A Systematic Review and Meta-Analysis of the Association between Serotonergic Gene Polymorphisms and Obstructive Sleep Apnea Syndrome

Background 5-Hydroxytryptamine receptor (5-HTR) and 5-hydroxytryptamine transporter (5-HTT) gene polymorphisms have been reported to be associated with susceptibility to obstructive sleep apnea syndrome (OSAS). The associations, derived from sporadic, inconsistent, small-sample-size studies, need to be evaluated further in a meta-analysis. Methods Relevant studies were identified by searching PubMed, Embase, The Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang, and Weipu. Eligible data were extracted from each included study. Odds ratios (ORs) were calculated using a fixed-effects or a random-effects model. The ORs and 95% confidence interval (CI) were used to assess the strength of the association between serotonergic gene polymorphisms and OSAS in the dominant and recessive models, as well as alleles. The Q statistic was used to evaluate homogeneity and Begg’s test was used to assess publication bias. Results Eight studies were finally included in the meta-analysis of the association between 5-HTR2A gene variants (including 102T/C and 1438G/A), 5-HTT gene polymorphisms (including 5-HTT gene-linked promoter region (5-HTTLRP), and serotonin transporter intron 2 variable number tandem repeat (STin2VNTR) and OSAS risk. The G allele of 5-HTR2A 1438G/A, long 5-HTTLPR, and 10-tandem-repeats STin2VNTR were shown to increase OSAS susceptibility, with ORs of 2.33 (A vs. G, 95% CI 1.48–3.66), 1.24 (L vs. S, 95% CI: 1.04–1.49), and 2.87 (10 vs. 12, 95% CI: 1.38–5.97), respectively. These significant differences were determined in both dominant and recessive models. Of the 5-HTR2A 1438G/A gene polymorphism, the AA genotype increased the OSAS risk, with an OR of 4.21 (95% CI: 2.83–6.25) in a recessive model in male OSAS patients, but no significant association was found in females. Conclusions Our meta-analysis demonstrated that polymorphisms in the 5-HTR2A 1438G/A and 5-HTT genes contributed to susceptibility to OSAS. The A allele of the 1438G/A gene polymorphism is predominantly distributed in males and increased the OSAS risk significantly.


Introduction
Because of its high morbidity and mortality, obstructive sleep apnea syndrome (OSAS) has recently become an important public health issue. It affects about 2% of females and 4% of males in middle-aged groups in Western counties [1]. Repeated partial or complete collapse of the pharynx during sleep, which results in apnea or hypopnea associated with oxygen desaturation and arousal from sleep, is the most characteristic feature of OSAS [2]. The pathophysiology of OSAS is associated with activation of a number of neural, humoral, thrombotic, metabolic, and inflammatory disease mechanisms [3].
Serotonergic neurons innervating motoneurons increase their firing rates in response to respiratory challenges, and long-term facilitation of respiratory activity in response to hypoxia is 5-hydroxytryptamine (5-HT, serotonin)-dependent [4]. Thus, it is reasonable that the receptor and transporter of 5-HT are also involved in the pathophysiological process of OSAS.
Although the various studies have provided important serotonergic genetic information regarding susceptibility to OSAS [9][10][11][12][13][14]19,20], because of the limitations of small sample sizes in each study, low statistical power, and inconsistent results, they are incapable of demonstrating a reliable association. Also, the associations hinted at from sporadic cases and populations of differing ethnic origins need to be re-analyzed comprehensively. In this report, to further evaluate the associations between 5-HT receptor and transporter gene polymorphisms and the risk of OSAS, we performed a meta-analysis by combining the genotyping data from eligible studies published to date.

Materials and Methods
This study was strictly followed the guidelines of the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis).

