Metformin for olanzapine-induced weight gain: a systematic review and meta-analysis

Abstract

Olanzapine is an atypical antipsychotic that is useful in schizophrenia and bipolar affective disorder, but its use is associated with troublesome weight gain and metabolic syndrome. A variety of pharmacological agents has been studied in the efforts to reverse weight gain induced by olanzapine, but current evidence is insufficient to support any particular pharmacological approach. We conducted a systematic review and meta-analysis of randomized controlled trials of metformin for the treatment of olanzapine-induced weight gain. Systematic review of the literature revealed 12 studies that had assessed metformin for antipsychotic-induced weight gain. Of these, four studies (n= 105) met the review inclusion criteria and were included in the final analysis. Meta-analysis was performed to see the effect size of the treatment on body weight, waist circumference and body-mass index (BMI). Weighted mean difference (WMD) for body weight was 5.02 (95% CI 3.93, 6.10) kg lower with metformin as compared with placebo at 12 weeks. For waist circumference, the test for heterogeneity was significant (P= 0.00002, I²= 85.1%). Therefore, a random effects model was used to calculate WMD, which was 1.42 (95% CI 0.29, 3.13) cm lower with metformin as compared with placebo at 12 weeks. For BMI, WMD was 1.82 (95% CI 1.44, 2.19) kg m⁻² lower with metformin as compared with placebo at 12 weeks. Existing data suggest that short term modest weight loss is possible with metformin in patients with olanzapine-induced weight gain.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Olanzapine in an atypical antipsychotic agent which is associated with significant weight gain.

• Metformin, an anti-hyperglycaemic agent, has been used to treat or prevent weight gain associated with olanzapine.

• Meta-analyses on studies that have examined the use of metformin for treatment of antipsychotic-induced weight gain report significant heterogeneity.

WHAT THIS STUDY ADDS

• Systematic review and meta-analysis showed that metformin is useful for the short-term treatment of olanzapine-induced weight gain.

Introduction

Body weight gain and metabolic alterations are clinically relevant side effects of atypical antipsychotics which are evident after approximately 10 weeks of treatment [1]. Olanzapine is linked to clinically significant body weight gain ranging from 0.9 kg month⁻¹ up to 6 to 10 kg or more after 1 year of treatment [2]. The Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study found that 30% of olanzapine-treated schizophrenia patients gained >7% of their baseline body weight [3]. Weight gain with olanzapine has been found to be significantly higher than with other atypical antipsychotics, except for clozapine [4]. It is hypothesized that appetite stimulation and insulin resistance are the underlying mechanisms for olanzapine-induced weight gain [5, 6]. A possible cause of appetite stimulation has been suggested to involve serotonin 5-HT_2C and histamine H₁-receptor antagonism, resulting in food craving and binge eating [7].

Therefore, effective pharmacological and nonpharmacological strategies are urgently required for optimal body weight control during olanzapine treatment [8]. Various medications such as amantadine, nizatidine, ranitidine, famotidine, topiramate, fenfluramine, reboxetine, fluoxetine, fluvoxamine, sibutramine, dextroamphetamine, d-fenfluramine, orlistat, phenylpropanolamine, rosiglitazone and metformin have been reported to counteract effectively antipsychotic-induced body weight gain [9–11]. The antidiabetic agent metformin, a biguanide, is particularly attractive because of its dual mechanism of decreasing body weight gain and improving insulin sensitivity, both of which are affected by olanzapine. Metformin has been demonstrated to improve glycaemic control and promotes a moderate weight loss in both diabetic and non-diabetic subjects [12, 13]. Studies that have assessed metformin for olanzapine-induced weight gain report weight loss, though the effect varies across studies. In the systematic review by Bushe et al. [14], studies that have evaluated the effect of metformin on antipsychotic-induced weight gain were included. They were not able to combine studies for meta-analysis due to the high level of heterogeneity. Further, they noted that the effect of dietary and lifestyle interventions could not be teased out. In another recent systematic review and meta-analysis [15], metformin treatment caused a significant body weight reduction in adult non-diabetic patients treated with atypical antipsychotics (4.8%, 95% CI 1.6, 8.0) and in children (4.1%, 95% CI 2.2, 6.0) compared with placebo. Ehret et al. [16] in their meta-analysis on six studies (n= 336) reported that metformin significantly reduced weight (WMD 3.16 kg, P= 0.0002), BMI (WMD 1.21 kg m⁻²; P= 0.0001), waist circumference (WMD 1.99 cm; P= 0.005), and HOMA-IR (WMD 1.71; P= 0.004) as compared with placebo in those receiving atypical antipsychotics. Maayan et al. [11] in their meta-analysis comparing 15 agents for antipsychotic-induced weight gain, reported metformin producing greatest weight loss (seven studies, n= 334, −2.94 kg, 95% CI −4.89, −0.99), that remained significant when metformin was initiated after occurrence of weight gain, but not when started concomitantly with antipsychotics. As the weight gain pattern is different for olanzapine as compared with other atypical antipsychotics and it is one of the most commonly used antipsychotics, we conducted a systematic review and meta-analysis with an objective to determine the effects of metformin for reducing or preventing weight gain associated with olanzapine.

