A, Hemodynamic assessment. Simultaneous measurement of left ventricular (LV) and aortic pressure time curves during the transcatheter aortic valve replacement (TAVR) procedure before the prosthetic valve deployment. Time (TLV-Ao) was defined as the time between LV peak systolic and aortic peak systolic pressure. B-D, Quantification of aortic valve calcification by multidetector computed tomography (MDCT). B, MDCT image of severely calcified aortic valve. Software automatically identifies the calcium in the scan volume (yellow), and calcification corresponding to the aortic valve leaflets was identified by reader (pink) (C and D). Agaston calcium scores were calculated using the computerized software.
eTable 1. Comparison of Baseline Clinical, Echocardiographic, and MDCT Characteristics According to Time Interval of LV and Ao Systolic Pressure Peak (n=123)
eTable 2. Logistic Regression Analysis for Diagnostic Ability of TLV-Ao to Predict Significant Aortic Valve Calcification (AVC) Using Gender Specific Threshold for AVC (n = 123)
eTable 3. Logistic Regression Analysis for Diagnostic Ability of TLV-Ao to Predict Significant Aortic Valve Calcification in Low-Gradient AS (n = 48)
eFigure 1. Association Between TLV-Ao and AV Calcification by MDCT in A) All Patients (n = 123), B) Patients With Low-Gradient AS (n = 48), and C) Patients With High-Gradient AS (n = 75)
eFigure 2. Impact of TLV-Ao on Survival in Patients With Low-Gradient Aortic Stenosis (AS) Who Underwent Transcatheter Aortic Valve Replacement (TAVR)
Customize your JAMA Network experience by selecting one or more topics from the list below.
Sato K, Kumar A, Jobanputra Y, et al. Association of Time Between Left Ventricular and Aortic Systolic Pressure Peaks With Severity of Aortic Stenosis and Calcification of Aortic Valve. JAMA Cardiol. Published online May 01, 20194(6):549–555. doi:10.1001/jamacardio.2019.1180
Is the time between left ventricular and aortic systolic pressure peaks associated with aortic stenosis (AS) severity in diagnosing severe low-gradient AS?
In this cohort study of 123 patients, greater time between left ventricular and aortic systolic pressure peaks was independently associated with higher aortic valve (AV) calcification and demonstrated an incremental value to detect patients with significant AV calcification vs conventional echocardiographic variables. In a subgroup of patients with low-gradient AS, only time between left ventricular and aortic systolic pressure peaks was associated with AV calcification, while other echocardiographic parameters were not associated with AV calcification.
Time between left ventricular and aortic systolic pressure peaks is a new index to grade AS severity and may act as a useful surrogate to differentiate low-gradient true severe AS among patients undergoing transcatheter aortic valve replacement.
Diagnosis of low-gradient severe aortic stenosis (AS) is challenging. We hypothesized that the time between left ventricular (LV) and aortic systolic pressure peaks (TLV-Ao) is associated with aortic stenosis (AS) severity and may have additive value in diagnosing severe AS, especially in patients with low-gradient AS.
To investigate the diagnostic utility of measuring catheter-based TLV-Ao in patients with severe AS.
Design, Setting, and Participants
We studied 123 patients with severe AS at the Cleveland Clinic Foundation, a tertiary referral center, who underwent transcatheter aortic valve replacement (TAVR) via femoral access and had pre-TAVR cardiac computed tomography assessment and hemodynamic measurements recorded during a TAVR procedure. All patients received hemodynamic evaluation, echocardiographic assessment, and quantification of aortic valve calcification (AVC) by multidetector computed tomography. Hemodynamic data were collected via left heart catheterization done just before TAVR, and TLV-Ao was calculated offline. Data were analyzed between October 5, 2015, and July 20, 2016.
Main Outcomes and Measures
The association between TLV-Ao and AVC or other conventional imaging parameters was analyzed.
