Skip to content


  • Research
  • Open Access

Genetic variation at the CYP2C19 gene associated with metabolic syndrome susceptibility in a South Portuguese population: results from the pilot study of the European Health Examination Survey in Portugal

  • 1,
  • 1,
  • 2,
  • 3,
  • 3,
  • 2,
  • 1,
  • 4,
  • 4,
  • 1 and
  • 1Email author
Diabetology & Metabolic Syndrome20146:23

  • Received: 4 November 2013
  • Accepted: 29 January 2014
  • Published:



Metabolic syndrome (MetS) is a cluster of conditions that occur together, increasing the risk of heart disease, stroke and diabetes. Since pathways implicated in different diseases reveal surprising insights into shared genetic bases underlying apparently unrelated traits, we hypothesize that there are common genetic components involved in the clustering of MetS traits. With the aim of identifying these common genetic components, we have performed a genetic association study by integrating MetS traits in a continuous MetS score.


A cross-sectional study developed in the context of the Portuguese Component of the European Health Examination Survey (EHES) was used. Data was collected through a detailed questionnaire and physical examination. Blood samples were collected and biochemical analyses were performed. Waist circumference, blood pressure, glucose, triglycerides and high density lipoprotein cholesterol (HDL) levels were used to compute a continuous MetS score, obtained by Principal Component Analysis. A total of 37 single nucleotide polymorphisms (SNPs) were genotyped and individually tested for association with the score, adjusting for confounding variables.


A total of 206 individuals were studied. Calculated MetS score increased progressively with increasing number of risk factors (P < 0.001). We found a significant association between CYP2C19 rs4244285 and the MetS score not detected using the MetS dichotomic approach. Individuals with the A allelic variant seem to be protected against MetS, displaying a lower MetS score (Mean difference: 0.847; 95%CI: 0.163-1.531; P = 0.015), after adjustment for age, gender, smoking status, excessive alcohol consumption and physical inactivity. An additive genetic effect of GABRA2 rs279871, NPY rs16147 and TPMT rs1142345 in the MetS score variation was also found.


This is the first report of a genetic association study using a continuous MetS score. The significant association found between the CYP2C19 polymorphism and the MetS score but not with the individual associated traits, emphasizes the importance of lipid metabolism in a MetS common etiological pathway and consequently on the clustering of different cardiovascular risk factors. Despite the sample size limitation of our study, this strategy can be useful to find genetic factors involved in the etiology of other disorders that are defined in a dichotomized way.


  • Metabolic syndrome
  • CYP2C19 gene
  • Genetic association study
  • Continuous MetS score


Metabolic syndrome (MetS) is a cluster of conditions — increased blood pressure, high blood glucose level, excess body fat around the waist and abnormal cholesterol levels — that occur together. It is strongly associated with cardiovascular diseases (CVD) and Type 2 Diabetes, increasing the risk of developing these disorders 2 and 5 fold, respectively [1]. MetS incidence and prevalence have clearly been rising worldwide, largely because of the increase in obesity rates, sedentary lifestyles and aging populations, and it is currently considered a significant public health problem [2]. In Portugal, the MetS prevalence, estimated for 2008, was 27.5% with regional variations, being highest in the Alentejo (30.99%) and lowest in the Algarve (24.42%) [3].

Due to the existence of multiple definitions considering different categorical cut-points, a consensus definition for MetS clinical diagnosis has been recently proposed. According to this definition, MetS is diagnosed when there are present at least three of the following five MetS features: abdominal obesity, elevated blood pressure, dyslipidemia (elevated triglycerides and low levels of high-density lipoprotein cholesterol), and hyperglycemia. Medication for any of these features is also considered as an indicator in the criteria for clinical diagnosis of MetS [4].

As a complex condition, MetS results from a complex interplay between many genetic and environmental factors. Lifestyle risk factors, particularly caloric excess diet and physical inactivity, seem to play an important role in MetS condition [5] but there is also evidence that its traits are highly heritable [6]. In a recent review of genome-wide association studies (GWAS), most of the single nucleotide polymorphisms (SNPs) associated with MetS traits are SNPs involved in lipid metabolism, like FTO rs9939609, TCF7L2 rs7903146, IL6 rs1800795, APOA5 rs662799, APOC3 rs2854117 and CETP rs708272 [7]. However, the identified variations explain only a very small fraction of disease burden in the population at large, suggesting that other genetic variants and interacting environmental factors are contributing to MetS susceptibility. Another important issue is that GWAS have identified numerous loci influencing metabolic risk traits individually, but to date, no loci have been found affecting the entire spectrum of MetS traits [8, 9]. These limitations may be due to the highly heterogeneous groups originated by the traditional dichotomic MetS approach, reflecting associations with particular individual traits. In this context, and given that comparison of pathways and processes implicated in different diseases are revealing surprising insights into the shared genetic bases underlying apparently unrelated traits [10], our main hypothesis is that there is a common genetic component underlying the clustering of MetS traits.

Although the consensus dichotomized definition previously described remains useful for clinical practice, it loses statistical power and information when performing association studies. There are multiple evidences that a continuous outcome increases the statistical power in genetic association studies instead of a dichotomous phenotype [8], and consequently for genetic epidemiological approaches, a continuous MetS score, obtained by integrating all MetS traits, would be a more appropriate and valid alternative to study the underlying risk factors responsible for that condition [11]. Therefore, taking these issues into account, the purpose of this study was to identify genetic factors associated with MetS, using a Principal Component Analysis (PCA) derived continuous MetS score, which has been previously validated [11], to perform a genetic association study using SNPs in candidate genes related to MetS features, like glucose/insulin homeostasis, cardiovascular regulation, body mass index and lipid/drug metabolism.


