Valve Replacement for Severe Aortic Stenosis With Low Transvalvular Gradient and Left Ventricular Ejection Fraction Exceeding 0.50

Valve Replacement for Severe Aortic Stenosis With Low Transvalvular Gradient and Left Ventricular Ejection Fraction Exceeding 0.50 Giuseppe Tarantini, MD, PhD, Elisa Covolo, MD, Renato Razzolini, MD, Claudio Bilato, MD, PhD, Anna Chiara Frigo, MS, Massimo Napodano, MD, Enrico Favaretto, MD, Chiara Fraccaro, MD, Giambattista Isabella, MD, Gino Gerosa, MD, Sabino Iliceto, MD, FACC, and Alain Cribier, MD, FACC Sections of Cardiology and Cardiac Surgery, Department of Cardiac, Thoracic and Vascular Sciences, University of Padua, Pauda; Department of Environmental Medicine and Public Health, Padua, Italy; and Department of Cardiology, Charles Nicolle University Hospital, University of Rouen, Rouen, France Background. Severe aortic stenosis with a low transvalvular gradient and preserved left ventricular ejection fraction (LVEF) is often misdiagnosed, leading to undertreatment of such patients with no clear indication for surgical intervention. This study investigated the outcome of aortic valve replacement (AVR) in patients with severe aortic stenosis and a low transvalvular gradient despite normal LVEF. Methods. Between 1985 and 2008, we evaluated 73 patients who underwent AVR compared with 29 patients who did not. Overall, aortic valve area was 1.0 cm 2 or smaller, LVEF was 0.50 or higher, and transvalvular gradient was 30 mm Hg or less. Multivariate and Cox analyses were used to compare these two groups according to AVR. Results. Compared with controls, AVR patients were younger and with higher body mass index. Coronary artery bypass grafting was performed simultaneously in 38 AVR patients (52%). At follow-up (median, 42 months; interquartile range, 23 to 75 months), survival was longer in AVR patients. By Cox analysis, AVR remained a major predictor of lower mortality (hazard ratio, 0.237; 95% confidence interval, 0.119 to 0.470; p < 0.0001). Conclusions. In patients with severe aortic stenosis and low transvalvular gradient despite a normal LVEF, AVR was associated with significant improvement in longterm survival and functional status and with a low operative mortality. (Ann Thorac Surg 2011;91:1808 15) 2011 by The Society of Thoracic Surgeons Most patients with severe aortic stenosis (AS) show a mean transvalvular pressure gradient (TVG) that exceeds 40 mm Hg, even the when left ventricular (LV) ejection fraction (LVEF) is low [1 3]. However, a relevant proportion of elderly patients, classified as severe AS on the basis of aortic valve area (AVA) of less than 1.0 cm 2, have low TVG despite a preserved LVEF ( 0.50) [4 7]. Patients with severe AS and low TVG, despite a normal LVEF, show a poor prognosis if treated medically. Moreover, they are frequently misjudged as having milder degrees of AS and are not treated effectively, although symptomatic. We reviewed our institutions experience to test the hypothesis that among patients with pure severe AS and low TVG despite normal LVEF, aortic valve replacement (AVR) improves long-term survival and functional status. Accepted for publication Feb 17, 2011. Address correspondence to Dr Tarantini, Department of Cardiac, Thoracic and Vascular Sciences, University Hospital, via Giustiniani 2, 35128 Padua, Italy; e-mail: giuseppe.tarantini.1@unipd.it. Patients and Methods This study was approved by Institutional Review Committee and was performed in accordance with the 2000 revised Declaration of Helsinki. Study Population We used database to retrospectively identify all patients who, from March 1985 to May 2008, showed an AVA of less than 2.0 cm 2 evaluated invasively (Fig 1). The study population consisted of 190 patients selected because of an AVA of 1.0 cm 2 or less, an LVEF exceeding 0.50, and a peak-to-peak TVG of 30 mm Hg or less, measured at cardiac catheterization. Of these, 88 patients were excluded because they were aged younger than 18 years, had more than moderate ( 2 /4) aortic or mitral regurgitation or any mitral stenosis, underwent previous valve replacement or repair, or required any valve replacement in addition to the aortic valve. Among the remaining 103 patients, 73 underwent operations for AVR with or without coronary artery bypass grafting (CABG) and were considered cases, whereas 29 did not and were considered controls. All patients gave written informed consent. 2011 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2011.02.057

