Article Text
PDF
Abstract
Objective To determine whether echocardiography-derived left ventricular filling pressure influences survival in patients with severe aortic stenosis (AS) undergoing aortic valve replacement (AVR).
Methods We retrospectively reviewed 1383 consecutive patients with severe AS, normal ejection fraction and interpretable filling pressure undergoing AVR. Left ventricular filling pressure was determined according to current guidelines using mitral inflow, mitral annular tissue Doppler, estimated right ventricular systolic pressure and left atrial volume index. Cox proportional hazards regression was used to assess the influence of various parameters on mortality.
Results Age was 75±10 years and 552 (40%) were female. Left ventricular filling pressure was normal in 325 (23%), indeterminate in 463 (33%) and increased in 595 (43%). Mean follow-up was 7.3±3.7 years, and mortality was 1.2%, 4.2% and 18.9% at 30 days and 1 and 5 years, respectively. Compared with patients with normal filling pressure, patients with increased filling pressure were older (78±9 vs 70±12, p<0.001), more often female (45% vs 35%, p=0.002) and were more likely to have New York Heart Association class III–IV symptoms (35% vs 24%, p=0.004), coronary artery disease (55% vs 42%, p<0.001) and concentric left ventricular hypertrophy (63% vs 37%, p<0.001). After correction for other factors, increased left ventricular filling pressure remained an independent predictor of mortality after successful AVR (adjusted HR 1.45 (95% CI 1.16 to 1.81), p=0.005).
Conclusions Preoperative increased left ventricular filling pressure is common in patients with AS undergoing AVR and has important prognostic implications, regardless of symptom status. Future prospective studies should consider whether patients with increased filling pressure would benefit from earlier operation.
- aortic stenosis
- echocardiography
- valve disease surgery
Statistics from Altmetric.com
Introduction
Aortic valve stenosis (AS) is a common valve disease that increases in prevalence with ageing.1 Progressive valve calcification results in left ventricular pressure overload, hypertrophy, fibrosis and diastolic dysfunction. Diastolic dysfunction is prevalent in heart failure and community populations, where it has important prognostic implications.2 3 Available data suggest that left ventricular remodelling, fibrosis and diastolic dysfunction have important prognostic implications4–6 and may persist for extended periods after successful aortic valve replacement (AVR).7 8 Recent data have indicated that preoperative diastolic dysfunction and subsequent reverse remodelling are important correlates of postoperative outcome in patients undergoing transcatheter AVR.9–12
Current class I indications for AVR include severe AS with concurrent symptoms or a left ventricular ejection fraction of <50%.13 The development of symptoms is associated with worse outcome and is the most common indication for AVR.14 Increasingly, patients with AS are elderly with multiple comorbidities, making it difficult to determine whether patients are symptomatic from their AS or their comorbid conditions.15 More specific, objective echocardiographic parameters would also be helpful to guide management in this challenging population.
Much of the literature in AS has focused on minimising preoperative risk, defining the severity of AS and identifying risk factors for the development of symptoms that would necessitate an operation.14–20 There are only limited data evaluating determinants of postoperative survival after successful surgical AVR.21–25 Doppler echocardiography is routinely performed in the clinical evaluation of severe aortic stenosis (AS) and recent guidelines allow for standardised assessment of filling pressures using Doppler echocardiography. We hypothesised that preoperative assessment of filling pressure, as a marker of left ventricular adverse remodelling and fibrosis, would correlate with postoperative outcomes. In a cohort of patients with severe AS considered for AVR, we sought to determine (1) the prevalence of increased left ventricular filling pressure based on the most recent diastolic function guidelines26 and (2) whether preoperative left ventricular filling pressure independently predicts long-term survival in this group.
Methods
We retrospectively identified 2140 patients with severe AS who underwent aortic valve intervention with or without coronary artery bypass grafting for severe native AS between 1 January 2002 and 31 December 2012. Among this group, we excluded 143 patients who underwent transcatheter AVR, 262 patients with a left ventricular ejection fraction of <50%, 149 patients with coexistent valve disease who met the exclusion criteria (see further), 95 patients who underwent concurrent other operation (ascending aorta replacement, septal myectomy, pericardiectomy or other), 35 patients who were in atrial fibrillation at the time of the echocardiogram and 73 patients with ≥2 missing diastolic function parameters (medial e’, E/e’ ratio, left atrial volume index or right ventricular systolic pressure). The final cohort included 1383 patients. Exclusion criteria for coexistent valve disease were ≥moderate mitral regurgitation, aortic regurgitation or tricuspid regurgitation. Patients were also excluded if they had coexistent mild or greater mitral stenosis, medial mitral annulus calcification, or if they were in atrial fibrillation at the time of the echocardiogram, given the difficulty grading diastolic function in these groups.