Study design and participants

We have performed a cross-sectional study as designed for the pilot study of the Portuguese Component of the European Health Examination Survey (EHES) project [12, 13]. This pilot study was conducted between 2010 and 2011, in the population covered by the São Brás de Alportel (Algarve) Health Center, constituted by 11089 individuals (2.6% of the total Algarve population). It consisted on an observational and descriptive epidemiological study with data collected through a detailed questionnaire (including socio-demographic factors, health state and health determinants) and physical examination. A blood sample for further biochemical analysis was also collected. Participants were selected using a simple random sampling scheme from the National Health System card number database, which covers over 99% of the total population from the São Brás de Alportel Health Center users. All participants were given a brief description of the objectives of the study, after which they signed an informed consent form. The study protocol was approved by the Ethics Committee of National Health Institute Doctor Ricardo Jorge and by the National Commission for Data Protection.

Measurements and blood sample collection

Blood pressure, anthropometric (weigh, height and waist circumference) and biochemical (high density lipoprotein cholesterol (HDL), triglycerides (TG) and glucose) measurements were performed in accordance with the recommendations proposed by FEHES – “Feasibility of a European Health Examination Survey” [14]. Genomic DNA was isolated from whole blood containing EDTA according to standard methods [15].

Calculation of the continuous MetS score

The MetS score was calculated by principal component analysis (PCA) with varimax rotation, as previously described [11]. We have considered six quantitative MetS risk factors (waist circumference, diastolic blood pressure (DBP), systolic blood pressure (SBP), glucose, TG and HDL plasma levels) assuming the newly harmonized diagnostic criteria of the MetS [4].

Shapiro-Wilk test was used to assess the normality of the MetS quantitative risk factors. Those non-normal distributed were normalized (SBP, Glucose, HDL and TG were normalized using log10[log10(SBP)], 1/[log10(Glucose)]10, [ln(HDL)]2 and log10[log10(TG)] respectively). We have used the first and second principle components (PC) which represent large fractions of MetS variance (considering eigenvalues >1.0). A final MetS score was computed by summing the first two individual PC scores, weighted by the relative contribution of each principal component in the explained variance. A higher MetS score indicates a less favorable MetS profile.

Validity of the MetS score was tested using the ANOVA for trend analysis. T-Test and ANOVA were also used to test mean difference values of the MetS score between groups of risk factor categories. The consensus definition recently proposed for MetS clinical diagnosis [4] was used to categorize individuals. According to this definition, MetS is diagnosed when there are present at least three of the following five MetS features: elevated waist circumference (men ≥94 cm, women ≥80 cm); low HDL cholesterol (men < 40 mg/dL, women < 50 mg/dL); elevated TG (≥150 mg/dL); elevated BP (≥130/85 mmHg); and elevated glucose levels (≥100 mg/dL). Medication for any of these features is also considered as alternative indicator in the criteria for diagnosis of MetS.

SNP selection and genotyping

To test the association between the MetS score variation and SNPs in different candidate genes potentially involved in the MetS etiology, we have selected SNPs based on their involvement in metabolic-related phenotypes. In order to capture the maximum variation with minimum genotyping effort, we have selected those that were haplotype-tagging SNPs and that represented a validated functionally relevant variation, reported in the OMIM database. Consequently, we have genotyped 37 SNPs: 13 SNPs in genes involved in glucose/insulin homeostasis CDKAL1 rs7754840, CDKN2A/Brs10811661, HHEX rs1111875, IGF2BP2 rs4402960, IL6 rs1800795, KCNJ11 rs5219, KCNQ1 rs2237892, MTNR1B rs10830963, PPARG rs1801282, SLC30A8 rs13266634, TCF7L2 rs7903146, ADCY5 rs11708067 and KCNQ1 rs231362), 10 SNPs in genes involved in body mass index (GNPDA2 rs10938397, MTCH2 rs10838738, NPC1 rs1805081, PTER rs10508503, SH2B1 rs7498665, FTO rs9939609, ADRB3 rs4994, GABRA2 rs279871, NPY rs16147, TMEM18 rs6548238), 7 SNPs in genes involved in cardiovascular system regulation (ACE rs4646994, NOS1AP rs12143842, ADRB1 rs1801252, ADRB2 rs1042714, NOS3 rs1799983, NOS3 rs2070744) and 7 involved in drug/lipid metabolism (APOE rs7412, LDLR rs2228671, CYP2C8 rs10509681, CYP2D6 rs16947, CYP2C19 rs4244285, TPMT rs1142345). All SNPs were genotyped by Sequenom MassARRAY platform except for seven (ACE rs464699, ADRB1 rs1801252, ADRB2 rs1042713, CYP2C8 rs10509681, CYP2C9 rs1799853, CYP2D6 rs16947 and NOS3 rs2070744) that were genotyped by Polymerase Chain Reaction (PCR) followed by Restriction Fragment Length Polymorphism (RFLPs) analysis (Additional file 1: Tables S1 and S2).

Statistical analysis

The statistical analysis was performed using IBM SPSS statistics 20. P-values < 0.05 were considered to be statistically significant. T-test and Mann-Whitney test were used to access differences of quantitative variables according to their adherence to the normal distribution, evaluated by the Shapiro-Wilk test. Proportions were compared using χ2 tests.

