Statistical analysis
The standardized mean difference (SMD) in Table 1 provides a quantitative assessment of the imbalance between the study and control groups. In general, if the SMD value is less than 0.1, no imbalance exists between the two groups. To assess the relationship between OSA and incident BVFL, Cox proportional hazard regression analysis were used. Unadjusted hazard ratios (HR) were not adjusted in Table 2, but adjusted HR were adjusted for eight clinical factors. Further 10000-person year (PYs) is defined as the number of new disease cases in one year in the population at risk of disease occurrence.