Abstract 433: Role And Importance Of Antecedent Blood Pressure In Predicting Outcomes In Cardiovascular Disease
Background: Blood pressure levels have been shown to be predictive of subclinical and clinical cardiovascular disease (CVD).There is little published information concerning the influence of antecedent blood pressure levels and HTN on CVD risk. Our hypothesis was that adding antecedent HTN to predictive model will increase model quality and consequently our ability to tailor treatment to patients needs.
Methods: We used data for 3344 participants from Framingham Heart Study Offspring who attended exam 3 and were followed up to 20 years for a first CVD event and analyzed prediction improvement reflected in changes in C-index, Net Reclassification Index (NRI) and the Integrated Discrimination Increment (IDI) values. Being exposed to serious critique in last 3 years, caused by improper overuse, this technique remains scientifically valid.
Antecedent BP is important independent predictor of CVD outcomes, surprisingly replacing current BP and BP treatment. Change in C-statistic was not statistically significant (p=0.19), which can be expected in saturated models with high c-index values. Both the IDI test (p=0.004) and the NRI test (p=0.046) indicate a significant improvement in the number of properly reclassified patients.
Discussion and conclusion: Using information from personal medical history to properly diagnose patients is well established, while using that information in epidemiology was neglected, almost forbidden. Our hypothesis that adding antecedent BP to predictive model will help to properly classify patients was confirmed, so in reverse of roles, epidemiology result confirms practice, underlying need to use information from personal medical history.
Author Disclosures: A. Bonifonte: None. T. Ayer: None. E. Veledar: None. A. Clark: None. P. Wilson: None.
- © 2014 by American Heart Association, Inc.