Of the included patients, the mean (SD) age was 81 (9) years, and 65 (54%) were men (54%). Among 123 patients, 48 patients (39%) had low-gradient AS (<40 mm Hg) and mean (SD) TLV-Ao was 69 (39) milliseconds. In multivariable logistic regression analyses, higher TLV-Ao (odds ratio [OR], 1.02; 95% CI, 1.01-1.04; P = .002) and higher peak aortic valve (AV) velocity (OR, 1.01; 95% CI, 1.00-1.02; P = .008) were independently associated with severe AVC (AVC >1000 AU). Adding TLV-Ao to the peak AV velocity and AV area showed significant incremental value to be associated with AVC, with a net reclassification improvement of 0.61 (95% CI, 0.23-0.99; P = .002) and integrated discriminatory improvement of 0.09 (95% CI, 0.03-0.16; P = .003). In a subgroup of patients with low-grade AS, higher TLV-Ao was the only parameter associated with severe AVC (OR, 1.02; 95% CI, 1.001-1.04; P = .03).
Conclusions and Relevance
Prolonged TLV-Ao was associated with severe AVC. This catheter-based hemodynamic index may be an additional surrogate to differentiate low-gradient true severe AS. Larger, prospective studies investigating the role of TLV-Ao as a marker of clinical outcomes in patients undergoing TAVR are required.
Accurate grading of aortic stenosis (AS) severity is essential to determine the appropriate therapy. Aortic stenosis is defined as severe when the peak aortic valve (AV) velocity is at least 4 m/s, mean AV pressure gradient at least 40 mm Hg, and aortic valve area (AVA) less than 1.0 cm2 or indexed AVA less than 0.6 cm2/m2 on echocardiographic evaluation.1 Because these echocardiographic parameters are highly flow dependent, patients with severe AS with reduced ejection fraction (EF) or low-flow status may demonstrate a lower pressure gradient and hence not meet this diagnosis. Low-gradient AS (LGAS) is characterized by the presence of small AVA (<1.0cm2) and yet low mean AV pressure gradient (<40 mm Hg).1 Among patients with severe AS, up to 50% may show low AV pressure gradient. However, coexistent low-flow states in patients with moderate AS also may present similar to LGAS on echocardiography, and hence, differentiation of true severe LGAS is challenging. While guidelines recommended use of dobutamine stress echocardiography or quantification of aortic valve calcification (AVC) by multidetector computed tomography (MDCT) to assess AS severity in LGAS, there is a paucity of data in the utility of hemodynamic evaluation to detect true LGAS.
Previous studies have shown that prolonged ejection time (ET) and delayed peak velocity measured by echocardiography is associated with severity of AS.2,3 We hypothesized that an invasively measured time between left ventricular (LV) and aortic systolic pressure peaks (TLV-Ao) would be associated with higher severity of AS and may have an incremental value in diagnosing severe AS, especially in patients with AS with low mean pressure gradient. Therefore, we investigated the diagnostic utility of measuring the TLV-Ao in patients with severe AS.
We studied all consecutive patients who had cardiac CT assessment before transcatheter aortic valve replacement (TAVR) and hemodynamic measurements recorded during a TAVR procedure via femoral access between October 5, 2015, and July 20, 2016. The TAVR procedures were performed by experienced operators with Edwards SAPIEN XT and SAPIEN 3 (S3) valves (Edwards Lifesciences), Medtronic CoreValve and Corevalve Evolut R (Medtronic), Direct Flow Medical (Direct Flow Medical), Portico device (St Jude Medical), and Lotus Valve System (Boston Scientific). Hemodynamic assessment was performed prior to the valve deployment to obtain the time between left ventricle and aortic systolic peaks. Pre-TAVR echocardiogram and cardiac CT were used to assess AS severity, and the association between hemodynamic variables and imaging data was investigated. We also obtained demographic and clinical data. Mortality data were obtained from medical records or publicly available online sources (last queried December 1, 2017). The study protocol was approved by the Cleveland Clinic institutional review board, and patient informed consent was waived owing to the retrospective nature of the study and use of deidentified data.