All SNPs were tested for Hardy Weinberg Equilibrium using the Hardy Weinberg R package [16], based on the χ2-test. The association between the MetS score and the isolated SNPs was tested by T-test. Subsequent correction for multiple comparisons was performed using the Bonferroni method. The association between the MetS categorized groups (participants with MetS versus participants without MetS) with selected SNPs was also performed using χ2-test. Individual association analysis between each MetS quantitative risk factor and the selected SNPs were also performed using either T-test or Mann-Whitney test, according to their adherence to the normal distribution. Lifestyle risk factors (smoking status, excessive alcohol consumption, physical inactivity and unhealthy diet) were also individually tested for association with the MetS score, using the T-test.

ANOVA for trend (P < 0.05) was used to assess linearity between the MetS score and the number of genetic risk factors, testing for additive genetic effects of risk variants in the increasing of MetS score values.

General linear model analysis (GLM) was used to test MetS score differences between subjects with different genotypes after adjusting for confounding variables such as age, gender, smoking status, excessive alcohol consumption, physical inactivity and unhealthy diet. Smoking status and excessive alcohol consumption were defined as previously described [17]. Inadequate physical activity was defined as practice of a regular physical activity such as running, cycling or other, in order to feel tired, less than once a week. Unhealthy diet was defined as absence of fruit or vegetables consumption on the day before the interview. Only variables contributing more than 5% for the MetS score variation were considered in the model.


Population characteristics

From the 221 recruited participants, we have excluded those with missing blood samples or missing values for the analyzed parameters (n = 15). The final study population consisted on 206 participants, 87 (42.2%) men and 119 (57.8%) women. The participants’ age ranged from 26 to 91 years, being the mean value 56.43 ± 16.23. The general characteristics of the participants included in this study are shown in Table 1.
Table 1

General characteristics of the participants


Participants (n = 206)


42.23% (35.49-48.98%)


57.77% (51.03-64.50%)

Age (years ± SD)

56.43 ± 16.23

MetS prevalence 1

46.12% (39.31-52.92%)


0.00 ± 1.41

MetS risk factors (Mean ± SD):


Waist circumference (cm)

95.50 ± 12.56

DBP (mmHg)

80.67 ± 9.96

SBP (mmHg)

131.72 ± 20.02

HDL (mg/dL)

53.51 ± 13.33

TG (mg/dL)

107.71 ± 60.29

Glucose (mg/dL)

103.29 ± 33.91

MetS related diseases



26.21% (20.21-32.22%)

Type 2 Diabetes

7.3% (3.73-10.83%)


12.6% (8.09-17.16%)


46.12% (39.31-52.92%)

Medication 2

43.20% (36.44-49.97%)

Lyfestyle risk factors



17.96% (12.72-23.20%)

Excessive alcohol consumption

8.74% (4.88-12.59%)

Inadequate physical activity

59.71% (53.01-66.41%)

Unhealthy diet

2.92% (29.37-42.47%)

Data is presented as mean ± standard deviation for continuous variables and % (95%CI) for proportions.

1For the metabolic syndrome definition, the newly harmonized diagnostic criteria was used [4].

2Medication for hypertension, type 2 diabetes and hypercholesterolemia was considered.

Abbreviations: CI Confidence interval, MetS metabolic syndrome, DBP diastolic blood pressure, SBP systolic blood pressure, HDL high density lipoprotein cholesterol, TG triglycerides.

We found differences between the frequencies of auto-reported MetS related chronic diseases confirmed by the family general practitioner (Table 1) and those obtained if we considered the correspondent parameters measured in this study, although they did not reach significance. We have found 26.21% auto-reported hypertensive participants, based on information collected by the questionnaire, compared to 39.32% of participants who have values of measured blood pressure above 140/90 mmHg, the cut-off value for hypertension definition. There are also 7.3% auto-reported diabetic participants, but 17.48% of the participants have levels of glucose above 110 mg/dL, the cut-off used in pre-diabetes diagnosis. The prevalence of MetS was 46.12%, assuming the newly harmonized definition (Figure 1). The MetS components prevalences’ are present in Figure 1, taking into account the most restrictive cut-off points [4].
Figure 1
Figure 1

MetS and its components prevalence. Participants medicated for hypertension, hypercholesterolemia and diabetes were also accounted. Error bars represent the 95% confidence intervals. Abbreviations: MetS, metabolic syndrome; DBP, diastolic blood pressure; SBP, systolic blood pressure; HDL, high density lipoprotein cholesterol; TG, triglycerides. (*For the MetS prevalence calculation, the newly harmonized definition was considered [4]).

Continuous MetS score calculation by PCA

In the total sample, from PCA considering the six quantitative MetS risk factors, we are able to explain 63.35% of these six components variance (PC1 and PC2 explained 35.42% and 27.43% of the variance, respectively). The measured correlations between each MetS risk factor and both principal components are presented in Table 2. BP, waist circumference and glucose levels are the components contributing more to PC1. For PC2, the main contributors are TG and HDL levels. No significant differences were found when subjects treated with medication for hypertension, hypercholesterolemia and diabetes were excluded from the analysis. In fact, medicated participants maintain altered values for the six quantitative MetS score risk factors and have higher MetS score values in comparison with non medicated participants, suggesting a medication inefficacy or absence of medication compliance.
Table 2

Correlation coefficients between the normalized components of MetS and the two principal components obtained from PCA

MetS normalized components

Coeficient correlation




Waist circumference


















Abbreviations: DBP diastolic blood pressure, SBP systolic blood pressure, HDL high density lipoprotein cholesterol, TG triglycerides, PC1 principal component 1, PC2 principal component 2.