Publication Search
To identify all eligible studies that address the association between 5-HT receptor and transporter gene polymorphisms and OSAS, we conducted key word searches in the biomedical electronic databases, including PubMed, Embase, The Cochrane Library, China National Knowledge Infrastructure (CNKI), Wanfang, and Weipu: (obstructive sleep apnea hypopnea syndrome or OSAHS or obstructive sleep apnea syndrome or OSAS or obstructive sleep apnea or OSA or sleep apnea or apnea) and (5-hydroxytryptamine receptor or 5-HTR or serotonin receptor or 5-hydroxytryptamine transporter or 5-HTT or serotonin transporter) in combination with (gene or polymorphism or variants or alleles) were used as research terms. The final search was updated in June 2013. The search was limited to studies published in English and Chinese.

Inclusion and Exclusion Criteria
Case-control studies that met the following criteria were included: 1) published in peer-reviewed journals, 2) contained independent data, 3) had enough data to evaluate the relationship between gene polymorphisms and OSAS susceptibility, and 4) had enough data to determine the genotype distributions in both the case and control groups. The authors were contacted if there were queries regarding the studies.
Studies were excluded for the following reasons: 1) lack of a suitable control group, 2) no usable data reported, 3) data were contradictory, 4) repeated data, and 5) abstracts or reviews.

Data Extraction
Two investigators (Drs. Xu and Guan) performed the data search, data screening, and data extraction independently. Disagreements regarding data screening and extraction were resolved by discussion. If there was still disagreement, a third reviewer (Prof. Yin) participated to resolve the issue. The following data were extracted from the original studies: first author, year of publication, country, sample size, age, gender, body mass index (BMI), apnea-hypopnea index (AHI), and genotype frequency.

Statistical Methods
The Review Manager software (ver. 5.0, Cochrane Collaboration, Oxford, UK), the SAS software (ver. 9.1.3, SAS Institute Inc., Cary, NC), and Stata (ver. 11.0, Stata Corporation, College Station, TX) were used for the meta-analysis and statistical analyses. For genetic frequency distributions, Hardy-Weinberg equilibrium (HWE) was calculated with the SAS software to assess goodness of fit. Homogeneity was tested using Cochran chisquared tests. The Mantel-Haenszel fixed-effects model or the random-effects model was used for the meta-analysis in the comparison of serotonergic gene polymorphisms [21][22]. Begg's test was used to evaluate publication bias.

Search Results
In total, 52 studies were identified by electronic and manual searching. At the stage of screening by the inclusion criteria, 38 studies were excluded for the following reasons: animal studies, cell studies, reviews, or not correlated with OSAS. At the stage of screening by the exclusion criteria, after reading the full articles, six studies were excluded for the following reasons: two articles were repeated studies [23][24], one was an abstract [25], two articles lacked complete data [26][27], and one article had contradictory data [28]. Finally, eight studies met the exclusion criteria and were included in our systemic review and metaanalysis ( Figure 1). Of the eight studies, six had data for 5-HTR2A 102 T/C, five for 5-HTR2A 1438 G/A, three for 5-HTTLRP, and three for STin2VNTR ( Figure 1, Table S1).

Study Characteristics
All eligible studies were published between 2005 and 2013. Five studies were from China [10,13,14,19,20], one from Turkey [9], one from Brazil [11], and one from Japan [12] (Table S1). The studied populations were mainly middle-aged and slightly overweight in cases and controls, although one study lacked this information (Table S1). One study did not provide gender data [12]; the male proportion in the remaining populations ranged from 40.0% to 88.6% (Table S1). The BMI ranged from 21.367.5 kg/m 2 to 29.363.7 kg/m 2 , although one study lacked these data [9]. The AHI of OSAS populations ranged from $5 to 53.9616.4 times per h; the AHI of control groups were all below 5 (Table S1). All polymorphisms were identified by polymerase chain reaction (PCR) or PCR-restriction fragment length polymorphism (PCR-RFLP) techniques.

