The aim of the present study was to evaluate whether assessment of stroke volume index (SVI) can be used to improve risk stratification among patients with low-gradient severe aortic stenosis and preserved ejection fraction (EF). Study population comprised 409 patients with aortic valve area ≤1.00 cm 2 , mean gradient <40 mm Hg, and a normal EF (≥50%) who were followed up in a tertiary referral center from 2004 to 2012. Echocardiographic parameters and clinical data were collected. Multivariate Cox proportional hazards regression modeling was used to evaluate the association between SVI and the risk of all-cause mortality. Mean age of study patients was 78 ± 11 years, and 42% were men. The mean SVI was 39 ± 7 ml/m 2 (tertile 1 = 32 ± 4 ml/m 2 ; tertile 2 = 39 ± 1 ml/m 2 ; tertile 3 = 47 ± 4 ml/m 2 ). Multivariate analysis showed that the SVI was the most powerful echocardiographic parameter associated with long-term outcome: each 5 ml/m 2 reduction in SVI was associated with a 20% increase in adjusted mortality risk (p = 0.01). Consistently, Kaplan-Meier analysis showed that the cumulative probability of survival during 3 years of follow-up was 60%, 72%, and 73% among patients in the low-, intermediate-, and high-SVI groups, respectively (p = 0.012). Our findings suggest that in patients with low-gradient severe aortic stenosis and preserved EF, there is a graded inverse relation between SVI and the risk of long-term mortality.
Severe aortic stenosis (AS) is a common valvular heart disease, which carries a poor prognosis in symptomatic patients. Severe AS is defined by aortic valve area (AVA) ≤1 cm 2 , indexed AVA <0.6 cm 2 , and mean pressure gradient ≥40 mm Hg. However, reduced left ventricular (LV) ejection fraction (EF) and reduced transvalvular flow (with preserved EF) are conditions associated with low-gradient severe AS (a mean pressure gradient <40 mm Hg). This group comprise up to 30% of the severe AS population. The echocardiographic confirmation of low-gradient severe AS with preserved EF is challenging as discrepancy between pressure gradient and AVA occurs. Furthermore, recent studies showed conflicting data regarding the outcome of patients with severe AS who present with low-gradient and reduced stroke volume index (SVI; now termed “paradoxical” low-flow low-gradient severe AS); although some studies suggest that this group suffers from an advanced disease and poor clinical outcome, others suggest that this population has a similar prognosis to moderate AS. These studies used a strict SVI cutoff of ≤35 ml/m 2 , whereas currently there are limited data on risk stratification of the wider population of patients with low-gradient severe AS and preserved EF. We hypothesized that the degree of reduction in SVI may provide incremental prognostic information to other echocardiographic parameters in patients with low-gradient severe AS and preserved EF.
Methods
Echocardiographic and Doppler studies of patients with severe AS and preserved EF were reviewed from Sheba Medical Center echocardiography (a tertiary referral center) database, comprising patients who were referred to the echocardiography laboratory during the years 2004 to 2012 and were followed up for data regarding aortic interventions and long-term mortality. Inclusion criteria included AVA ≤1 cm 2 , mean gradient <40 mm Hg, and LV EF ≥50%. Exclusion criteria included any other significant valvular disease. Patients with moderate, moderate-to-severe, and severe aortic regurgitation were excluded. The study was approved by the respective human research review board of our institute.