The MetS score adequacy and validity is shown in Figure 2A. As expected, this score increases progressively with increasing numbers of risk factors (ANOVA test for linear trend, P < 0.001). As observed, the MetS score is clearly higher in subjects with MetS versus subjects without (0.97 ± 1.10 versus -0.83 ± 1.09; T-test P <0.001), when we categorize individuals based on the dichotomic MetS definition (Figure 2B).
Figure 2
Figure 2

Mets score validity. A MetS score variation according to the number of risk factors (ANOVA for trend P < 0.001). B Comparison between the affected versus unaffected participants (T-test P < 0.001). The consensus MetS definition was considered to define the affection status MetS groups [4].

Genotyping data

The descriptive statistics regarding the tested SNPs are described in Table 3. Minor allele frequencies (MAF) range from 0.027 to 0.491. No significant differences were found between the European MAF described in NCBI database [18] and the obtained MAF of the markers for the population under study. All SNPs are in Hardy-Weinberg equilibrium (P >0.05).
Table 3

List of SNPs selected in the present study











P-value 2

Glucose/insulin homeostasis



C → G


0.336 (226)

0.286 (206)





C → T


0.199 (226)

0.201 (206)





A → G


0.416 (226)

0.371 (206)





G → T


0.280 (118)

0.272 (206)





C → G


0.465 (226)

0.337 (206)





C → T



0.333 (206)





C → T


0.075 (226)

0.051 (206)





C → G


0.300 (120)

0.223 (206)





C → G


0.076 (118)

0.093 (205)





C → T


0.239 (226)

0.286 (206)





C → T


0.279 (226)

0.303 (205)





A → G


0.226 (226)

0.199 (206)





C → T


0.482 (112)

0.234 (128)


Cardiovascular regulation






0.420 (206)





C → T


0.188 (224)

0.265 (206)





A → G



0.108 (206)





C → G


0.467 (120)

0.407 (204)





A → G


0.358 (226)

0.362 (206)





G → T


0.342 (120)

0.417 (206)





C → T


0.106 (226)

0.124 (206)





C → T



0.451 (206)


Body mass Index



A → G


0.446 (112)

0.481 (206)





A → G


0.363 (226)

0.282 (206)





A → G


0.467 (120)

0.288 (206)





C → T


0.092 (218)

0.075 (206)





A → G


0.382 (226)

0.303 (206)





A → T


0.449 (118)

0.361 (205)





C → T


0.088 (226)

0.090 (205)





A → G



0.434 (206)





A → G


0.491 (226)

0.450 (206)





C → T


0.150 (220)

0.127 (205)


Lipid/Drug metabolism



C → T



0.027 (161)





C → T


0.137 (226)

0.129 (206)





C → T


0.104 (106)

0.138 (206)





A → G



0.393 (206)





G → A


0.155 (116)

0.129 (206)





A → G


0.027 (226)

0.032 (205)


1Values referring to the HAPMAP-CEU population, available in the NCBI database [18].

2P-values were obtained by “HardyWeinberg” R package, based on the χ2 -test.

Association analysis

Assuming a dominance model, a significant association was found between the MetS score and the following SNPs: CYP2C19 rs4244285 (P = 4.9×10-4), GABRA2 rs279871 (P = 0.018), NPY rs16147 (P = 0.029) and TPMT rs1142345 (P = 0.003). Considering the sample size and the allelic frequency of the four significant associated SNPs: CYP2C19 rs4244285, GABRA2 rs279871, NPY rs16147, TPMT rs1142345, this study has 80% power to detect a mean difference of 0.642, 0.598, 0.612 and 1.132 respectively, in the MetS score variation. Regarding the CYP2C19 rs4244285 SNP, individuals included in the GA + AA genotype group seem to be protected against MetS, displaying a lower MetS score (Mean difference: 0.792; 95%CI: 0.351-1.233; P < 0.001) (Table 4). After Bonferroni correction for multiple testing, the association between the MetS score and the CYP2C19 rs4244285 remains significant (P = 0.018). Within both genotype groups, medicated individuals display higher MetS score values than non medicated ones, although not reaching significance.
Table 4

SNPs significantly associated with MetS score




MetS score

Mean difference


P-value a

Corrected P-valueb




0.192 ± 1.380






GA + AA1


-0.600 ± 1.362





0.350 ± 1.374






GA + GG2


-0.154 ± 1.409





0.342 ± 1.606






GA + GG3


-0.134 ± 1.313





-0.080 ± 1.375








1.119 ± 1.601


aT-test was used to compare MetS score mean values between the two groups.

bCorrected P-values were obtained using the Bonferroni test to multiple testing correction.

1The GA + AA group consists on 3 AA and 47 GA individuals.

2The GA + GG group consists on 36 AA and 107 GA individuals.

3The GA + GG group consists on 37 AA and 111 GA individuals.

The MetS score is presented as mean ± SD.

Despite the fact that associations of the three SNPs (rs279871, rs16147, rs1142345) in the GABRA2, NPY and TPMT genes, with the MetS score do not remain significant after multiple testing correction, we can observe an additive genetic effect of these variants in the MetS score (Figure 3A), since it increases with the growing number of genetic risk factors (ANOVA for trend P < 0.001). Moreover, this additive genetic effect is age independent, as shown in Figure 3B.
Figure 3
Figure 3

Additive genetic effects of the SNPs ( GABRA2 rs279871, NPY rs16147 , TPMT rs1142345 ) . A- Influence of the number of risk genotypes in the MetS score values. MetS score increases with increasing number of genetic risk factors (ANOVA for trend P < 0.001). B- Additive genetic effect representation. Each line represents a different number of risk genotypes, considering sufficient the presence of one allele for each variant to be at risk. We have considered the 4 significantly associated SNPs previous reported: CYP2C19 rs4244285, GABRA2 rs279871, NPY rs16147 and TPMT rs1142345. No individuals with 4 risk genotypes for the 4 SNPs were identified in this population.

