Effectiveness of Metformin for Weight Reduction in Children and Adolescents Treated with Mixed Dopamine and Serotonin Receptor Antagonists: A Naturalistic Cohort Study

Abstract

Objectives:

Mixed dopamine and serotonin receptor antagonists (DSRAs) are associated with significant weight gain and its complications. Our aim was to evaluate the effectiveness of metformin in reducing body mass index (BMI) and metabolic parameters in children treated with DSRAs.

Methods:

We report a naturalistic study of 49 children and adolescents (mean age 14.9 ± 3.7 years), with BMI >85 percentile for age, treated with DSRAs during 2018–2020 in a child psychiatry clinic. Clinical data, anthropometric measurements, and laboratory tests were compared between those who were (study group, n = 31) and were not (control group, n = 18) treated with metformin.

Results:

The mean study duration was 9.7 ± 5.9 months. The BMI standard deviation scores (BMI-SDS) of the study group declined significantly (from 2.08 ± 0.40 to 1.81 ± 0.54, p < 0.001), while the BMI-SDS of the control group did not change (from 2.03 ± 0.45 to 2.04 ± 0.47, p = 0.838). In the study group, the decline in the delta BMI-SDS/month was more robust among those with good than poor adherence to metformin (−0.047 ± 0.039 vs. −0.004 ± 0.017, p = 0.003). The decrease in BMI-SDS was greater for patients treated with risperidone and clothiapine than with other DSRAs. Fasting insulin and insulin resistance index (homeostasis model assessment of insulin resistance [HOMA-IR]) declined in the study group (from 25.4 ± 13.8 to 19.9 ± 10.7, p = 0.033 and from 5.4 ± 2.7 to 4.2 ± 2.1, p = 0.028, respectively).

Conclusions:

Metformin treatment was associated with significant decreases in BMI, fasting insulin, and HOMA-IR. The effect of metformin seems to be dependent on adherence and type of DSRAs.

Introduction

Over the past two decades, prescriptions of mixed dopamine and serotonin receptor antagonists (DSRAs) for children and adolescents have increased remarkably (Ellul et al. 2018). DSRAs (commonly termed antipsychotic medications) have demonstrated effectiveness for psychosis, severe impulsivity, and mood instability, but these drugs are associated with significant weight gain (Bak et al. 2014). In a review of 34 studies in youth with psychotic and bipolar disorders (Maayan and Correll 2011), weight gain ranged from 0 to 16.2 kg, depending on the particular drug used. Weight gain was reported highest in the first half-year after commencing treatment, yet continued to rise in subsequent months (Bushe et al. 2012). In young patients, weight gain may impair adherence to DSRA medication (Klein et al. 2020).

Managing risks for weight gain and the metabolic syndrome of patients treated with DSRAs is challenging. One option is the addition of a pharmacological therapy for weight reduction, such as metformin (de Silva et al. 2016). Metformin is a biguanide that is currently registered only for treating type 2 diabetes in adults and in children aged ≥10 years. Several randomized controlled trials (RCTs) have shown metformin to be effective for weight reduction in the pediatric population (Lentferink et al. 2018). A few studies specifically included children with DSRA-induced weight gain (Ellul et al. 2018). Most of them were RCT or open-label trials of short periods (12–16 weeks).

We aimed to assess the effectiveness of metformin in a naturalistic setting. The primary outcome was the change in body mass index standard deviation scores (BMI-SDS) during metformin treatment. Secondary outcomes were changes in fasting glucose, fasting insulin, glycosylated hemoglobin (HbA1C), and lipid profile.

Methods

Patients

This is a cohort study of children and adolescents treated with DSRAs at the Child Psychiatry Outpatient Clinic at Sheba Medical Center between 2018 and 2020. All the patients underwent psychiatric evaluations based on the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (American Psychiatric Association 2013) criteria. Inclusion criteria were as follows: age 8–21 years, DSRA treatment, and BMI >85 percentile for age. Exclusion criteria were as follows: other medical conditions that could affect body weight, such as hypothyroidism, diabetes, and glucocorticoid treatment, and renal impairment, due to the risk of lactic acidosis during metformin treatment.

The 49 patients who met study eligibility criteria received recommendations for physical activity and nutritional consulting from a dietitian. Those who failed to reduce weight were offered metformin treatment, at a dosage of 850 mg: once daily for patients aged 8–10 years and twice daily for patients aged 10–21 years. The 31 patients who initiated metformin treatment constituted the study group. The control group included the 18 patients who were not treated with metformin because they or their parents chose not to initiate the treatment or the treatment was not offered to them by their physician.