The aortic valve area was calculated using the continuity equation, and multiple imaging windows (eg, apical, suprasternal notch, right supraclavicular and right parasternal) were interrogated to determine the peak aortic jet velocity.27 Patients with an aortic valve systolic mean Doppler gradient of ≥40 mm Hg or with an aortic valve area of ≤1.0 cm2 were considered to have severe AS and were included in the analysis.13
Echocardiographic data
Echocardiographic data were obtained through electronic query of our institution’s echocardiographic database. All patients had a comprehensive transthoracic echocardiogram within 1 year prior to their surgery. Preoperative echocardiograms were performed within 2 months of AVR in 1215/1383 (88%) patients and within 4 months in 1365/1383 (99%) patients. The echocardiographic data were interpreted in accordance with current guidelines for evaluation of cardiac chamber size, function, left ventricular chamber remodelling, AS severity and diastolic function.26–28 The E/e’ ratio was calculated as the early transmitral inflow velocity by pulse wave Doppler (E wave) divided by the medial early mitral annular tissue velocity (medial e’). Left atrial volume was calculated using the area–length technique and was indexed to body surface area.26 Right ventricular systolic pressure was estimated using the peak tricuspid systolic regurgitation velocity added to the estimated right atrial pressure based on the size and collapsibility of the inferior vena cava.29
A simplified diastolic function approach to assess left ventricular filling pressure was used according to current guidelines28 based on an E/e’ of ≥15 (medial e’ tissue velocity), a septal e’ velocity of <7 cm/s, a left atrial volume index of >34 mL/m2 and an estimated right ventricular systolic pressure of >35 mm Hg. Patients with less than two abnormal parameters were classified as normal left ventricular filling pressure; those with two abnormal parameters were classified as indeterminate; and those with more than two abnormal parameters were classified as increased left ventricular filling pressure.
Clinical comorbidities
Clinical comorbidities were determined by query of the Rochester Epidemiology Project database based on ICD-9 diagnosis data, described previously,30 and our institution’s cardiovascular surgery database. Only patient diagnoses documented prior to the time of the cardiac surgery were included. Manual chart review was performed to reconcile discrepancies between the two databases, to obtain information on coronary artery disease from the coronary angiogram reports and to determine preoperative New York Heart Association (NYHA) functional class. An epicardial coronary artery stenosis of ≥70% in the left anterior descending, left circumflex or right coronary artery or a left main lesion of ≥50% was considered significant. Surgical procedures performed were obtained through our institution’s cardiovascular surgery database. Mortality data were determined by query of the Rochester Epidemiology Project database and the National Death Index on 4 January 2016, and time to mortality was defined as the number of days after AVR.
Analysis
Statistical analysis was performed with JMP V.9.0.1 (SAS) and SAS V.9.4. A p value of <0.05 was considered significant. Data were displayed as mean±SD or SE, where appropriate. Continuous variables were compared with the use of t-test, and nominal variables were compared using contingency tables and χ2 analysis. Logistic regression was used to correlate patient symptom status with clinical and echocardiographic variables. The primary endpoint was all-cause mortality following AVR. The influence of various clinical and echocardiographic parameters on mortality was evaluated with Cox proportional hazards regression. For all variables, the proportional hazard assumption was evaluated both visually by plotting residuals versus time and formally by testing for correlation between residual and time. No strong violations were noted. Results of these analyses are presented as HR and associated 95% CIs. All clinical and echocardiographic variables with a p value of <0.20 on univariate analysis were included as candidate variables in the multivariate analysis. In the final multivariate model, multicollinearlity was assessed for correlated variables by fitting several models with combinations of these variables to ensure that parameter estimates did not change significantly. Kaplan-Meier methods were used for survival estimates. The goodness of fit for the multivariate model was examined using the χ2 model fit data. A secondary outcome analysis was performed using inverse propensity weighting. Variables found to be associated with mortality in univariate analysis were used entered into a logistic regression model for abnormal filling pressure. Each patient’s probability of having increased filling pressure was then used to create a propensity weight. Distributions were found to be similar after propensity weighting, and the outcome analysis was repeated with these weights.
