- Email: [email protected]
Grading Aortic Valve Stenosis With Dimensionless Index During Pre-cardiopulmonary Bypass Transesophageal Echocardiography: A Comparison With Transthoracic Echocardiography
Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
Contents lists available at ScienceDirect
Journal of Cardiothoracic and Vascular Anesthesia journal homepage: www.jcvaonline.com
Original Article
Grading Aortic Valve Stenosis With Dimensionless Index During Pre-cardiopulmonary Bypass Transesophageal Echocardiography: A Comparison With Transthoracic Echocardiography 1 George B. Whitener, MD, FASE*, , Paul C. Shanahany, Bethany J. Wolfz, Alan C. Finley, MD, FASE* *
Department of Anesthesia and Perioperative Medicine, Medical University of South Carolina, Charleston, SC y University of Rochester School of Medicine and Dentistry, Rochester, NY z Department of Public Health Sciences, Medical University of South Carolina, Charleston, SC
Objective: The authors hypothesized that grading valvular aortic stenosis (AS) with dimensionless index (DI) during intraoperative pre-cardiopulmonary bypass (pre-CPB) transesophageal echocardiography (TEE) would match the grade of AS during preoperative transthoracic echocardiography (TTE) for the same patients more often than when using peak velocity (Vp), mean pressure gradient (PGm), or aortic valve area (AVA). Design: Retrospective, observational. Setting: Single university hospital. Participants: The participants in this study included 123 cardiac surgical patients with any degree of AS, who underwent open cardiac surgery between 2010 and 2016 at the Medical University of South Carolina and had Vp, PGm, AVA, and DI values available from reporting databases or archived imaging. Interventions: None. Measurements and Main Results: When using DI, pre-CPB TEE grading of AS severity was 1 grade higher 21.1% of the time and 1 grade lower 13.0% of the time compared with TTE, for an overall disagreement rate of 34.1%. The overall disagreement rates between pre-CPB TEE and TTE for Vp, PGm, and AVA were 39.8%, 33.3%, and 33.3%, respectively. Conclusions: The authors could not demonstrate that DI was better than Vp, PGm, or AVA at matching AS grades between intraoperative preCPB TEE and preoperative TTE. When DI was used, pre-CPB TEE was more likely to overestimate than underestimate the severity of AS compared with TTE. However, when Vp or PGm was used, pre-CPB TEE was more likely to underestimate the severity of AS compared with TTE. A comprehensive approach without overemphasis on 1 parameter should be used for AS assessment by intraoperative TEE. Ó 2019 Elsevier Inc. All rights reserved. Key Words: aortic stenosis; dimensionless index; grading; pre-cardiopulmonary bypass transesophageal echocardiography; intraoperative transesophageal echocardiography
G. Whitener helped design the study, conduct the study, analyze the data, and write the manuscript. P. Shanahan helped conduct the study, acquire the data, and write the manuscript. B. Wolf helped design the study, analyze the data, provide statistical analysis, and write the manuscript. A. Finley helped design the study and edit the manuscript. 1 Address reprint requests to George Whitener, MD, FASE, Department of Anesthesia and Perioperative Medicine, Medical University of South Carolina, 25 Courtenay Dr, 4210 Ashley River Tower, Charleston, SC 29425. E-mail address: [email protected] (G.B. Whitener). https://doi.org/10.1053/j.jvca.2019.03.046 1053-0770/Ó 2019 Elsevier Inc. All rights reserved.