Ann Thorac Surg TARANTINI ET AL 2011;91:1808 15 LOW-GRADIENT AORTIC STENOSIS 1809 Abbreviations and Acronyms AS aortic stenosis AVA aortic valve area AVR aortic valve replacement BMI body mass index CABG coronary artery bypass graft CAD coronary artery disease CI confidence interval COPD chronic obstructive pulmonary disease DM diabetes mellitus EuroSCORE European System for Cardiac Operative Risk Evaluation F/U follow-up HR hazard ratio LV left ventricle/ventricular LVEF left ventricular ejection fraction NA not available OR odds ratio SVI stroke volume index TEE transesophageal echocardiogram TTE transthoracic echocardiogram TVG transvalvular gradient gradient was measured after the pullback, with the catheter measurements performed a few centimeters downstream of the valve, where pressure recovery is complete. Mitral and aortic regurgitation were semiquantitated from 0 (none) to 4 (severe) [11]. Low stroke volume index (SVI) was defined as a SVI of 35 ml/m 2 or less [4]. Total hemodynamic load was estimated by the valvular-arterial impedance, which was calculated by dividing LV systolic pressure by SVI. The SVI/aortic pulse pressure ratio was used as an index of total systemic arterial compliance. Coronary artery disease (CAD) was defined as 50% or more lumen narrowing of the left main artery or 70% or more lumen narrowing of major epicardial vessels by visual assessment. Multivessel CAD was defined when either the left main or at least 2 major epicardial vessels were involved. Because of the type of measured variables (i.e. valvular-arterial impedance, which is quantified Data Collection Baseline clinical data, echocardiographic results, invasive hemodynamic data, and operative data were obtained by reviewing medical records and databases. Survival and functional status was obtained by telephone interview of the patient or his or her physician, or by reviewing medical records. Death was classified as cardiac or noncardiac. When a noncardiac cause of death was not clearly documented, the death was considered to be cardiac-related. The last evaluation of survival status of the patients was performed in April 2009. Status could not be confirmed only in 1 patient, who was excluded from the survival analyses. Within 30 days before cardiac catheterization, 94 patients (92%) underwent echocardiographic assessment, 77 (82%) in our institution and 17 (18%) in other hospitals. Eight patients without echocardiographic data underwent cardiac catheterization between 1985 and 1988. The mean TVG was calculated using the simplified Bernoulli equation, and the AVA was calculated with the continuity equation. Mitral and aortic regurgitation were semiquantitated from 0 (none) to 4 (severe) by analysis of color-flow Doppler imaging [8 10]. Because of the well-established tradition of relying on the invasive pressure gradient in our center [3, 11], this strategy was used independently of whether AVA was calculated by noninvasive methods. All patients underwent right and left cardiac catheterization and coronary angiography, as reported elsewhere [11]. Briefly, cardiac output and index were measured by the Fick method during cardiac catheterization, and AVA was calculated by the Gorlin equation. Peak-to-peak TVG was achieved by measuring the LV pressure and then the ascending aortic pressure after the catheter was pulled back. The Fig 1. Selection of the patient population. (AVA aortic valve area; LVEF left ventricular ejection fraction; TVG transvalvular gradient.)