Adherence to metformin was assessed at the follow-up visits by asking: “Over the last 7 days, how many days were you able to take metformin exactly as prescribed?” An answer of 5 days or above was defined as good adherence (Anghel et al. 2019). The study group also received vitamin B12 supplementation.

Demographic, clinical, and laboratory data were obtained from the patients' medical charts. Height, weight, and BMI-SDS were calculated using age and sex-specific growth data (based on the growth charts by the Centers for Disease Control and Prevention). Delta BMI-SDS/month was calculated as BMI-SDS at the end of the study minus BMI-SDS at study initiation, divided by study duration (months).

The study protocol was approved by the local institutional review board and was performed in accordance with the Helsinki Declaration (IRB smc-6485-19).

Laboratory tests

Blood samples for glucose, insulin, HbA1C, lipid profile, and vitamin B12 were obtained between 07:00 and 09:00 AM after 12 hours fast, at the initiation of metformin treatment and every 3–6 months thereafter. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as: (fasting glucose [mg/dL] × fasting insulin [IU/mL])/405.

Insulin resistance is usually defined as a HOMA-IR value greater than the 75th percentile for persons without diabetes according to the World Health Organization. The cutoff values reported in the literature vary widely between countries. Values above 1.7–3.8 are considered to indicate insulin resistance (Tang et al. 2015).

Statistical analysis

Data were analyzed with the IBM SPSS software version 25 (IBM, Corp.). Continuous variables were presented as means and standard deviations, or as medians and interquartile ranges, as appropriate. Categorical variables were presented as numbers and percentages and analyzed using Fisher's exact test or the Chi-Square test. The Student's t-test was applied to compare continuous variables between patients in the study and control groups, between females and males, and between patients with good and poor adherence. The paired t-test was applied to assess changes over time in laboratory results and anthropometric parameters.

The one-way analysis of variance test was applied to compare between four types of DSRAs. Correlations were analyzed by univariate analysis. Variables significantly associated with delta BMI-SDS/month were included in a regression model to examine independent predictors of BMI reduction. The results were considered significant if the two-sided p-value was <0.05.

Results

The mean age of the 49 patients included was 14.9 ± 3.7 years. No differences were found between the 35 boys and 14 girls regarding any clinical or laboratory parameter examined (data not shown). The psychiatric diagnoses were: disruptive disorders (n = 34), autism spectrum disorder (n = 15), schizophrenia (n = 6), major depressive disorder (n = 4), obsessive compulsive disorder (n = 4), bipolar affective disorder (n = 1), anxiety disorder (n = 9), and borderline personality disorder (n = 1). The mean time lapse from the initiation of DSRA treatment to the beginning of the study was 3.8 ± 3.6 years. Thirty-two patients (65%) had a history of one to five different DSRA medications before the current treatment. Seventeen patients (35%) were treated with only one type of DSRAs; of them, only three patients (6%) initiated DSRA treatment <3 months before the study and were thus defined as DSRA naive.

The mean study duration was 9.7 ± 5.9 months. The study and control groups did not differ significantly in sex distribution, anthropometric parameters, age at first DSRA initiation, age at study initiation, the type of DSRA used during the study period, and study duration (Table 1).

Table 1.

Demographic and Clinical Characteristic of the Cohort at Baseline

	All (n = 49)	Study group (metformin users; n = 31)	Control group (metformin nonusers; n = 18)	p
Age at study initiation, years	14.9 ± 3.7	15.5 ± 3.6	13.9 ± 3.7	0.128
Sex
Male	35 (71%)	21 (68%)	14 (78%)	0.527
Female	14 (29%)	10 (32%)	4 (22%)
Study duration, months	9.7 ± 5.9	10.2 ± 6.5	8.8 ± 4.7	0.434
Current DSRA
Risperidone	19 (39%)	8 (26%)	11 (61%)	0.099
Aripiprazole	18 (37%)	13 (42%)	5 (28%)
Quetiapine	5 (10%)	5 (16%)	0
Clothiapine	6 (12%)	4 (13%)	2 (11%)
Clozapine	1 (2%)	1 (3%)	0
Age at first DSRA initiation, years	11.1 ± 4.3	10.6 ± 4.7	11.9 ± 3.4	0.324
Age at current DSRA initiation, years	13.0 ± 3.3	13.2 ± 3.6	12.6 ± 2.9	0.583
Anthropometric data
Height, cm	161.2 ± 12.1	162.7 ± 11.7	158.7 ± 12.7	0.264
Height SDS	0.12 ± 1.18	0.02 ± 1.2	0.31 ± 1.2	0.399
Weight, kg	82.4 ± 22.7	86.2 ± 23.1	75.9 ± 20.8	0.123
Weight SDS	2.02 ± 0.71	2.05 ± 0.73	1.96 ± 0.70	0.666
BMI, kg/m²	31.2 ± 5.3	32.1 ± 5.5	29.6 ± 4.5	0.110
BMI-SDS	2.06 ± 0.41	2.08 ± 0.40	2.03 ± 0.45	0.678

The data are presented as number (%) or as mean ± standard deviation. The p-values compare the study and control groups.

