Ulm, Kurt; Barthel, Petra; Rolnitzky, Linda; Schneider, Raphael; Malik, M.; Bigger Jr., J. Thomas; Schmidt, Georg
Risk Stratification in Post-MI Patients Based on Left Ventricular Ejection Fraction and Heart-Rate Turbulence.
Sonderforschungsbereich 386, Discussion Paper 365
Objectives: Development of risk stratification criteria for predicting mortality in post-infarction patients taking into account LVEF and heart-rate turbulence (HRT). Methods: Based on previous results the two parameters LVEF (continuously) and turbulence slope (TS) as an indicator of the HRT were combined for risk stratification. The method has been applied within two independent data sets (the MPIP-trial and the EMIAT-study). Results: The criteria were defined in order to match the outcome of applying LVEF ( 30 % in sensitivity. In the MPIP trial the optimal criteria selected are TS normal and LVEF ( 21 % or TS abnormal and LVEF ( 40 %. Within the placebo group of the EMIAT-study the corresponding criteria are: TS normal and LVEF ( 23 % or TS abnormal and LVEF ( 40 %. Combining both studies the following criteria could be obtained: TS normal and LVEF ( 20 % or TS abnormal and LVEF ( 40 %. In the MPIP study 83 out of the 581 patients (= 14.3 %) are fulfilling these criteria. Within this group 30 patients have died during the follow-up. In the EMIAT-trial 218 out of the 591 patients (= 37.9 %) are classified as high risk patients with 53 deaths. Combining both studies the high risk group contains 301 patients with 83 deaths (ppv = 27.7 %). Using the MADIT-criterion as classification rule (LVEF ( 30 %) a sample of 375 patients with 85 deaths (ppv = 24 %) can be selected. Conclusions: The stratification rule based on LVEF and TS is able to select high risk patients suitable for implanting an ICD. The rule performs better than the classical one with LVEF alone. The high risk group applying the new criteria is smaller with about the same number of deaths and therefor with a higher positive predictive value. The classification criteria have been validated within a bootstrap study with 100 replications. In all samples the rule based on TS and LVEF (= NEW) was superior to LVEV alone, the high risk group has been smaller (( s: 301 ( 14.5 (NEW) vs. 375 ( 14.5 (LVEF)) and the positive predictive value was larger (( s: 27.2 ( 2.6 % (NEW) vs. 23.3 ( 2.2 % (LVEF)). The new criteria are less expensive due to a reduced number of high risk patients selected.