Association between the HTR2C rs1414334 C/G gene polymorphism and the development of the metabolic syndrome in patients treated with atypical antipsychotics

Setting

The proposed study was conducted in the Elche, Crevillente and Santa Pola Department of Health of the Valencian Community, which is a Mediterranean region located in southeastern Spain. In its mental health services, coverage is provided to approximately 325,000 inhabitants (almost entirely Caucasian). Health care is universal and free for all patients, except those with an irregular immigration status.

Study design and participants

In this cross-sectional, observational study, a sample of inpatients and outpatients treated for at least three months with atypical antipsychotics (clozapine, olanzapine, risperidone, quetiapine, aripiprazole, ziprasidone, amisulpride, asenapine and paliperidone) was analyzed. For the sample, consecutive patients attending Mental Health Services of the Valencian Community at the Elche, Crevillente and Santa Pola Department of Health between December 2012 and June 2013 were selected. All patients had to be of legal age (>18 years) and have a diagnosis of a psychotic disorder. This was defined as having a psychiatrist’s diagnosis of schizophrenia, schizophreniform disorder, schizoaffective disorder or other psychotic disorders, or bipolar disorder, and continuous treatment with antipsychotic medication (American Psychiatric Association, 2000). To comply with the current legislation, all patients were adequately informed about the study (written form) and, if agreeing to participate, signed a consent form before inclusion.

The study was approved by the Clinical Research Ethics Committee of the General University Hospital of Elche on November 20, 2012.

Variables and measurements

The primary variable was the diagnosis of the metabolic syndrome using the ATPIII criteria (Grundy et al., 2004), i.e., the presence of three or more of the following metabolic criteria: abdominal obesity, hypertriglyceridemia, low levels of high density lipoprotein cholesterol (HDL-C), elevated blood pressure and insulin resistance. All parameters of the metabolic syndrome criteria were determined using appropriate clinical guidelines (NHLBI Obesity Education Initiative Task Force, 1998; NCEP, 2002; Mancia et al., 2007; American Diabetes Association, 2012).

Secondary clinical variables were: age (years), gender, diagnosis (schizophrenia or other) (American Psychiatric Association, 2000), years with the diagnosis, treatment duration with the current antipsychotic (years), continuing therapy with a favorable (aripiprazole, ziprasidone or palperidone) or unfavorable (olanzapine, clozapine, risperidone, quetiapine or asenapine) (American Diabetes Association et al., 2004; Stahl, Mignon & Meyer, 2009; Schreiner et al., 2012; Stoner & Pace, 2012; Fleischhacker et al., 2013; Young, Taylor & Lawrie, 2015) metabolic profile, use of antidepressants, mood stabilizers and number of pack-years (number of cigarettes smoked per day × number of years smoked)/20). These variables were obtained by clinical interview with the patient and corroborated by the electronic medical record. It is obligatory that all drug prescriptions be made through this computerized system (Agencia Valenciana de Salud, 2011).

The following variables related to lifestyle were obtained by conducting the National Survey of Health and the European Health Survey (Ministerio de Sanidad, Política Social e Igualdad & Instituto Nacional de Estadística, 2009; Ministerio de Sanidad, Servicios Sociales e Igualdad & Instituto Nacional de Estadística, 2011): usual number of hours of sleep, physical activity during leisure time and habitual activity (work and home) and consumption of food (fruit, meat, eggs, fish, pasta, rice and potatoes, bread and cereals, vegetables, legumes, cold meats, milk products and sweets). Regular physical activity was assessed on a scale as follows: 0, sitting; 1, standing; 2, walking or moving around; 3, great effort. Physical activity during leisure time was measured with the following scale: 0, never; 1, less than once per month; 2, once or several times per month; 3, one or more times per week. Finally, the scale of all the food (fruit, meat, eggs, fish, pasta, rice and potatoes, bread and cereals, vegetables, legumes, cold meats, milk products and sweets) was rated as: 0, daily; 1, three or more times per week; 2, once or twice per week; 3, less than once per week; 4, never or almost never.

DNA isolation and genotyping

Genomic DNA was extracted from EDTA blood samples using the semi-automated QIAcube system (Qiagen, Hilden, Germany). The rs1414334 C/G polymorphism was analyzed by real-time PCR using a pre-developed assay (C_7455701_10, Applied Biosystems, Madrid, Spain) for allelic discrimination in a 7300 Real-Time PCR System (Applied Biosystems). The reaction was carried out with a TaqMan genotyping PCR master mix (Applied Biosystems), according to the protocol provided by Applied Biosystems.

The rs1414334 polymorphism allele C/G is described in the SNP database (http://www.ncbi.nlm.nih.gov/snp).

Sample size

The sample size during the study period was 166 patients, 68 of whom had the metabolic syndrome. As the sample was collected a posteriori, the power of the test was calculated to establish an association between an exposure factor and a disease; that is, whether an odds ratio (OR) is different from 1. For this calculation we assumed C allele in 30% of cases and 13% of controls, a control/case ratio of 2:1 and a type I error of 5% (Risselada et al., 2012). With these parameters, the power of the test was 81.7%.

