118 Chapter 4 * sex) in all models. From the resulting model, regression lines per subgroup could be obtained following the standard regression equation for a linear model: y= a + β*X, with the intercept (a) and slope (β) specified for each subgroup. The slopes (β) of these regression equations were compared between the patient groups to illustrate the estimated difference in the evolution of a specific echocardiographic parameter over a 10-year period. To study which of the abnormalities in the echocardiography features occur when (early or late in disease evolution), the differences in absolute values of the echocardiographic parameters between FD patients and the healthy control subjects per age decade were assessed by a Wilcoxon rank test. For these comparisons, an adjusted p-value using the Bonferroni method was displayed to correct for multiple testing. For the comparison of these absolute values, only the last available echocardiogram per patient was selected to ensure that the influence of repeated measurements in single individuals was absent. Values of the five main echocardiographic parameters on the first echocardiogram of each patient were assessed for their potential relation to development of AF during the follow-up period using a Cox proportional-hazards model, that corrected for unequal follow-up duration for each patient. To determine the relation of the echocardiographic changes in FD patients to other markers of myocardial injury and dysfunction we assessed the relation between three comprehensive echocardiographic features (LVMi, E/e’ and GLS) and the levels of cardiac enzymes in plasma (NT- proBNP and hs-TnT), using a Spearman correlation analysis. We regarded a p-value ≤0.05 as statistically significant in all analyses. For the following echocardiographic variables only descriptive statistics were given: LV EF, TAPSE (too much nonlinear variation in measurements of FD patients for a valid GLM approach), TR Vmax, SoV (data only available for FD patients) and GLS (different measurement algorithms in FD patients vs healthy controls, which makes the comparison between these two groups invalid).
RkJQdWJsaXNoZXIy MTk4NDMw