No association was found between the individual MetS quantitative risk factors (waist circumference, BP, glucose, TG and HDL blood levels) and the selected SNPs. Similarly, no association was found between MetS using the dichotomic definition (participants with MetS versus participants without MetS) and the analyzed SNPs. No association was found between the MetS score and the selected lifestyle risk factors.

General linear model analysis

Using general linear model analysis, we found that differences on the MetS score between subjects with GG genotype and GA + AA genotype on the CYP2C19 rs4244285 remain significant after adjustment for age, gender, smoking status, excessive alcohol consumption and physical inactivity (Mean difference: 0.847; 95%CI: 0.163-1.531; P = 0.015). The variable unhealthy diet was discarded since it contributes less than 5% for the MetS score variation. General linear model analysis was also performed for the other associated SNPs: GABRA2 rs279871 (Mean difference: 0.597; 95%CI: 0.053-1.247; P = 0.071), NPY rs16147 (Mean difference: 0.854; 95%CI: 0.175-1.532; P = 0.014) and TPMT rs1142345 (Mean difference: 0.850; 95%CI: 0.086-1.787; P = 0.075), assuming the adjustment for the same variables.


GWAS have identified common variants of modest and small-effect size at hundreds of loci for common chronic disorders but a substantial fraction of heritability remains unexplained [34]. Given that GWAS performed on MetS traits, using a traditional dichotomic definition indicates that only a small portion of the variation in these traits can be explained by the reported common genetic variants, our aim was to identify genetic risk factors involved in MetS etiology using a quantitative phenotype. To achieve our goal we have tested the association between a group of biologically relevant SNPs in order to seek positive genetic associations with a continuous MetS score, obtained by PCA, instead of the dichotomized MetS definition. We believe that, using a continuous MetS score based on PCA to address the issue of genetic susceptibility to MetS and its clustering associated traits, will increase the statistical power and unravel the missing heritability. In fact, the MetS score based on this approach was able to explain over 63.0% of the phenotype.

In this study, we have found a significant association between the MetS score and the CYP2C19 rs4244285. Individuals carrying the A allelic variant have a lower MetS score, suggesting that these genotypes are conferring protection against MetS. Most importantly, the differences on the MetS score between subjects with GG genotype and GA + AA genotype on the CYP2C19 remained significant after adjustment for age, gender, smoking status, excessive alcohol consumption and physical inactivity, which are important confounding factors, suggesting a pivotal role of this gene in metabolic regulation.

Regarding the CYP2C19 gene polymorphisms, most of the studies performed to date focus on the altered drug metabolism [35] and little importance has been given to its role on lipid metabolism involvement. It is known that the functional enzyme product of the CYP2C19 gene also metabolizes important endogenous substrates, namely arachidonic acid, a n-6 unsaturated fatty acid, to produce epoxyeicosatrienoic (EETs) compounds which generally possess vasodilating, anti-inflammatory, anti-apoptotic, anti-thrombotic, natriurectic and cardioprotective effects [36]. On the other hand, CYP2C19 rs4244285 consists in an aberrant splice site, generating a truncated non-functional protein, without catalytic activity [30] which has been shown to be associated with increased adverse cardiovascular events in patients medicated with clopidogrel [37]. This hypothesis is in accordance with our results, given that, within the GA + AA genotype group, medicated individuals display a higher MetS score than non medicated ones. However, further studies will be necessary to clarify whether this variant is causal or if it is in linkage disequilibrium with the true causal variant.

Given that CYP2C19 is a highly polymorphic gene, with at least 19 allelic variants reported [38], we have to consider the potential coexistence of other variants that could compensate this defective allele by overexpressing it. Actually, a recently described gain of function allele in CYP2C19 gene (rs12248560) is a regulatory polymorphism enhancing CYP2C19 expression with potential to compensate the CYP2C19 rs4244285 defective variant [39]. We expect that, with the use of next generation sequencing technologies, we will be able to further dissect genetic variation present in this gene and clarify the relative contribution of each variant to the CYP2C19 actual function.

We have also found an association between the MetS score and SNPs at the GABRA2, NPY and TPMT genes. Despite the fact that these associations do not remain significant after multiple testing correction, likely due to the lack of statistical power, they might represent an additive genetic effect that should be taken in consideration in the etiology of MetS. The genes NPY and GABRA2 have been previously found associated with obesity and food intake, being involved in multiple central nervous system functions regulation [4042], while TPMT has an important role in drug metabolism [43].

Given these results, we hypothesize that CYP2C19 rs424428 might be involved in a common pathway, the deregulation of which, in addition to other specific genetic factors, may lead to the different MetS associated traits. This hypothesis is corroborated by the fact that, when we test associations between each MetS quantitative risk factor (waist circumference, BP, glucose, TG and HDL blood levels) and the selected candidate SNPs, no association was found. This finding suggests that there is a common basic pathway involving variants in CYP2C19 alone or in association with other genes that may lead to the development of MetS and associated traits, similarly to what has been described for the HLA or CTLA4 genes in the development of different autoimmune disorders [44].