Hemodynamic evaluation was performed during the TAVR procedure before the prosthetic valve deployment by a standard transfemoral approach. The LV and aortic pressures were recorded using 2 separate fluid-filled catheters simultaneously. Hemodynamic data were retrieved and re-evaluated in an offline system, the AXIOM Sensis P hemodynamic software, version VC11D (Siemens Healthineers) to obtain LV and aortic pressure and time. Software automatically detected the LV peak systolic and aortic peak systolic pressure, then time (TLV-Ao) was calculated as the time between LV peak systolic and aortic peak systolic pressure (Figure, A).
Quantification of AVC was performed using noncontrast and contrast-enhanced CT imaging with commercially available calcium-scoring software (TeraRecon Aquarius Workstation). In noncontrast enhanced images, we used a standard threshold of 130 Hounsfield units (HU) for calcium detection.4-6 In contrast-enhanced CT images, the threshold to detect calcification was decided individually. A region of interest was placed in the ascending aorta at the transaxial level to measure the maximum aortic attenuation value in HU. We applied a threshold cutoff value of maximum aortic attenuation 100 HU for defining calcium.4,5 After setting the threshold, the program automatically identified calcium in the scan volume; a single blinded reader identified and marked the calcification corresponding to the aortic valve leaflets (Figure, B-D). Calcification extending into the left ventricular outflow tract, coronary arteries, and aorta were excluded if they were contiguous with the calcification on the valve, and only the calcium on the leaflets and the annulus was included in the analysis. Using the computerized software, Agatston calcium scores were calculated. Total calcium scores were described in Agatston units (AU).
All patients underwent a comprehensive echocardiographic assessment using commercially available ultrasonography systems. All echocardiographic measurements were reviewed and measured by experienced readers who were blinded to clinical information according to the guidelines.1,7 Echocardiographic parameters included the following variables: peak aortic valve velocity, peak and mean transvalvular gradient, aortic jet velocity-time integral (VTI), left ventricular outflow tract (LVOT) diameter, mean LVOT flow velocity, LVOT VTI, AVA, LV ejection fraction, LV end-diastolic volume, LV end-systolic volume, and stroke volume. Peak aortic valve velocity was measured using continuous-wave Doppler, and multiple acoustic windows were checked to determine the highest velocity at baseline. Transvalvular aortic gradients were calculated from velocity using the simplified Bernoulli equation. Acceleration time (AT) was defined as the time between the beginning of systolic flow to its peak velocity, and ET was measured as the time from onset to end of systolic flow.3 The AVA was calculated using the continuity equation. Stroke volume (SV) was calculated from the cross-sectional area of the LVOT and VTI of LVOT flow: SVLVOT = CSALVOT × VTILVOT, where CSALVOT is the cross-sectional area of LVOT. The LGAS was defined as mean transvalvular gradient <40 mm Hg.
Continuous data are explained as mean and standard deviation when normally distributed or median (interquartile range). Categorical data are presented as absolute numbers and percentages. We used the unpaired t test, Mann-Whitney test, or χ2 test to compare data between the 2 groups as appropriate. We performed univariable linear regression analysis to assess the association between TLV-Ao or ln AVC with other imaging parameters. We also constructed a multivariable model to detect better determinants, and all possible confounders and relevant parameters to detect dependent variable (TLV-Ao and ln AVC) were entered into the model with stepwise selection. We compared linear regression models with the likelihood ratio test to see whether additional terms significantly reduced the error sum of squares using R package FSA8 (R Foundation for Statistical Computing). Incremental value to predict severe AVC was defined as the significant increase in continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI).9,10 The NRI is a statistic index to quantify improvement in prediction performance by adding a new parameter to baseline predictors for binary outcomes. The IDI is a similar measure to indicate that adding a new parameter to a binary prediction model improved the discrimination slope. The NRI and IDI were evaluated using an R package, PredictABEL11 (R Foundation for Statistical Computing). A 2-sided P value of less than .05 was considered statistically significant. All statistical analyses were performed with JMP, version 10.0 (SAS Institute Inc), and R software, version 3.2.2 (R Foundation for Statistical Computing).