ACCURATELY GRADING VALVULAR AORTIC stenosis (AS) during pre-cardiopulmonary bypass (pre-CPB) transesophageal echocardiography (TEE) is important, especially in the clinical context of incidental AS, when a plan has not been made prior to coming to the operating room to address a pathological valve. The rate of incidental AS is low, but not insignificant. In a cohort of 3,835 patients undergoing coronary artery bypass grafting, 3.3% of patients had an additional mitral or
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
aortic valve procedure added to the operation due to pre-CPB TEE findings.1 In another large retrospective study, the rate of incidental mild or moderate AS in 8,202 patients undergoing coronary artery bypass grafting was reported to be 2.4%.2 The ability to delineate mild, moderate, and severe AS in the setting of another cardiac surgical procedure is important because, in this context, the grade of AS affects surgical decision-making.3 Even under ideal conditions, inconsistencies between AS grading parameters complicate overall severity assignment. Additionally, the presence of general anesthesia complicates intraoperative pre-CPB TEE because of its impact on these grading parameters. For instance, the discordance rate between aortic valve area (AVA) and mean pressure gradiant (PGm) to grade AS is worse during pre-CPB TEE compared with transthoracic echocardiography (TTE).4 Moreover, PGm preferentially underestimates the degree of AS severity during pre-CPB TEE compared with TTE.5 Aortic stenosis grading cutoffs were derived from patients without general anesthesia or positive pressure ventilation.3,69 Nevertheless, these cutoffs are the only grading standard available for the intraoperative echocardiographer. Thus, an
2377
ideal grading parameter for assessing incidental AS during pre-CPB would be affected little by hemodynamic loading conditions, would be reproducible while limiting intrinsic measurement errors, and would have a high correlation between pre-CPB TEE and TTE to allow translation of cutoff values. The dimensionless index (DI) is a potentially ideal AS grading parameter because it does not require measuring the left ventricular outflow tract, as AVA does, and, unlike peak velocity (Vp) and PGm, DI maintains stability across varying loading conditions and angles of Doppler interrogation.10 Therefore, the authors hypothesized that grading AS with DI during pre-CPB TEE would match the grade of AS during TTE more often than when using Vp, PGm, or AVA. Methods Patient Selection After submitting a waiver of consent for a retrospective review and obtaining approval from the Medical University of South Carolina institutional review board, the authors obtained
Fig 1. Flow diagram illustrating selection of study population. AI, aortic insufficiency; AS, aortic stenosis; AV, aortic valve; MR, mitral regurgitation; RV, right ventricle; TEE, transesophageal echocardiography; TR, tricuspid regurgitation; TTE, transthoracic echocardiography.
2378
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
the records for all patients undergoing any cardiac surgical procedure from January 1, 2010 to August 11, 2016, who had any degree (mild, moderate, or severe) of AS listed on the institution’s intraoperative TEE report. The authors simultaneously selected for patients with normal right ventricular function, mild or less aortic regurgitation, mild or less mitral regurgitation, and mild or less tricuspid regurgitation. Once these patients were identified, the following AS grading parameters were obtained from the report: Vp, PGm, AVA, and DI. Of these patients, the authors cross-referenced their respective preoperative TTE reports to compare the aforementioned grading parameters. The authors excluded any patients with TTE reports more than 5 months from the time of surgery. TTE reports closest to the time of surgery were used in situations where multiple studies were available. Subsequently, all matching pre-CPB TEE and TTE examinations were reviewed by an echocardiographer certified by the National Board of Echocardiography in advanced perioperative echocardiography to obtain any missing values and corroborate the reports with imaging. Patients with missing, poor imaging, or inadequate Doppler-derived waveforms were excluded. During review, the authors also found and excluded patients who had a bioprosthetic aortic valve, pericardial effusion, abnormal right ventricular function by TTE, or greater than mild valvular regurgitation by TTE (Fig 1). Echocardiographic Variables and Acquisition The Vp, PGm, and AVA values were recorded routinely and available in both reporting systems. Examinations were reviewed to ensure that measurements were obtained according to guidelines by the American Society of Echocardiography.11,12 The Vp and PGm values were obtained by integrating
