People with a severe mental illness (SMI) are more likely to have physical health conditions such as diabetes, hypocholesteraemia and cardiovascular disease which can lead to premature mortality compared with the general population (Hayes 2015, Olfson 2015).

Such health conditions can be caused by obesity and overweight, in part due to the side effects of antipsychotic medication, which for many people are a necessity for managing their severe mental illness, as well as modifiable lifestyle choices such as poor diet and lack of exercise.

To explore whether targeting one of these factors (poor diet) could improve weight management, Teasdale and colleagues carried out a systematic review of nutrition interventions for people with severe mental illness. The review is published today in the British Journal of Psychiatry.

Methods

Evidence gathering included searching a number of electronic databases (Medline, EMBASE, Central PsychINFO and CINAHL), a manual search of reference lists of related systematic reviews and Internet searches using Google Scholar.

Inclusion criteria

Study design	Randomised controlled trial (RCT)
Participants	18+ years and a diagnosis of schizophrenia spectrum disorder, bipolar affective disorder or depression with psychotic features
Intervention	Any nutrition intervention delivered either alone or as part of a multidisciplinary intervention
Setting	Inpatient services, outpatient programmes and community volunteer services
Outcomes	Anthropomorphic parameters (primary outcomes): ·       weight ·       body mass index (BMI) ·       waist circumference Biochemical and/or nutritional parameters (secondary outcomes): ·       cholesterol ·       low- and high-density lipoprotein ·       triglycerides ·       blood glucose ·       insulin ·       dietary intake

Statistical analysis

Where data were available the authors carried out random effects meta-analyses for the primary and secondary outcomes, calculating Hedge’s g with 95% confidence intervals (CIs).

For the primary outcomes, a further set of analyses were carried out:

• Sub-group analyses for:
  • who delivered the intervention (dietician vs. other) and
  • timing of intervention delivery (at initiation of antipsychotic medication vs. subsequent to antipsychotic medication)
• Meta-regression analyses to explore the potential moderating effects of:
  • proportion of men in the study sample
  • mean age of participants
  • proportion receiving antipsychotic medication
  • duration of intervention
  • profession of individual delivering intervention

Heterogeneity and publication bias were also assessed.

Results

In total, 26 studies met the inclusion criteria for this review of which 20 reported data that could be pooled for meta-analysis.

Sample sizes ranged from 15-291 participants, whereas the age of participants ranged from 26-55 years. Most studies recruited participants from outpatient settings (k=21). Other settings were inpatient services (k=3) or a mix of outpatients and inpatients (k=2).

Interventions included manual-based lifestyle interventions, general nutrition or healthy-eating education, weight management guidance, cooking classes, meal replacements and free fruit and vegetables.

Seven studies were delivered by a dietician, five studies included a nutrition professional as part of the intervention team, whereas in 14 studies no input from a nutritional professional was reported. Comparator groups included treatment as usual or written nutritional information.

Primary outcomes

Pooled analyses indicated a beneficial effect of nutritional interventions on:

• Weight: g= -0.39 (95% CI, -0.56 to -0.21; p<0.001; I²= 55%; k=19)
• BMI: g= -0.39 (95% CI, -0.56 to -0.22; p<0.001; I²= 51%; k=17)
• Waist circumference: g= -0.27 (95% CI, -0.42 to -0.12; p<0.001; I²= 17%; k=11)

For ease of interpretation, the pooled mean difference between the intervention and control group was:

• a reduction in weight of 2.7 kg (just under half a stone)
• a reduction in BMI of 0.8 points
• a reduction of 2.3 cm for waist circumference

When adjusting for publication bias, the results remained statistically significant.

Subgroup analyses

Interventions delivered around the initiation of antipsychotic medication had larger effect sizes than those delivered after antipsychotic medication had been taken for some time, although this difference was not significant.

Interventions delivered by dieticians were significantly more effective than interventions delivered by other health professionals or mental health clinicians.

Meta-regressions

Meta-regressions indicated that the delivery of interventions by dieticians significantly predicted weight change and BMI, but not for waist circumference (although the trend was in the same direction). The proportion of men in intervention and control groups also positively predicted an inability to lose weight.

