Impact of amlodipine or ramipril treatment on left ventricular mass and carotid intima-media thickness in nondiabetic hemodialysis patients

This paper provides an overview of problems in multivariate modeling of epidemiologic data, and examines some proposed solutions. Special attention is given to the task of model selection, which involves selection of the model form, selection of the variables to enter the model, and selection of the form of these variables in the model. Several conclusions are drawn, among them: a) model and variable forms should be selected based on regression diagnostic procedures, in addition to goodness-of-fit tests; b) variable-selection algorithms in current packaged programs, such as conventional stepwise regression, can easily lead to invalid estimates and tests of effect; and c) variable selection is better approached by direct estimation of the degree of confounding produced by each variable than by significance-testing algorithms. As a general rule, before using a model to estimate effects, one should evaluate the assumptions implied by the model against both the data and prior information.

To test the predictive values of and independent contributions to cardiovascular and all-cause mortality of various arterial parameters exploring characteristics of the arterial wall at different sites, we studied prospectively 110 stable end-stage renal disease patients on hemodialysis. These parameters involved carotid diameter, carotid intima-media thickness, carotid compliance, carotid distensibility, carotid incremental elastic modulus, aortic diameter, aortic pulse wave velocity, and the presence of arterial calcifications measured at the sites of the carotid artery, abdominal aorta, iliofemoral axis, and legs. The presence of calcifications was analyzed semiquantitatively as a score (0 to 4) according to the number of arterial sites with calcifications. During a follow-up of 53+/-21 months (mean+/-SD), 25 cardiovascular and 14 noncardiovascular deaths occurred. In univariate analysis, the carotid incremental elastic modulus was the most closely related to prognosis. Risk of death increased with the number of vascular sites involved by calcifications. Moreover, information (in terms of prediction) given by carotid elastic incremental modulus was additive to the presence and extent of vascular calcification-related prediction value. Adjusted hazard ratios of all-cause and cardiovascular mortality for an increase of 1 unit in calcification score were 1.9 (95% confidence interval [CI], 1.4 to 2.6) and 2.6 (95% CI, 1.5 to 4.4), respectively (P<0.001 for both). Adjusted hazard ratios of all-cause and cardiovascular mortality for a 1-SD increase in carotid incremental elastic modulus were 1.6 (95% CI, 1.2 to 2.2) and 1.7 (95% CI, 1.2 to 2.4), respectively (P<0.01 for both). The results of this study showed that the presence and extent of vascular calcifications were strong predictors of cardiovascular and all-cause mortality. Carotid incremental elastic modulus gave additional predictive value.