a continuous-wave Doppler signal of flow across the aortic valve. For pre-CPB TEE, Vp and PGm were obtained from a deep transgastric long-axis view with the sample gate aligned as parallel to blood flow as possible. If unavailable, a transgastric long-axis view was used. For TTE, aortic valve flow was interrogated from one of multiple viewing angles, including the apical, right parasternal, suprasternal, subcostal, or supraclavicular viewing windows, and the view yielding the highest-velocity jet was used. The AVA was calculated by using the continuity equation6,12,13 for both TTE and pre-CPB TEE: ðCSALVOT Þ ðVTILVOT Þ AVA cm2 ¼ ; VTIAV where CSALVOT represents the cross-sectional area (CSA) of the left ventricular outflow track (LVOT), VTILVOT is the velocity-time integral (VTI) of flow through the LVOT, while VTIAV is the VTI of flow through the aortic valve. To calculate the CSALVOT, an LVOT diameter was obtained in the midesophageal long-axis view in pre-CPB TEE and by the parasternal long-axis view during preoperative TTE. The LVOT diameters were measured 0.5 to 1.0 cm to the ventricular side of the valve annulus at the site of LVOT VTI measurement as was previously recommended by the European Association of Echocardiography and the American Society of Echocardiography.11 The DI was calculated by the following equation12: DI ¼
ðVTILVOT Þ ðVTIAV Þ
The DI routinely was not reported; therefore, the TTE images were reviewed for each patient by a certified echocardiographer to confirm the waveform values for LVOT
Fig 2. TTE images showing (A) LVOT diameter measurement in a parasternal long-axis aortic valve view, (B) LVOT-VTI by pulsed-wave Doppler in an apical view, and (C) AV-VTI by continuous-wave Doppler in an apical view. AV, aortic valve; LVOT, left ventricular outflow tract; TTE, transthoracic echocardiography; VTI, velocity-time integral.
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
VTI by pulsed-wave Doppler and aortic valve VTI by continuous-wave Doppler. Patients with missing or poor imaging were excluded. Of note, the authors assured that the VTIs used to calculate the reported AVA by continuity equation were the same waveforms used to calculate DI. Examples of VTI acquisition for DI calculation along with LVOT measurement for both TTE and TEE are shown in Fig. 2 and 3. Statistical Analysis Descriptive statistics were calculated for all patient characteristics and for the 4 TTE and TEE measures described in the previous section. Patient severity was determined using the grading severity cutoffs taken from current American Heart Association, American College of Cardiology, and American Society of Echocardiography guideline recommendations and applied to each grading parameter for each patient for both pre-CPB TEE and TTE.3,12 The used cutoffs are summarized in Table 1. The authors examined the agreement in severity grading for each metric as determined by TTE and pre-CPB TEE across all patients and performed additional stratification of participants by left ventricular ejection fraction <50% versus 50%. Agreement for each measure was evaluated using the weighted kappa statistic and tested whether the kappa statistic took a value of 0 (indicating no agreement). The weighted kappa statistic ranges between 1 (perfect agreement) and ¡1 (completely contradictory) and accounts for the magnitude of the difference between the 2 methods in estimating agreement.14 Guidelines for interpreting the kappa statistic indicate that values between 0.4 and 0.59 suggest weak agreement, values between 0.6 to 0.79 suggest moderate agreement, and values greater than 0.8 suggest strong agreement.15 The authors also calculated 95% confidence
2379
Table 1 Aortic Stenosis Grading Limits Measurement
Mild
Moderate
Severe
Peak velocity (m/s) Mean gradient (mmHg) Aortic valve area (cm2) Indexed valve area (cm2/m2) Dimensionless index
<3.0 <20 >1.5 >0.85 >0.50
3.0-4.0 20-40 1.0-1.5 0.60-0.85 0.25-0.50
4.0 40 <1.0 <0.6 <0.25
Adapted from Baumgartner et al.11 and Nishimura et al.3
intervals (CIs) for the weighted kappa statistics to compare the level of agreement between TTE and pre-CPB TEE for DI with the agreement for Vp, PGm, and AVA, where nonoverlapping CIs for kappa statistics indicate that the metric with higher agreement between TTE and pre-CPB TEE is significantly better. The authors also estimated the kappa statistic stratified by Simpson’s LVEF of <50% or 50%. As a secondary measure of agreement, the authors looked at percent agreement with 95% CIs for each metric defined as the percent of participants for whom the TTE and TEE severity for each metric was equivalent. Finally, the authors examined the correlation between TTE and TEE measures for each metric using Pearson’s correlation. All analyses were conducted in SAS, version 9.4 (SAS Institute, Cary, NC). Results A total of 319 patients met initial inclusion criteria. Once all pre-CPB and TTE examinations were reviewed, 123 patients were left to compare AS grading parameters between pre-CPB
Fig 3. TEE images in the same patient as Fig 2 showing (A) LVOT diameter measurement in a midesophageal long-axis view, (B) LVOT-VTI by pulsed-wave Doppler in a deep transgastric long-axis view, and (C) AV-VTI by continuous-wave Doppler in a deep transgastric long-axis view. Note the lower Vp (labeled AV Vmax on the image), PGm, LVOT-VTI, and AV-VTI values compared with transthoracic echocardiography in Fig 2. AV, aortic valve; LVOT, left ventricular outflow tract; PGm, mean pressure gradient; TEE, transesophageal echocardiography; Vp, peak velocity; VTI, velocity-time integral.
