholarship from Coventry University
- Dr Andrew Turner, UK.Andrew Turner's salary is paid by Coventry University
- Professor Gill Furze, UK.Gill Furze's salary is paid for by Coventry University
- Amanda Farley, UK.Amanda Farley's salary is paid for by the University of Birmingham
- Dr Nicola Lindson-Hawley, UK.Nicola Lindson-Hawley's salary is funded by a grant from the National Institute of Health Research (NIHR).