Obstructive sleep apnea (OSA) is a prevalent sleep disorder characterized by the recurrent partial or complete obstruction and collapse of the upper airway during sleep, leading to episodes of apneas and hypoventilation. 1 2 Research studies have reported that the prevalence of OSA in the adult population ranges from 9 to 38%, with a higher prevalence observed in males (13–33%) compared to females (6–19%). Moreover, the prevalence of OSA tends to increase with age and is closely associated with the prevalence of obesity. 3 4

There is mounting evidence indicating that OSA serves as an independent risk factor for several cardiovascular diseases, including hypertension, 5 stroke, 6 pulmonary hypertension, 7 and heart failure. 8 Venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolism (PE), is recognized as the third most common cardiovascular disease worldwide. 9 There is evidence suggesting that OSA may also be linked to an increased risk of VTE. 10 For instance, a prospective study involving 15,664 subjects (1,424 subjects with OSA) observed a twofold higher incidence of VTE in patients with OSA compared to non-OSA patients. 11 Similarly, findings from a national retrospective cohort study conducted by Peng and his colleagues indicated that patients with OSA had a 3.50-fold higher risk of DVT and a 3.97-fold higher risk of PE compared to the general population. 12 However, the results of observational studies remain somewhat controversial. A 5-year prospective study involving 2,109 subjects concluded that OSA did not increase the risk of VTE recurrence. 13 Another retrospective analysis involving 1,584 patients, of which 848 were women, revealed an intriguing discovery suggesting that OSA may serve as an independent risk factor for VTE solely in women, rather than in men. 14 Moreover, patients with VTE were found to have a higher prevalence of OSA, 15 suggesting a potential bidirectional relationship.

Although previous observational studies have investigated the potential association between OSA and VTE, elucidating aspects of the association from these studies is challenging due to the limitations of potential confounders and reverse causality bias. Mendelian randomization (MR) is a genetic epidemiological methodology that utilizes genetic variants, such as single-nucleotide polymorphisms (SNPs), as instrumental variables (IVs) to infer the genetic association between exposure and outcome. 16 The advantage of MR analysis lies in the random assignment of genetic variants during meiosis, which effectively circumvents the effects of potential confounders and reverse causality encountered in classical epidemiologic studies. 17

At present, the nature of the association between OSA and VTE remains inconclusive, and there is a dearth of pertinent studies comprehensively exploring the genetic association between OSA and VTE. Therefore, this study aimed to conduct a bidirectional two-sample MR analysis using publicly available summary statistics from large-scale genome-wide association studies (GWAS) to genetically assess the exact association between OSA and VTE, including PE and DVT.

In this study, we conducted a comprehensive two-sample MR analysis to explore the genetic association between OSA and VTE. Our MR findings did not yield evidence of a significant association between OSA and VTE from a genetic standpoint.

Our findings contradict some previous observational studies suggesting a link between susceptibility to OSA and an increased risk of VTE. 29 30 31 32

However, these studies were hindered by inadequate consideration of confounding factors, particularly obesity, along with methodological flaws and small sample sizes. Obesity is widely recognized as a significant risk factor for both OSA 33 and VTE. 34 Therefore, it is crucial not to overlook the impact of obesity in striving for a deeper understanding of the potential association between OSA and VTE. Notably, a cohort study involving 31,309 subjects indicated a higher likelihood of VTE development among patients with more severe OSA. Yet, this association disappeared upon adjusting for confounders, notably obesity levels. 35 Thus, it is plausible that the observed association between OSA and VTE could be attributed to obesity confounding. Additionally, Aman and his colleagues' report yielded consistent results, suggesting that OSA does not elevate the risk of VTE after adjusting for obesity confounding. 36

MR is a robust analytical method that employs genetic variation as IVs to deduce the genetic association between exposure and outcome. Consequently, it effectively controls for confounders induced by environmental factors and mitigates reverse causality bias. In this study, we meticulously screened genetic variants and thoroughly accounted for the effects of obesity levels to procure reliable IVs for inferring the genetic association between OSA and VTE. To mitigate bias and enhance the reliability of our MR findings, we devised initial and validation MR analyses supplemented by a series of sensitivity analyses, drawing upon datasets sourced from various origins. Notably, neither MR analysis provided evidence supporting a genetic association between OSA and VTE. Moreover, a succession of sensitivity analyses served to bolster the robustness of our MR results. These findings indicate that, although diverging from some previous observational studies, our results are reliable and corroborate the conclusions drawn from the MR study.

While our MR study did not find evidence supporting a genetic association between OSA and VTE, it remains possible that OSA could influence the onset or progression of VTE. Virchow's triad depicts three major factors inducing VTE: endothelial injury, venous stasis, and hypercoagulability. 37 The pathophysiologic mechanism linking OSA and VTE remains unknown but may be associated with OSA's capacity to affect the three classical mechanistic pathways of Virchow's triad. 38 Intermittent hypoxia, a signature feature of OSA, can induce oxidative stress and activate inflammatory markers, further damaging the vascular endothelium. 39 40 OSA-associated hemodynamic alterations and reduced physical activities may result in venous stasis. 41 A growing number of studies have demonstrated a strong correlation between OSA and hypercoagulability. A retrospective cohort study aimed at assessing coagulation in patients with OSA suggested that patients with moderate to severe OSA experienced elevated markers of blood coagulability, primarily evidenced by shortened prothrombin time, compared to healthy individuals. 42 Two additional studies of thrombotic parameters found that patients with OSA possessed higher levels of the thrombin–antithrombin complex. 43 44 Furthermore, several coagulation factors, such as fibrinogen, coagulation factor VII, coagulation factor XII, and vascular hemophilic factor, which play a crucial role in the coagulation process, are elevated in patients with OSA. 45 Collectively, this evidence supports that patients with OSA are in a state of hypercoagulability, facilitating our understanding of the underlying pathophysiologic mechanisms between OSA and VTE. Considering these potential mechanisms, future large-scale studies are necessary to thoroughly explore the potential association between OSA and VTE, delving into greater depth.

The greatest strength of this study is that the bidirectional two-sample MR analysis designed based on summary data from large-scale GWAS was used for the first time to investigate the genetic association between OSA and VTE. Furthermore, to bolster the robustness of the findings and mitigate bias, we conducted initial and validated MR analyses using two independent OSA GWAS datasets. Subsequently, a series of sensitivity analyses provided further validation and affirmed the robustness of the results. However, our study also has several limitations. First, it was exclusively centered on European individuals, thereby constraining the generalizability of our findings to other ethnicities or ancestries. Second, the lack of individual-level data in the summary-level statistics prevented us from stratifying the study population by important factors such as age or sex. Lastly, there is a possibility of sample overlap between the exposure and outcome datasets, but the F-statistics of the IVs selected in the MR analysis were sufficiently strong to mitigate the potential effects of weak instrumental bias.

In conclusion, our MR study did not uncover genetic evidence supporting an association between OSA and VTE, including DVT and PE. This implies that the association between OSA and VTE reported in some previous observational studies may rely on alternative pathways to function, rather than being directly linked to the diseases themselves.