Acute and maintenance effects of non-pharmacologic interventions for antipsychotic associated weight gain and metabolic abnormalities: A meta-analytic comparison of randomized controlled trials

Abstract

Objective

To evaluate the efficacy of non-pharmacological interventions for antipsychotic-associated weight gain.

Methods

Systematic literature search and meta-analysis of randomized controlled trials comparing behavioral interventions with control groups to ameliorate antipsychotic-associated weight gain.

Results

Across 17 studies (n = 810, mean age: 38.8 years, 52.7% male, 40.8% White, 85.6% with schizophrenia-spectrum disorders), non-pharmacological interventions led to a significant reduction in weight (− 3.12 kg; CI: − 4.03, − 2.21, p < 0.0001) and body mass index (BMI) (− 0.94 kg/m²; CI: − 1.45, − 0.43, p = 0.0003) compared with control groups. Intervention benefits extended to all secondary outcomes, except for high density-lipoprotein-cholesterol and systolic blood pressure. Compared to controls, intervention patients experienced significant decreases in waist circumference (WMD = − 3.58 cm, CI: − 5.51, − 1.66, p = 0.03), percent body fat (WMD = − 2.82%, CI: − 5.35, − 0.30, p = 0.03), glucose (WMD = − 5.79 mg/dL, CI: − 9.73, − 1.86, p = 0.004), insulin (WMD = − 4.93 uIU/mL, CI: − 7.64, − 2.23, p = 0.0004), total cholesterol (WMD = − 20.98 mg/dL, CI: − 33.78, − 8.19; p = 0.001), low density-lipoprotein-cholesterol (WMD = − 22.06 mg/dL, CI: − 37.80, − 6.32, p = 0.006) and triglycerides (WMD = − 61.68 mg/dL, CI: − 92.77, − 30.59, p = 0.0001), and less weight gain > 7% (29.7% vs. 61.3%; RR = − 0.52, CI: − 0.35, − 0.78, p = 0.002; number-needed-to-treat = 4). Up to 12 months after the intervention ended (mean = 3.6 months), benefits endured regarding weight (WMD = − 3.48 kg, CI: − 6.37, − 0.58, p = 0.02), but not BMI (p = 0.40). Subgroup analyses showed superiority of non-pharmacological interventions irrespective of treatment duration, individual or group, cognitive behavioral or nutritional interventions, or prevention versus intervention trials. However, weight and BMI were significantly improved only in outpatient trials (p < 0.0001), but not in inpatient or mixed samples (p = 0.09–0.96).

Conclusion

Behavioral interventions effectively prevented and reduced antipsychotic-associated weight gain and cardiometabolic perturbations, at least in outpatients agreeing to participate in trials aimed at improving physical health. Effective treatments ranged from nutritional interventions to cognitive behavioral therapy.

Introduction

Antipsychotic efficacy for psychotic disorders (Leucht et al., 2009), bipolar disorder (Correll et al., 2010), major depressive disorders (Nelson and Papakostas, 2009) and irritability/aggression associated with autism (Correll et al., 2011a), as well as off-label use in other psychiatric conditions (Maher et al., 2011) is counterbalanced by significant weight gain and cardiometabolic risk (Allison et al., 1999, American Diabetes Association et al., 2004, Lieberman et al., 2005, Kahn et al., 2008, De Hert et al., 2012). This weight gain is problematic as it may adversely affect adherence, quality of life (Allison et al., 2003), and especially, cardiovascular morbidity and mortality (Newcomer, 2005, Correll et al., 2011b, De Hert et al., 2011).

Pharmacologic interventions to ameliorate antipsychotic weight gain have had moderate success. Out of 15 agents examined in a recent meta-analysis, only five showed significant benefit versus placebo, and three were already taken off the market due to adverse effects (fenfluramine, sibutramine, reboxetine). Metformin and topiramate reduced antipsychotic-related weight gain compared to placebo by 2.5–3 kg after 8–12 weeks of treatment (Maayan et al., 2010). Pharmacologic interventions can cause additional side effects (Maayan and Correll, 2010). Metformin carries a risk of lactic acidosis, particularly in the elderly and in those with compromised renal function, and its use can be limited by nausea, vomiting and diarrhea. Topiramate has been associated with cognitive blunting (Narula et al., 2010).

Conversely, non-pharmacological interventions do not have such side effects and have shown promise. In a meta-analysis of 10 randomized controlled (RCTs) (n = 482) Alvarez-Jimenez et al. (2008) reported that non-pharmacological interventions led to 2.56 kg less weight gain and 0.91 kg/m² less BMI increase than the control condition and that nutritional counseling was equivalent to cognitive behavioral therapy (CBT). While this meta-analysis provided support for non-pharmacological interventions, the small number of studies precluded secondary analyses regarding metabolic outcomes, and the mediating impact of treatment duration. A more recent, systematic review included the same 10 RCTs plus 6 additional, but non-randomized studies (Gabriele et al., 2009), reporting similar results.

The current study expands upon the prior publications (Alvarez-Jimenez et al., 2008, Gabriele et al., 2009) by (1) including additional RCTs , and (2) analyzing effects on insulin, glucose, high density lipoprotein (HDL), low density lipoprotein (LDL), total cholesterol, triglycerides, and systolic blood pressure. In addition we also aimed to assess maintenance effects after the behavioral interventions ended.