Despite the fact that, further association and functional studies are necessary to highlight the role of CYP2C19 in MetS etiology, we think that these results may contribute to the identification of new therapeutic targets that may be useful in the treatment of the different clustered traits instead of treating them individually. Novel therapies targeted at these newly identified genes may be developed and consequently improve the outcome of patients affected by this disorder.


Our study represents an integrative approach to identifying genetic risk factors involved in MetS etiology, through a continuous MetS score obtained by PCA. This score alone explains over 63.0% of the phenotype, supporting the usefulness of a continuous MetS risk score, instead of the dichotomized MetS definition traditionally used in case-control studies.

A significant corrected association between a CYP2C19 rs4244285 and the MetS score was clearly found. The less frequent allele of this variant seems to be conferring a protective effect to MetS susceptibility. In addition, we found that these differences remained significant after adjustment for age, gender, smoking status, excessive alcohol consumption and physical inactivity. We hypothesize that CYP2C19 rs4244285 is involved in a common pathway, the deregulation of which, in addition to other specific genetic factors, may lead to the different MetS associated traits. This hypothesis is supported by the fact that no association was found between this SNP and each quantitative risk factor. Other variants in GABRA2, NPY and TPMT might represent additive genetic factors of modest effect that should be taken into consideration to understand the complete etiology of MetS, as well as environmental factors such as smoking status, diet and physical activity. Despite the need of further studies to confirm this association in a larger population with better statistical power, we are confident that these results will lead to the identification of new therapeutic targets. These may be useful in the treatment of the different clustered traits instead of treating them individually and consequently improving the outcome of patients affected by both MetS and associated chronic diseases.



Blood pressure


Cardiovascular diseases


Diastolic blood pressure


European Health Examination Survey


Feasibility of a European Health Examination Survey


General practitioner


Genome-wide association study


High density lipoprotein cholesterol


Hardy Weinberg Equilibrium


Metabolic syndrome


National Center for Biotechnology Information


Principal component


Principal component analysis


Restriction fragment length polymorphism


Systolic blood pressure


Single nucleotide polymorphism





The authors are grateful to the Association for Research and Development of the Faculty of Medicine (AIDFM), Lisbon. The authors are also grateful to J. Costa for assistance with genotyping (Instituto Gulbenkian de Ciência/Genomics Unit).


The pilot study of the Portuguese Component of the European Health Examination Survey (EHES) project has received funding from the European Commission/DG Sanco (Agreement number: 20092301 – EHES JA – EAHC). This study has also received funding from the Portuguese Foundation for Science and Technology (FCT) (Project Reference: PTDC/SAU-ESA/101743/2008).

Authors’ Affiliations

Departamento de Epidemiologia, Instituto Nacional de Saúde Doutor Ricardo Jorge, Lisboa, Portugal
Laboratório de Saúde Pública Dra. Laura Ayres, Faro, Portugal
Administração Regional de Saúde do Algarve, Faro, Portugal
Departamento de Promoção da Saúde e Prevenção das Doenças Não Transmissíveis, Instituto Nacional de Saúde Doutor Ricardo Jorge, Lisboa, Portugal


