Author + information
- Jonathan C. Hsu, MD⁎ (, )
- Scott D. Solomon, MD,
- Scott McNitt, MS,
- Arthur J. Moss, MD and
- Elyse Foster, MD
- ↵⁎Division of Cardiology, Electrophysiology Section, University of California, San Francisco, 500 Parnassus Avenue, MU-434, Box 1354, San Francisco, California 94143
We appreciate the interest of Dr. Lim in our study of super-response to cardiac resynchronization therapy involving patients enrolled in the MADIT-CRT trial (Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy) (1). We thank him for his thoughtful review of our paper and his inquiry involving overall mortality in our study.
Dr. Lim is correct that 74 of 1,089 patients (6.8%) randomized to CRT died during an average follow-up of 29 months in the original MADIT-CRT trial (2). In our sub-analysis (1), there were 25 deaths in the subgroup of 752 patients (3.3%) studied over a median follow-up of 15.2 months. However, Dr. Lim does not have access to the primary data when he calculated crude mortality rates to draw comparisons between hypo-responders from our study versus implantable cardioverter-defibrillator (ICD) recipients from the original MADIT-CRT trial. To clarify, in the ICD-only trial arm, 30 of 623 patients (4.8%) with left ventricular ejection fraction (LVEF) measurements at both baseline and 12-month echocardiograms subsequently died, compared with 25 of 752 patients (3.3%) in our study of CRT-D recipients. Therefore, to compare a crude mortality risk of 4.8% in the ICD arm (the relevant comparator group) and 6.3% in the CRT-D hypo-responder group and claim that hypo-responders had “almost double the annualized mortality rate” is incorrect and statistically unsound.
We do not agree with the statements of Dr. Lim that our study “underestimated the adverse effects of CRT in ‘hypo-responders'” or that the 49 deaths not accounted for in our cohort infers a higher proportion/mortality rate of “hypo-responders.” We would like to clarify any misunderstandings from the inferences of Dr. Lim. Of the other 49 deaths in the CRT-D arm not included in our analysis, 23 of 337 excluded patients (6.8%) died during the first year of follow-up, and 26 of 337 excluded patients (7.7%) died after the first year. By definition, the 337 excluded patients had higher rates of death, because some patients died before or otherwise did not meet the inclusion criteria of our cohort of both baseline and 12-month follow-up echocardiograms. These data indicate that patients who had not died but missed their 12-month echocardiogram might be at greater risk of subsequent poor outcomes, perhaps from factors associated with incomplete follow-up (e.g., illness). Although selection bias might have been introduced, we found no other feasible way to perform an analysis with LVEF change, because serial LVEF measurements were conditional on having survived to 1 year. We acknowledged this potential limitation in our paper (1).
Finally, our study identified characteristics including left bundle branch block and longer QRS duration (≥150 ms) associated with super-response to CRT-D therapy, and thus we agree with Dr. Lim that these findings highlight the possibility that patients with the converse might not derive similar benefits, despite incurring the risk and costs of this therapy. However, to recommend withholding CRT-D therapy in a subgroup of otherwise eligible patients without prospective studies to support this practice might be ill-advised. As a result, we believe that the hypotheses generated by these observations should be tested in future studies.
- American College of Cardiology Foundation
- Hsu J.C.,
- Solomon S.D.,
- Bourgoun M.,
- et al.