Metformin for Prevention of Weight Gain and Insulin Resistance With Olanzapine: A Double-Blind Placebo-Controlled Trial

Transcription

1 Brief Communication Metformin for Prevention of Weight Gain and Insulin Resistance With Olanzapine: A Double-Blind Placebo-Controlled Trial Trino Baptista, MD, PhD 1, Jessan Martínez, MD 2, Anny Lacruz, MSci 3, Nairy Rangel, MD 4, Serge Beaulieu, MD, PhD 5, Ana Serrano, MD 2, Yinet Arapé, MD 2, Maritza Martinez, MD 2, Soaira de Mendoza, MD 6, Luis Teneud, MD, PhD 1, Luis Hernández, MD 1 Objective: To assess whether metformin prevents body weight gain (BWG) and metabolic dysfunction in patients with schizophrenia who are treated with olanzapine. Method: Forty patients taking olanzapine (10 mg daily) were randomly allocated to a metformin (n = 20; 850 to 1700 mg daily) or placebo (n = 20) group in a 14-week double-blind study. Waist circumference (WC), BWG, body mass index (BMI) fasting glucose, insulin, and lipids were evaluated at baseline and at Weeks 7 and 14 of treatment. Results: At Week 14, BWG (kg) was similar in the metformin group (5.5 kg) and the placebo group (6.3 kg), P = 0.4. There were no differences between the changes in BMI, WC, glucose, insulin, insulin resistance index (HOMA-IR), and plasma lipid levels observed in the treatment group and the placebo group; however, glucose levels decreased significantly after metformin administration (P = 0.02). The HOMA-IR decreased significantly in both groups, but 3 subjects from the placebo group developed fasting glucose levels greater than 5 mmol/l. After taking metformin, triglyceride levels increased, but the cholesterol profile improved significantly. Conclusions: Metformin did not prevent olanzapine-induced BWG. While some lipid parameters worsened during placebo, the HOMA-IR improved in both the placebo and the metformin groups. Carbohydrate metabolism impairment was not systematically observed during short-term olanzapine administration. (Can J Psychiatry 2006;51: ) Information on funding and support and author affiliations appears at the end of the article. Clinical Implications As previously reported, olanzapine significantly increases BMI. Impairment in carbohydrate metabolism, however, was only observed in a few patients. Metformin may prevent metabolic dysfunction but not weight gain during olanzapine administration. Limitations The sample studied may not represent most patients with schizophrenia who are taking atypical antipsychotic drugs. Metformin administration must be studied during prolonged administration of higher olanzapine dosages and with larger patient samples. Key Words: atypical antipsychotic drugs, diabetes mellitus, hyperlipidemia, pharmacologic intervention 192

2 Metformin for Prevention of Weight Gain and Insulin Resistance With Olanzapine: A Double-Blind Placebo-Controlled Trial Excessive BWG, hyperglycemia, type 2 diabetes mellitus, and hyperlipidemia are clinically relevant side effects of APD treatment, mainly with olanzapine and clozapine, and to a lesser extent, with quetiapine and risperidone (1). It is hypothesized that appetite stimulation and insulin resistance are underlying mechanisms (1). Amantadine, nizatidine, ranitidine, famotidine, topiramate, fenfluramine, reboxetine, sibutramine, and metformin are among the drugs reported to effectively counteract APDinduced BWG (1,2). The antidiabetic agent metformin is particularly attractive since it decreases BWG and improves insulin sensitivity (3). Two studies evaluated metformin s effects on APD-induced BWG. In the first, a crossover, placebo-controlled study, metformin was ineffective in reversing BWG induced by conventional APDs in 5 adults with schizophrenia (4). In the second, an open label, before-after study, metformin significantly decreased weight in 19 children treated with diverse atypical APDs (5). Both trials intended to ameliorate preexisting drug-induced BWG. The present study is aimed at preventing BWG during olanzapine administration in patients with severe schizophrenia or related disorders who have been stabilized for more than 5 years with conventional APDs. Method The study was approved by the human ethics committees of Los Andes University. Written informed consent was obtained from the head of the institution and from the patients. Sample size calculation was based on an expected difference of 40% in BWG between the 2 groups, a power of 80%, and a Type I error of 5%. During gradual switching from conventional to atypical APDs, 40 clinically stable inpatients with severe schizophrenia or schizoaffective disorders, free of any other chronic disease and hormone replacement therapy, were selected. Each Abbreviations used in