A total of 123 patients were studied. Table 1 shows the baseline characteristics of these patients. Most patients underwent TAVR with the SAPIEN 3 valve (103 [84%]), and the remaining received the Evolut R (13 [11%]), SAPIEN XT (1 [1%]), Direct Flow Medical (n = 4; 3%), Portico device (n = 1; 1%), or Lotus Valve System (1 [1%]). Thirty-five patients (28%) received a 29-mm valve, 46 patients (37%) had a 26-mm valve, and 37 patients (30%) had a 23-mm valve. None had history of SAVR or surgical AV repair. Among 123 patients, 48 (39%) had LGAS (mean gradient <40 mm Hg) and 75 patients (61%) had high-gradient AS.
The mean (SD) value of TLV-Ao in our patients was 69 (39) milliseconds. In a linear regression analysis, higher TLV-Ao was associated with a higher mean AV pressure gradient (β, .28; P = .002), peak AV velocity (β, .25, P = .006), AT/ET ratio (β, .18; P = .046), advanced AVC (β, .39, P < .001 for ln AVC), and higher LVEF (β, .19, P = .03) (Table 2; eFigure 1A in the Supplement). In a multivariable model, only higher AVC (β, .37; P < .001) and LVEF (β, .20, P = .045) were associated with prolonged TLV-Ao.
We divided the patients into 2 groups using the median value of TLV-Ao. As shown in eTable 1 in the Supplement, patients with longer TLV-Ao (>66 milliseconds) were more likely to be men, have a higher body surface area, lower STS score, higher LVEF, higher peak AV velocity, mean pressure gradient, and higher AVC. Interestingly, flow-independent echocardiographic parameters for AS severity, such as AVA or DVI, did not differ between these 2 groups.
In the subgroup of patients with LGAS, higher TLV-Ao was the only parameter associated with higher ln AVC (β, .30; P = .03) (Table 3; eFigure 1B in the Supplement). The AVC was still associated with TLV-Ao in high-gradient AS (β, .42; P < .001) (eFigure 1C in the Supplement).
Because AVC is considered an alternative way to differentiate true severe AS, we performed a logistic regression analysis to assess the association of TLV-Ao and severe AVC (>1000 AU).12 In a univariable logistic regression analysis, higher TLV-Ao was significantly associated with severe AVC (odds ratio [OR], 1.02; 95% CI, 1.01-1.04; P < .001), with highest C statistics (0.72) (Table 4). In multivariable logistic regression analysis, higher TLV-Ao (OR, 1.02; 95% CI, 1.01-1.04; P = .002) and higher peak AV velocity (OR, 1.01; 95% CI, 1.003-1.02; P = .008) were found to be significantly associated with severe AVC after adjusting for all relevant conventional echocardiographic indicators of AS severity. The optimal cutoff value of TLV-Ao greater than 63 milliseconds showed a sensitivity of 66% and specificity of 72% (positive predictive value, 87%; negative predictive value, 43%). Adding TLV-Ao to conventional AS severity parameters (AVA and peak AV velocity) showed an incremental value in determining association with severe AVC, with an NRI of 0.61 (95% CI, 0.23-0.99; P = .002) and IDI of 0.09 (95% CI, 0.03-0.16; P = .003). In the logistic regression model, neither AT nor AT/ET was associated with severe AVC, suggesting the limited ability of noninvasive indices to detect severe AS. Adding TLV-Ao showed significant incremental value vs AT/ET to diagnose severe AVC, with NRI of 0.76 (95% CI, 0.39-1.12; P < .001) and IDI of 0.13 (95% CI, 0.07-0.20; P < .001).
Because some studies have suggested sex-based differences in AVC severity, we repeated the analysis using sex-specific thresholds to identify severe AVC (>1200 AU for women and >2000 AU for men).13,14 Higher TLV-Ao was still significantly associated with severe AVC (OR, 1.13; P = .01) (eTable 2 in the Supplement). When we subdivided TLV-Ao by median, patients with TLV-Ao greater than 66 milliseconds were 2.3 times more likely to demonstrate severe AVC. Adding TLV-AO to the AVA and AV peak velocity showed significant incremental value to determine association with severe AVC, with an NRI of 0.41 (95% CI, 0.06-0.75; P = .02) and IDI of 0.03 (95% CI, 0.003-0.06; P = .03).
