The aims of this study were to investigate the clinical outcomes of patients with low-gradient aortic stenosis despite preserved left ventricular ejection fraction and to assess reliable prognostic clinical-instrumental features in patients experiencing or not experiencing aortic valve replacement (AVR). Clinical-laboratory and echocardiographic data from 167 patients (median age 78 years, interquartile range 69 to 83) with aortic valve areas <1.0 cm 2 , mean gradients ≤30 mm Hg, and preserved left ventricular ejection fraction (≥55%), enrolled from 2005 to 2010, were analyzed. During a mean follow-up period of 44 ± 23 months, 33% of patients died. On multivariate analysis, independent predictors of death were baseline New York Heart Association functional class III or IV (hazard ratio 2.16, p = 0.038) and atrial fibrillation (hazard ratio 2.00, p = 0.025). Conversely, AVR was protective (hazard ratio 0.25, p = 0.01). The magnitude of the protective effect of AVR seemed to be relatively more important in patients with atrial fibrillation than in those in sinus rhythm, independently of the severity of symptoms. Age >70 years showed a trend toward being a prognostic predictor (p = 0.082). In conclusion, in patients with low-gradient aortic stenosis despite a preserved left ventricular ejection fraction, AVR was strongly correlated with a better prognosis. Patients with atrial fibrillation associated with advanced New York Heart Association class had the worst prognosis if treated medically but at the same time a relative better benefit from surgical intervention.
The objectives of the present study were to comprehensively assess clinical and echocardiographic characteristics of a large cohort of patients with low-gradient (LG) aortic stenosis (AS) and normal left ventricular (LV) ejection fractions (LVEF) and to identify reliable and feasible clinical-instrumental features capable of improving the clinical management of this challenging disease.
Methods
We analyzed our database to evaluate clinical, echocardiographic, and biohumoral data from patients with LG AS despite normal LVEF observed in our department from January 2005 to December 2010. Inclusion criteria were (1) aortic valve area (AVA; measured by the continuity equation) <1.0 cm 2 , (2) mean gradient (MG) ≤30 mm Hg, and (3) LVEF ≥55%. We arbitrarily decided to adopt strict enrollment criteria to minimize the burden of measurement errors. Moreover, for the same reason, the inclusion criteria had to be satisfied in ≥2 consecutive echocardiographic evaluations. We excluded patients in whom the LVEF and/or aortic transvalvular gradient and functional AVA could not be evaluated quantitatively because of inadequate echocardiographic windows, patients who had previously undergone aortic valve replacement (AVR), patients with severe aortic regurgitation, and patients with severe mitral regurgitation and/or significant mitral structural alteration (e.g., mitral stenosis, valvular endocarditis). Clinical, echocardiographic, and biohumoral data were accurately collected from our database and medical records.
Evaluation of clinical data included several variables: age, gender, biometric data, medical history (documented diagnosis of hypertension, hypercholesterolemia, diabetes, coronary artery disease, renal disease, or chronic obstructive pulmonary disease), symptomatic status (specifically the presence of dyspnea classified as absent/mild, New York Heart Association [NYHA] functional class I or II, or moderate or severe, NYHA functional class III or IV; angina; and syncope) and ongoing pharmacologic treatment. Moreover we collected patients’ electrocardiographic findings and laboratory biochemical profiles (hemoglobin, creatinine, glycated hemoglobin, triglycerides, total cholesterol, high-density lipoprotein cholesterol, aspartate transaminase, and alanine transaminase).
All Doppler echocardiographic examinations were performed or reviewed by a level 3 trained echocardiographer. M-mode and 2-dimensional echocardiographic measurements included LV dimensions, end-diastolic, end-systolic, and stroke volumes, and LVEF (with biplane method of discs from apical 4- and 2-chamber views), aortic annular dimensions, and left and right end-systolic atrial areas. LV diameters and volumes were indexed to body surface area. Significant LV hypertrophy was defined as interventricular septal thickness >1.2 cm.
Aortic flow velocity was measured by continuous-wave Doppler parallel to the stenotic jet using the best of multiple windows, and peak gradient and MG were calculated using the simplified Bernoulli equation according to current guidelines. Peak systolic flow velocity curve of the LV outflow tract was measured by pulsed-wave Doppler just below the aortic valve. AVA was calculated from the simplified continuity equation. Valvuloarterial impedance (Zva), a measure of global LV hemodynamic load, was calculated as Zva = (SBP + MG)/SVI, where SBP is the systolic arterial blood pressure, MG the mean transvalvular pressure gradient, and SVI the stroke volume indexed to body surface area. Mitral and aortic regurgitation were assessed semiquantitatively by using a polyparametric approach. All measurements were averaged on 3 to 5 consecutive beats.
For patients who underwent AVR, all surgical records were reviewed.