Methods

Data sources and search strategy

Studies were identified using online searches of PUBMED/MEDLINE and Cochrane database (CENTRAL). Searches were conducted using the combination of terms ‘atypical antipsychotics’, ‘olanzapine’, ‘body weight gain’, ‘obesity’ and ‘metformin’. We inspected the reference list of all identified studies, including existing reviews for relevant citations. The search was restricted to publications in the English language.

Study selection: inclusion criteria

One reviewer (SKP) initially evaluated the abstracts from the literature search. The following criteria were used to identify the studies:

1. Randomized, double-blind clinical trials comparing metformin with placebo for olanzapine-induced weight gain,

2. Outcome measures to include body weight, waist circumference and BMI,

3. Study duration of at least 12 weeks.

Data extraction

Two reviewers (SKP and NG) decided, independently, whether individual studies met the inclusion criteria. Disagreements were resolved by discussion with a third reviewer (AKJ). We used a standardized form, and extracted data which included patient and study characteristics, outcome measures and study results.

Assessment of methodological quality of studies

The methodological quality of included trials in this review was done using the Jadad Scale [17]. It includes three items: (i) Was the study described as randomized? (ii) Was the study described as double-blind? (iii) Was there a description of withdrawals and drop outs? Scoring was done as follows: one point for a positive answer and one point deducted if either the randomization or the blinding/masking procedures were inadequate. Cut-off of two points on the Jadad scale was considered.

Quantitative data synthesis

Meta-analyses were undertaken to estimate overall treatment effects where the trials were considered to be similar enough to combine using Rev Man 4.2 version. This decision was based on assessing similarity of trial characteristics as well as results. Separate meta-analyses were undertaken for each outcome (body weight, waist circumference and BMI). Treatment effects were expressed as weighted mean differences as outcomes were continuous with 95% confidence intervals (95% CI). Homogeneity among studies was tested using Cochran's Q-test and I² statistic, in which greater than 50% indicates a moderate amount of heterogeneity [18]. If significant statistical heterogeneity was detected (Cochran Q-test P < 0.1 or I² value > 50%), random effects estimates were calculated using DerSimonian & Laird methods [19]. Otherwise, a fixed-effect model was used for analysis.

Results

Studies included

The combined search strategies identified 12 papers on use of metformin in olanzapine-induced weight gain. Three studies were excluded because of the following reasons: Baptista et al. [20] used a metformin and sibutramine combination, whereas Klein et al. [21] and Shin et al. [22] had reported combined results for atypical antipsychotics. Four other papers were commentaries on other studies, whereas Morrison et al. [23] conducted an open-label study. Finally, four studies met the review inclusion criteria (total 105 subjects) and were included in the final analysis [24–27]. Characteristics of included studies are summarized in Table 1. One of the included studies [25] had results from four groups (metformin only, metformin and lifestyle intervention, lifestyle intervention only or placebo), of which only comparison between metformin and placebo was included and the other two groups were excluded in our meta-analysis. In the study of Baptista et al. [26], 14 week values were included in the meta-analysis.

Table 1.

Characteristics of included studies

Study	Methods	Paticipants	Intervention	Outcome
Baptista et al. [26]	Allocation: randomized Blinding: double Duration: 14 weeks	Diagnosis: schizophrenia and schizoaffective n= 40	1.Olanzapine 10 mg plus metformin 850 to 1750 mg daily, n= 19 2.Olanzapine 10 mg plus placebo, n= 18	Body weight, WC, BMI, BPRS
Baptista et al. [27]	Allocation: randomized Blinding: double Duration: 12 weeks	Diagnosis: schizophrenia and bipolar disorder n= 80	1.Olanzapine 5–20 mg plus metformin 850 to 2550 mg daily. n= 36 2.Olanzapine plus placebo, n= 36	Body weight, WC, BMI, glucose, insulin, HOMA-R, leptin, cortisol, growth hormone, lipid profile
Wu et al. [24]	Allocation: randomized Blinding: double Duration: 12 weeks	Diagnosis: schizophrenia n= 40	1.Olanzapine 15 mg plus metformin 750 mg daily, n= 20 2.Olanzapine 15 mg plus placebo, n= 20	Body weight, WC, BMI, waist-to-hip ratio
Wu et al. [25]	Allocation: randomized Blinding: double Duration: 12 weeks	Diagnosis: schizophrenia n= 64	1.Olanzapine 15 mg plus metformin 750 mg daily, n= 32 2.Olanzapine 15 mg plus placebo, n= 32	Body weight, WC, BMI, glucose, insulin,