BMI, body mass index; DSRA, mixed dopamine and serotonin receptor antagonist; SDS, standard deviation scores.

The mean metformin dose was 19.9 ± 5.0 mg/kg/day. Nine patients (29%) reported gastrointestinal symptoms (abdominal pain, nausea, or diarrhea). Four of them stopped metformin treatment due to gastrointestinal symptoms, and five patients continued the treatment, with attenuation or resolution of the symptoms. Three patients stopped metformin treatment because they felt the treatment was ineffective.

During the study period, the BMI of the study group declined significantly (from 32.1 ± 5.5 to 30.2 ± 5.7 kg/m², p = 0.003), as did the BMI-SDS (from 2.08 ± 0.40 to 1.81 ± 0.54, p < 0.001). The respective changes in the control group were not statistically significant (from 29.6 ± 4.5 to 31.1 ± 6.4 kg/m², p = 0.11 and from 2.03 ± 0.45 to 2.04 ± 0.47, p = 0.84). Delta BMI-SDS/month was −0.033 for the study group and 0.001 for the control group (p = 0.007).

The laboratory results of the study group at the beginning and end of the study are presented in Table 2. Fasting insulin and HOMA-IR declined significantly during the study period (from 25.4 ± 13.8 to 19.9 ± 10.7, p = 0.033 and from 5.4 ± 2.7 to 4.2 ± 2.1, p = 0.028, respectively).

Table 2.

Laboratory Blood Results of the Study Group (Metformin Users)

	Normal range	First visit	Last visit	p
Glucose, mg/dL	70–100	89.5 ± 8.4	88.4 ± 9.4	0.207
HbA1C, %	4–5.7	5.3 ± 0.3	5.2 ± 0.4	0.056
Insulin, IU/mL	2–20	25.4 ± 13.8	19.9 ± 10.7	0.033
HOMA-IR	^a	5.4 ± 2.7	4.2 ± 2.1	0.028
Total cholesterol, mg/dL	75–170	169.7 ± 27.9	169.8 ± 32.8	0.987
LDL, mg/dL	50–110	107.7 ± 25.3	108.1 ± 29.7	0.913
HDL, mg/dL	40–85	41.1 ± 8.6	41.5 ± 8.6	0.497
Triglycerides, mg/dL	35–150	132.0 ± 65.4	131.7 ± 66	0.915
Vitamin B12, pg/mL	187–1060	368.0 ± 139.2	393.7 ± 194.9	0.322

Data are presented as mean ± standard deviation.

HOMA-IR: 0.5–1.5 optimal range, values above 1.7 are considered insulin resistance.

HbA1C, glycosylated hemoglobin; HDL, high density lipoprotein; HOMA-IR, homeostasis model assessment of insulin resistance; LDL, low density lipoprotein.

Twenty-one patients (68%) were defined as having good adherence and 10 (32%) as having poor adherence. The delta BMI-SDS/month showed greater reduction among patients with good than with poor adherence (−0.047 ± 0.039 vs. −0.004 ± 0.017, p = 0.003). The delta BMI-SDS/month also positively correlated with age at study initiation (R = 0.392, p = 0.029). The delta BMI-SDS/month did not correlate with the following parameters: patient sex, age at first DSRA initiation, age at the current DSRA initiation, duration of treatment with DSRAs, BMI-SDS at study initiation, the laboratory test results considered at baseline, and HOMA-IR. In the study group, the BMI-SDS at the last visit correlated positively with the BMI-SDS at study initiation (R = 0.795, p < 0.001) and negatively with the metformin dose mg/kg (R = −0.437, p = 0.014).

To further explore the predictors of change in BMI-SDS, we stratified the study group according to the DSRAs used during the study. We found statistically significant differences between the subgroups, including in the delta BMI-SDS/month, which was significantly lower among patients treated with risperidone or clothiapine (Table 3). A multiple regression model that controlled for age at initiation of treatment with DSRAs, age at entry to the study, and BMI-SDS at study initiation yielded a delta BMI-SDS/month of patients treated with risperidone of −0.035 (p = 0.046) and of patients treated with clothiapine of −0.046 (p = 0.029).

Table 3.

