Article Text
Abstract
Objectives The timing of aortic valve replacement (AVR) remains controversial in symptomatic patients with normal flow, low-gradient severe aortic stenosis (AS) and preserved LVEF. We sought to compare long-term mortality of early AVR versus a watchful observation strategy.
Methods From 2000 to 2011, we prospectively evaluated 284 consecutive symptomatic patients (136 men, age 68±10 years) with normal flow, low-gradient severe AS and preserved LVEF who were potential candidates for early AVR. Normal flow, low-gradient severe AS was defined as indexed aortic valve area <0.6 cm2/m2 with mean gradient <40 mm Hg and stroke volume index ≥35 mL/m2. Early AVR was performed on 98 patients (early AVR group), while the watchful observation strategy was selected for 186 patients (watchful observation group). Patients in the watchful observation group were referred for AVR if mean gradient was ≥40 mm Hg during follow-up.
Results There were no significant differences between the early AVR and the watchful observation groups for the risk of overall mortality (HR 0.94 for the early AVR; 95% CI 0.51 to 1.73) or for the estimated actuarial 8-year mortality rates (17±5% vs 27±5%, p=0.84) in the overall cohort. Society of Thoracic Surgeons score, comorbidity index, age, coronary artery disease, aetiology of AS and performance of AVR were associated with overall survival. For 83 propensity-score-matched pairs, the risk of overall death was not significantly different between the two groups (HR 1.13 for the early AVR, 95% CI 0.55 to 2.35, p=0.74).
Conclusions Early AVR and watchful observation strategy show similar survival in symptomatic patients with normal flow, low-gradient severe AS and preserved LVEF. Watchful observation with timely performance of AVR should be considered a therapeutic option.
Statistics from Altmetric.com
Introduction
Although the echocardiographic grading of aortic stenosis (AS) should combine the assessment of aortic valve area (AVA) with a mean gradient or aortic jet velocity,1–3 one-fourth to one-third of the patients diagnosed with severe AS based on AVA criteria have a mean gradient below 40 mm Hg,4 ,5 and it remains a clinical conundrum whether such patients should be classified into severe or moderate AS. Low-gradient severe AS with preserved LVEF can arise from measurement errors, inconsistencies between cut-off values in current guidelines and paradoxical low flow (LF).6 The timing of aortic valve replacement (AVR) had not been reported in previous clinical studies,5 ,7–9 and it is not known whether early AVR performed at the time of diagnosis could decrease mortalities related to low-gradient severe AS.6 Watchful waiting and timely performance of AVR may be a treatment option for symptomatic patients with normal flow (NF), low-gradient severe AS.10 Using our prospectively collected registry data on symptomatic patients with NF, low-gradient severe AS and preserved LVEF, we sought to compare long-term mortality of early AVR versus a watchful observation strategy.
Methods
Study population
A prospective registry has included all consecutive patients with severe AS undergoing echocardiography at our hospitals. Case report forms, including patient demographics, clinical presentation and echocardiographic data, were stored in an electronic database.11 Clinical and echocardiographic follow-up data of study patients were collected annually and entered into the database. Of the 2195 patients diagnosed with severe AS based on severe restriction of systolic opening and AVA <1.0 cm2 from 2000 to 2011, 394 patients with concomitant valve disease, 1037 patients with aortic jet velocity >4.0 m/s or mean gradient >40 mm Hg and 180 patients with LVEF <0.50 were excluded. Among 584 patients with low-gradient severe AS and preserved EF, 74 patients with an AVA indexed to the body surface area (BSA) ≥0.6 cm2/m2, 98 asymptomatic patients, 39 patients with severe symptoms requiring surgery (dyspnoea of New York Heart Association functional class III or IV, angina of Canadian Cardiovascular Society class III or IV and syncope) and 45 patients who were not candidates for AVR due to poor medical condition such as age ≥90 years (n=15), advanced malignant disease or serious comorbidities (n=30), were excluded from the study. All patients were evaluated by cardiologists or cardiac surgeons, and classified as having dyspnoea of New York Heart Association functional class II or angina of Canadian Cardiovascular Society class I or II. Among 328 mildly symptomatic patients, 44 patients with LF were excluded, and the primary cohort was comprised of 284 patients (136 men, mean age±SD 68±10 years) with NF, low-gradient severe AS who were potential candidates for early AVR (figure 1). Because the benefit of early AVR remains controversial in mildly symptomatic patients, the choice of treatment strategy for each patient was at the discretion of the attending physician who explained the potential benefits and procedural risks of early AVR in detail, and most importantly, took the preferences of the individual patients into account. Early elective AVR was performed on 98 patients (early AVR group) within 6 months of the initial echocardiographic evaluation, and the watchful observation strategy was chosen for 186 patients (watchful observation group). Informed consent was obtained from each patient, and the study protocol was approved by the institutional review board at our institution.