1810 TARANTINI ET AL Ann Thorac Surg LOW-GRADIENT AORTIC STENOSIS 2011;91:1808 15 more precisely by invasive methods) and considering the completeness and the well-established standardization of all the invasive data in our referral center, hemodynamic data were used for statistical analysis. All surgical records were reviewed to determine the type and size of aortic valve prosthesis and to show whether CABG was performed concomitantly with AVR. The surgical procedures, along with cross-clamp and cardiopulmonary bypass times, were included for the analysis. AS severity was also assessed visually during the operation. Statistical Analysis Data are expressed as median (range) for continuous variables and as count (percentage) for categoric variables. The Wilcoxon rank sum or 2 tests were used, as appropriate, to compare clinical, echocardiographic, cardiac catheterization, and surgical data between the patients who did and did not undergo AVR. The Wilcoxon rank sum test was used to compare preoperative and postoperative functional classes. Logistic regression analysis was used to investigate the univariate predictors of AVR, considering as independent variables observed and plausible correlates of AVR the following: age, body mass index (BMI), New York Heart Association and Canadian Cardiovascular Society functional classes, syncope, diabetes mellitus, hypertension, peripheral vascular disease, chronic airway obstruction, estimated creatinine clearance rate, CAD, previous myocardial infarction, LVEF, European System for Cardiac Operative Risk Evaluation (EuroSCORE), peak-topeak TVG, AVA, mean pulmonary artery pressure, wedge pulmonary capillary pressure, pulmonary vascular resistance, LV end-diastolic pressure, and systemic diastolic pressure. All the predictors with p 0.2 in the univariate analysis were entered in the multivariate analysis where the significance level was set to 5%. The results are expressed as odds ratio with relative 95% confidence intervals and p values. The Kaplan-Meier log-rank test was used to analyze the effect of AVR on survival. The relationship between type of treatment, baseline factors, and death was evaluated with Cox proportional hazard analysis. The effect of AVR on outcome was assessed by intention-to-treat analysis. Because of the limited sample size of the study and the need to limit the number of covariates to have sufficient statistical power, we considered six models plus the unadjusted one with only treatment as covariate. The following models were considered adjusting for (1) age, (2) age and BMI, (3) age, BMI, and creatinine, (4) Euro- SCORE, (5) CAD, (6) CABG, and (7) diabetes mellitus, peripheral vascular disease, chronic obstructive pulmonary disease, and renal insufficiency. The results are expressed as hazard ratios and relative 95% confidence intervals; values of p 0.05 were considered statistically significant. We also calculated the propensity score for AVR status, built with a nonparsimonious approach (see previous paragraph for the variables included). The area under the receiver operating characteristics curve was 0.82, indicating good discrimination between patients who did or did not receive AVR. Propensity scores were used to divide the study group into quintiles, and those in propensity quintiles 2 to 4 showed reasonable matching of propensity scores, variances of propensity scores, and baseline characteristics. Overall, mean propensity scores were 0.76 0.16 in the AVR group and 0.61 0.20 in the control group (p 0.0001); mean propensity scores in propensity-matched patients were 0.75 0.08 in AVR group and 0.73 0.10 in the control group (p 0.46). Kaplan-Meier estimates were used to generate survival curves between matched patients who did or did not receive AVR. Data were analyzed with SAS 9.1.3 software (SAS Institute Inc, Cary, NC). Results Baseline Characteristics of the Study Population Baseline clinical and electrocardiographic data are reported in Table 1. Baseline echocardiographic, hemodynamic, and angiographic data are reported in Table 2. The value of AVA indexed for body surface area in our study population was not different between AVR and medically treated groups (p 0.3). Similarly, when the AVA values were indexed for BMI, no significant differences between the AVR and the medically treated groups were observed (p 0.2). Low SVI was present in 20 AVR patients (27%) and in 6 controls (21%; p 0.6). Low SVI status inversely correlated with end-diastolic volume index (OR, 0.75; 95% CI, 0.61 to 0.91; p 0.005). Compared with those with preserved SVI, the patients with low SVI had an higher systolic pressure (OR, 1.04; 95% CI, 1.002 to 1.08; p 0.041), higher heart rate (OR, 1.05; 95% CI, 1.012 to 1.09; p 0.009), higher valvular-arterial impedance (OR, 9.50; 95% CI, 