WC, Waist circumference; BMI, Body-mass index; HOMA-R, Homeostatic model assessment for insulin resistance.

Study quality

All of the four studies were described as randomized. Two studies [24, 25] were randomly assigned in blocks of four/eight to ensure approximately equal numbers of participants within the two treatment groups. Concealment of allocation was adequately reported in two studies [24, 25]. All the studies were double-blind. The dropout rate ranged from 7.5 to 10% across these studies.

Meta-analysis

Forest plots for meta-analyses for body weight gain, waist circumference and BMI are presented in Figures 1–3. For body weight gain, the test for heterogeneity was not significant (P= 0.45, I²= 0%). Weighted mean difference for body weight was 5.02 (95% CI 3.93, 6.10) kg lower with metformin as compared with placebo at 12 weeks. For waist circumference, the test for heterogeneity was significant (P= 0.00002, I²= 85.1%). Therefore, a random effects model was used to calculate WMD, which was 1.42 (95% CI 0.29, 3.13) cm lower with metformin as compared with placebo at 12 weeks. For BMI, the test for heterogeneity was not significant (P= 0.52, I²= 0%) and WMD was 1.82 (95% CI 1.44, 2.19) kg m⁻² lower with metformin as compared with placebo at 12 weeks.

Figure 1.

Forest Plot showing body weight (kg) at 12 weeks in randomized controlled trials comparing metformin and placebo for olanzapine-induced weight gain

Figure 3.

Forest plot showing BMI (kg m⁻²) at 12 weeks in randomized controlled trials comparing metformin and placebo for olanzapine-induced weight gain

Figure 2.

Forest plot showing waist circumference (cm) at 12 weeks in randomized controlled trials comparing metformin and placebo for olanzapine-induced weight gain

Discussion

Existing data suggest that short term modest weight loss is possible with metformin in patients with olanzapine-induced weight gain. In our meta-analysis, weight reduction was 5.02% with metformin, which was higher that for the approved drugs for weight reduction such as orlistat and sibutramine, which reduced body weight by only 2.9% and 4.3%, respectively [28]. The adverse events reported with metformin were similar to placebo groups. These results encourage additional studies in more homogenous populations on the potential use of metformin in assisting olanzapine-treated patients in the long-term control of body weight and BMI. Baseline screening and a monitoring plan must be initiated on commencement of antipsychotic treatment [29]. With significant weight gain or emerging metabolic effects, the risks and benefits of antipsychotic choice and concomitant medications should be re-evaluated. Although metformin may not be routinely indicated for all patients on olanzapine as prophylaxis based on the current evidence, its use is justified in patients having olanzapine-induced weight gain with no contraindication to metformin.

Our study is limited by the number of studies included for meta-analysis. The small number of studies did not allow us to conduct tests for publication bias. Also, sensitivity analysis was not performed in our study. Nevertheless, heterogeneity was not seen for body weight gain and BMI in the studies. For waist circumference, a random effects model was used to curb the problem of heterogeneity. Although, statistically homogenous, there were differences in the study population in four studies. Weight loss in the Wu et al. studies [24, 25] was greater than that in the Baptista et al. studies [26, 27], which may be because of younger participants in the former who responded better than the older persons [30]. Furthermore, two of the studies [24, 26] have evaluated prevention of olanzapine-induced weight gain, whereas the other two [25, 27] have studied reversal of weight gain. Although these studies were combined in our review, it has been suggested that metformin appears to benefit more when started early in the course of treatment [30]. Nevertheless, the mechanism of action of metformin does not seem to vary in both these situations, which justifies the combination of these studies in our meta-analysis. We suggest that all future studies should respect standards of measuring outcomes and of reporting data in order to enhance the comparability of study results. Also, binary outcomes (number of patients losing >7% initial body weight as in CATIE study [3]) should also be reported as they are easier to interpret and clinically relevant.

Competing interests

There are no competing interests to declare.

Sources of funding

None.