this article APD antipsychotic drug BMI body mass index BPRS Brief Psychiatric Rating Scale BWG body weight gain HDL high density lipoprotein (cholesterol) HOMA-IR Insulin Resistance Index IM intramuscular LDL low density lipoprotein (cholesterol) NS nonsignificant SD standard deviation VLDL very low density lipoprotein (cholesterol) WC waist circumference patient was treated with APDs for an average of 30.7 years, SD 10.1 years. Before the computer-based random allocation of patients to either the olanzapine (10 mg daily at bedtime) plus metformin group (850 to 1750 mg daily, n = 20) or the olanzapine plus placebo group (same number of pills as metformin, n = 20) the patients were required to have normal physical and laboratory examination results (liver, renal, hematology, and thyroid function). Previous treatment was a standard schedule of fluphenazine decanoate or haloperidol depot (25 and 100 mg/im monthly), respectively, plus chlorpromazine of levomepromazine (100 mg orally at bedtime). Depot APDs were maintained, but olanzapine replaced the oral medication. A balanced diet, 2500 to 3000 Kcal daily, was provided. Tobacco use, snacks consumption, and degree of physical exercise were not controlled. Appetite could not be quantified owing to the severity of the mental disorder. The following outcome variables were assessed at baseline and at Weeks 7 and 14 in carefully monitored fasting conditions: body weight, BMI (weight / height 2 ), WC, glucose, and BPRS. Lipids, postload glucose levels (75 grams orally), and HOMA-IR (glucose insulin / 22.5) were also calculated at baseline and at Week 14. Two-tailed t tests for unpaired and paired samples were conducted for between- and within-group comparisons, respectively. Bivariate correlations were calculated between the BPRS scores, anthropometric variables, and chemical variables. When P < 0.05, values were considered significant. Results Two patients taking the placebo and one taking metformin dropped out of the study owing to change in residence. Twenty-two men (mean age 47.9 years, SD 10.6) and 15 women (mean age 47.4 years, SD 5.9) completed the study. The metformin group comprised 9 women and 10 men. The placebo group comprised 6 women and 12 men. Neither group significantly differed from the other on any variable at baseline. The main side effect of metformin was mild gastrointestinal discomfort. The highest dosage was well tolerated by all the patients. No outcome variable differed significantly from one group to the other. Most patients were moderately symptomatic at baseline. Olanzapine was associated with significant clinical improvement, as is reflected by the BPRS score (withingroup comparisons for Week 14: metformin t 18 = 5.1, P < 0.001; placebo t 17 = 2.8, P < 0.02) (Table 1). Body weight increased similarly in the metformin and the placebo groups respectively. At Week 7, mean BWG increase was 2.2 kg, SD 2.4 kg, compared with mean 2.6 kg, SD 2.1 kg; unpaired t test, NS. At Week 14, mean BWG increase was 5.5 kg, SD 3.3 kg, compared with mean 6.3 kg, SD 2.3 kg, NS. The within-group analysis (paired t test) was statistically Can J Psychiatry, Vol 51, No 3, March

3 The Canadian Journal of Psychiatry Brief Communication Table 1 Effects of metformin (n = 19) or placebo (n = 18) on anthropometric variables and BPRS during olanzapine administration Baseline Week 7 Week 14 BPRS Metformin 18.3 (8.2) 12.8 (6.9)* 11.7 (6.2)** Placebo 17.4 (6.9) 14.7 (7.5) 13.5 (8.2)** Body weight (kg) Metformin 58.3 (10.3) 60.5 (10.2)** 63.8 (10.2)** Placebo 59.4 (8.5) 61.9 (8.5)** 65.6 (8.5)** BMI Metformin 23.1 (2.8) 23.9 (2.8)** 25.3 (2.9)** Placebo 23.0 (3.3) 24.0 (3.6)** 25.5 (3.9)** Waist circumference Metformin 86.0 (8.3) 90.6 (10.2)** 91.2 (9.1)** Placebo 84.1 (8.3) 88.9 (8.1)** 87.8 (7.6)** No significant between-group differences were observed with any variable (P > 0.05). For within-group comparison details, see text. *P < 0.05; **P < 0.01 compared with baseline significant for Weeks 7 and 14 respectively. For the metformin group, t 18 = 5.9, P < 0.001, and t 18 = 7.3, P < For the placebo group, t 17 = 6.2, P = 0.001, and t 17 = 6.3, P = The BMI and WC also increased significantly in both groups (within-group comparisons), but there were no between-group differences (data not shown) (Table 1). There was a positive correlation between changes in weight and changes in WC. For the metformin group, r 19 = 0.58, P = For the placebo group, r 18 = 0.53, P = There was an inverse correlation between weight gain and changes in BPRS after metformin administration (r = 0.53, P = 0.01) but not after placebo administration (P = 0.3). Nurses and patients reported strong increase in appetite, but