We performed similar analysis in a subgroup of patients with LGAS (mean gradient <40 mm Hg; n = 48). In this cohort, longer TLV-Ao (OR, 1.02; 95% CI, 1.001-1.04, P = .03) was the only parameter associated with severe AVC (eTable 3 in the Supplement). Adding TLV-Ao to conventional AS severity parameters increased the C statistics from 0.68 to 0.74 (P = .27) and showed an incremental value to determine association with severe AVC, with an NRI of 0.30 (95% CI, −0.27 to 0.86; P = .30) and IDI of 0.08 (95% CI, −0.003 to 0.17; P = .06).
A total of 15 patients (12%) died during the first year after TAVR. There were no significant associations between TLV-Ao and mortality (hazard ratio, 0.996; 95% CI, 0.99-1.007; P = .45). In the subgroup of patients with LGAS, 20 patients showed prolonged TLV-Ao (>66 milliseconds) and 28 patients had TLV-Ao of 66 milliseconds or less. Of those 25 patients with TLV-Ao greater than 66 milliseconds, 4 (20%) died during the first year of follow-up, while 3 (11%) died among those with TLV-Ao of 66 milliseconds or less. When we performed survival analysis in the subgroup of LGAS, higher TLV-Ao demonstrated a trend toward having higher mortality (hazard ratio, 0.99; 95% CI, 0.97-1.006; P = .22) (eFigure 2 in the Supplement).
In this study, we demonstrated that (1) higher TLV-Ao was associated with greater AS severity as measured by conventional echocardiography and higher AVC on MDCT; (2) greater TLV-Ao was independently associated with higher AVC and demonstrated an incremental value to detect patients with severe AVC vs conventional echocardiographic variables; and (3) in a subgroup of patients with LGAS, only TLV-Ao was associated with AVC, while other echocardiographic parameters were not associated with AVC.
To our knowledge, this is the first study to show that a directly measured TLV-Ao is significantly associated with AV calcification calculated on MDCT. The delay in aortic pressure upstroke, which can be detected on palpitation as pulsus tardus et parvus, slow rising, and late peaking pulse, has been reported in several prior studies.2,3,10,11 Worsening severity of AS increases the velocity across the stenotic valve, along with prolongation of ET, causing delay in reaching the maximal velocity across the valve that could be detected by carotidography, phonocardiography, and echocardiography.2,3,15,16 Although these prolongations in AV peak velocity are highly affected by cardiac function or heart rate, time to the maximal velocity of the AV flow is not sufficient as a surrogate of AS severity, especially in LGAS. Several efforts have been made to overcome this limitation, such as adjusting time to peak velocity by ET.2,3 As invasive hemodynamic assessment enables us to evaluate time to peak velocity in LV and aortic systolic pressure separately, it follows that TLV-Ao could be an independent surrogate of AS severity. Furthermore, perhaps our results could be reproduced using noninvasive methods by measuring time to peak velocity of LVOT and AV. However, it may be difficult to measure TLV-Ao with flow because peak velocity by echocardiography is not a simultaneous measurement of peak pressure.17 Further, variability in image quality also makes it difficult to find peak Doppler gradient by echocardiography to enable us to apply this concept using echocardiography. Indeed, we show that TLV-Ao profile by invasive hemodynamic assessment may be a strong discriminator of true severe AS by detecting severe AVC.
Of note, we used AVC for a reference of AS severity because it is an important anatomical marker of AS severity.13,18 A significant association between TLV-Ao and AVC suggests that TLV-Ao might serve as a marker of AS severity irrespective of hemodynamic status or cardiac function. Although our results are limited by small sample size, it is still important to note that TLV-Ao was the only parameter associated with AVC in patients with LGAS.
We have demonstrated that a novel catheter-based hemodynamic index, ie, TLV-Ao, is associated with severe AS. Doppler-based hemodynamic assessment of AS severity could be riddled with discrepancies, especially in patients with LGAS. Our results suggest that a time between LV and aortic systolic pressure peaks is a more sensitive parameter to evaluate anatomic deformation of AV than conventional echocardiographic parameters, especially in patients with LGAS.
