Study Summary

The RECOVERY trial was a multicenter, randomized, parallel group, open-label study designed to compare medical management with early surgery in patients with asymptomatic very severe aortic stenosis (AS; valve area ≤0.75 cm2, and peak velocity ≥4.5 m/s or mean transvalvular gradient ≥ 50 mmHg). Patients were excluded if they were symptomatic or had an ejection fraction <50%. Of the >1,000 patients screened, 144 were randomly assigned 1:1 to early surgery (within 2 months of randomization) or guideline-directed medical management. The primary endpoint was a composite of surgical or cardiovascular mortality. Data were analyzed on an intention-to-treat basis, and median follow-up was >6 years for each group.

The mean age of patients was 64.2 years. The primary outcome occurred in 15% of the medically managed group and only 1% of the early surgery group (HR 0.09, 95% CI 0.01–0.67, P = 0.003). No early operative deaths occurred. The incidence of the secondary outcome — death from any cause — was also statistically lower in the early surgery group (7% vs 21%; HR 0.33, 95% CI 0.12–0.90). In the medical management group, the incidence of sudden cardiac death was 4% at 4 years and 14% at 8 years.


Study Strengths: This was a well-designed trial. Despite the small patient population, the observed effect size was large. Patients in the early surgery group received surgery quickly (median 23 days from randomization). However, 74% of those in the medical care group went on to receive late surgery (median 700 days from randomization). The primary endpoint of operative or cardiovascular mortality is not effected by the unblinded trial design.

Study Limitations: The study was performed in South Korea, and the patient population may be dissimilar to the rest of the world. For example, patients in this trial were relatively young and the most common cause of AS was a bicuspid valve (54% of medically-managed patients and 67% of those in the surgical group). Only 24 patients underwent stress testing, which could have led to overestimation of the number of asymptomatic patients. Transcatheter aortic valve replacements (TAVR) were not included in this trial, which could have lowered the periprocedural risk further and increased the effect size. Only included patients with ‘very severe’ AS were included. Medically managed patients with asymptomatic ‘severe’ AS (peak velocity, 4.0–4.5 m/s) would ostensibly have lower mortality than the patients in this trial.

Next Steps/Clinical Perspective: The RECOVERY trial is an important study that challenges current practice, particularly regarding the incidence of sudden death. The study shows that, even in patients with asymptomatic AS, early valve replacement should be considered. The trial should be repeated with a more generalizable population, including patients with ‘severe’, rather than just ‘very severe’ AS, and with the inclusion of TAVR. Practitioners should conduct carefully monitored stress testing to discern whether patients are truly asymptomatic. This perhaps represents one of the few circumstances that a stress test should be considered in “asymptomatic” patients - not only to evaluate valvular hemodynamics under stress, but to ascertain a patient’s functional status, who may be unconsciously diminishing their behavior due to worsening valvular pathology.

Trial Reference

Kang,  D.-H. et al. Early surgery or conservative care for asymptomatic aortic stenosis. N. Engl. J. Med. doi: 10.1056/NEJMoa1912846.
[PubMed: 31733181]