  1. Eckel RH, Grundy SM, Zimmet PZ: The metabolic syndrome. Lancet. 2005, 365 (9468): 1415-1428. 10.1016/S0140-6736(05)66378-7.View ArticlePubMedGoogle Scholar
  2. Cornier M-A, Dabelea D, Hernandez TL, Lindstrom RC, Steig AJ, Stob NR, Van Pelt RE, Wang H, Eckel RH: The Metabolic Syndrome. Endocr Rev. 2008, 29 (7): 777-822. 10.1210/er.2008-0024.View ArticlePubMedGoogle Scholar
  3. Fiuza M, Cortez-Dias N, Martins S: Síndrome metabólica em Portugal: prevalência e implicações no risco cardiovascular-Resultados do Estudo VALSIM. Rev Port Cardiol. 2008, 27 (12): 1495-1529.PubMedGoogle Scholar
  4. Alberti K, Eckel R, Grundy S, Zimmet P, Cleeman J, Donato K, Fruchart J, James W, Loria C, Smith S: Harmonizing the Metabolic Syndrome A Joint Interim Statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circ. 2009, 120: 1640-1645. 10.1161/CIRCULATIONAHA.109.192644.View ArticleGoogle Scholar
  5. Wijndaele K, Duvigneaud N, Matton L, Duquet W, Delecluse C, Thomis M, Beunen G, Lefevre J, Philippaerts RM: Sedentary behaviour, physical activity and a continuous metabolic syndrome risk score in adults. Eur J Clin Nutr. 2007, 63 (3): 421-429.View ArticlePubMedGoogle Scholar
  6. Bosy-Westphal A, Onur S, Geisler C, Wolf A, Korth O, Pfeuffer M, Schrezenmeir J, Krawczak M, Muller MJ: Common familial influences on clustering of metabolic syndrome traits with central obesity and insulin resistance: the Kiel obesity prevention study. Int J Obes. 2006, 31 (5): 784-790.Google Scholar
  7. Povel CM, Boer JMA, Reiling E, Feskens EJM: Genetic variants and the metabolic syndrome: a systematic review. Obes Rev. 2011, 12 (11): 952-967. 10.1111/j.1467-789X.2011.00907.x.View ArticlePubMedGoogle Scholar
  8. Frazer KA, Murray SS, Schork NJ, Topol EJ: Human genetic variation and its contribution to complex traits. Nat Rev Genet. 2009, 10 (4): 241-251.View ArticlePubMedGoogle Scholar
  9. Kraja AT, Vaidya D, Pankow JS, Goodarzi MO, Assimes TL, Kullo IJ, Sovio U, Mathias RA, Sun YV, Franceschini N: A Bivariate Genome-Wide Approach to Metabolic Syndrome: STAMPEED Consortium. Diabetes. 2011, 60 (4): 1329-1339. 10.2337/db10-1011.PubMed CentralView ArticlePubMedGoogle Scholar
  10. Torkamani A, Topol EJ, Schork NJ: Pathway analysis of seven common diseases assessed by genome-wide association. Genomics. 2008, 92: 265-272. 10.1016/j.ygeno.2008.07.011.PubMed CentralView ArticlePubMedGoogle Scholar
  11. Wijndaele K, Beunen G, Duvigneaud N, Matton L, Duquet W, Thomis M, Lefevre J, Philippaerts R: A continuous metabolic syndrome risk score: utility for epidemiological analyses. Diabetes Care. 2006, 29: 2329-View ArticlePubMedGoogle Scholar
  12. Kuulasmaa K, Tolonen H, Koponen P, Kilpeläinen K, Avdicová M, Broda G, Calleja N, Dias C, Gösswald A, Kubinova R, et al: An overview of the European Health Examination Survey Pilot Joint Action. Arch Public Health. 2012, 70 (1): 20-10.1186/0778-7367-70-20.PubMed CentralView ArticlePubMedGoogle Scholar
  13. Instituto Nacional de Saúde Doutor Ricardo Jorge: Inquérito de Saúde com Exame Físico: Relatório do estudo realizado em São Brás de Alportel. 2012, Lisboa: INSA. Departamento de Epidemiologia. Departamento da Promoção de Saúde e Doenças Crónicas,,Google Scholar
  14. FEHES - Feasibility of a European Health Examination Survey.,
  15. Lahiri DK, Numberger JI: A rapid non-enzymatic method for the preparation of HMW DNA from blood for RFLP studies. Nucleic Acids Research. 1991, 19 (19): 5444-10.1093/nar/19.19.5444.PubMed CentralView ArticlePubMedGoogle Scholar
  16. Graffelman J: HardyWeinberg: Graphical tests for Hardy-Weinberg equilibrium. R package version 1.4.1. 2012,,Google Scholar
  17. Matias Dias C, Briz T: Distribuição conjunta de determinantes de saúde relacionados com comportamentos na população portuguesa. Observações_Boletim epidemiológico. Vol.1, N°1, Jul-set. 2012, Lisboa: Instituto Nacional de Saúde Doutor Ricardo Jorge,,Google Scholar
  18. National Center for Biotechnology Information.,
  19. Mihaescu R, Meigs J, Sijbrands E, Janssens AC: Genetic risk profiling for prediction of type 2 diabetes. PLoS Curr. 2011, 3: 1208-View ArticleGoogle Scholar
  20. Dupuis J, Langenberg C, Prokopenko I, Saxena R, Soranzo N, Jackson AU, Wheeler E, Glazer NL, Bouatia-Naji N, Gloyn AL: New genetic loci implicated in fasting glucose homeostasis and their impact on type 2 diabetes risk. Nat Genet. 2010, 42 (2): 105-116. 10.1038/ng.520.PubMed CentralView ArticlePubMedGoogle Scholar
  21. Crisan D, Carr J: Angiotensin I-converting enzyme: genotype and disease associations. J Mol Diagn. 2000, 2 (3): 105-115. 10.1016/S1525-1578(10)60624-1.PubMed CentralView ArticlePubMedGoogle Scholar
  22. Eijgelsheim M, Aarnoudse ALHJ, Rivadeneira F, Kors JA, Witteman JCM, Hofman A, van Duijn CM, Uitterlinden AG, Stricker BHC: Identification of a common variant at the NOS1AP locus strongly associated to QT-interval duration. Hum Mol Genet. 2009, 18 (2): 347-357.View ArticlePubMedGoogle Scholar
  23. Ranade K, Jorgenson E, Sheu WHH, Pei D, Hsiung CA, Chiang F-T, Chen Y-DI, Pratt R, Olshen RA, Curb D, et al: A Polymorphism in the β1 Adrenergic Receptor Is Associated with Resting Heart Rate. Am J Hum Genet. 2002, 70 (4): 935-942. 10.1086/339621.PubMed CentralView ArticlePubMedGoogle Scholar