Patient and public involvement
Patients and/or the public were not involved in the design, conduct, reporting or dissemination of this research.
Results
Baseline characteristics
The final cohort included 1383 subjects. The mean age was 75±10 years and 552 (40%) were female. Bicuspid AS was present in 406 patients (29%), and patients with bicuspid aortic valve were younger (68±10 vs 78±8 years, p<0.001) and were less likely to have increased left ventricular filling pressure (27% vs 50%, p<0.001). Clinical characteristics according to left ventricular filling pressure are shown in table 1. Patients with increased left ventricular filling pressure were older (78±9 vs 70±12 years, p<0.001), were more likely to be women (45% vs 35%, p<0.001) and had increased prevalence of multiple cardiac and non-cardiac comorbidities.
Clinical characteristics
Echocardiographic characteristics according to left ventricular filling pressure are shown in table 2. Left atrial volume index was available in 1367 (99%) of patients, mitral annular e’ in 1383 (100%), E/e’ in 1343 (97%), right ventricular systolic pressure in 1167 (84%) and left ventricular mass index in 1323 (96%). A left atrial volume index of >34 mL/m2 was present in 809 (59%); E/e’ of ≥15 was present in 726 (54%); and estimated right ventricular systolic pressure of >35 mm Hg was present in 474 (41%). Compared with those with normal filling pressure, patients with increased filling pressure were more likely to have concentric left ventricular hypertrophy and had a higher stroke volume index (50±9 vs 48±8 mL/m2, p<0.001), a smaller aortic valve area (0.78±0.14 vs 0.82±0.14 cm2, p<0.001) and a higher mean systolic gradient (58±15 vs 54±12 mm Hg, p<0.001).
Echocardiographic characteristics
Correlates of NYHA functional class
NYHA functional status at the time of surgery was available in 1371 patients (99%). Prior to operation, 293 patients (21%) had minimal or no symptoms; 660 (48%) had class II NYHA functional status; and 418 (30%) had NYHA class III–IV symptoms. Independent correlates of symptoms (≥NYHA class II) at the time of surgery included age (OR 1.35 (95% CI 1.16 to 1.57) per 10-year increase, p<0.001), female sex (OR 1.93 (95% CI 1.42 to 2.62), p<0.001), body mass index (OR 1.04 (95% CI 1.02 to 1.08) per one unit increase, p=0.001), LVEF (OR 1.21 (95% CI 1.08 to 1.36) per 5% decrease, p<0.001), increased left ventricular filling pressure (OR 1.77 (95% CI 1.24 to 2.52), p=0.002), coronary artery disease (OR 1.78 (95% CI 1.34 to 2.37), p<0.001), a previous diagnosis of congestive heart failure (OR 2.25 (95% CI 1.58 to 3.20), p<0.001) and aortic valve area (OR 1.16 (95% CI 1.04 to 1.30) per 0.1 cm2 reduction, p=0.007). The full univariate and multivariate analyses are shown in online supplementary table 1. Among the 293 patients with minimal or no symptoms, filling pressure was increased in 104 (35%), indeterminate in 103 (35%) and normal in 86 (29%).
Supplemental material
Correlates of mortality after AVR
There were 642 (46%) deaths over a follow-up time of 7.3±3.7 years. Operative mortality occurred in three patients (0.2%). Mortality was 1.2%, 4.2% and 18.9% at 30 days and 1 and 5 years, respectively. Considered separately, abnormal E/e’, medial e’, estimated right ventricular systolic pressure and increased left atrial volume were each associated with increased postoperative mortality (figure 1). Pre-AVR left ventricular filling pressure correlated with postoperative survival (figure 2). Univariate and multivariate clinical and echocardiographic correlates of postoperative survival are shown in table 3. Independent predictors of survival post-AVR included advanced age, NYHA functional class III–IV symptoms, congestive heart failure, diabetes, lung disease, elevated creatinine, indexed stroke volume and increased left ventricular filling pressure (table 3). Results were similar when inverse propensity weighting was used (online supplementary table 2). When stratified by symptom status, left ventricular filling pressure remained a significant predictor of postoperative mortality in patients with and without preoperative symptoms (figure 3).
