Background
Patients with severe aortic stenosis (AS) are known to have increased left ventricular apical rotation (ApRot) during systole, but its clinical relevance is unknown. The aim of this study was to assess the association of ApRot with patient symptoms and total mortality.
Methods
A retrospective analysis was performed on 82 patients (mean age, 77 ± 14 years; 40% men) with newly diagnosed severe AS with indexed aortic valve areas ≤ 0.6 cm 2 /m 2 and left ventricular ejection fractions ≥ 50%. Sixty-three percent of patients were symptomatic. ApRot was calculated using speckle-tracking echocardiography. Patients were divided into two groups on the basis of ApRot: high ApRot (>4.0°, n = 41) and low ApRot (≤4.0°, n = 41).
Results
There were 33 deaths and 30 aortic valve replacement procedures after 33 ± 17 months of follow-up. Patients in the high-ApRot group had smaller indexed aortic valve areas ( P = .021) and increased valvuloarterial impedance ( P = .014). There was no difference in overall symptoms, but the low-ApRot group experienced more syncope ( P = .020). Patients in the high-ApRot group had reduced survival with medical therapy (log-rank P = .018) after aortic valve replacement (log-rank P = .039) and overall (log-rank P = .009). Asymptomatic patients with low ApRot had the best survival, while asymptomatic patients with high ApRot had similar survival to that of symptomatic patients (log-rank P = .008). On adjusted Cox regression, ApRot ≥ 6.0° was independently associated with death (hazard ratio, 3.06; P = .003). On receiver operating characteristic curve analysis, ApRot added incremental prognostic value to indexed aortic valve area, symptom status, and aortic valve replacement status.
Conclusion
Increased ApRot is independently associated with poor survival and may represent a compensatory mechanism to preserve cardiac output against severe obstruction to flow and high systolic load.
Highlights
- •
Patients with severe AS and increased ApRot have worse survival.
- •
Asymptomatic patients with high ApRot have similar outcomes to those of symptomatic patients.
- •
Increased ApRot in independently associated with mortality in patients with severe AS.
In patients with severe aortic stenosis (AS), left ventricular (LV) myocardial deformation measured by speckle-tracking echocardiography (STE) has emerged as a reliable measure of subtle LV systolic dysfunction. Longitudinal strain has been the most thoroughly investigated of these parameters. Basal longitudinal strain impairment has been shown to be associated with symptoms, while global longitudinal strain (GLS) impairment has been shown to be associated with symptoms, likelihood of symptom recovery after aortic valve implantation, and overall survival. Another parameter of myocardial deformation that can be measured by STE, and is receiving increasing attention, is apical rotation (ApRot). Previous studies have shown ApRot to be increased in patients with severe AS, but much less is known about the association of ApRot with symptoms and survival in these patients. Increased ApRot in patients with severe AS was associated with chest pain or electrocardiographic LV strain pattern in one study, but another study reported decreased ApRot to be more associated with the presence of symptoms overall. The association between ApRot and survival in severe AS is yet to be reported. ApRot is key to cardiac performance, and understanding the interplay between ApRot, symptoms, and survival is key to defining a comprehensive model of myocardial deformation and its impact on clinical outcomes. Therefore, this study was undertaken to assess ApRot in patients with severe AS and determine its association with symptoms and survival.
Methods
Patient Population
A retrospective review was performed of our institution’s electronic echocardiography database and electronic health record for all patients referred for transthoracic echocardiography in 2010 and found to have indexed aortic valve area (AVA i ) ≤ 0.6 cm 2 /m 2 and LV ejection fractions (LVEFs) ≥ 50%. Three hundred thirty-eight patients met these criteria. Patients were excluded if they were diagnosed previously with severe AS, were in atrial fibrillation during echocardiography, had regional wall motion abnormalities, had any congenital heart lesion, or had echocardiograms of insufficient quality ( Figure 1 ). The final study cohort consisted of 82 patients with newly diagnosed severe AS and preserved LVEFs. Clinical data for patients were obtained from chart review. Severe, life-limiting noncardiac diseases were also recorded. Coronary heart disease was defined as >70% stenosis on coronary angiography, positive noninvasive stress test results, previous myocardial infarction, or previous coronary revascularization. Heart failure with preserved ejection fraction (HFpEF) was identified from International Classification of Diseases, Ninth Revision, codes on prior discharge summaries as the reason for admission and confirmed with chart review.