Patient Age and Anthropomorphic Characteristics According to the Mixed Dopamine and Serotonin Receptor Antagonist Used

	Risperidone (Risperdal; n = 8)	Aripiprazole (Ariply; n = 13)	Quetiapine (Seroquel; n = 5)	Clothiapine (Etumine; n = 4)	p
Age at first DSRA initiation, years	8.0 ± 3.4	12.2 ± 4.6	13.9 ± 3.3	8.4 ± 4.4	0.040
Age at study initiation, years	14.4 ± 4.4	15.9 ± 3.1	18.1 ± 1.1	14.0 ± 3.8	0.159
Weight-SDS at study initiation	2.58 ± 0.85	2.14 ± 0.56	1.39 ± 0.39	1.53 ± 0.56	0.010
BMI-SDS at study initiation	2.35 ± 0.39	2.15 ± 0.25	1.67 ± 0.30	1.79 ± 0.47	0.005
Delta BMI-SDS/month	−0.050 ± 0.048	−0.019 ± 0.029	−0.014 ± 0.029	−0.074 ± 0.029	0.029

BMI, body mass index; DSRA, mixed dopamine and serotonin receptor antagonist; SDS, standard deviation scores.

Discussion

We report a significant decrease in BMI-SDS, of 0.27, among children and adolescents treated with DSRAs and metformin, compared to no significant change in a control group treated with DSRAs only. The mean BMI-SDS reduction in our study group is above the threshold that has been suggested as clinically meaningful in children (Ells et al. 2018). Our findings concur with RCTs that showed decreases in BMI-SDS of 0.08–0.14, after 16–24 weeks treatment with metformin in young DSRA users (Klein et al. 2006; Anagnostou et al. 2016; Correll et al. 2020). A naturalistic study of youth with autism spectrum disorder showed a decrease in BMI-SDS of 0.07, after a mean duration of 2 years (Wink et al. 2017). We also demonstrated improvement in fasting insulin and HOMA-IR with agreement to previous study (Klein et al. 2006).

In the current study, the reduction in BMI-SDS was associated with younger age at study initiation, good adherence to metformin, and the use of certain DSRAs. Patients treated with risperidone and clothiapine had significantly greater reduction in BMI-SDS than did those treated with other DSRAs, after controlling for other factors that could have confounded this association. The association between reduction in BMI-SDS and good adherence to metformin concurs with an RCT of adolescents with insulin resistance without psychiatric disorders (Love-Osborne et al. 2008). To our knowledge, the association with type of DSRA has not been previously reported.

Our study included mostly patients with a long duration of DSRA treatment before the study and only three were DSRA naive. Previous studies showed a much larger metformin-induced reduction in weight among DSRA-naive patients (de Silva et al. 2016). Therefore, early administration of metformin after initiation of DSRAs may be most effective, as is currently recommended for adults (Hendrick et al. 2017).

The safety and tolerability of metformin were good; no serious side effects were reported, and only four patients stopped metformin treatment due to gastrointestinal symptoms. In other patients, the gastrointestinal symptoms resolved or attenuated with time, as was previously described (McDonagh et al. 2014).

Some limitations of our study should be acknowledged. Most of the cohort was treated with DSRAs for a long period before study entry, and their BMI at DRSA initiation was unavailable. The naturalistic nature precluded a strict protocol for obtaining anthropometric measurements, and blood tests were not uniformly taken during the study and the follow-up periods. We addressed this caveat by calculating the change in BMI-SDS per month. Another limitation is the small size of the cohort. The advantages of our study include its relatively long duration and the naturalistic design that assessed change in a real-life setup.

Conclusions

Following the initiation of metformin treatment, mean values of BMI, fasting insulin, and HOMA-IR decreased significantly in pediatric patients treated with DSRAs. The safety and tolerability of metformin were good. The effect of metformin seems to be dependent on adherence and the type of DSRAs.

Clinical Significance

Metformin treatment was associated with significant decreases in BMI, fasting insulin, and HOMA-IR among children and adolescents with a BMI >85 percentile for age who were treated with DSRAs. The results of our naturalistic cohort study support offering metformin by child psychiatrists to individuals meeting these characteristics who failed lifestyle intervention. Metabolic monitoring and evaluation of adherence should be maintained.

Footnotes

Disclosures

The authors have nothing to disclose.

References

American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 5th ed. Washington, DC: American Psychiatric Association; 2013.