3.55 to 25.45; p 0.001), and lower LVEF (OR, 0.92; 95% CI, 0.87 to 0.97; p 0.005). The preserved SVI subgroup showed a prolonged systolic ejection period compared with the low SVI subgroup, with a median of 0.33 seconds (range, 0.20 to 0.58 seconds) vs a median of 0.28 seconds (range, 0.21 to 0.52 second; p 0.01), but a similar respective median transvalvular flow rate of 201 ml/s (range, 139 to 243 ml/s) vs 198 ml/s (range, 129 to 242 ml/s; p 0.6). Type of Treatment Of the 102 patients, 73 (72%) underwent AVR operations, of whom 38 (52%) received a concomitant CABG, and 6 (8%) underwent a concomitant aortic root or ascending aorta replacement. Mean aortic prosthesis size was 22 2 mm (range 19 to 29 mm), and 61 patients received a bioprosthesis. Operative mortality rate was 2.7%. Significant valve calcification was observed at the operative examination in all patients. AVR was performed after baseline study with a median of 2 months (interquartile range, 21 days to 5.6 months). Only one patient died while waiting for AVR. Univariate and multivariate predictors of AVR are reported in Table 3. Reasons for nonsurgical management were physical disability (3%),

Ann Thorac Surg TARANTINI ET AL 2011;91:1808 15 LOW-GRADIENT AORTIC STENOSIS 1811 Table 1. Baseline Clinical and Electrocardiographic Data Total AVR No AVR Variable a (n 102) (n 73) (n 29) p Value Age, years 78 (72 81) 76 (72 80) 79 (76 84) 0.02 Females 59 (58) 40 (55) 19 (65) 0.32 Body mass index, kg/m 2 26 (23 28) 26 (24 28) 24 (21 28) 0.01 Body surface area, m 2 1.76 (1.60 1.91) 1.78 (1.61 1.93) 1.71 (1.54 1.84) 0.06 Comorbidities Systemic hypertension 97 (95) 68 (93) 29 (100) 0.32 Diabetes mellitus 31 (30) 22 (30) 9 (31) 0.93 Peripheral vascular disease 72 (71) 53 (73) 16 (66) 0.48 COPD 16 (16) 12 (16) 4 (14) 0.74 Hypercholesterolemia 72 (71) 55 (75) 17 (59) 0.1 Renal insufficiency b 14 (14) 9 (13) 5 (18) 0.5 Preoperative symptoms Symptoms 90 (88) 64 (88) 26 (90) 0.78 Dyspnea 64 (63) 49 (67) 15 (52) 0.18 Angina 62 (61) 42 (58) 20 (69) 0.37 Syncope 15 (15) 12 (16) 3 (10) 0.43 Prior revascularization 11 (11) 7 (10) 4 (14) 0.72 Prior myocardial infarction 18 (18) 10 (14) 8 (28) 0.1 Electrocardiogram Left ventricular hypertrophy 59 (58) 43 (59) 16 (55) 0.73 Rhythm 0.31 Sinus 77 (75) 58 (80) 19 (66) Atrial fibrillation 22 (23) 14 (19) 9 (31) Paced 2 (2) 1 (1) 1 (3) EuroSCORE 9 (7 11) 8 (7 11) 9 (7 11) 0.30 a Values are expressed as number (%) or median (interquartile [25 75] range). b Creatinine level 130 mol/l. AVR aortic valve replacement; COPD chronic obstructive pulmonary disease; EuroSCORE European System for Cardiac Operative Risk Evaluation. advanced age (7%), unconfirmed indication by the surgeon (66%), and patient refusal (24%). Survival Clinical follow-up was complete in 101 of 102 patients; the median follow-up for in-hospital survivors was 42 months (interquartile range, 23 to 75 months). Overall mortality was 37%: 19 patients in AVR group (26%) and 18 (62%) in those treated medically (p 0.001; Fig 2A). Cardiac death occurred in 13 AVR patients (18%) and in 15 control patients (52%; p 0.001). AVR remained a major predictor of lower mortality in all the multivariate models considered (Table 4). Of note, AVR remained an independent predictor of outcome by Cox analysis using AVR and CABG as covariates. Overall survival was similar in patients with low SVI and normal SVI (p 0.47); AVR was performed with similar prevalence in both groups (p 0.2). Moreover, AVR improved survival in both groups (p 0.001). When the analysis was limited to the 78 patients with AVA between 0.8 and 1 cm 2, AVR remained a predictor of better survival compared with conservative management (p 0.001). Among the propensity-matched patients, clinical follow-up was complete, with a median time for in-hospital survivors of 53 months (interquartile range, 25 to 79 months). AVR was done in 44 of the 61 patients (72%). Total mortality rate was 33%, comprising 11 AVR patients (25%) and 9 control patients (53%; p 0.06; Fig 2B). Cardiac death rate was 18% in AVR and 41% in control patients (p 0.09). No statistically significant differences were observed in survival between patients treated by AVR with or without CABG (p 0.6, Fig 2C). Figure 3 illustrates the functional and anginal status changes in operated-on patients on a case-by-case basis. Comment The principal findings of our study, spanning more than 20 years, are: 1. isolated, degenerative, severe AS/low TVG despite normal LVEF is a small but a significant proportion (5%) of all patients with severe AS; 2. patients with low TVG/normal LVEF have a poor prognosis if treated conservatively; and 3. AVR is associated with low perioperative mortality and is a major predictor of improved survival and functional status independently from the SVI.

1812 TARANTINI ET AL Ann Thorac Surg LOW-GRADIENT AORTIC STENOSIS 2011;91:1808 15 Table 2. Baseline Echocardiographic, Hemodynamic, and Angiographic Data Total AVR No AVR Variable (n 102) (n 73) (n 29) p Value Echocardiographic data Aortic valve area, cm 2 0.80 (0.70 0.89) 0.90 (0.80 0.98) 0.90 (0.75 1.00) 0.95 Mean TVG, mm Hg 33 (27 38) 33 (27 39) 33 (27 37) 0.85 LV ejection fraction 0.61 (0.56 0.66) 0.61 (0.56 0.67) 0.60 (0.50 0.66) 0.42 Hemodynamic and angiographic data Aortic valve area, cm 2 0.90 (0.80 0.99) 0.90 (0 80 0.98) 0.90 (0.75 1.00) 0.76 Peak-to-peak TVG, mm Hg 30 (20 30) 30 (23 30) 30 (20 30) 0.39 Cardiac index, L min 1 m 2 2.6 (2.3 2.8) 2.5 (2.3 2.8) 2.7 (2.3 2.9) 0.50 LV end-diastolic volume index, ml/m 2 63 (55 83) 62 (54 84) 67 (57 80) 0.99 Transvalvular flow rate, ml/s 198 (172 218) 198 (178 218) 198 (170 220) 0.45 Stroke volume index, ml/m 46 13 46 13 47 11 0.68 LV ejection fraction 0.69 (0.61 0.74) 0.68 (0.61 0.74) 0.69 (0.62 0.76) 0.42 Mean pulmonary pressure, mm Hg 18 (15 26) 20 (15 26) 18 (12 26) 0.29 Wedge pulmonary capillary pressure, mm Hg 10 (6 16) 12 (8 16) 8 (5 16) 0.14 LV end-diastolic pressure, mm Hg 15 (10 20) 15 (10 20) 17 (12 20) 0.18 Diastolic arterial pressure, mm Hg 65 (60 80) 70 (60 80) 60 (53 70) 0.03 Valvular-arterial impedance, mm Hg ml 1 m 2 4.0 (3.2 5.2) 4.1 (3.3 5.3) 3.9 (3.1 4.8) 0.47 Systemic arterial compliance, ml mm Hg 1 m 2 0.51 (0.39 0.62) 0.51 (0.39 0.64) 0.54 (0.40 0.60) 0.74 Coronary artery disease 63 (62) 42 (58) 21 (72) 0.18 Multivessel coronary artery disease 45 (44) 31 (42) 14 (48) 0.59 Values are expressed as number (%) or median (interquartile [25 75] range). AVR aortic valve replacement; LV left ventricular; TVG transvalvular gradient. Only a few, heterogeneous, observational studies are available with follow-up on this topic (Table 5). The discrepancies in the prevalence of patients with severe AS/low TVG/normal LVEF among our study and others [4, 5] might be explained by differences in patient selection (ie, inclusion of combined valve disease) or in selection of the methods (invasive or not) and of the criteria (ie, indexed AVA or not) used for grading the severity of AS. Indeed, we included only patients with isolated and degenerative AS to avoid confounders that might blur the reliability of the hemodynamic measurements and the prognostic effect of AS on its own. Moreover, we used invasive measures for our analyses, because although echocardiography is currently the main diagnostic strategy, the direct measurements of the pressures of the heart chambers by catheterization are also less biased in the calculation of clinically significant variables (valvular-arterial impedance, cardiac output). We observed an improved survival in AVR patients compared with controls consistently with the findings of others [7, 12]. AVR was performed 72% of our patients, compared with other studies where the AVR rate was 33% to 46% [5, 7, 12]. The clinical decision in our study was supported by the onset of symptoms, which were present in 88% of patients. Symptoms status was not available in the studies of Hachicha and colleagues [4] and Pai and colleagues (7), and only in 24% of enrolled patients in the study of Barasch and colleagues [5], even Table 3. Independent Predictors of Aortic Valve Replacement Univariate Analysis Multivariate Analysis Predictor OR (95% CI) p Value OR (95% CI) p Value Age/10 years 0.397 (0.182 0.864) 0.020 0.335 (0.125 0.896) 0.029 Body mass index/5 units 2.268 (1.213 4.239) 0.010 2.375 (1.134 4.974) 0.022 Creatinine clearance 1.029 (1.004 1.055) 0.023...... Coronary artery disease 0.516 (0.202 1.318) 0.167...... Aortic diastolic pressure 1.038 (1.002 1.075) 0.038...... Wedge pressure 1.049 (0.984 1.118) 0.140...... LV end-diastolic pressure 0.958 (0.899 1.021) 0.183...... Pulmonary vascular resistance 0.717 (0.467 1.101) 0.128...... Previous myocardial infarction 0.417 (0.145 1.194) 0.103...... CI confidence interval; LV left ventricle; OR odds ratio.