Echocardiographic evaluation
Echocardiographic evaluation was performed at baseline and annually during follow-up. Comprehensive 2D and Doppler echocardiographic examinations were performed on all patients with commercially available imaging systems. End-systolic volume, end-diastolic volume and EF of the LV were obtained using the biplane Simpson's method,12 and LV mass was calculated using the formula validated by Devereux and associates.13 The maximal aortic jet velocity was recorded with the apical, right parasternal or suprasternal window that yielded the highest velocity signal. The maximal and mean pressure gradients across the aortic valve were calculated using a modified Bernoulli equation, and the AVA was estimated from the continuity equation14 and indexed to the BSA. Low-gradient severe AS was defined as indexed AVA <0.6 cm2/m2 with mean gradient <40 mm Hg.2 Stroke volume was calculated as the product of time-velocity-integral and cross-sectional area of LV outflow tract and indexed to the BSA. NF was defined as stroke volume index (SVi) ≥35 mL/m2.7 For each measurement, at least three cardiac cycles were averaged.
AVR procedures
In the early AVR group, surgical and transcatheter AVR was performed on 93 and 5 patients, respectively, and the median interval from the diagnosis to AVR was 16 days (IQR 6–42 days). Surgical AVR were performed with mechanical (n=36) or bioprosthetic (n=57) valves, and concomitant coronary artery bypass grafting procedures were performed at the time of AVR on 27 patients. In the watchful observation group, late AVR, including 13 transcatheter AVR, was performed on 101 (54%) patients at the median interval of 1054 days (IQR 653–1532 days).
Follow-up
All the study patients regularly visited their attending physicians at 3–6-month intervals, with a clinical examination and query about changes in specific symptoms. They were educated to report aggravation of symptoms promptly, and echocardiographic examination was repeated earlier than scheduled if symptoms worsened. Data were collected until June 2014. Operative and procedural mortality was defined as death within 30 days of the index procedure or in-hospital death before discharge. Deaths were classified as cardiovascular or non-cardiovascular on the basis of the medical records. Follow-up information was complete for 275 patients (97%), and for the 9 (3%) patients who were lost to follow up, the data on the vital status, dates and causes of death were obtained from the Korean national registry of vital statistics.
In the watchful observation group, progression of AS was monitored by change in aortic jet velocity and mean transaortic pressure gradient on follow-up echocardiographic examinations, and patients were referred for AVR if aortic jet velocity ≥4.0 m/s or mean gradient ≥40 mm Hg.
The primary end point of the study was defined as death from any cause, and the secondary end point was cardiovascular death during follow-up. Cardiovascular deaths included deaths from myocardial infarction, cardiogenic shock, heart failure, infective endocarditis, intracerebral haemorrhage, cerebral infarction, anticoagulation-related bleeding and thromboembolism, sudden deaths and operative and procedural mortality.
Statistical analysis
Categorical variables are presented as numbers and percentages, and were compared using the χ2 test and Fisher's exact test. Continuous variables are expressed as mean±SD, and were compared using the Student's unpaired t test or the Mann–Whitney U test. Analysis of clinical end point was performed on an intention-to-treat basis. Event-free survival curves were constructed with Kaplan–Meier estimates, and were compared using log-rank test. For Kaplan–Meier analysis, we analysed all clinical events by time to first event.
Differences between the early AVR and watchful observation groups in long-term rates of outcomes were assessed with unadjusted Cox proportional hazards analyses. To reduce the effect of treatment selection bias and potential confounding in this observational study, and to ascertain the results of main univariate analysis in the overall cohort, we performed rigorous adjustment for the differences in baseline characteristics using propensity-score matching.15 Propensity scores were estimated without regard to outcome variables, using multiple logistic regression analysis. All prespecified clinical and echocardiographic covariates were included in the full non-parsimonious models for treatment with early surgery versus watchful observation strategy (tables 1 and 2). The propensity-score-matched pairs were created by matching early surgery and watchful observation strategy subjects on the logit of the propensity score using callipers of width equal to 0.2 of the SD of the logit of the propensity score.16 Propensity-score matching for the entire population yielded 83 matched pairs of patients. The discrimination and calibration ability of the propensity-score model was assessed by means of the C statistic (C=0.734) and the Hosmer–Lemeshow statistic (p=0.09). Matching balance was also assessed with standardised differences for each of the baseline variables, and all of the standardised differences for each of the baseline variables were <0.15 (15%) after matching (see online supplementary table S1). In the propensity-score-matched cohort, the risks of clinical end points were compared using univariate Cox regression models with robust SEs that accounted for the clustering of matched pairs. All reported p values were two-sided, and a value of p<0.05 was considered statistically significant. SAS software V.9.1 (SAS Institute, Cary, North Carolina, USA) was used for statistical analyses.