The primary end point of our analysis was overall mortality. Death was also classified as cardiac or noncardiac. The study was approved by the institutional review committee and was performed in accordance with the 2000 revised Declaration of Helsinki.
Summary statistics of clinical and instrumental variables are expressed as mean ± SD, medians and interquartile ranges (IQRs), or counts and percentage, as appropriate. Comparisons between groups were made by analysis of variance for continuous variables, using the Brown-Forsythe statistic when the assumption of equal variances did not hold, or by the nonparametric median test; chi-square statistics were calculated for discrete variables. Univariate Cox regression was then carried out to investigate the association of each of the observed variables with mortality. Starting from the list of statistically significant and clinically relevant variables, a multivariate Cox model was built, using the backward-conditional stepwise algorithm as a tool for covariate selection (p <0.10). Variables were entered on the original scale, without submitting them to any transformation or cutting off, except for age (see “ Results ”). Receiver-operating characteristic curve analysis was used to evaluate the global accuracy of the multivariate Cox model with or without age, and corresponding areas under the curve were computed and compared using Delong’s method. Finally, to present the final model estimates, a nomogram was built. The entire analysis was performed using the SPSS version 19 (SPSS, Inc., Chicago, Illinois) and R version 2.14.1 (R Foundation for Statistical Computing, Vienna, Austria).
Results
Over the considered 5-year period of analysis, within the cohort of 1,149 consecutive patients with severe AS (AVA <1 cm 2 ) observed in the echocardiography laboratory of our department, we retrospectively identified 397 patients (35%) with LG (≤30 mm Hg); among this group, we included 167 patients (15%) with LG AS and normal LVEF (≥55%), matching the above-mentioned enrollment criteria (see Figure 1 ).
Complete baseline clinical and laboratory data are listed in Table 1 . The median age of our population was 78 years, and 65% were women. Two-thirds of patients had history of hypertension, while 25% had diabetes and 52% had dyslipidemia. Coronary artery disease (defined as the presence of previous myocardial infarction and/or coronary artery revascularization) was present in 33% of cases. Among major extracardiac co-morbidities, chronic obstructive pulmonary disease, chronic renal insufficiency, and peripheral vascular disease were present in 10%, 16%, and 28% of patients, respectively. An anemic status (hemoglobin <13 g/dl in men and <12 g/dl in women) was present in 34% of patients. Two, 3, and 4 simultaneous co-morbidities were present in 21%, 10%, and 3.6% of patients, respectively (see Figure 2 ).
| Variable | Total (n=167) |
|---|---|
| Age (year) | 78 (69 – 83) |
| Females | 65% |
| Weight (Kg ± SD) | 71 ±13.5 |
| Height (m ± SD) | 1.64 ± 0.09 |
| Body mass index (Kg/m 2 ± SD) | 26.15 ± 4.1 |
| Body surface Area (m 2 ± SD) | 1.64 ± 0.37 |
| Hypertension (History) | 65% |
| Dyslipidaemia | 52.2% |
| Diabetes mellitus | 24.7% |
| Peripheral vascular disease | 28.1% |
| Chronic obstructive pulmonary disease | 9.7% |
| Renal Insufficiency ∗ | 16.3% |
| Coronary artery disease † | 33.2% |
| Revascularization | 17.8% |
| Prior myocardial infarction (STEMI or NSTEMI) | 15.8% |
| Angina pectoris | 31.8% |
| Syncope | 6% |
| Dyspnea (NYHA III – IV) | 15% |
| Haemoglobin (g/dL) | 12.9 (11.2 – 13.6) |
| Anaemia (n) ‡ | 34.1% |
| Creatinine (mg/dL) | 1.11 (0.88 – 1.22) |
| Glycated Haemoglobin (%) | 6.7 (5.9 – 7.5) |
| Triglycerides (mg/dL) | 105 (80 – 132) |
| Total Cholesterol (mg/dL) | 184 (157 – 218) |
| HDL Cholesterol (mg/dL) | 49 (41 – 58) |
| Aspartate transaminase (U/L) | 21 (18-26) |
| Alanine transaminase (U/L) | 17 (12-23) |
| Sinus rhythm, | 75.4% |
| Atrial Fibrillation, (persistent and permanent) | 21% |
| Paced | 3,6% |
| Left Bundle Branch Block | 7.6% |
| ECG Left ventricular hypertrophy | 22.7% |
| Therapy | |
| Acetylsalicylic acid | 56% |
| ACE inhibitor | 45.4% |
| Angiotensin receptor blocker | 18.6% |
| β-blocker | 41.4% |
| Nitroglycerin | 37.1% |
| Calcium-channel blocker | 32.9% |
| Digoxin | 16.4% |
| Diuretic | 51.1% |
| Antiarrhythmic drug § | 17.9% |
| Anticoagulant therapy | 24.1% |
| Statin | 41.4% |
| Allopurinol | 7,1% |
∗ Creatinine level >1.2 mg/dL.