Echocardiographic Data
Aortic valve area (AVA) was determined using the continuity equation. LVEF was determined using the Simpson biplane method. Stroke volume was determined from the product of LV outflow tract area and the velocity-time integral. Relative wall thickness (RWT) and indexed LV mass (LVM i ) were determined by using the following equations : RWT = 2 × LVPWd/LVEDD and LVM i (g/m 2 ) = (0.8{1.04[(IVSd + LVPWd + LVEDD) 3 − (LVEDD)]} + 0.6)/BSA, where IVSd is interventricular septal thickness in diastole, LVPWd is LV posterior wall thickness at end-diastole, LVEDD is LV end-diastolic diameter, and BSA is body surface area. Total LV hydraulic load was calculated from valvuloarterial impedance as Z va = (MG + SAP)/SV i , where SAP is systolic arterial pressure, MG is the mean transvalvular pressure gradient, and SV i = indexed stroke volume. Energy loss index (ELI) was calculated as [(ELI = aortic area × AVA)/(aortic area − AVA)]/BSA.
STE
Two-dimensional echocardiographic images were obtained with Philips iE33 systems (Philips Medical Systems, Andover, MA) according to the Intersocietal Accreditation Commission Echocardiography–approved protocol ( http://www.intersocietal.org/echo ). A study of sufficient quality was defined as permitting strain analysis in 90% of all segments of the apical two-chamber, three-chamber, and four-chamber views, the parasternal short-axis basal, middle, and apical views, as well as 100% of all apical segments in the short-axis apical view, at a minimum of 40 frames/sec. Short-axis apical views were typically located distal to the papillary muscles and proximal to the obliteration of the LV cavity. Strain analysis was performed by two blinded readers. Both readers performed duplicate strain analyses for 16 randomly selected patients to enable analysis of reproducibility. Strain analysis was performed using QLAB version 10.2 (Philips Medical Systems). One complete cardiac cycle was defined by using electrocardiography–echocardiography synchronization and selecting the first end-diastolic frame of the P-R segment. In all views, a region of interest was defined by manually demarcating the borders of the endocardium and epicardium. The software then segmented the myocardium within this region of interest. Careful review and adjustment of the region of interest were performed manually to ensure the optimal myocardial tracking. Peak GLS and global longitudinal strain rate were calculated from averaging all segments of the four-chamber, three-chamber, and two-chamber views. Peak global circumferential strain (GCS) and global circumferential strain rate were calculated by averaging all segments of the short-axis apical, middle, and basal views. Peak ApRot and basal rotation were calculated from the short-axis apical and basal views, respectively, along the midwall of the myocardium. Counterclockwise rotation was considered positive in accordance with engineering notation.
Clinical End Points
The primary end point was all-cause mortality. Surgical aortic valve replacement (AVR) was defined as a secondary end point (our institution did not perform transcatheter AVR until July 2012). Outcome data were obtained from our institution’s electronic medical record. When a patient’s status could not be confirmed, the patient, or his or her documented health care agent, was contacted via telephone. This study was approved by the Institutional Review Board of the Albert Einstein College of Medicine and Montefiore Medical Center.
Statistical Analysis
Study subjects were divided into two groups on the basis of the median ApRot for the sample population and compared on the basis of their baseline characteristics, symptom status, and survival. Continuous variables are reported as mean ± SD. Categorical variables are reported as number (percentage). Distribution of continuous variables was tested using the D’Agostino K test with Royston’s revision. Comparison of continuous variables between the two patient groups was performed using the two-tailed Student’s t test or the Mann-Whitney U test as appropriate. Comparison of categorical variables was performed using the two-tailed Fischer exact test. Comparative survival analysis was performed using the Kaplan-Meier test. To assess the association between ApRot and total mortality, a regression analysis was performed using a Cox proportional hazard model. Univariate analysis was performed first using all patient variables. Then multivariate analysis was performed using the variables found to be significant in the univariate model. To enable direct comparison of hazard ratios, continuous variables were dichotomized according to their optimal cutoff values on the basis of their Youden J statistics, derived from receiver operating characteristic (ROC) analysis. An ROC analysis was performed to assess the incremental information of ApRot over established predictors of all-cause mortality, including AVA i , symptoms, and AVR status. To quantify test reproducibility and absolute agreement between measurements, interobserver and intraobserver interclass correlation coefficients (ICC) were calculated for all strain parameters. Statistical analysis was performed using Stata 12 (StataCorp LP, College Station, TX). Statistical significance was defined as P < .05.