Ann Thorac Surg TARANTINI ET AL 2011;91:1808 15 LOW-GRADIENT AORTIC STENOSIS 1813 Table 4. Impact of Aortic Valve Replacement on Mortality Model HR (95% CI) p Value Unadjusted model 0.237 (0.119 0.470 0.0001 Model adjusted for Age 0.275 (0.134 0.563 0.0004 Age and BMI 0.306 (0.143 0.653 0.0022 Age, BMI, and creatinine 0.279 (0.129 0.607 0.0013 EuroSCORE 0.266 (0.132 0.537 0.0002 Coronary artery disease 0.234 (0.117 0.470 0.0001 Coronary artery bypass graft 0.258 (0.113 0.591 0.0014 DM, vascular disease, COPD, and renal insufficiency a 0.220 (0.109 0.445 0.0001 a Creatinine level 130 mol/l. BMI body mass index; CI confidence interval; COPD chronic obstructive pulmonary disease; DM diabetes mellitus; EuroSCORE European System for Cardiac Operative Risk Evaluation; HR hazard ratio. though symptoms are relevant in the natural history of AS and influence the time of AVR. Other points of strength of our study are (1) the longer follow-up period of our study (median, 3.5 years), compared with others [5 7, 12], and (2) the comprehensive characterization of the patients. Although these patients should have indication for operation, they are not uncommonly overlooked in clinical practice because of the low TVG with preserved LVEF, and often the finding of a small AVA is considered inconsistent and therefore a measurement error. On the basis of our observations, patients in the grey zone, with an AVA of 0.8 to 1 cm 2, such as those reported by Minners and colleagues [13], also had a significant advantage from AVR. Thus, the cutoff of AVA of less than 0.8 cm 2 might be very specific but is not very sensitive for severe AS. Overall, our findings support the guidelines recommendation of a value for AVA of less than 1 cm 2 for defining severe AS and also patients with low TVG and severe AS but with a normal LVEF. The low TVG in our patients derived mainly from a reduced TV flow rate. In 26% of patients, this was related to a reduced SVI, reduced LV end-diastolic volume index, and increased valvular-arterial impedance. However, the remaining patients (74%) had a normal SVI but a significantly longer systolic ejection period. Of note, low SVI was not a predictor of death in our study, in contrast to others [4]. This discrepancy might be due to the inclusion of high TVG patients; indeed, our findings are consistent with the recent reanalysis of the Hachicha data by Dumesnil and colleagues [12], where survival was similar in low TVG patients irrespectively from the Fig 2. Kaplan-Meier estimates of the probability of survival are shown according to aortic valve replacement (AVR) in the (A) overall study population and in (B) propensity-matched patients according to (C) AVR with ( ) or without ( ) coronary artery bypass grafting (CABG).