† History of myocardial infarction and/or coronary artery revascularization.
‡ Hb in Men < 13 g/dL, women < 12 g/dL.
Symptoms attributable to AS were present in 73% of patients; moderate to severe dyspnea (NYHA class III or IV) was present in 15% of patients. Persistent or permanent atrial fibrillation (AF) on electrocardiography was present in 21% of patients (see Table 1 ).
Baseline echocardiographic data are listed in Table 2 . Median AVA and indexed AVA were 0.88 cm 2 and 0.51 cm 2 /m 2 , respectively. Peak gradient and MG were 38 and 22 mm Hg, respectively. The median LVEF of our population was 66%, the median LV end-diastolic diameter was 4.7 cm, the median LV end-diastolic volume index was 43 ml/m 2 , the median SVI was 28 ml/m 2 , and the median Zva was 5.7 mm Hg/(ml/m 2 ). Sixty-five percent of our patients had mild to moderate mitral regurgitation, and 79% had mild to moderate aortic regurgitation.
| Variable | Total (n=167) |
|---|---|
| Aortic valve area (cm 2 ) | 0.88 (0.8 – 0.93) |
| Indexed aortic valve area (cm 2 /m 2 ) | 0.51 (0.46-0.61) |
| Aortic peak jet velocity (m/sec) | 3.20 (2.9 – 3.39) |
| Peak transvalvular pressure gradient (mmHg) | 38 (29 – 44) |
| Mean transvalvular pressure gradient (mmHg) | 22 (16 – 28) |
| Aortic annulus (cm) | 2 (1.8 – 2.1) |
| Left ventricle ejection fraction (%) | 66 (62 – 71) |
| Left ventricle end-diastolic diameter (cm) | 4.7 (4.3 – 5.1) |
| Left ventricle end-systolic diameter (cm) | 2.6 (2.3 – 3.2) |
| Left ventricle end-diastolic volume (ml) | 68 (52 – 86) |
| Indexed Left ventricle end-diastolic volume (ml/m 2 ) | 43 (35– 53) |
| Left ventricle end-systolic volume (ml) | 22 (17 – 30) |
| Interventricular septum thickness (cm) | 1.3 (1 – 1.5) |
| Left atrial area (cm 2 ) | 26 (21 – 32) |
| Mild to moderate mitral regurgitation (n) | 65% (108) |
| Mild to moderate aortic regurgitation (n) | 79% (132) |
| Stroke volume (ml) | 45 (36-57) |
| Indexed stroke volume (ml/m 2 ) | 28 (23 – 35) |
| Valvulo-arterial impedance, mmHg/(ml/m 2 ) | 5.7 (4.3 – 7.2) |
| Pulmonary arterial pressure (mmHg) | 32 (27 – 42) |
After diagnosis, 43 patients (26%) underwent AVR, 16 (37%) of them simultaneously underwent coronary artery bypass grafting, and 8 (18.6%) had mitral valve surgery as well upon surgical intraoperative decision, even if nonsevere mitral disease was documented on preoperative echocardiography. The operations confirmed significant AS in all patients. Perioperative mortality was 2.3% (1 case of death within 30 days after AVR). Patients who underwent AVR with respect to those treated medically were significantly younger (72 vs 79 years, p <0.001) and had higher body weight, height, body surface area, body mass index, and and hemoglobin levels. Evaluating co-morbidities, operated patients with respect to those treated medically had a lower profile (mean number of co-morbidities 0.8 ± 0.9 vs 1.3 ± 1.14 in operated and nonoperated patients, respectively, p = 0.03). On Doppler echocardiographic assessment, non–clinically significant differences were found (see Table 3 ). At 1-year echocardiographic follow-up in operated patients, we observed a trend in the reduction of interventricular septal thickness (median 1.4 cm [IQR 1.2 to 1.5] to 1.3 cm [IQR 1.1 to 1.35], p = 0.06).
| Variable | Aortic Valve Replacement | P-value | |
|---|---|---|---|
| YES (n=43) | NO (n=124) | ||
| Age (years) | 72 (65-78) | 79 (72-84) | 0.003 |
| Females | 56% | 68% | 0.159 |
| Weight (kg ± SD) | 76,3± 14.9 | 69.4±12.5 | 0.004 |
| Height (m ± SD) | 1.67±0.9 | 1.64±0.9 | 0.05 |
| Body mass index (Kg/m 2 ± SD) | 27.1±4 | 25.7±4.1 | 0.06 |
| Body surface Area (m 2 ± SD) | 1.78±0.41 | 1.59±0.34 | 0.003 |
| Hypertension (history) | 58.1% | 68.5% | 0.21 |
| Diabetes mellitus | 30.6% | 22.7% | 0.34 |
| Peripheral vascular disease | 29.7% | 27.5% | 0.79 |
| Chronic obstructive pulmonary disease | 8.3% | 10.2% | 0.74 |
| Dyslipidaemia | 63.9% | 48.6% | 0.11 |
| Renal Insufficiency ∗ | 11.4% | 18% | 0.36 |
| Prior myocardial revascularization | 16.7% | 18.2% | 0.84 |
| Atrial Fibrillation | 18.6% | 21.8% | 0.66 |
| Anaemia † | 16.7% | 41% | 0.01 |
| Haemoglobin (g/dL) | 13.3 (12.5-14.2) | 12.6 (11.25-13.7) | 0.047 |
| Angina | 51.4% | 25.4% | 0.03 |
| Syncope | 6.9% | 4.8% | 0.27 |
| Dyspnea (NYHA I-II) | 83.8% | 85.1% | 0.85 |
| Dyspnea (NYHA III-IV) | 16,2% | 14.9 % | 0.85 |
| Left ventricle end-diastolic diameter (cm) | 4.8 (4.6-5) | 4.7 (4.3-5.2) | 0.42 |
| Indexed left ventricle end-diastolic volume (ml/m 2 ) | 40 (34-53) | 43 (35-53) | 0.45 |
| Aortic valve area (cm 2 ) | 0.89 (0.8-0.93) | 0.88 (0.8-0.93) | 0.84 |
| Peak transvalvular pressure gradient (mmHg) | 41 (38-46) | 38 (34-46) | 0.24 |
| Mean transvalvular pressure gradient (mmHg) | 21 (14-17) | 22 (17.2-26) | 0.13 |
| Indexed stroke volume (ml/m2) | 28.2 (23-35) | 29.4 (23-34) | 0.99 |
| Valvulo-arterial impedance mmHg/(ml/m 2 ) | 5.5 (5-7.3) | 5.7 (4.2-7.2) | 0.57 |
| Aortic annulus (cm) | 2 (1.9-2.1) | 2 (1.8-2.0) | 0.21 |
| Interventricular septum thickness (cm) | 1.4 (1.2-1.5) | 1.3 (1.1-1.5) | 0.15 |
| Left ventricle ejection fraction (% ) | 68 (64-75) | 66 (61-71) | 0.01 |
| Left atrial area (cm 2 ) | 26 (20-35) | 25 (21-32) | 0.57 |
| Moderate mitral regurgitation (n) | 6.9% (3) | 15.3% (19) | 0.33 |
| Moderate aortic regurgitation (n) | 27.9% (12) | 20% (25) | 0.12 |
| Pulmonary arterial pressure (mmHg) | 36 (31-48) | 32 (26-41) | 0.15 |
| Coronary artery bypass graft | 37.2% | 0 | |
| Mitral valve surgery | 18.6% | 0 | |
∗ Creatinine level > 1.2 mg/dL.
During a mean follow-up period of 44 ± 23 months, 54 patients (32%) died, 35 (65%) of them of cardiovascular-related causes. Patients who died during the observation period were older (81 years [IQR 74 to 85] vs 76 years [IQR 67 to 82], p = 0.051) and more frequently women (76% vs 60%, p = 0.04). Moreover, they presented a higher prevalence of persistent or permanent AF at enrollment (33% vs 17%, p = 0.01) and a larger left atrial area (27.5 vs 25 cm 2 p = 0.014). With respect to NYHA functional class and co-morbidities (mean number of co-morbidities 1.10 ± 1.04 vs 1.3 ± 1.17 in dead and living patients, respectively) no significant differences were found. Univariate predictors of death in our population are listed in Table 4 . History of systemic hypertension, diabetes, and/or dyslipidemia was not related to propensity for surgery or clinical outcomes on univariate analysis. On multivariate Cox modeling, the predictive factors independently associated with all-cause mortality were NYHA functional class III or IV (hazard ratio [HR] 2.16, 95% confidence interval [CI] 1.05 to 4.57, p = 0.038) and presence of AF (HR 2.00, 95% CI 1.09 to 3.68, p = 0.025). Conversely, AVR was associated with significantly better survival (HR 0.25, 95% CI 0.09 to 0.71, p = 0.01) (see Table 4 ). Age showed a trend toward being a prognostic predictor, only after 70 years (HR for each decade increase after 70 years 1.02, 95% CI 1.0004 to 1.04, p = 0.082). Receiver-operating characteristic curves in Figure 3 show the significantly better global accuracy of the prognostic model that included age with respect to the same model without age (area under the curve 0.74 vs 0.69, p = 0.018).
