DISCUSSION
This study demonstrates that a considerable fluctuation of EF occurs among patients with EF ≤ 35%, changing their ICD eligibility status over time. After one year, almost 40% of patients with EF ≤35% experienced a >10% improvement in EF. But over the next 1.4 years, approximately 25% of those who had an improvement in EF lost their gains returning back to EF ≤35%. While the fluctuation in EF was present in both ischemic and non-ischemic cardiomyopathy, a larger proportion of those with non-ischemic cardiomyopathy had an improvement in EF and a lower proportion had subsequent reduction. These results may explain the root cause of continued SCD risk after EF improvement.
Several previous studies have shown that 30-40% of patients with ICD experience a substantial improvement in EF post-implantation. Our results support these previous observations.2–4Presumed to be from continued medical therapy, improvement in EF was associated with a lower incidence of appropriate ICD therapy in comparison to those whose EF did not change.1,3–6Similar observations were also made among patients with cardiac resynchronization therapy, suggesting that EF improvement is associated with better outcomes regardless of its etiology.11These studies raised the question of whether ICD generator should be replaced at the end of battery life among patients with EF improvement.
Recently, in a post hoc analysis of SCD-HeFT we demonstrated that patients with a substantial improvement in EF appeared to have a similarsurvival benefit from ICD in comparison to those who did not demonstrate EF improvement.7 To date, this is the only comparison of survival in relation to ICD among patients with EF improvement. The results from the present investigation suggest that recurrence of cardiomyopathy among patients with recovered EF may be the culprit for the continued benefit of ICD among these patients. Cumulatively these results support ICD generator replacement at the end of battery life regardless of EF.
Recently, Lupón et al. examined the prognostic impact of the dynamic changes in EF over a 15-year follow-up in a prospective, consecutive, observational registry of outpatient heart failure patients.12 They found that the majority of patients had a marked rise in EF during the first year, maintained up to a decade, followed by a slow EF decline thereafter. Our findings add to this literature by showing that fluctuation in EF around the cut-off 35% may influence clinical decisions on ICD therapy. The present data also showed that while the mean EF was stable between the 1st and 2nd follow-up (31.7% vs. 32.6, respectively). there remained significant fluctuation in ~25% of the cohort, which had potential treatment implications for those patients.
Patients with non-ischemic cardiomyopathy were more likely to experience an improvement in EF in this and previous cohorts.2,3,12 Absence of large myocardial scar volume may be responsible from this observation.12,13 Lupón and colleagues showed that while ischemic heart failure patients showed a modest EF increase during the first year of medical therapy those with non-ischemic heart failure had a more pronounced improvement.12 Further, those with non-ischemic heart failure had a more prolonged increase during follow-up after one year than ischemic heart failure patients.12 Collectively, these results may help us understand the lack of survival improvement with ICD therapy among patients with non-ischemic cardiomyopathy.14,15
While medical therapy for heart failure is responsible for improvement in EF, less is known about factors that lead to re-worsening of EF after an improvement. Recently, the TRED-HF study examined whether patients with dilated cardiomyopathy could discontinue heart failure medications after recovery of cardiac function, which was defined as being asymptomatic, EF increasing from < 40% to > 50%, normalization of left ventricular end diastolic volume, and decreasing of N-terminal pro-B-type natriuretic peptide to < 250 ng/L.16 These researchers found that cardiomyopathy relapsed in 44% of the patients following withdrawal of medical therapy versus in no patients who continued heart failure medications, suggesting that heart failure treatment should be continued indefinitely.16 Our analysis adds to these data by showing that EF trajectory may be volatile even with continued medical therapy.
Patients with normalization of EF constitute a special subgroup. Based on limited data the risk of appropriate ICD therapy is very low in these patients.4,11 However, in a small prospective study, Cioffi et al. showed EF normalization to be a transient finding in 55% of their patients, with the effect lasting a mean duration of 15 ± 5 months.17 In our study, only 3% of patients experienced a rise in EF to >55%. Similar to the Cioffi study, 56% of our patients had a subsequent drop in EF of > 10%. This suggests that normalization of EF is not equivalent to a cure of HF, but rather remission as a significant proportion of these patients remain at risk for relapse and SCD.18–20