Secondary outcomes

Evidence from five studies indicated a reduction in blood glucose in the intervention group compared with the control group, g= -0.37 (95% CI, -0.69 to -0.05; p=0.02; I²= 68%).

For all other secondary outcomes there was no evidence of benefit from nutritional interventions when compared with the control interventions.

Conclusions

The authors concluded that:

We found that nutrition interventions improved anthropometric measures by reducing weight, BMI and waist circumference. Importantly, our review provides evidence about the most effective goals and delivery methods, including preventing weight gain from the initiation of antipsychotic therapy and the use of qualified health professionals such as dietitians to deliver individualised interventions. Our results indicated that nutrition interventions were most effective when delivered by a dietitian, with meta-regression analyses confirming this in multivariate models. These findings show a clear and important role for dietitians as part of the multidisciplinary mental health team.

Strengths and limitations

This is a well conducted systematic review: the authors reported in line with the PRISMA statement, carried out a comprehensive search of multiple databases supplemented by hand-searching and the quality of included studies was assessed. By carrying out sub-group analyses and meta-regressions the authors were also able to provide useful information regarding the timing of nutritional interventions as well as who should be delivering them.

There were, however, several limitations. Many different types of interventions were included in the same meta-analysis, for example, both individual and group interventions were combined, as well as interventions which solely focused on nutrition and those which also included exercise and psychological approaches. Whilst the inclusion of many different types of interventions does still answer the question – whether or not nutritional interventions are effective, it does risk muddying the waters in that it doesn’t outline the duration, intensity, mode (group or individual) and components of what makes an effective intervention. As previously mentioned the authors have done a good job in exploring some of these factors, however the relatively high levels of heterogeneity, especially in the subgroup analyses, may be due to the pooling of very diverse interventions. This warrants further examination to ensure confidence in the pooled estimate.

Another limitation lies in the lack of information about the time-points included in the analysis. It is unclear how long the interventions were delivered for and whether reductions in weight were only seen at post-intervention or whether they held at longer follow-ups.

Finally, as we know that people with severe mental illness are often difficult to engage in treatment, it would have been useful to know what participant adherence rates to the interventions were and whether there was any unequal drop-out between groups.

Summary

Overall, this was a well conducted review that found small to moderate improvements in weight, BMI and waist circumference following nutritional interventions. There was limited evidence of benefit for biochemical and nutritional outcomes such as total cholesterol and dietary intake, except for blood glucose which showed significant reductions after nutritional interventions.

Whilst there were a few limitations in the analysis, this review provides promising and encouraging evidence that something can be done to help people with severe mental illness improve their diet and subsequently tackle overweight and obesity.

As mentioned by the authors in their discussion, to ensure people with SMI do in fact receive important information and help with dietary management, it is important that dieticians are included in multidisciplinary mental health teams.

Physical health monitoring of people with SMI in primary care, especially those receiving antipsychotic medication, would also help flag any weight issues.

Links

Primary paper

Teasdale SB, Ward PB, Rosenbaum S, Samaras K, Stubbs B. (2016) Solving a weighty problem: systematic review and meta-analysis of nutrition interventions in severe mental illness. The British Journal of Psychiatry 1–9. doi: 10.1192/bjp.bp.115.177139 [Abstract]

Other references

Hayes JF, Miles J, Walters K, King M, Osborn DP. (2015) A systematic review and meta-analysis of premature mortality in bipolar affective disorder. Acta Psychiatr Scand. 2015 Jun;131(6):417-25.

Olfson M, Gerhard T, Huang C, Crystal S, Stroup TS. (2015) Premature mortality among adults with schizophrenia in the United States. JAMA Psychiatry. 2015 Dec;72(12):1172-81. [PubMed abstract]

Photo credits

• Dan Zen CC BY 2.0
• Garry Knight CC BY 2.0
• Greg O’Connell CC BY 2.0
• Julian Burgess CC BY 2.0
• Charles Hutchins CC BY 2.0