Results
Study Population
Eighty-two patients with newly diagnosed severe AS and preserved LVEFs constituted the study population. Their average age was 77 ± 14 years, and 40% were men. The mean ApRot was 5.1 ± 3.5° (range, −2.0° to 15.7°), and the median ApRot was 4.1°. The cohort was divided into two groups on the basis of median ApRot: high ApRot (>4.0°, n = 41) and low ApRot (≤4.0°, n = 41). There was no significant difference in age, comorbidity profile, or overall rate of symptoms between the two groups ( Table 1 ). However, patients in the low-ApRot group experienced more syncope (29% vs 7%, P = .020). Forty-nine patients were excluded because of lack of an adequate short-axis apical view, and they were not significantly different in age, AVA i , MG, SV i , LVEF, survival (log-rank P = .537), or time to AVR (log-rank P = .938) compared with the study population.
| Variable | Total ( n = 82) | ApRot ≤ 4.0° ( n = 41) | ApRot > 4.0° ( n = 41) | P |
|---|---|---|---|---|
| Demographics | ||||
| Age (y) | 77 ± 14 | 75 ± 13 | 78 ± 15 | .255 |
| Men | 33 (40%) | 15 (37%) | 18 (44%) | .635 |
| BSA (m 2 ) | 1.84 ± 0.22 | 1.84 ± 0.25 | 1.84 ± 0.19 | .977 |
| Comorbidities | ||||
| Hypertension | 70 (85%) | 37 (90%) | 33 (80%) | .349 |
| Hyperlipidemia | 47 (57%) | 24 (59%) | 23 (56%) | .999 |
| Diabetes mellitus | 37 (45%) | 20 (49%) | 17 (42%) | .657 |
| Coronary heart disease | 27 (33%) | 15 (37%) | 12 (30%) | .639 |
| HFpEF | 9 (11%) | 5 (12%) | 4 (10%) | .999 |
| History of AF | 18 (22%) | 6 (15%) | 12 (30%) | .181 |
| Symptoms | ||||
| Symptomatic | 52 (63%) | 28 (68%) | 24 (59%) | .492 |
| Angina | 23 (28%) | 8 (20%) | 15 (36%) | .139 |
| Syncope | 15 (18%) | 12 (29%) | 3 (7%) | .020 |
| Dyspnea | 18 (22%) | 9 (22%) | 9 (22%) | .999 |
Echocardiography and Strain Analysis
The study population had a mean AVA i of 0.45 ± 0.11 cm 2 /m 2 , a mean MG of 36 ± 16 mm Hg, and a mean LVEF of 66 ± 9% ( Table 2 ). Patients in the high-ApRot group were found to have significantly smaller AVA i (0.42 ± 0.11 vs 0.47 ± 0.11 cm 2 /m 2 , P = .021) and significantly lower ELI (0.48 ± 0.14 vs 0.55 ± 0.14 cm 2 /m 2 , P = .046). There was no difference in MG or peak gradient. LV dimensions, including LVPWd, IVSd, LVEDD, LV volumes, and LVM i , were also similar between study groups. Patients with high ApRot had significantly higher Z va (5.03 ± 1.09 vs 4.29 ± 1.09 mm Hg · m 2 /mL, P = .014).
| Variable | Total ( n = 82) | ApRot ≤ 4.0° ( n = 41) | ApRot > 4.0° ( n = 41) | P |
|---|---|---|---|---|
| Aortic valve | ||||
| AVA (cm 2 ) | 0.81 ± 0.21 | 0.86 ± 0.19 | 0.76 ± 0.21 | .026 |
| AVA i (cm 2 /m 2 ) | 0.45 ± 0.11 | 0.47 ± 0.11 | 0.42 ± 0.11 | .021 |
| MG (mm Hg) | 36 ± 16 | 35 ± 15 | 37 ± 17 | .468 |
| Peak gradient (mm Hg) | 62 ± 29 | 57 ± 23 | 66 ± 33 | .183 |
| ELI (cm 2 /m 2 ) | 0.51 ± 0.14 | 0.55 ± 0.14 | 0.48 ± 0.14 | .046 |
| Left ventricle | ||||
| IVSd (cm) | 1.1 ± 0.2 | 1.1 ± 0.2 | 1.2 ± 0.2 | .608 |
| Posterior wall diameter (cm) | 1.1 ± 0.2 | 1.1 ± 0.2 | 1.1 ± 0.2 | .862 |
| LVEDD (cm) | 4.9 ± 3.5 | 4.4 ± 0.6 | 5.5 ± 0.7 | .300 |
| End-diastolic volume (mL) | 92 ± 29 | 91 ± 30 | 94 ± 29 | .704 |
| End-systolic volume (mL) | 32 ± 13 | 33 ± 14 | 31 ± 13 | .632 |
| LVEF (%) | 66 ± 9 | 64 ± 9 | 67 ± 9 | .231 |
| SV i (LVOT) (mL/m 2 ) | 37 ± 8 | 39 ± 8 | 35 ± 9 | .056 |
| RWT | 0.51 ± 11 | 0.52 ± 0.12 | 0.51 ± 0.10 | .700 |
| LVM i (g/m 2 ) | 100 ± 27 | 99 ± 30 | 101 ± 26 | .687 |
| Afterload | ||||
| Z va (mm Hg · m 2 /mL) | 4.65 ± 1.09 | 4.29 ± 1.09 | 5.03 ± 1.09 | .014 |
High-quality STE was achievable in 2,910 of 3,034 myocardial segments (96%). Per protocol, 100% of all apical segments were suitable for speckle-tracking analysis. The mean ApRot was 5.1 ± 3.5°, mean basal rotation was −3.8 ± 2.8°, mean GLS was −18.8 ± 3.7°, and mean GCS was −24.7 ± 6.3° for the study population. No difference in GLS, GCS, or basal rotation was detected between study groups ( Table 3 ).
| Variable | Total ( n = 82) | ApRot ≤ 4.0° ( n = 41) | ApRot > 4.0° ( n = 41) | P |
|---|---|---|---|---|
| GLS (%) | −18.8 ± 3.7 | −18.3 ± 2.8 | −19.3 ± 3.8 | .204 |
| Global longitudinal strain rate (sec −1 ) | −1.12 ± 0.22 | −1.11 ± 0.23 | −1.12 ± 0.22 | .771 |
| GCS (%) | −24.7 ± 6.3 | −24.8 ± 6.2 | −24.6 ± 5.6 | .909 |
| Global circumferential strain rate (sec −1 ) | −1.63 ± 0.47 | −1.57 ± 0.46 | −1.69 ± 0.47 | .239 |
| ApRot (°) | 5.1 ± 3.5 | 2.1 ± 1.2 | 7.4 ± 3.0 | — |
| Basal rotation (°) | −3.8 ± 2.8 | −3.5 ± 2.8 | −4.0 ± 2.8 | .463 |
ApRot and Survival
During a mean follow-up period of 33 ± 17 months, there were 33 deaths and 30 surgical AVR procedures. Two patients were lost to follow-up, one of whom underwent AVR. Coronary angiography was performed on all patients who underwent AVR and showed no suitable targets for coronary artery bypass grafting. Similarly, no patient underwent concurrent mitral valve replacement. There were significantly more deaths in the high-ApRot group (54% vs 29%, P = .022) and an equal number of AVR procedures. Patients who died were significantly older (82 vs 72 years, P = .001), had higher ApRot (6.1 ± 4.2° vs 3.8 ± 2.7°, P = .010), and had larger left ventricles (LVM i = 108 ± 27 vs 95 ± 28 g/m 2 , P = .034). Nineteen patients were found to have significant life-limiting noncardiac disease, including end-stage renal disease, cirrhosis, dementia, and a diagnosis of previous or active cancer. There was no difference in the prevalence of these severe noncardiac diseases between study groups ( P = .980), between patients who died and survivors ( P = .213), or between symptomatic and asymptomatic patients ( P = .787). On overall survival analysis, patients with high ApRot were found to have worse survival than patients with low ApRot (log-rank P = .009) ( Figure 2 A). Time to AVR was similar between the two groups (log-rank P = .135) ( Figure 2 B). In AVR recipients, patients with high ApRot had worse survival (log-rank P = .039) ( Figure 2 C), as did patients with high ApRot who were managed with optimal medical therapy (log-rank P = .018) ( Figure 2 D).
On univariate regression analysis, age, dyspnea, HFpEF, LVM i , and ApRot were found to be independently associated with mortality ( Table 4 ). On multivariate regression analysis, after adjusting for age and HFpEF, ApRot ≥ 6.0° was significantly associated with death (hazard ratio, 3.06; P = .003), while a diagnosis of HFpEF had the strongest association with death (hazard ratio, 6.86; P = .001). Dyspnea and LVM i were considered colinear to a diagnosis of HFpEF and were therefore excluded from the multivariate model.