Results
Baseline characteristics
The baseline characteristics of the early surgery and watchful observation groups were compared, as shown in tables 1 and 2. The early AVR group had lower incidence of degenerative AS, smaller indexed AVA, higher aortic jet velocity, higher transaortic gradient (all p<0.05). Baseline clinical and echocardiographic characteristics of the LF and NF groups were listed in online supplementary tables S2 and S3.
Comparison of outcomes in the overall cohort
There were two cases of operative mortality and one mortality case of transcatheter AVR in the early AVR group. The median follow-up was 4.9 years (IQR 3.2–7.4 years) in the early AVR group and 5.3 years (IQR 3.5–7.6 years) in the watchful observation group (p=0.48). During follow-up, seven cardiovascular and five non-cardiovascular deaths occurred in the early AVR group, and 22 cardiovascular and 10 non-cardiovascular deaths in the watchful observation group. There was no significant difference between the two groups for the risk of all-cause death (HR 0.94 for the early AVR, 95% CI 0.51 to 1.73, p=0.84) or the risk of cardiovascular death (HR 0.91 for the early AVR, 95% CI 0.43 to 1.92, p=0.81). The estimated actuarial 8-year mortality rates were 17±5% in the early AVR group and 27±5% in the watchful observation group, and the estimated actuarial 8-year cardiovascular mortality rates were 11±4% in the early AVR group and 18±4% in the watchful observation group (figure 2). The causes of cardiovascular deaths in the early AVR group were AVR procedural mortality in three patients, tissue valve failure in two, infective endocarditis in one, stroke in two, constrictive pericarditis in one and cardiogenic shock in one. In the watchful observation group, 16 and 6 cardiovascular deaths occurred during medical follow-up and after late AVR, respectively. The causes of cardiovascular deaths were sudden death in nine patients, cardiogenic shock in three, acute myocardial infarction in three, congestive heart failure in two, infective endocarditis in two, late AVR mortality in two and anticoagulation-related bleeding in one. Progression to high-gradient AS occurred in 145 patients in the watchful observation group, but 44 (30%) patients refused to undergo AVR because of no change in symptoms or the high operative risks, and 15 cardiovascular deaths occurred among those who refused to undergo AVR. In the additional analysis in which those who refused to undergo AVR were censored at the time of progression to high-gradient AS, the estimated actuarial mortality rates were 5±2% at 4 years and 16±4% at 8 years, and the estimated actuarial cardiovascular mortality rates were 3±1% at 4 years and 6±2% at 8 years.
Comparison of outcomes in the propensity-score-matched cohort
For the 83 propensity-score-matched pairs, there was no significant difference between the early AVR and the watchful observation groups for the risk of all-cause death (HR 1.13 for the early AVR, 95% CI 0.55 to 2.35, p=0.74) or the risk of cardiovascular death (HR 1.22 for the early AVR, 95% CI 0.53 to 2.81, p=0.65). The estimated actuarial 8-year mortality rates were 20±5% in the early AVR group and 22±7% in the watchful observation group, and the estimated actuarial 8-year cardiovascular mortality rates were 13±5% in the early AVR group and 12±4% in the watchful observation group (figure 3).
Predictors of outcome
Using unadjusted Cox proportional hazards analysis, we found that younger age, absence of coronary artery disease, lower comorbidity index, lower Society of Thoracic Surgeons (STS) score,17 non-degenerative AS and early or late performance of AVR in the early AVR and the watchful observation groups were associated with improved survival. The treatment groups, AS severity parameters, LVEF or valvuloarterial impedance were not associated with overall mortality (table 3).
In the watchful observation group, AVA significantly decreased from 0.86±0.09 to 0.73±0.15 cm2 (p<0.001), mean gradient increased from 33.8±3.7 to 45.9±9.8 mm Hg (p<0.001) without a change in SVi (from 45.9±6.4 to 45.5±8.0 mL/m2, p=0.55) during a median echocardiographic follow-up of 1.9 years (IQR 1.1–3.2 years). The estimated actuarial rate of cardiovascular death or need for AVR was 31±3% at 2 years and 65±4% at 4 years.
Discussion
The present study comparing early AVR with watchful observation demonstrated that early AVR did not improve long-term survival in symptomatic patients with NF, low-gradient severe AS and preserved LVEF. Furthermore, watchful observation strategy showed excellent outcomes if AVR was promptly performed at the time of progression to high-gradient AS.
In contrast to the present study, Ozkan et al9 recently reported that AVR was associated with better survival than medical therapy in symptomatic patients with NF as well as LF low-gradient severe AS and preserved LVEF. Differences in the inclusion criteria of study patients and the study protocol may explain these conflicting results. The timing of AVR was also not reported,9 and their AVR group might have included a substantial proportion of patients who had undergone late AVR after the symptoms had worsened or progression to high-gradient AS had occurred. There was only one prospective trial with a protocol of watchful observation, in which, regular clinical and echocardiographic follow-ups were performed on 435 asymptomatic patients with low-gradient severe AS, and excellent outcomes, similar to those of the present study, were observed: overall death occurred in 13%, and AVR was performed on 42% of patients during mean follow-up of 46 months.18 Although adherence to medical therapy alone is associated with poor outcome, watchful observation and timely performance of AVR appear to be a safe and effective treatment option in mildly symptomatic as well as asymptomatic patients with low-gradient severe AS and preserved LVEF.
Predictors of outcome
Previous studies reported that performance of AVR was associated with better survival than medical therapy alone in symptomatic patients with low-gradient severe AS and preserved LVEF, but significant differences in clinical characteristics were also observed between medical and surgical groups7–9 where AVR might have been associated with lower mortality because medical therapy was tended to be chosen for older patients with comorbidities.6 ,19
Low-gradient severe AS may result from normal or reduced stroke volume despite preserved LVEF.20 ,21 It has been reported that NF low-gradient AS is associated with a better prognosis22 ,23 and favourable survival with medical management alone.5 Compared with NF low-gradient AS, LF low-gradient AS is characterised by slightly lower LVEF, smaller AVA and higher valvuloarterial impedance,5 ,20 ,21 and it has a high prevalence of atrial fibrillation, heart failure and reduced survival.5 ,8 ,23 The 2014 American Heart Association/American College of Cardiology guidelines recommend AVR for symptomatic patients with low-gradient severe AS and preserved LVEF only if LF is present, and suggest that symptomatic patients with AVA between 0.8 and 1.0 cm2 and mean gradient <40 mm Hg at NF rates should be closely evaluated to determine whether they would benefit from AVR.1 Performance of AVR is associated with an improved survival, but optimal timing of AVR remains controversial,24 and a prospective randomised trial is required to compare outcomes of early AVR versus watchful observation in low-gradient severe AS with preserved LVEF.
We also found that as AVA decreased during watchful observation, the gradient significantly increased without change of SVi. Although progression to high-gradient AS was previously observed in 41% of asymptomatic patients during mean follow-up of 46 months,18 our patients with mild symptoms showed a higher rate of progression (65±4% at 4 years). Thus, patients with low-gradient AS with even mild symptoms should be monitored more frequently for changes in symptoms or progression of AS, and referred for AVR without delay once they reach criteria for high-gradient AS on a regular echocardiographic follow-up.
Study limitations
Comparison of treatment strategies was subject to the limitations inherent to non-randomised assignment, and such limitations may have significantly affected our results due to selection bias and unmeasured confounders. To control for the inherent biases related to treatment selection and heterogeneity in baseline factors, we performed propensity analysis, which consistently showed that watchful observation strategy was not inferior to early AVR.
In the watchful observation group, 30% of patients refused to undergo AVR, and a significant proportion of cardiovascular deaths may have been prevented if AVR had been performed earlier. In the Euro Heart Survey, surgery was also denied in as many as 33% of elderly patients despite the presence of severe symptomatic AS,25 and refusal of surgery might be a serious risk to the strategy of watchful observation. Although the prevalence of comorbidity was low in the present study, other comorbid conditions might have been related to development of symptoms. As in the previous studies,5 ,22 the prevalence of LF was much lower than NF in the present study, and the statistical power was insufficient to compare outcomes of treatment strategies according to flow pattern.
Conclusions
Early AVR and watchful observation strategy show similar survival in mildly symptomatic patients with NF, low-gradient severe AS and preserved LVEF. Watchful observation with timely performance of AVR should be considered a therapeutic option for low-gradient severe AS.
Key messages
What is already known on this subject?
Performance of aortic valve replacement (AVR) is associated with better survival than medical therapy alone in symptomatic patients with low-flow, low-gradient severe aortic stenosis (AS) and preserved EF.
What might this study add?
Watchful observation strategy is associated with good outcome, and early AVR does not improve survival in symptomatic patients with normal flow, low-gradient severe AS and preserved EF.
How might this impact on clinical practice?
Although adherence to medical therapy alone is associated with poor outcome, watchful observation with timely performance of AVR would be a safe and effective treatment option for symptomatic patients with low-gradient severe AS and preserved EF.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
Footnotes
D-HK and JYJ contributed equally to this study.
Contributors All authors contributed significantly to the completion of the study and the manuscript, including reading and approval of the manuscript in its current form.
Competing interests None declared.
Patient consent Obtained.
Ethics approval ASAN Medical Center Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.