Two-dimensional transthoracic echocardiographic and Doppler studies were obtained with clinical ultrasound machines equipped with 3.5-MHz transducers using standard views. The studies were digitally stored (McKesson’s Horizon Cardiology Medical Software, Tel Aviv, Israel). Parasternal long-axis view was used for measuring the aortic annulus diameter in early systole. Pulsed Doppler in the LV outflow tract (LVOT) from the apical window allowed evaluating flow. This represents an average of 3 cardiac cycles in patients with sinus rhythm and 6 cardiac cycles for patients with atrial fibrillation. Continuous-wave Doppler recording of flow through the aortic valve was performed from the apical, right parasternal, suprasternal, and subcostal windows to minimize the effect of Doppler angulation with flow. LV stroke volume (SV) was derived using the time-velocity integral of LVOT, assuming a circular geometry of the LVOT. An indexed stroke volume was calculated as SV divided to body surface area (BSA). Multiplying heart rate to SV allowed calculating cardiac output; indexed cardiac output to BSA was calculated as well. AVA was derived from the continuity equation; an indexed AVA to BSA was than calculated. Using the continuous wave jet recording, peak and mean velocity and time-velocity integral were measured. LV function was estimated by the reader. Using the Simpson method (single plane), LV end-diastolic volume, end-systolic volume, SV, and EF were estimated. These were available in 340 patients.
Clinical data were obtained from computerized patient files; age and gender were recorded, as well as height and weight. The diagnosis of hypertension, hyperlipidemia, diabetes mellitus, renal failure, history of smoking, ischemic heart disease, and cerebrovascular disease was documented. Intervention was defined as surgical or transcatheter aortic valve replacement. This was obtained based on the hospital intervention database and patient files. Complete clinical data were available for 87% of the study population. The primary end point of the present study was all-cause mortality. Evaluation of the patient survival status was performed in December 2012 using the Israeli Ministry of Interior National Registry and was confirmed for all patients.
Categorical data were compared using the chi-square test or Fisher’s exact test. Continuous data were compared with Student t test and 1-way analysis of variance. In the primary analysis, the SVI was assessed as a continuous measure and categorized into approximate tertiles. Cut-off points were 37 and 42 ml/m 2 . In a secondary analysis, the SVI was also dichotomized at 35 ml/m 2 , as previously reported. Data by SVI tertiles are presented as mean ± SD or as counts with proportions (%). The Kaplan-Meier method was used to compare cumulative survival. Survival was compared among the 3 tertiles, and the overall difference was compared by the log-rank test. In this analysis, patients were censored on aortic valve intervention to avoid a possible survival bias associated with aortic valve replacement. Univariate and multivariate Cox proportional hazards models were used to calculate hazard ratio for survival. The multivariate model included adjustment for age, gender, ischemic heart disease, body mass index (BMI), as well as aortic valve intervention as a time-dependent covariate.
Forward stepwise regression modeling was used to identify echocardiographic parameters that were independently associated with all-cause mortality. Candidate prespecified echocardiographic covariates included in the model were LV hypertrophy (interventricular septum thickness >12 mm), estimated LV mass (>96 g for women and >116 g for men), AVA (≤0.8 cm 2 ), SVI (for each 5 ml/m 2 decrease), left atrium enlargement (>20 cm 2 ), cardiac index (<2.73 L/min/m 2 [median]), and LV diastolic dimensions (<4.6 cm [median]). Additional prespecified covariates in the multivariate models included age (1-year incremental), gender, ischemic heart disease, BMI, and aortic valve intervention, assessed as a time-dependent covariate. Statistical significance was accepted for a 2-sided p <0.01. The statistical analyses were performed with IBM SPSS, version 20.0 (Chicago, Illinois).
Results
A total of 409 patients were included in this study; baseline clinical characteristics of all patients and by SVI tertiles are listed in Table 1 . Compared with the high-SVI tertile, patients in the lower and intermediate-SVI tertiles had greater BMI and BSA and a greater frequency of women ( Table 1 ). No differences in risk factors among the tertiles were noted.
| Parameter | All (N = 409) | Low Tertile (N = 136) | Mid Tertile (N = 137) | High Tertile (N = 136) |
|---|---|---|---|---|
| Male ∗ | 172 (42%) | 62 (46%) | 67 (49%) | 43 (32%) |
| Age (years) | 78 ± 11 | 79 ± 10 | 76 ± 12 | 78 ± 10 |
| Body mass index (kg/m 2 ) ∗ | 28 ± 5 | 30 ± 7 | 28 ± 4 | 27 ± 4 |
| Body surface area (m 2 ) ∗ | 1.8 ± 0.19 | 1.85 ± 0.22 | 1.82 ± 0.18 | 1.70 ± 0.15 |
| Diabetes mellitus | 132 (37%) | 45 (39%) | 47 (39%) | 40 (33%) |
| Hypertension | 245 (69%) | 78 (67%) | 80 (66%) | 87 (75%) |
| Dyslipidemia | 176 (49%) | 53 (46%) | 56 (46%) | 67 (56%) |
| Active smokers | 22 (6%) | 7 (6%) | 10 (8%) | 5 (4%) |
| Chronic renal failure | 82 (23%) | 35 (30%) | 23 (19%) | 24 (20%) |
| Ischemic heart disease | 164 (46%) | 57 (50%) | 56 (46%) | 51 (43%) |
| Cerebrovascular disease | 65 (18%) | 27 (24%) | 23 (19%) | 15 (13%) |
SVI was normally distributed among study population; mean SVI was 39 ± 7 ml/m 2 ( Figure 1 ). There were 110 patients (27%) with SVI ≤35 ml/m 2 and 299 patients (73%) with SVI >35 ml/m 2 . The echocardiographic characteristics of the study population by SVI tertiles are summarized in Table 2 . Patients among all 3 SVI tertiles had similar LV dimensions, LV mass (and LV mass index), and right ventricle systolic pressure. Similarly, LV EF was comparable, despite a statistically significant difference among the 3 groups. However, patients with low SVI presented with a significantly smaller AVA, lower aortic valve mean gradient, and lower end-diastolic volume. Estimation of SV and EF with the Simpson method has shown good correlation with the SV measured by the continuity equation (r = 0.75, p <0.001) and with the EF estimated by the reader (r = 0.35, p <0.0001).
| Parameter | All (N = 409) | Low Tertile (N = 136) | Mid Tertile (N = 137) | High Tertile (N = 136) |
|---|---|---|---|---|
| Valve area (cm 2 ) ∗ | 0.83 ± 0.12 | 0.78 ± 0.14 | 0.84 ± 0.11 | 0.87 ± 0.08 |
| Indexed valve area (cm 2 ) ∗ | 0.46 ± 0.07 | 0.42 ± 0.07 | 0.46 ± 0.06 | 0.51 ± 0.05 |
| Left ventricle outflow diameter (cm 2 ) ∗ | 2.00 ± 0.12 | 1.97 ± 0.12 | 2.03 ± 0.13 | 2.01 ± 0.11 |
| Mean gradient (mm Hg) ∗ | 31 ± 6 | 29 ± 7 | 31 ± 6 | 32 ± 5 |
| Peak gradient (mm Hg) ∗ | 53 ± 11 | 49 ± 12 | 53 ± 11 | 56 ± 10 |
| Peak velocity (m/s) ∗ | 3.6 ± 0.4 | 3.5 ± 0.4 | 3.6 ± 0.4 | 3.7 ± 0.4 |
| Ejection fraction (%) ∗ | 60 ± 5 | 59 ± 5 | 60 ± 5 | 60 ± 5 |
| Left ventricle outflow velocity time integral (cm) ∗ | 22 ± 4 | 19 ± 3 | 22 ± 3 | 25 ± 3 |
| Aortic velocity time integral (cm) ∗ | 85 ± 13 | 77 ± 14 | 86 ± 11 | 92 ± 10 |
| Stroke volume (continuity, ml) ∗ | 70 ± 12 | 58 ± 9 | 71 ± 8 | 80 ± 9 |
| Stroke volume index (ml/m 2 ) ∗ | 39 ± 7 | 32 ± 4 | 39 ± 1 | 47 ± 4 |
| Cardiac index (l/min/m 2 ) ∗ | 2.7 ± 0.5 | 2.4 ± 0.5 | 2.7 ± 0.5 | 3.1 ± 0.4 |
| Left ventricle mass (gr) | 193 ± 47 | 193 ± 50 | 196 ± 46 | 190 ± 46 |
| Left ventricle mass index (gr/m 2 ) | 108 ± 24 | 104 ± 24 | 107 ± 23 | 111 ± 26 |
| Left ventricle diastolic dimension (cm) | 4.6 ± 0.5 | 4.6 ± 0.5 | 4.7 ± 0.4 | 4.5 ± 0.5 |
| Left ventricle systolic dimension (cm) | 2.70 ± 0.5 | 2.7 ± 0.5 | 2.7 ± 0.5 | 2.6 ± 0.6 |
| Interventricular septum width (cm) | 1.23 ± 0.19 | 1.24 ± 0.18 | 1.22 ± 0.2 | 1.23 ± 0.20 |
| Posterior wall thickness (cm) | 1.06 ± 0.16 | 1.06 ± 0.16 | 1.05 ± 0.15 | 1.05 ± 0.15 |
| Left ventricle end diastolic volume (ml) ∗ | 116 ± 19 | 110 ± 20 | 118 ± 18 | 120 ± 18 |
| Left ventricle end systolic volume (ml) | 47 ± 12 | 48 ± 12 | 47 ± 12 | 45 ± 12 |
| Stroke volume (Simpson, ml) ∗ | 70 ± 11 | 62 ± 10 | 71 ± 9 | 76 ± 9 |
| Ejection fraction (Simpson, %) ∗ | 65 ± 10 | 60 ± 10 | 65 ± 8 | 71 ± 9 |
| Left atrium area (cm 2 ) | 23 ± 7 | 24 ± 10 | 22 ± 6 | 22 ± 4 |
| Estimated right ventricle systolic pressure (mm Hg) | 40 ± 12 | 39 ± 12 | 40 ± 14 | 41 ± 10 |
Mean follow-up duration was 35 ± 26 months. During follow-up, 94 patients (23%) had aortic valve intervention (88 underwent surgical aortic valve replacement and 6 underwent transcatheter aortic valve replacement). Eighteen patients (13%) had an aortic intervention in the lower SVI tertile, 43 (31%) in the middle tertile, and 33 (24%) in the upper tertile.
Overall, during long-term follow-up, 141 (35%) of study patients died, of whom 61 (45%) were in the lower SVI tertile, 40 (29%) died in the middle SVI tertile, and 40 patients (29%) died in the upper SVI tertile. Kaplan-Meier survival analysis showed that the cumulative probability of survival at 3 years was 57% among patients in the lower SVI tertile compared with 72% among patients in the intermediate tertile and 73% among those in the upper SVI tertile (log-rank p value = 0.012 for the overall difference during follow-up, Figure 2 ).
When dichotomized at 35 ml/m 2 , Kaplan-Meier survival analysis showed that the cumulative probability of survival at 3 years was 59% among patients with SVI ≤35 ml/m 2 compared with 70% among patients with SVI >35 ml/m 2 (log-rank p value = 0.046 for the difference during follow-up, Figure 3 ). Results of the multivariate Cox proportional hazards models are summarized in Table 3 . When examined as a continuous measure, each 5 ml/m 2 decrease in SVI was associated with a 20% increase in all-cause mortality (p = 0.01). Tertile analysis yielded consistent results: compared with the upper tertile of SVI, patients in the lower SVI tertile experienced a significant increase in the adjusted risk for all-cause mortality, whereas patients in the middle tertile did not experience a statistically significant risk increase ( Table 3 ). In contrast, the SVI dichotomized at 35 ml/m 2 was not shown to be significantly associated with long-term mortality after multivariate adjustment. The results of the forward stepwise regression model are summarized in Table 4 . After adjustment for the prespecified covariates (age, gender, ischemic heart disease, BMI, and aortic valve intervention as a time-dependent covariate), the SVI and age were independently associated with an increased risk for all-cause mortality. The SVI was the most powerful echocardiographic parameter independently associated with increased mortality among the 7 candidate echocardiographic parameters, demonstrating that each 5 ml/m 2 reduction in the SVI was associated with an increase in mortality ( Table 4 ). In contrast, none of the other prespecified echocardiographic parameters were shown to be significantly associated with long-term mortality after multivariate adjustment for SVI.
