We appreciate the insightful comments of Shah regarding geometry and function for prognostic assessment in ischaemic cardiomyopathy.1 We agree with him about the prognostic importance of end-systolic volume (ESV). As Shah points out, in our investigation, there was a trend towards larger baseline ESV before revascularization in patients with cardiac death (P= 0.052).2 For the purposes of responding to Shah's query, we assessed the prognostic value of ESV following revascularization. Larger ESV on follow-up echocardiography was a univariate predictor (P= 0.003) of cardiac death. However, neither baseline nor follow-up ESV was an independent predictor of outcome on multivariate analysis. Wall motion scores, ejection fraction (EF) improvement, and digoxin use in follow-up remained as the only independent predictors of outcome.
While we recognize the prognostic importance of ventricular volumes, a smaller ESV has not always predicted improved outcome, most notably in the STICH trial where surgical reduction in ventricular volume did not improve outcome.3 Additionally, accurate, reproducible measurement of ESV utilizing two-dimensional echocardiography and the biplane method of discs is a challenging exercise. Jacobs et al.4 showed that the interobserver variability in the measurement of ESV was 24% compared with 14% for EF. Jacobs et al. also showed that the limits of agreement between two-dimensional echocardiography and magnetic resonance imaging for the determination of ESV was two times larger than that for EF. In our study, the interobserver variability for the measurement of ESV was 22% compared with 12% for EF (absolute difference of 3.8 EF points).
Shah importantly points out that improvement of ventricular function may occur late (>12 months) after revascularization in the hibernating myocardium presumably because of the extended time required to restore degraded myocardial contractile structures.5 One of the limitations of our study was that follow-up echocardiography was performed a mean of 4.8 months after revascularization, a time at which some patients with the hibernating myocardium may not have achieved functional recovery. Thirty-four patients in our study who had echocardiograms early in follow-up also had exams between 1 and 2 years following revascularization. In 7 of these 34 subjects (20%), significant EF improvement was observed at the time of the second follow-up exam, and in those individuals, EF data were used from the second follow-up exam to limit the problem of underestimating EF improvement. In the 34 subjects who had repeat follow-up studies, the mean change in EF from early to late follow-up was modest (34.5% early follow-up to 36.9% late follow-up).
Another perceived limitation of our study was that revascularization was incomplete in a quarter of patients. Since the goal of our study was to determine the prognostic impact of improvement in global ventricular systolic function (EF), rather than improvement in regional function, we did not exclude these patients.
We appreciate the careful reading of our study by Shah and the critique of our methods which highlights for readers of the journal the clinical value and challenges of serial assessment of left ventricular geometry and function.