Supplemental material
Survival according to echocardiographic parameters of left ventricular diastolic function. Abnormal diastolic function parameters including increased E/e’ (A), reduced e’ velocity (B), increased RVSP (C) and increased LAVI (D) were associated with increased mortality after aortic valve replacement. LAVI, left atrial volume index; RVSP, right ventricular systolic pressure.
Long-term mortality according to echocardiographically derived LV filling pressure. There was a progressive decrease in survival in those with normal, indeterminate and increased LV filling pressure (p<0.001). LV, left ventricular.
Long-term mortality according to LV filling pressure in patients without and with preoperative symptoms. After stratification of patients by symptom status, preoperative left ventricular filling pressure remained a significant correlate of postaortic valve replacement mortality in asymptomatic (A) and symptomatic (B) patients. LV, left ventricular.
Univariate and multivariable Cox proportional HRs for risk of mortality after aortic valve replacement
The χ2 statistic for the multivariate model, including only the clinical data (excluding all echocardiographic data), was 275. The χ2 statistic increased to 282 when including left ventricular ejection fraction and did not change significantly with the sequential addition of stroke volume index (281). With the addition of left ventricular filling pressure to the model, the χ2 increased to 292. The χ2 statistic was slightly higher (298) when including E/e’ alone instead of left ventricular filling pressure. When including left atrial volume index of >34 mL/m2, medial mitral e’ velocity of <7 cm/s and right ventricular systolic pressure of >35 mm Hg instead of left ventricular filling pressure, the χ2 statistic was 287, 282 and 283, respectively.
Discussion
Much of the previous work in AS has focused on identifying preoperative risk factors in patients with severe AS to minimise preoperative morbidity and mortality.15–19 However, there are limited data evaluating the impact of preoperative variables, including those that can be obtained at the time of routine Doppler echocardiography, on postoperative survival after AVR.21–25 We identified multiple clinical and echocardiographic parameters independently associated with increased risk of death after AVR. After correction for clinical factors associated with mortality such as age, symptom status, coronary artery disease, congestive heart failure, lung disease and kidney function, echocardiography-derived left ventricular ejection fraction, left ventricular filling pressure and stroke volume index remained independent correlates of postoperative mortality in a cohort of patients with normal ejection fraction undergoing AVR for severe AS. Either integrated diastolic function grading according to current guidelines28 or E/e’ alone offers important incremental prognostic importance in patients with severe AS undergoing AVR. In our model, E/e’ alone had similar or potentially more incremental prognostic power as compared with full diastolic function grading by integrating E/e’, left atrial size, e’ velocity and estimated right ventricular systolic pressure.
Current guidelines give a class I indication for AVR in symptomatic AS. However, in an increasingly elderly population with high rates of comorbidities, determining whether symptoms are attributable to AS versus coexisting cardiopulmonary disease, excess weight, sedentary lifestyle or advanced age can be challenging. Our data also suggest that left ventricular filling pressure correlates with symptom status at the time of AVR, but symptom status appears multifactorial, and even asymptomatic patients were found to have evidence of increased filling pressure preoperatively. Increased age, body mass index and coronary artery disease were independent correlates of symptom status. However, 35% of patients without clear symptoms also had increased filling pressure on echocardiography, and these patients appear to be at increased risk of mortality post-AVR (figure 3). This finding highlights important pathophysiological mechanisms of symptom development in AS and suggests that left ventricular filling pressure, independent of symptom status, could be used as a marker of increased risk in patients with severe AS considered for AVR.
Patients with increased filling pressure undergoing operation had slightly more advanced AS at the time of AVR and more frequently had concentric left ventricular hypertrophy, which could have contributed to the more advanced diastolic dysfunction seen. Increased filling pressure was more common in patients with trileaflet AS, and patients with increased filling pressure were older and had more frequent hypertension, diabetes and coronary artery disease. Increased age and overall worse cardiovascular health could also accelerate diastolic dysfunction, contribute to more advanced symptoms and hinder reverse remodelling postoperatively. Regardless, our data indicate that increased left ventricular filling pressure at the time of operation portends a worse prognosis postoperatively. These findings indicate that, ideally, AVR should occur prior to the onset of significant diastolic dysfunction and that optimising comorbidities postoperatively may also have an important role in improving outcomes.