This is a single-center, observational study conducted at a tertiary referral center. In addition, we performed this evaluation only in patients who underwent TAVR. Further, multicentric, prospective studies are required to validate our findings, including larger populations with varying severity of AS. Regarding the predictive ability of TLV-Ao, higher TLV-Ao showed a trend toward having higher mortality in LGAS. While we failed to confirm the predictive value of TLV-Ao, our data might be underpowered owing to small sample size and low event rates. Hence, future studies are required to investigate the role of this novel parameter in predicting patient outcomes among patients undergoing TAVR and surgical AVR in larger cohorts of patients with AS.
Because we have a limited number of patients who underwent dobutamine stress echocardiography, we only used AVC, which is an established flow-independent parameter of AS severity, as a gold standard of true severe AS. Moreover, using contrast-enhanced studies in AVC investigation could lead to the underestimation of AV calcification. Prior studies showed that CT attenuation values for calcification were much higher than luminal contrast and feasible to assess AS severity using contrast enhanced CT.5,19
Furthermore, TLV-Ao can only be obtained invasively with simultaneous LV and aortic pressure measurements, which may limit its availability. Nonetheless, our results showed incremental value vs conventional echocardiographic parameters, suggesting utility to estimate TLV-Ao in LGAS or those with inconclusive echocardiographic results. In addition, because we used fluid-filled catheters for evaluation, catheter whip artifact may affect the reliability of our measurements in some cases. Hence, we measured TLV-Ao automatically using an offline system to minimize the effect of catheter-related artifact and reduce the concern in reproducibility. Finally, prior studies revealed that the hemodynamic condition and arterial stiffness might affect the timing of the aortic pressure waves.20 Increased arterial stiffness leads to earlier systolic timing owing to earlier arrival of reflected wave and might affect TLV-Ao. Reduced arterial compliance frequently occurs in elderly patients with AS, which contributes to increased afterload, decreased LV function, and low mean pressure gradient.21
Higher TLV-Ao is associated with severe AVC and demonstrates an incremental diagnostic value to detect severe AVC vs conventional echocardiographic variables. In our study, TLV-Ao was the only variable associated with severe AVC in patients with LGAS. The TLV-Ao may be a new index to grade AS severity and may act as a useful surrogate to differentiate low-gradient true severe AS among patients undergoing TAVR. Prospective studies, including larger populations of patients with varying severity of AS, are needed to confirm our findings and investigate the potential role of TLV-Ao as a marker of diagnosis and clinical outcome.
Corresponding Author: Samir Kapadia, MD, Catheterization Laboratory, Cleveland Clinic, J2-3, 9500 Euclid Ave, Cleveland, OH 44195 (firstname.lastname@example.org).
Accepted for Publication: March 1, 2019.
Published Online: May 1, 2019. doi:10.1001/jamacardio.2019.1180
Correction: This article was corrected on July 3, 2019, to fix errors in Panel A of the Figure.
Author Contributions: Dr Kapadia had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Sato, Kumar, Menon, Harb, Wael, Kapadia.
Acquisition, analysis, or interpretation of data: Sato, Kumar, Jobanputra, Betancor, Halane, George, Krishnaswamy, Tuzcu, Mick, Svensson.
Drafting of the manuscript: Sato, Kumar, Jobanputra, Betancor, Halane, George, Menon, Wael.
Critical revision of the manuscript for important intellectual content: Kumar, Jobanputra, Krishnaswamy, Tuzcu, Harb, Wael, Mick, Svensson, Kapadia.
Statistical analysis: Sato, Jobanputra, George, Menon.
Obtained funding: Svensson.
Administrative, technical, or material support: Jobanputra, Harb, Wael, Svensson.
Supervision: Tuzcu, Wael, Mick, Svensson, Kapadia.
Conflict of Interest Disclosures: Dr Wael reported serving as a member of the Board of Directors of American Society of Nuclear Cardiology and on the ACC/AHA guidelines for chest pain writing group. Dr Svensson reported serving as an unpaid member of the Executive Committee of the PARTNER Trial and Chairman of the PARTNER Publications Committee. No other disclosures were reported.
Create a personal account or sign in to: