Although snoring is common in the general population, it has been largely understudied from an individual genetic and lifestyle perspective. In particular, the clinical relevance of PRS for snoring has not been fully elucidated in Asian populations. In this study, we calculated the most recent PRS for snoring and (1) showed the difference explained by the PRS between UK Biobank participants of European ancestry and Korean participants of Asian ancestry, and (2) analyzed its relationship with lifestyle risk factors such as smoking and alcohol use, physical activity, and sleeping time features to investigate the degree to which these individual factors modify the inherited susceptibility to snoring.
An early cohort study on the genetic characteristics of snoring was conducted in a cardiovascular disease study cohort consisting of 3,387 men aged 54–74 [9] years. A total of 3,308 participants answered the survey, and habitual snoring strongly correlated with the family history of grandparents, parents, brothers, and children. The largest difference between the group that complained of habitual snoring and the control group was the family history of self-reported snoring. Another previous study reported genetic results on snoring (n = 408,000; snorers = 152,000) using data from the UK Biobank [8]. In total, 37% of all study subjects had snoring, and snorers had a higher rate of diagnosed sleep apnea than subjects without snoring (2.88% of snoring patients vs. 0.63% of controls). Snoring was correlated with age (OR = 1.011 [1.009–1.012]) and sex (OR males = 2.264 [2.212–2.316]) and showed a positive correlation with BMI, smoking, and alcohol intake frequency, and a negative correlation with socioeconomic status. They identified 42 significant genome-wide loci with an SNP-based heritability estimate of approximately 10% on the liability scale. Genetic correlations with body mass index, alcohol intake, smoking, schizophrenia, anorexia nervosa, and neuroticism were observed in the European population. Polygenic scores predicted recent snoring and probable obstructive sleep apnea in an independent Australian sample (n = 8000). A potential causal relationship between high BMI and snoring was suggested based on Mendelian randomization analysis results.
Several studies have compared PRSs by applying genetic analysis of specific phenotypes to other independent group data [19–24]. According to a previous study that assessed PRSs for coronary artery disease and type 2 diabetes as predictive factors for cardiovascular (CV) mortality [21], both CAD PRS (low vs. very high genetic risk groups, CAD PRS hazard ratio [HR] 2.61 [2.02–3.36]) and T2DM PRS (HR 2.08 [1.58–2.73]) were significantly correlated with CV mortality risk. These associations remained significant even after adjusting for a wide range of demographic and clinical characteristics. Adherence to an unhealthy lifestyle was also significantly linked to an elevated risk of CV mortality in the very high genetic risk group (favorable vs. unhealthy lifestyle with very high genetic risk for CAD PRS, HR 8.31 [5.12–13.49]; T2DM PRS, HR 5.84 [3.39–10.04]). In all genetic risk categories, the population attributable fraction (PAF) for CV mortality was 32.1% (95% CI 28.8–35.3%), while the PAF for smoking was 14.1% (95% CI 12.4–15.7%). Age, sex, and lifestyle factors did not significantly interact with PRSs predicting the risk of CV mortality.
Another study utilized a comprehensive health checkup database from the Korean population in conjunction with genotyping to generate PRS for BMI [24]. This study conducted a phenome-wide association (PheWAS) analysis, and a longitudinal association between BMI and PRS-BMI was observed. A model that predicted ten-year BMI based on age, sex, and baseline BMI was more accurate after including PRS-BMI (p = 0.003). Higher deciles of PRS were directly correlated with changes in BMI in a linear mixed model evaluating longitudinal changes in BMI with age. Significant correlations were found between PheWAS and metabolic syndrome, bone density, and fatty liver disease.
In our study, the genetic snoring risk score was calculated for 3526 snorers and 1939 non-snorers by applying the PRS based on the snoring GWAS results of a European study. Because of analyzing the various stages of significance showed, the GWAS p-value of 1.8e-08 cut-off showed the highest R2 (0.5403%). Still, it did not reach the explanatory power of the previous study. This means that the genetic explanatory power of the snoring GWAS study did not reach that of the snoring PRS calculated in the evaluation group of the same ethnic group.
Overall, the results of our study suggest that the odds ratio for snoring in the PRS high group was high. Thus, the effect of PRS can be interpreted as a genetic risk factor. Lifestyle variables interpreted as having genetic risk factors were alcohol consumption, sleeping late (derived by sleeping mid-time), and smoking. In the case of BMI, individuals with a low BMI avoid snoring, which seems to be due to genetic influences. Our results showed that the risk of snoring was high when exposed to risk factors such as PRS, sex, age, drinking experience, BMI, and sleep middle time. The risk factors that PRS influenced were male sex, older age, alcohol consumption, smoking, lower BMI, low physical activity, and late sleep mid-time.
To the best of our knowledge, this is the first study to investigate the associations between lifestyle habits and the genetic risk of snoring in the Korean population derived from European PRSs. To date, most large-scale genetic studies have been conducted in European populations. However, in the case of snoring, the effect size was very small, with an odds ratio of 0.99 to 1.01, as seen in the European GWAS results. Hence, a large-scale cohort study is required to develop other racial populations, including the Korean PRS model.
The limitations of this study are as follows: First, there was no Korean snoring GWAS data that could be applied to the PRS model in this study. Therefore, it was impossible to conduct a preliminary analysis to verify the difference in the basic genetic structure of the European group used as a reference and the Korean group. Second, explanatory power may have decreased because of the small sample size.