  24. Dias R, Gowdak M, Pereira A: Genetics and cardiovascular system: influence of human genetic variants on vascular function. Genes Nutr. 2011, 6 (1): 55-62. 10.1007/s12263-010-0193-7.PubMed CentralView ArticlePubMedGoogle Scholar
  25. Corella D, Ordovas JM: Single Nucleotide Polymorphisms that influence lipid metabolism. Interaction with Dietary Factors. Annu Rev of Nutr. 2005, 25: 341-390. 10.1146/annurev.nutr.25.050304.092656.View ArticleGoogle Scholar
  26. Hinney A, Vogel CG, Hebebrand J: From monogenic to polygenic obesity: recent advances. Eur Child Adolesc Psychiatr. 2010, 19 (3): 297-310. 10.1007/s00787-010-0096-6.View ArticleGoogle Scholar
  27. Marti A, Corbalán MS, Martínez-Gonzalez MA, Martinez JA: TRP64ARG polymorphism of the β3-adrenergic receptor gene and obesity risk: effect modification by a sedentary lifestyle. Diabetes Obes Metab. 2002, 4 (6): 428-430. 10.1046/j.1463-1326.2002.00227.x.View ArticlePubMedGoogle Scholar
  28. Bauer L, Dick D, Bierut L, Bucholz K, Edenberg H, Kuperman S, Kramer J, Nurnberger J, O’Connor S, Rice J, et al: Obesity, Smoking, and Frontal Brain Dysfunction. Am J Addict. 2010, 19 (5): 391-400. 10.1111/j.1521-0391.2010.00069.x.PubMed CentralView ArticlePubMedGoogle Scholar
  29. Zhang L, Bijker MS, Herzog H: The neuropeptide Y system: Pathophysiological and therapeutic implications in obesity and cancer. Pharmacol Ther. 2011, 131 (1): 91-113. 10.1016/j.pharmthera.2011.03.011.View ArticlePubMedGoogle Scholar
  30. Goldstein JA: Clinical relevance of genetic polymorphisms in the human CYP2C subfamily. Br J Clin Pharmacol. 2001, 52 (4): 349-355. 10.1046/j.0306-5251.2001.01499.x.PubMed CentralView ArticlePubMedGoogle Scholar
  31. Božina N, Bradamante V, Lovrić M: Genetic Polymorphism of Metabolic Enzymes P450 (CYP) as a Susceptibility Factor for Drug Response, Toxicity, and Cancer Risk. Arch Ind Hyg Toxicol. 2009, 60: 217-Google Scholar
  32. Bertrand-Thiébault C, Berrahmoune H, Thompson A, Marie B, Droesch S, Siest G, Foernzler D, Visvikis-Siest S: Genetic Polymorphism of CYP2C19 Gene in the Stanislas Cohort. A link with Inflammation. Ann Hum Genet. 2008, 72 (2): 178-183. 10.1111/j.1469-1809.2007.00417.x.View ArticlePubMedGoogle Scholar
  33. Pavlovic S, Zukic B, Nikcevic G: Pharmacogenomics of Thiopurine S Methyltransferase: Clinical Applicability of Genetic Variants, Clinical Applications of Pharmacogenetics. Clinical Applications of Pharmacogenomics. Edited by: Sanoudou D. 2012, Rijeka: Tech, 75-94.Google Scholar
  34. Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K: A comprehensive review of genetic association studies. Genet Med. 2002, 4 (2): 45-61. 10.1097/00125817-200203000-00002.View ArticlePubMedGoogle Scholar
  35. Zabalza M, Subirana I, Sala J, Lluis-Ganella C, Lucas G, Tomás M, Masiá R, Marrugat J, Brugada R, Elosua R: Meta-analyses of the association between cytochrome CYP2C19 loss- and gain-of-function polymorphisms and cardiovascular outcomes in patients with coronary artery disease treated with clopidogrel. Heart. 2012, 98 (2): 100-108. 10.1136/hrt.2011.227652.View ArticlePubMedGoogle Scholar
  36. Zordoky BNM, El-Kadi AOS: Effect of cytochrome P450 polymorphism on arachidonic acid metabolism and their impact on cardiovascular diseases. Pharmacol Ther. 2010, 125 (3): 446-463. 10.1016/j.pharmthera.2009.12.002.View ArticlePubMedGoogle Scholar
  37. Giusti B, Gori AM, Marcucci R, Abbate R: Relation of CYP2C19 loss-of-function polymorphism to the occurrence of stent thrombosis. Expert Opin Drug Metab Toxicol. 2010, 6 (4): 393-407. 10.1517/17425251003598878.View ArticlePubMedGoogle Scholar
  38. CYP2C19 allele nomenclature.,
  39. Li-Wan-Po A, Girard T, Farndon P, Cooley C, Lithgow J: Pharmacogenetics of CYP2C19: functional and clinical implications of a new variant CYP2C19*17. Br J Clin Pharmacol. 2010, 69: 222-230. 10.1111/j.1365-2125.2009.03578.x.PubMed CentralView ArticlePubMedGoogle Scholar
  40. Mutschler J, Abbruzzese E, Wiedemann K, von der Goltz C, Dinter C, Mobascher A, Thiele H, Diaz-Lacava A, Dahmen N, Gallinat J, et al: Functional Polymorphism in the Neuropeptide Y Gene Promoter (rs16147) Is Associated with Serum Leptin Levels and Waist-Hip Ratio in Women. Ann Nutr Met. 2013, 62 (4): 271-276. 10.1159/000346799.View ArticleGoogle Scholar
  41. Hohmann S, Buchmann AF, Witt SH, Rietschel M, Jennen-Steinmetz C, Schmidt MH, Esser G, Banaschewski T, Laucht M: Increasing association between a neuropeptide Y promoter polymorphism and body mass index during the course of development. Pediatr Obes. 2012, 7 (6): 453-460. 10.1111/j.2047-6310.2012.00069.x.View ArticlePubMedGoogle Scholar
  42. Morton GJ, Cummings DE, Baskin DG, Barsh GS, Schwartz MW: Central nervous system control of food intake and body weight. Nat. 2006, 443 (7109): 289-295. 10.1038/nature05026.View ArticleGoogle Scholar
  43. Blaker PA, Arenas-Hernandez M, Marinaki AM, Sanderson JD: The pharmacogenetic basis of individual variation in thiopurine metabolism. Pers Med. 2012, 9 (7): 707-725. 10.2217/pme.12.85.View ArticleGoogle Scholar
  44. Castiblanco J, Arcos-Burgos M, Anaya JM: What is next after the genes for autoimmunity?. BMC Med. 2013, 11: 197-10.1186/1741-7015-11-197.PubMed CentralView ArticlePubMedGoogle Scholar


© Gaio et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.