Prior studies have identified the severity of AS as an important marker of outcome prior to AVR.15–20 Conversely, our data indicate that survival after AVR is not linked to the severity of preoperative AS but instead to parameters of myocardial systolic and diastolic dysfunctions in conjunction with coronary artery disease and other comorbidities. The degree to which the left ventricle reverse remodels after AVR is not well defined, but our data suggest that preoperative diastolic abnormalities may have long-lasting effects with important clinical implications after AVR. Increased left ventricular filling pressure correlates with preoperative symptom status and postoperative survival. By operating before the onset of significant diastolic dysfunction, there is potential to decrease perioperative morbidity and mortality and to improve long-term survival by avoiding the deleterious effects of adverse left ventricular remodelling. Future prospective studies should consider randomisation of patients with severe AS and increased filling pressure to AVR versus observation to determine whether patients with increased filling pressure would benefit from earlier operation.
Limitations
Given the retrospective nature of this study, it is subject to potential bias and confounding factors. We attempted to minimise the impact of confounders by excluding patients with significant concurrent valve disease and by using multivariate methods. Due to the large number of events, we were able to correct for multiple potential clinical and echocardiographic factors. Nevertheless, one limitation of our study is the lack of an independent validation sample which will be important to confirm our findings.
Because our clinical and echocardiographic data represent a single snapshot in time, we do not know in whom and to what degree there is left ventricular reverse remodelling after AVR. This will be important to evaluate in future studies as it may provide important insight into how adverse remodelling impacts postoperative survival and whether there are subgroups of patients who are more likely to have beneficial reverse remodelling.
Finally, our dataset does not allow us to definitively determine whether operating on asymptomatic patients with signs of abnormal diastolic dysfunction would result in improved outcomes. This will be an important question to address with future research.
Conclusions
Postoperative mortality following AVR for severe AS is influenced by a range of clinical and echocardiographic parameters. Preoperative increased left ventricular filling pressure is common in patients with preserved ejection fraction and is independently associated with postoperative mortality. Incorporation of echocardiographic assessment of filling pressure improves preoperative risk stratification; future prospective studies should consider whether patients with increased filling pressure would benefit from earlier operation.
Key messages
What is already known on this subject?
Symptom development, small aortic valve area, high aortic transvalvular gradient and reduced ejection fraction are associated with increased need for aortic valve replacement (AVR) and worse preoperative outcomes in patients with severe AS. Available data indicate that left ventricular diastolic abnormalities may persist for an extended period after AVR. We hypothesised that preoperative assessment of filling pressure, as a marker of left ventricular adverse remodelling and fibrosis, would correlate with survival after AVR.
What might this study add?
In a series of 1383 consecutive patients undergoing surgical AVR, the presence of increased left ventricular filling pressure derived from echocardiography was an independent risk factor for mortality (HR 1.45) during long-term follow-up.
How might this impact on clinical practice?
Our results indicate that routine preoperative assessment of left ventricular filling pressure contains incremental prognostic information that may enhance shared decision-making surrounding AVR. This highlights the need for prospective studies evaluating the benefit of early AVR in patients with severe AS and increased filling pressure.
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.
Footnotes
Twitter @pattypellikka
Contributors JJT: study planning, data analysis, manuscript writing and review. MB: study planning, data acquisition and analysis. JS and ROA: study planning, data acquisition, and analysis. VTN, MFE, SVP, JKO, HVS and PAP: study design and critical review of the manuscript. JD: data acquisition, analysis and critical review of the manuscript. CS: data acquisition and analysis.
Funding This study was made possible using the resources of the Rochester Epidemiology Project, which is supported by the National Institute on Aging of the National Institutes of Health under award number R01AG034676.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval The study was approved by the Mayo Clinic Institutional Review Board. Informed consent was waived, given the retrospective nature of this study.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data available from corresponding author upon reasonable request.