1814 TARANTINI ET AL Ann Thorac Surg LOW-GRADIENT AORTIC STENOSIS 2011;91:1808 15 Fig 3. Changes in (A) New York Heart Association functional class symptoms (p 0.003) and (B) Canadian Classification Society Class symptoms (p 0.0001) in the 53 patients who underwent operations. SVI, and AVR remained a major protective index. Thus, our study confirms and extends the concept that AVA is the most important measure of AS severity and should always be calculated in the evaluation of AS, regardless of low TVG or the presence of lower or normal SVI. AVR-related factors were younger age and higher BMI. Sex was not an AVR predictor, and there was no sex prevalence in the AVR patients; therefore, the lower BMI in medically managed patients was due to patient characteristics. This is a retrospective study of consecutive patients admitted to a single center. The treatment assignment was not randomized but was by clinical decision. Although we used different models for confounder adjustment, we were not able to correct for unmeasured variables that might affect survival. Moreover, because of the small sample size, we might have underestimated the effect of CABG on survival in the operated-on group. We did not correct AVA for the body surface area to assess the severity of AS. However, the value of the indexed AVA in our study population was similar between patients treated or not by AVR. Inconsistent grading of the severity of AS may occur using data obtained both by catheterization and by echocardiography [13, 14], explaining the different prevalence of low TVG, normal LVEF, and severe AS observed among the studies. Notwithstanding, we found a good agreement between the Gorlin AVA and the Doppler AVA by the Bland-Altman test (data not shown, available on request). The use of the peak-to-peak TVG as surrogate of mean TVG might be suboptimal because the peak LV systolic pressure and the peak aortic pressure do not occur at the same point and may be influenced by systemic hemodynamic, especially hypertension. Nevertheless, the specific systolic blood pressure cutoff value influencing this relationship is not clear. By our data, 40% of the patients had a systolic blood pressure exceeding 160 mm Hg during catheterization. Although systemic arterial compliance was lower in patients with systolic pressure higher than 160 mm Hg compared with patients with a systolic blood pressure of less than 160 mm Hg (p 0.001), there were no differences of the TVG measured by echocardiogram and catheterization between the two groups (p 0.6). Thus, it is likely that only Table 5. Characteristics of Studies on Severe Aortic Stenosis/Low Transvalvular Gradient/Preserved Ejection Fraction Subgroup with Severe AS, Low TVG, Preserved EF Events, % First author (year) Pts No. No. AVA (cm 2 ) TVG (mm Hg) EF Evaluation AVR (%) Median F/U (years) AVR Control Symptoms, % Christensen 150 12 1.2 50 0.50 TTE, TEE, CC 25 2.5 100 30 100 (2004) [6] Hachicha a 512 199 0.6 b 30 0.50 TTE NA 1.5 NA NA NA (2007) [4] Barasch 215 47 0.6 b 30 0.50 TTE, CC 33 1.9 73 72 100 e (2008) [5] Pai (2008) [7] 740 52 0.8 30 0.55 TTE 35 2.4 90 c 20 c,d NA Dumesnil a 512 316 0.6 b 40 0.50 TTE 46 1.5 39 18 d NA (2009) [12] Tarantini (2011) [3] 102 102 1.0 30 0.50 TTE, CC 72 3.5 74 38 d 88 a Same study group. b AVA was indexed. c 5-year actuarial survival. d p 0.05 vs AVR group. e Data available in 24% of study population. AS aortic stenosis; AVA aortic valve area; AVR aortic valve replacement; CC cardiac catheterization; EF ejection fraction; F/U follow-up; NA not available; TEE transesophageal echocardiogram; TTE transthoracic echocardiogram; TVG transvalvular gradient.