as previously mentioned, it was not quantified. Basal glucose levels decreased under metformin (t 18 = 2.5, P < 0.02) and remained stable while taking the placebo. No significant changes were observed in the postload glucose levels. Insulin decreased significantly in both groups, which is reflected in a significant reduction in the HOMA-IR. For the metformin group, t 18 = 4.1, P < 0.001; for the placebo group, t 17 = 3.7, P < 0.01 (Table 2). At baseline, 3 patients taking metformin, but none taking the placebo, had fasting glucose levels greater than 5.5 mmol/l. At Week 14, 3 patients taking the placebo and no patients taking metformin had fasting glucose levels greater than 5.5 mmol/l. At Week 14, one placebo patient had fasting glucose levels of 7.94 mml/l, which reversed after olanzapine discontinuation. At Week 14, triglycerides increased after taking metformin (t 18 = 5.9, P < 0.001). They did not increase after taking the placebo (P = 0.1). Total cholesterol decreased after metformin administration: (t 18 = 4.2, P < 0.001) but not after taking the placebo (P = 0.2). LDL cholesterol decreased after metformin administation, but not significantly (P = 0.1), whereas it significantly increased after taking the placebo (t 17 = 2.2, P < 0.05). The HDL cholesterol levels increased in both groups: for the metformin group, t 18 = 2.8, P < 0.02; for the placebo group, t 17 = 9.1, P = VLDL levels did not significantly change in either group (Table 2). Discussion With a relatively low dosage of olanzapine (10 mg daily for 14 weeks), mean BWG in all patients was 5.9, SD 2.8 kg. This figure is congruent with Allison and others metaanalysis of a broad olanzapine dosage range showing a BWG of 4.1 kg in 10 weeks (6). Metformin did not prevent increases in BWG or WC. Triglycerides also significantly increased. Total cholesterol levels decreased and HDL levels significantly increased under metformin. HDL levels significantly decreased under the placebo as well. The HOMA-IR index decreased in both groups. Three patients with high fasting glucose improved under metformin, while glucose levels deteriorated in 3 placebo patients. Thus, with the exception of the triglyceride levels, metformin displayed a positive metabolic effect that appears dissociated from its ability to prevent olanzapine-induced BWG. Metformin was well tolerated and did not interfere with clinical improvement. 194

4 Metformin for Prevention of Weight Gain and Insulin Resistance With Olanzapine: A Double-Blind Placebo-Controlled Trial Table 2 Effects of metformin (n = 19) or placebo (n = 18) on glucose, insulin, HOMA-IR, and lipids during olanzapine administration Baseline Week 7 Week 14 Glucose, basal, mmol/l Metformin 4.8 (0.5) 4.7 (0.6) 4.4 (0.6)* Placebo 4.7 (0.3) 4.6 (0.5) 4.6 (1.0) Glucose, postload, mmol/l Metformin 7.7 (2.9) 6.3 (2.9) Placebo 7.3 (3.0) 5.8 (2.1)*** Insulin, basal, miu/ml Metformin 20.7 (5.3) 15.3 (8.2)** Placebo 23.5 (8.2) 14.4 (5.1)** Homa-IR Metformin 4.5 (1.3) 2.9 (1.3)** Placebo 4.9 (1.7) 3.1 (1.7)** Triglycerides (mg/dl) Metformin 112 (56.6) 123 (64.7)** Placebo 142 (87.7) 153 (86.8) Total cholesterol (mg/dl) Metformin 199 (33.2) 181 (38.6)** Placebo 194 (38.7) 205 (66.2) LDL cholesterol (mg/dl) Metformin 130 (33.4) 102 (40.3) Placebo 116 (30.7) 128 (65.7)* HDL cholesterol Metformin 50.1 (9.8) 54.4 (9.4)* Placebo 45.3 (10.3) 48.3 (10.1)** VLDL cholesterol (mg/dl) Metformin 21.9 (9.8) 24.6 (12.9) Placebo 27.8 (17.4) 30.5 (17.4) No significant between-group differences were observed with any variable (P > 0.05). For within-group comparison details, see text. *P < 0.05; **P < 0.01; ***P < 0.07 compared with baseline Can J Psychiatry, Vol 51, No 3, March

5 The Canadian Journal of Psychiatry Brief Communication In patients with type 2 diabetes, metformin improves insulin resistance and lipid profile and either has a neutral effect or decreases body weight (3). We thus hypothesize that olanzapine-induced BWG is mainly associated with direct appetite stimulation (1), a mechanism that is not primarily influenced by metformin (3). Appetite was, unfortunately, not measured. This hypothesis requires further evaluation. Since BWG during olanzapine treatment reaches a plateau after 39 weeks (7), it remains unclear whether metformin could exert a positive effect on BWG beyond 14 weeks with olanzapine dosages above 10 mg daily and a larger sample size. The apparent improvement in insulin resistance under olanzapine cannot be explained here. It may be related to stress decrease associated with the good clinical outcome. This finding appears to contradict the accepted assumption that olanzapine, at least on a short-term basis, induces global deleterious effects