Statistical methods

All variables were described using frequencies (absolute and relative), means and standard deviations. A stepwise logistic regression model was constructed to determine the relationship between the polymorphism and the metabolic syndrome. Since we had 68 events (metabolic syndrome), the model could only hold a maximum of six explanatory variables (one for every 10 events) (Ramírez-Prado et al., 2015). As we had a total of 25 variables, to determine which would be introduced in the model we first compared proportions and means between the metabolic syndrome (yes and no) by using the Chi-square test and the t-test, and then selected those variables with the smaller p-value. Thus the adjusted OR were obtained. The goodness-of-fit of the final model was assessed using the likelihood ratio test.

Since the HTR2C gene is X-linked, all the analyses were carried out for men and women separately. This is important because women encounter a very different situation than men. Women can be heterozygote, whereas men cannot, which creates a completely different biological scenario. This is also important with respect to previous articles published on this topic (Mulder et al., 2007; Mulder et al., 2009; Risselada et al., 2012).

All analyses were performed with a significance of 5% and for each relevant parameter its associated confidence interval (CI) was calculated. The statistical package used was IBM SPSS Statistics 19 (IBM, Armonk, NY, USA).

Summary

This study found no statistically significant association between the presence of the C allele of the rs1414334 polymorphism in the HTR2C gene and the metabolic syndrome.

Study strengths and limitations

The main strength of this study is that it provides greater knowledge of the association between the polymorphism analyzed and the metabolic syndrome. Moreover, all the variables that the Dutch group did not analyze in their studies were taken into account, such as physical activity, diet and smoking (Mulder et al., 2007; Mulder et al., 2009; Risselada et al., 2012). To minimize information bias, all measurements were taken in direct contact with the patient and according to current clinical guidelines. To minimize selection bias, data were taken from all the patients who attended mental health services during a specific period (six months).

Regarding limitations, we must bear in mind that our design is cross-sectional, thus we could not establish causal relationships between the factors analyzed and the presence of the metabolic syndrome. In other words, our study, as in the Dutch group’s study (Mulder et al., 2007; Mulder et al., 2009; Risselada et al., 2012; Ma et al., 2014), only assessed the association between the polymorphism studied and the metabolic syndrome; so to establish causality longitudinal studies should be performed. We must point out that the sample of our patients corresponds to a population treated with atypical antipsychotics, thus we cannot extrapolate the results to the general population (high heterogeneity). However, the aim of this study was to analyze the association between the metabolic syndrome and the polymorphism studied in patients with this type of treatment. Consequently, the results were obtained from a homogeneous sample of the population treated with atypical antipsychotics. Moreover, the development of the metabolic syndrome may have affected which medication the patients were given. However, Tables 1–3 show that patients with treatment of an unfavorable metabolic profile had a very similar proportion of the metabolic syndrome compared with the total sample.

Comparison with the existing literature

Although several studies address the possible associations between different HTR2C gene polymorphisms and the presence of the metabolic syndrome, only the Dutch studies focus on the rs1414334 polymorphism (Mulder et al., 2007; Mulder et al., 2009; Risselada et al., 2012; Ma et al., 2014). The main statistically significant finding of these studies was a direct association between the presence of the C allele of the polymorphism and the metabolic syndrome. As our results differ from those of these other authors, we cannot consider the association to be confirmed. Furthermore, the Dutch group did not stratify their results by gender, and since the HTR2C gene is X-linked, gender-specific analyses should have been carried out.

In agreement with Kuzman et al. (2008), it is possible that the variability of factors analyzed may influence the differences found. The prevalence of the C allele of the polymorphism analyzed in our study was similar to that of the Dutch studies. Conversely, the prevalence of the metabolic syndrome in our sample was slightly higher (Mulder et al., 2007; Mulder et al., 2009; Risselada et al., 2012; Ma et al., 2014). Additionally, previous studies did not assess atypical antipsychotics exclusively, as a percentage of the subjects had been prescribed traditional antipsychotics. Finally, the prevalence of schizophrenia in these papers was higher than in ours (Mulder et al., 2007; Mulder et al., 2009; Risselada et al., 2012). It is possible there were other differences on variables that were not reported in the Dutch studies. For example, our patients had their disease for an average of almost 15 years and had been taking the current antipsychotic medication for approximately five years. This information has not been analyzed by the Dutch group. On the other hand, there is a strong possibility that previous findings were false positives (Ioannidis et al., 2001). Therefore, the differences might be due to these factors.

Implications for research

The set of studies by the Dutch group indicates that the presence of the C allele of the HTR2C rs1414334 gene polymorphism appears to be associated with the metabolic syndrome. Our findings, however, do not support this association. As mentioned above, this could be due to differences between our patients and those of the Dutch group or could indicate type 1 error in the original findings. To distinguish these possibilities, we would need multicentre studies evaluating the association in different antipsychotic treatments, preferably using a prospective design. However, before investing in such research, it would be helpful to establish whether this SNP has a functional influence.