Ann Thorac Surg TARANTINI ET AL 2011;91:1808 15 LOW-GRADIENT AORTIC STENOSIS 1815 a blood pressure level much higher than 160 mm Hg will change significantly the relationship between the peakto-peak assessment and the continuity equation, but this point needs further studies. References 1. Ross J Jr, Braunwald E. Aortic stenosis. Circulation 1968;38:61 7. 2. Bonow RO, Carabello BA, Kanu C, et al. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2006;48:e1 148. 3. Tarantini G, Buja P, Scognamiglio R, et al. Aortic valve replacement in severe aortic stenosis with left ventricular dysfunction: determinants of cardiac mortality and ventricular function recovery. Eur J Cardiothorac Surg 2003;24:879 85. 4. Hachicha Z, Dumesnil JG, Bogaty P, et al. Paradoxical low-flow, low-gradient severe aortic stenosis despite preserved ejection fraction is associated with higher afterload and reduced survival. Circulation 2007;115:2856 64. 5. Barasch E, Fan D, Chukwu EO, et al. Severe isolated aortic stenosis with normal left ventricular systolic function and low transvalvular gradients: pathophysiologic and prognostic insights. J Heart Valve Dis 2008;17:81 8. 6. Christensen KL, Ivarsen HR, Thuesen L, et al. Aortic valve stenosis: fatal natural history despite normal left ventricular function and low invasive peak-to-peak pressure gradients. Cardiology 2004;102:147 51. 7. Pai RG, Varadarajan P, Razzouk A. Survival benefit of aortic valve replacement in patients with severe aortic stenosis with low ejection fraction and low gradient with normal ejection fraction. Ann Thorac Surg 2008;86:1781 9. 8. Helmcke F, Nanda NC, Hsiung MC, et al. Color Doppler assessment of mitral regurgitation with orthogonal planes. Circulation 1987;75:175 83. 9. Perry GJ, Helmcke F, Nanda NC, et al. Evaluation of aortic insufficiency by Doppler color flow mapping. J Am Coll Cardiol 1987;9:952 9. 10. Schiller NB, Shah PM, Crawford M, et al. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989;2:358 67. 11. Razzolini R, Gerosa G, Leoni L, et al. Transaortic gradient is pressure-dependent in a pulsatile model of the circulation. J Heart Valve Dis 1999;8:279 83. 12. Dumesnil JG, Pibarot P, Carabello B. Paradoxical low flow and/or low gradient severe aortic stenosis despite preserved left ventricular ejection fraction: implications for diagnosis and treatment. Eur Heart J 2010;31:281 9. 13. Minners J, Allgeier M, Gohlke-Baerwolf C, et al. Inconsistencies of echocardiographic criteria for the grading of aortic valve stenosis. Eur Heart J 2008;29:1043 8. 14. Jander N. Low-gradient severe aortic stenosis with preserved ejection fraction: new entity, or discrepant definitions? Eur Heart J 2008;10:E-11 15. INVITED COMMENTARY Traditionally, the severity of aortic stenosis has been determined by the magnitude of the transvalvular pressure gradient. It is undisputed, that in case of a severely reduced left ventricular function, even a moderate transvalvular pressure gradient may qualify a patient for valve replacement, as the impaired ventricle is no longer capable of building up a greater pressure differential. But how do we determine left ventricular function? It is common practice to use the left ventricular ejection fraction (LVEF) as a marker of left ventricular performance. In most, but not all patients, it reflects systolic function well enough to serve as an acceptable surrogate marker. In addition, diastolic dysfunction, which may affect left ventricular function just as severely as systolic dysfunction, is only poorly assessed by LVEF. In summary, a patient can suffer from severe left ventricular dysfunction, despite a normal LVEF. Driven by the advances of echocardiography, which allows a fairly accurate measurement of the aortic valve orifice area, more and more patients have been identified with severe aortic stenosis and a low transvalvular pressure gradient, in the presence of a normal LVEF. Although the issue has been discussed among cardiologists for some time, it has been addressed in a number of studies only recently. It is the great merit of this excellent article by Tarantini and coworkers, that it is among the first to report the outcome of these patients after aortic valve replacement [1]. The study shows clearly the superior survival of patients who had undergone valve replacement in comparison to those who had not and gives a strong argument in favor of surgery for this particular group of patients. Furthermore, the article gives us a ball park figure how many patients with severe aortic stenosis may be affected. As it is the surgeon who takes ultimately responsibility for the operative indication, it is important to be aware of these facts and to advocate valve replacement where it is appropriate. Hans J. Geissler, MD, PhD Department of Cardiovascular Surgery Helios Klinikum Siegburg Ringstrasse 49 53721 Siegburg, Germany e-mail: hans-joachim.geissler@helios-kliniken.de Reference 1. Tarantini G, Covolo E, Razzolini R, et al. Valve replacement for severe aortic stenosis with low transvalvular gradient and left ventricular ejection fraction exceeding 0.50. Ann Thorac Surg 2011;91:1808 15. 2011 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2011.04.015