Anagnostou

, Aman

, Handen

, Sanders

, Shui

, Hollway

, Brian

, Arnold

, Capano

, Hellings

, Butter

, Mankad

, Tumuluru

, Kettel

, Newsom

, Hadjiyannakis

, Peleg

, Odrobina

, McAuliffe-Bellin

, Zakroysky

, Marler

, Wagner

, Wong

, Macklin

, Veenstra-VanderWeele

: Metformin for treatment of overweight induced by atypical antipsychotic medication in young people with autism spectrum disorder: A randomized clinical trial. JAMA Psychiatry, 73:928–937, 2016.

Anghel

, Farcas

, Oprean

: An overview of the common methods used to measure treatment adherence. Med Pharm Rep, 92:117–122, 2019.

Bak

, Fransen

, Janssen

, van Os

, Drukker

: Almost all antipsychotics result in weight gain: a meta-analysis. PLoS One, 9:e94112, 2014.

Bushe

, Slooff

, Haddad

, Karagianis

: Weight change from 3-year observational data: Findings from the worldwide schizophrenia outpatient health outcomes database. J Clin Psychiatry, 73:e749–e755, 2012.

Correll

, Sikich

, Reeves

, Johnson

, Keeton

, Spanos

, Kapoor

, Bussell

, Miller

, Chandrasekhar

, Sheridan

, Pirmohamed

, Reinblatt

, Alderman

, Scheer

, Borner

, Bethea

, Edwards

, Hamer

, Riddle

: Metformin add-on vs. antipsychotic switch vs. continued antipsychotic treatment plus healthy lifestyle education in overweight or obese youth with severe mental illness: Results from the IMPACT trial. World Psychiatry, 19:69–80, 2020.

de Silva

, Suraweera

, Ratnatunga

, Dayabandara

, Wanniarachchi

, Hanwella

: Metformin in prevention and treatment of antipsychotic induced weight gain: A systematic review and meta-analysis. BMC Psychiatry, 16:341, 2016.

Ells

, Rees

, Brown

, Mead

, Al-Khudairy

, Azevedo

, McGeechan

, Baur

, Loveman

, Clements

, Rayco-Solon

, Farpour-Lambert

, Demaio

: Interventions for treating children and adolescents with overweight and obesity: An overview of Cochrane reviews. Int J Obes (Lond), 42:1823–1833, 2018.

Ellul

, Delorme

, Cortese

: Metformin for weight gain associated with second-generation antipsychotics in children and adolescents: A systematic review and meta-analysis. CNS Drugs, 32:1103–1112, 2018.

10.

Hendrick

, Dasher

, Gitlin

, Parsi

: Minimizing weight gain for patients taking antipsychotic medications: The potential role for early use of metformin. Ann Clin Psychiatry, 29:120–124, 2017.

11.

Klein

, Topalian

, Starr

, Welge

, Blom

, Starr

, Deetz

, Turner

, Sage

, Utecht

, Fornari

, Patino Duran

, Higdon

, Sutton

, Sorter

, Correll

, DelBello

: The importance of second-generation antipsychotic-related weight gain and adherence barriers in youth with bipolar disorders: Patient, parent, and provider perspectives. J Child Adolesc Psychopharmacol, 30:376–380, 2020.

12.

Klein

, Cottingham

, Sorter

, Barton

, Morrison

: A randomized, double-blind, placebo-controlled trial of metformin treatment of weight gain associated with initiation of atypical antipsychotic therapy in children and adolescents. Am J Psychiatry, 163:2072–2079, 2006.

13.

Lentferink

, Knibbe

CAJ

, van der Vorst

MMJ

: Efficacy of metformin treatment with respect to weight reduction in children and adults with obesity: A systematic review. Drugs, 78:1887–1901, 2018.

14.

Love-Osborne

, Sheeder

, Zeitler

: Addition of metformin to a lifestyle modification program in adolescents with insulin resistance. J Pediatr, 152:817–822, 2008.

15.

Maayan

, Correll

: Weight gain and metabolic risks associated with antipsychotic medications in children and adolescents. J Child Adolesc Psychopharmacol, 21:517–535, 2011.

16.

McDonagh

, Selph

, Ozpinar

, Foley

: Systematic review of the benefits and risks of metformin in treating obesity in children aged 18 years and younger. JAMA Pediatr, 168:178–184, 2014.

17.

Tang

, Li

, Song

, Xu

: Optimal cut-off values for the homeostasis model assessment of insulin resistance (HOMA-IR) and pre-diabetes screening: Developments in research and prospects for the future. Drug Discov Ther, 9:380–385, 2015.

18.

Wink

, Adams

, Pedapati

, Dominick

, Fox

, Buck

, Erickson

: Brief report: Metformin for antipsychotic-induced weight gain in youth with autism spectrum disorder. J Autism Dev Disord, 47:2290–2294, 2017.