Association of Plasma Homocysteine with Self-Reported Sleep Apnea Is Confounded by Age: Results from the National Health and Nutrition Examination Survey 2005-2006
Figure 3
Multivariate association of high plasma homocysteine and RBC folate levels with the risk of self-reported sleep apnea. Results are shown as point (diamonds) and 95% confidence intervals (error bars) for odds ratios estimated through a series of nested logistic regression commands. Twenty logistic regression models (indicated by # on the left) were run with varying combinations of covariates. The covariates included were G: male gender; R: non-Hispanic white race; U: birth in the United States; B: body mass index > 28 Kg/m2; E: high education; M: married; S: ever smoker; a: ever alcohol use; A: age > 46 years; H: hypertension; C: cardiovascular disease. Model 1 contained only high plasma homocysteine and high RBC folate as the independent variables. The results from models 19 and 20 cannot be directly compared with the remaining 18 models since the information for hypertension and cardiovascular disease was not available for a large number of subjects (shown under column titled ). Statistically significant associations (when the error bars did not straddle unity indicated by dashed vertical lines) are shown in red color for high plasma homocysteine and in blue color for high RBC folate. Statistically nonsignificant associations are shown in black color for plasma homocysteine and gray color for RBC folate. Statistical significance is shown on individual plots as color-coded values.