Meta-analysis of the Association between 5-HTR and 5-HTT Gene Polymorphisms and OSAS
The association between gene polymorphisms and susceptibility to OSAS was evaluated in the total case and control populations combined from eligible studies of each gene locus. All genetic loci were tested in a dominant model, a recessive model, and a susceptibility allele versus insusceptibility allele model.
Of the 5-HTR2A 102T/C polymorphisms, six studies containing 566 case subjects and 728 controls were included in this metaanalysis [9][10][11][12][13][14]. The I 2 of the dominant models (TT+TC vs. CC), recessive model (TT vs. TC+CC) and T allele vs. C allele were 0  and the p values of the Q test were higher than 0.1, suggesting no significant heterogeneity among the studies, so we used a fixedeffects model to analyze the data. As shown in Table 2, metaanalysis of the dominant model (TT +TC vs. CC), recessive model (TT vs. TC +CC), and T allele vs. C allele revealed that no genotype or allele was associated with OSAS risk (Table 2, Figure 2A, B, C). For the 5-HTR2A 1438G/A polymorphism, five studies had 466 case subjects and 551 controls [9][10][11]13,14]. The I 2 of the dominant models (AA+AG vs. GG) and the A allele vs. G allele were over 50%, and the p values of the Q test were 0.47 and ,0.01, respectively, suggesting strong heterogeneity among the studies. So, these data were analyzed using a random-effects model. The meta-analysis revealed that the A allele was associated with OSAS risk with an OR of 2.44 (95% CI: 1. 11-5.38) in the dominant model (AA+AG vs. GG). The A allele increased the OSAS risk with an OR of 2.33 (95% CI: 1.48-3.66) compared with the G allele (Table 2, Figure 2D, F). The I 2 of the recessive model (AA vs. AG+GG) was 0 and the p value of the Q test was less than 0.01, so we used a fixed-effects model to analyze the data. As shown in Table 2 and Figure 2E, the A allele was associated with OSAS risk with an OR of 3.31 (95% CI: 2.48-4.40) in the recessive model (AA vs. AG+GG).
Of the long or short 5-HTTLPR, three studies were pooled for totals of 593 and 479 subjects in the OSAS and control groups [10,19,20], respectively. The I 2 of the dominant model (LL+SL vs. SS), recessive model (LL vs. SL+SS), and L allele vs. S allele were 0 and p values of the Q test were all more than 0.1. Meta-analysis with a fixed-effects model showed that the LL genotype increased the OSAS risk at an OR of 1.52 (95% CI: 1.07-2.16) when compared with a population with LS or SS genotypes. A single allele L increased the OSAS risk at an OR of 1.24 (95% CI: 1.04-1.49) compared with the allele S (Table 2, Figure 2G and I). The dominant model (LL+SL vs. SS) showed no significant association with OSAS risk (Table 2, Figure 2H).
The genotype distribution of 5-HTR2A 102C/T, 5-HTTLPR L/ S, and STin2VNTR 10/12 polymorphisms showed no significant difference between males and females (Table 3), while the AA genotype of 5-HTR 2A 1438G/A in males was significantly more common than in females. The AA genotype increased the OSAS risk with an OR of 4.21 (95% CI: 2.83-6.25) in a recessive model in male OSAS patients; however no significant association was found in females ( Figure 3, Table 3).

Publication Bias
Begg's test was used to evaluate publication bias in the literature. The shapes of the funnel plots revealed no obvious asymmetry (not shown). We then used Begg's test to assess statistical evidence of funnel plot symmetry, but no significant publication bias was indicated (Table 4).