on glucose metabolism. Metabolic dysfunction may be concentrated in some patients whose predictive profile has yet to be identified (1,7). No double-blind, placebo-controlled study about the effects of metformin on olanzapine-induced BWG has been published. Our study comprised patients with severe mental disorders, chronically treated with typical APDs, who started atypical agents at a relatively advanced age. It remains unanswered whether metformin could display positive effects on BWG in a different clinical population. Funding and Support This study was supported by Grant M from CDCH-T-ULA (Consejo de Desarrollo Científico, Humanístico y Tecnológico, Universidad de Los Andes, Mérida, Venezuela); by Fundación Polar (Caracas, Venezuela); and by a NARSDA Young Investigator Award to Serge Beaulieu, Quebec, Canada. References 1. Baptista T, de Mendoza S, Beaulieu S, Bermúdez A, Martinez M. The metabolic syndrome during atypical antipsychotic drug treatment: mechanisms and management. Metab Syndrom Relat Dis 2004;2: Henderson DC, Copeland PM, Daley TB, Borba CP, Cather C, Nguyen DD, and others. A double-blind placebo-controlled trial of sibutramine for olanzapine-associated weight gain. Am J Psychiatry 2005;162: Seufert J, Lubben G, Dietrich K, Bates PC. A comparison of the effects of thiazolidinediones and metformin on metabolic control in patients with type 2 diabetes mellitus. Clin Ther 2004;26: Baptista T, Alvarez L, Cubillán E, Mendoza S, Hernandez L. Metformin in the obesity associated to antipsychotic drug-administration: a pilot study. J Clin Psychiatry 2001;62: Morrison JA, Cottingham EM, Barton BA. Metformin for weight loss in pediatric patients taking psychotropic drugs. Am J Psychiatry 2002;159: Allison DB, Mentore JL, Heo M, Chandler LP, Cappelleri JC, Infante MC, and others. Antipsychotic-induced weight gain: a comprehensive research synthesis. Am J Psychiatry 1999;156: Kinon BJ, Kaiser CJ, Ahmed S, Rotelli MD, Kollack-Walker S. Association between early and rapid weight gain and change in weight over one year of olanzapine therapy in patients with schizophrenia and related disorders. J Clin Psychopharmacol 2005;25: Manuscript received May 2005, revised, and accepted August Professor, Department of Physiology, Los Andes University Medical School, Mérida, Venezuela. 2 Psychiatry Resident, Los Andes University Medical School, Mérida, Venezuela. 3 Professor, Department of Biochemistry, Los Andes University Medical School, Táchira, Venezuela. 4 Attending Psychiatrist, Psychiatric Rehabilitation Center Dr Raul Castillo, Peribeca, Táchira, Venezuela. 5 Professor, McGill University School of Medicine, Montreal, Quebec. 6 Professor, Metabolism Center, Mérida, Venezuela. Address for correspondence: Dr T Baptista, PO Box 93, Mérida, 5101-A, Venezuela, trinbap@yahoo.com Résumé : La metformine pour la prévention de la prise de poids et l insulinorésistance avec olanzapine : un essai à double insu contrôlé contre placebo Objectif : Évaluer si la metformine prévient la prise de poids corporel (PPC) et la dysfonction métabolique chez les patients souffrant de schizophrénie qui sont traités à l olanzapine. Méthode : Quarante patients prenant de l olanzapine (10 mg par jour) ont été affectés au hasard à un groupe de metformine (n = 20; 850 à mg par jour) ou de placebo (n = 20) dans une étude à double insu de 14 semaines. Le tour de taille (TT), la PPC et l indice de masse corporelle (IMC), le glucose, l insuline et les lipides à jeun ont été évalués au départ et aux 7 e et 14 e semaines du traitement. Résultats :Àla14 e semaine, la PPC (kg) était semblable dans le groupe de metformine (5,5 kg) et le groupe de placebo (6,3 kg), P = 0,4. Il n y avait pas de différences non plus entre les changements d IMC, de TT, de glucose, d insuline, et d indice d insulinorésistance (HOMA-IR), et les niveaux de lipide plasmatique observés dans le groupe de traitement et le groupe de placebo. Toutefois, les niveaux de glucose diminuaient significativement après l administration de metformine (P = 0,02). L HOMA-IR diminuait significativement dans les deux groupes, mais 3 sujets du groupe de placebo ont développé des taux de glucose à jeun supérieurs à 5 mmol/l. Après avoir pris de la metformine, les niveaux de triglycérides ont augmenté, mais le profil du cholestérol s est significativement amélioré. Conclusions : La metformine n a pas prévenu la PPC induite par l olanzapine. Bien que certains paramètres lipides aient empiré durant le placebo, l HOMA-IR s est amélioré dans les groupes de placebo et de metformine. La dysfonction du métabolisme des glucides n a pas été systématiquement observée durant l administration d olanzapine à court terme. 196