In this issue of EuroIntervention, Witberg et al1 analyse data from the AMTRAC Registry to examine the incidence, predictors and outcomes of left ventricular (LV) recovery post-transcatheter aortic valve implantation (TAVI). Of 10,872 TAVI patients included in this 17-centre registry from Europe and Israel, 914 had severe LV dysfunction prior to TAVI (LV ejection fraction [LVEF] ≤30%). LVEF recovered by ≥10% in 59.5% of patients and normalised to ≥50% in approximately one-quarter of patients. Compared to patients with a baseline LVEF >30%, no LV recovery was associated with increased mortality, LV recovery with similar mortality, and LV normalisation with lower mortality at 3 years. The miraculous nature of these results deserves further examination.
While it may seem obvious that unloading the ventricle with an aortic intervention would lead to early imprÂoÂvÂeÂÂÂÂÂments in LV systolic function, these findings should not be taken for granted. For example, it also seems obvious that revascularisation of multivessel coronary artery disease (CAD) should improve LV systolic function in patients with ischaemic cardiomyopathy. Yet, revascularisation with either percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) was not associated with increased LV recovery for patients with extensive CAD and an EF <35% in 2 landmark trials, REVIVED-BCIS2 for PCI2 and STICH for CABG3.
Unlike the failure of coronary revascularisation to improve LV function, TAVI leads to significant LV recovery (improvement in LVEF ≥10%) for 32.8-62.2% of patients across multiple TAVI trial-based analyses456, including the surgical aortic valve replacement (SAVR) cohort in PARTNER Cohort A4 (Table 1). The current AMTRAC Registry is notable for being the first analysis demonstrating LV recovery in patients with severe cardiomyopathy only (LVEF ≤30%) and thus reassures clinicians about the value of TAVI in this high-risk group. LV recovery is associated with improved survival in all these analyses, and its impact can be inferred on pathophysiology beyond the heart. LV recovery likely leads to improvements in cardiac output post-TAVI; thus, the observation that patients in Northern New England were 3 times more likely to experience acute kidney recovery than injury after either SAVR or TAVI may be an indirect reflection of improved cardiac function7.
Witberg et al identify 4 key predictors of LV recovery after TAVI, including absence of prior myocardial infarction and high aortic valve gradients (>40 mmHg)1. These predictors are similar to those identified in prior analyses and emphasise that LV recovery may not occur for all patients after TAVI; a common predictor of LV recovery is the absence of prior myocardial infarction1456, suggesting that recovery requires a ventricle with minimal post-infarct fibrosis. The utility of imaging studies in assessing pre-TAVI myocardial fibrosis burden should be evaluated prospectively to improve patient selection and prognosis.
The strength of the AMTRAC Registry is the inclusion of patients that are poorly represented in prior analyses, i.e., those with an LVEF <30%. But, there are important limitations to this study. First, the lack of uniformity in post-TAVI echocardiogram timing complicates comparison with other studies (with standard trial-mandated 30-day echocardiogram) and potentially biases survival analysis – the 30-day landmark analysis may be insufficient, as the mean time-to-echocardiogram was 34 days. Second, the lack of serial echocardiograms means that this analysis cannot confirm early (prior to discharge)456 recovery and will miss some undefined percentage of patients with late recovery5. Finally, the exclusion of 112 patients for lack of post-TAVI LVEF data creates further ambiguity in the survival analysis. Despite these limitations, Witberg et al should be congratulated on furthering our understanding of the miracle of post-TAVI LV recovery; the exclusive focus on AS patients with an LVEF ≤30% extends and confirms this remarkable observation to an especially high-risk and poorly studied population.
Table 1. Recent studies on left ventricular recovery after TAVI.
| Study registry Authors (Publication year) | Patient inclusion | LV recovery (EF increase ≥10%) Patients, % | Timing of LV recovery assessment | Predictors of LV recoveryHR (95% CI) |
|---|---|---|---|---|
| PARTNER Cohort AElmariah et al4(2013) | N=108(EF <50%)Mean EF 37.1% | 51.6 | 30 days post-TAVI,echo timing per trial protocol | Prior PPM: 0.34 (0.15-0.77)Baseline EF: 0.91 (0.86-0.95)Mean AVG (per mmHg): 1.03 (1.01-1.06) |
| Medtronic CoreValve U.S. Pivotal TrialDauerman et al5 (2016) | N=156(EF ≤40%)Mean EF 32% | 62.2 | 30 days post-TAVI,echo timing per trial protocol | Prior MI: 0.44 (0.19-1.03)Mean AVG >40 mmHg: 4.59 (1.76-11.96) |
| PARTNER 1, 2 and S3Kolte et al6(2022) | N=659(EF <50%)Mean EF 37.8% | 32.8 | 30 days post-TAVI,echo timing per trial protocol | AVA: 0.19 (0.05-0.73)Cancer: 0.56 (0.37-0.86)LVEDD: 0.59 (0.44-0.78)Diabetes: 0.61 (0.4-0.92)Prior MI: 0.65 (0.42-0.98)Baseline EF: 0.93 (0.9-0.95)SVI: 1.03 (1-1.06)BMI: 1.06 (1.02-1.1) |
| AMTRAC RegistryWitberg et al1(2023) | N=914(EF ≤30%)Mean EF 27.3% | 59.5 | 0-60 days post-TAVI,echo timing per site practice: median 34 days (IQR 10-45 days) | Prior MI: 0.45 (0.28-0.71)GFR <60 mL/min: 0.49 (0.32-0.77)LF-LG AS: 0.50 (0.29-0.84)Mean AVG (per mmHg): 1.02 (1.01-1.04) |
| AVA: aortic valve area; AVG: aortic valve gradient; BMI: body mass index; CI: confidence interval; EF: ejection fraction; GFR: glomerular filtration rate; HR: hazard ratio; IQR: interquartile range; LF-LG AS: low-flow, low-gradient aortic stenosis; LV: left ventricular; LVEDD: left ventricular end-diastolic diameter; MI: myocardial infarction; PPM: permanent pacemaker; SVI: stroke volume index; TAVI: transcatheter aortic valve implantation | ||||
Conflict of interest statement
H.L. Dauerman is a consultant for Medtronic and Boston Scientific; and has research grants from Boston Scientific and Medtronic. R. Lahoud has research grants from Medtronic.