Discussion
Although the genetic basis of OSAS is poorly understood, studies to date have focused on relationships between gene polymorphisms and susceptibility to OSAS [9][10][11][12][13][14]19,20,29]. Among genetic variants, polymorphisms in 5-HTR and 5-HTT have been investigated most frequently. Thus, the associations suggested from sporadic, inconsistent, small-sample-size studies needed to be further evaluated in a systematic review and metaanalysis. In this report, the combined number of samples for 5-HT2A 102C/T, 5-HT2A 1438G/A, 5HTTLPR, and STin2VNTR polymorphisms in cases and controls were 566, 466, 593, and 593 and 728, 551, 479, and 479, respectively. Our results revealed that the A allele of 5-HT2A 1438G/A, the long 5-HTTLPR, and 10 tandem repeats of STin2VNTR increased the OSAS risk and that 5-HT2A 102T/C was not associated with OSAS. All risk alleles displayed increased susceptibility to OSAS in both dominant and recessive models except the long 5-HTTLPR, which displayed no significant association with OSAS risk in a dominant model (LL+SL vs. SS); this may be explained by an insufficient sample size or unknown factors. Of the 5-HTR2A 1438G/A polymorphism, the AA genotype increased the OSAS risk with an OR of 4.21 (95% CI: 2.83-6.25) in a recessive model in male OSAS patients; however no significant association was found in the female population. Other gene polymorphisms, such as 5-HTR2A 1421C/T, 1678A/C, 24563A/C, 61472A/C, and 5-HTR2C 796C/G, failed in the meta-analysis due to the limitations of eligible studies or insufficient data. We attempted to categorize the data according to BMI, AHI, and gender to further evaluate the association between gene polymorphisms and OSAS. However, this was not possible due to the limited quantity of data available. Although race is a major factor in the distribution of gene polymorphisms, and our study data were from Japan [12], China [10,13,14,19,20], Turkey [9], and Brazil [11], we found no distribution differences in the gene polymorphisms evaluated. We could not conduct a haplotype analysis in this study because we  During the preparation of this manuscript, a systematic review and meta-analysis of the association between 5-HT2A receptor polymorphisms and risk of OSAS was published [30]. The differences between this report and our study are as follows: our study included not only 5-HTR gene polymorphisms but also 5HTTLPR and STin2VNTR gene polymorphisms, and our study first demonstrated that gender was an important cofactor for 1438G/A allele distribution; the AA genotype increased the OSAS risk with an OR of 4.21 (95% CI: 2.83-6.25) in a recessive model in male OSAS patients, but no significant association was found in females.
5-HT is associated with circadian rhythm and breathing regulation, and the 5-HTR2A and 5-HTR2C subtypes play an important role in the maintenance of upper airway stability and normal breathing in obesity [31]. 5-HTT plays an important role in serotoninergic transmission. Thus, studies concerning polymorphisms in the 5-HTR2A and 5-HTT gene would seem to be important for determining the genetic basis of OSAS. In this study, 5-HT2A 1438G/A, the long 5-HTTLPR, and the 10 tandem repeat variant of STin2VNTR were demonstrated to be risk alleles for OSAS. Further evidence from large-scale population studies and bio-functional assessments of these alleles are essential to confirm the associations between the genotypes and phenotypes.
This meta-analysis had a number of limitations. First, the pooled subjects in our meta-analysis were relatively small. Second, some studies with negative results may not have been reported, which will likely increase the publication bias in a meta-analysis. Despite these limitations, meta-analysis is a cost-effective and reasonable approach to evaluating sporadic, inconsistent, and small-sample-size studies.
In conclusion, we demonstrated by a meta-analysis that the 5-HTR2A 1438G/A, 5-HTTLPR, and STin2VNTR polymorphisms confer susceptibility to OSAS. The AA genotype increased the OSAS risk with an OR of 4.21 (95% CI: 2.83-6.25) in a recessive model in male OSAS patients, although no significant association was found in females. Further studies using larger samples, of better designs, and that include additional ethnicities are warranted. Additionally, the biological mechanism(s) and function(s) of the 5-HTR and 5-HTT genotypes and haplotypes in the pathogenesis of OSAS should be investigated further.

Supporting Information
Table S1 Characteristics summary of publications included in meta-analysis.