PURPOSE: The vitamin K antagonist (VKA) warfarin is effective for the prevention of thromboembolisms. Maintenance doses differ greatly among patients and are known to be primarily determined by genetic polymorphisms. The relative impact of dietary vitamin K intake is still a matter of debate. We hypothesize that a multivariate model is more suitable for exploring the relation between dietary intake of vitamin K and warfarin dose than conventional uni- or bivariate analyses. METHODS: In a cross-sectional study, we interviewed 244 patients in the maintenance phase of warfarin therapy and detected polymorphisms in the VKORC1 and CYP2C9 genes. Dietary vitamin K intake was estimated from food frequency questionnaires. RESULTS: A univariate correlation analysis and the regression coefficient from the multivariate model showed a small but significant negative relation between vitamin K intake and warfarin dose. A loading plot of the partial least squares regression model illustrated this counter-intuitive observation, which might be explained by the latent structure between variables. The variation in warfarin dose could be divided into two significant latent variables, the so-called components. In component one, pharmacogenetics explained 52% of dose variation. Component two described health-related behavior (diet, physical activity and body weight) and explained 8% of dose variation. Here, vitamin K intake positively correlated with warfarin dose. DISCUSSION: This study highlights the importance of choosing a statistical method that reflects the complexity of data for interpretation of results from observational studies. The multivariate model appears to be well suited to describe the complex relationship between vitamin K intake and VKA dose.
PURPOSE: The vitamin K antagonist (VKA) warfarin is effective for the prevention of thromboembolisms. Maintenance doses differ greatly among patients and are known to be primarily determined by genetic polymorphisms. The relative impact of dietary vitamin K intake is still a matter of debate. We hypothesize that a multivariate model is more suitable for exploring the relation between dietary intake of vitamin K and warfarin dose than conventional uni- or bivariate analyses. METHODS: In a cross-sectional study, we interviewed 244 patients in the maintenance phase of warfarin therapy and detected polymorphisms in the VKORC1 and CYP2C9 genes. Dietary vitamin K intake was estimated from food frequency questionnaires. RESULTS: A univariate correlation analysis and the regression coefficient from the multivariate model showed a small but significant negative relation between vitamin K intake and warfarin dose. A loading plot of the partial least squares regression model illustrated this counter-intuitive observation, which might be explained by the latent structure between variables. The variation in warfarin dose could be divided into two significant latent variables, the so-called components. In component one, pharmacogenetics explained 52% of dose variation. Component two described health-related behavior (diet, physical activity and body weight) and explained 8% of dose variation. Here, vitamin K intake positively correlated with warfarin dose. DISCUSSION: This study highlights the importance of choosing a statistical method that reflects the complexity of data for interpretation of results from observational studies. The multivariate model appears to be well suited to describe the complex relationship between vitamin K intake and VKA dose.
Authors: Majid Moridani; Lei Fu; Rita Selby; Francisco Yun; Tatjana Sukovic; Betty Wong; David E C Cole Journal: Clin Biochem Date: 2006-04-21 Impact factor: 3.281
Authors: T E Klein; R B Altman; N Eriksson; B F Gage; S E Kimmel; M-T M Lee; N A Limdi; D Page; D M Roden; M J Wagner; M D Caldwell; J A Johnson Journal: N Engl J Med Date: 2009-02-19 Impact factor: 91.245
Authors: Erik Fung; Nikolaos A Patsopoulos; Steven M Belknap; Daniel J O'Rourke; John F Robb; Jeffrey L Anderson; Nicholas W Shworak; Jason H Moore Journal: Semin Thromb Hemost Date: 2012-10-06 Impact factor: 4.180