Predictors and incremental prognostic value of left ventricular mechanical dyssynchrony response during stress-gated positron emission tomography in patients with ischemic cardiomyopathy.

ABSTRACT

OBJECTIVES: The study sought to investigate whether there is stress-induced worsening of left ventricular mechanical dyssynchrony (LVMD) in patients with ischemic cardiomyopathy (ICM), determine the predictors of LVMD response (LVMDR) and its incremental prognostic value. BACKGROUND: The effect of stress physiology on LVMD in patients with ICM has not been adequately evaluated, and the prognostic value of abnormal LVMDR beyond traditional predictors of mortality remains uncertain. METHODS: 489 consecutive patients with ICM, LV ejection fraction (EF) <35% undergoing rest/stress Rb-82 gated PET were evaluated. LVMD was determined by phase analysis (SD) from gated rest and peak stress images; LVMDR was defined as stress SD - rest SD, and stratified by tertiles. All-cause mortality was the primary outcome. Linear regression was performed to determine the predictors of LVMDR, and Cox proportional hazard modeling to assess its independent prognostic value. Integrated discrimination improvement (IDI) and net reclassification improvement (NRI) were performed to determine incremental value of abnormal LVMDR. RESULTS: Independent predictors of worse LVMD at peak stress were perfusion defect size (PDS) and resting heart rate; while higher resting LVEF, LVEF reserve and rest phase SD were associated with lower LVMDR. Over a mean follow-up of 2.0 ± 1.4 years, 123 patients (25%) died. After multivariate analysis, LVMDR was an independent predictor of all-cause mortality (HR 1.19[1.01;1.38], per 10° increase, P = .04) and reclassified 18% of patients with IDI 1.4% (P = .02) and NRI 9% (P = .057). CONCLUSION: In patients with ICM, an increase of LVMD during peak gating stress as compared to rest was an independent predictor of all-cause mortality, and had a modest incremental prognostic value. Future studies are needed to validate our findings.