2380
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
Table 2 Patient Demographics Characteristic Age at surgery, y* Male sex, n (%)y Body surface area, m2* Mean time from TTE to TEE, d* Surgical procedure, n (%)y AVR CABG + AVR CABG TAVR AVR + ascending aortic repair Heart transplant Echo parameter Average LVOT diameter by TTE, cm* Average LVOT diameter by TEE, cm* Average SVI by TTE, mL/m2* Average SVI by TEE, mL/m2*
Average mean gradient by TTE, mmHg* Average mean gradient by TEE, mmHg* Average peak velocity by TTE, m/s* Average peak velocity by TEE, m/s* Average AVA by TTE, cm2* Average AVA by TEE, cm2* Average VTILVOT by TTE, cm* Average VTILVOT by TEE, cm* Average VTIAV by TTE, cm* Average VTIAV by TEE, cm* Average DI by TTE* Average DI by TEE*
n = 123 71.3 § 10.9 84 (68) 1.98 § 0.25 28.8 § 34.6 40 (32.5) 30 (24.4) 23 (18.7) 21 (17.1) 8 (6.5) 1 (0.8) n = 123 2.1 § 0.21 2.2 § 0.24 44.0 § 11 42.6 § 12 35.1 § 18.5 28.1 § 16.7 3.74 § 1.00 3.33 § 0.92 1.08 § 0.44 1.06 § 0.37 24.4 § 5.80 22.2 § 6.40 88.9 § 29.6 86.2 § 29.5 0.30 § 0.12 0.28 § 0.10
Abbreviations: AVA, aortic valve area; AVR, aortic valve replacement; CABG, coronary artery bypass grafting; CPB, cardiopulmonary bypass; DI, dimensionless index; SVI, stroke volume index; TAVR, transcutaneous aortic valve replacement; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; VTI, velocity-time integral. * Data presented as mean § standard deviation. y Data presented as frequency in the population.
and TTE (Fig 1). Baseline patient and echocardiographic characteristics for the study population are shown in Table 2.The impact of pre-CPB TEE for each grading parameter on AS severity compared with preoperative TTE is summarized in Table 3. When using DI to grade AS, pre-CPB TEE produced an AS severity 1 grade higher or 1 grade lower than preoperative TTE, 21.1% and 13.0% of the time, respectively. Any discrepancy in severity was seen 34.1% of the time for DI. When using Vp, pre-CPB revealed an AS severity 1 grade higher, 1 grade lower, or 2 grades lower than preoperative TTE, 4.1%, 30.9%, and 4.9% of the time, respectively. Any discrepancy in grading was seen 39.8% of the time for Vp. When using PGm to grade AS, pre-CPB TEE produced an AS severity 1 grade higher, 1 grade lower, or 2 grades lower than preoperative TTE 2.4%, 29.3%, and 1.6% of the time, respectively. Any discrepancy in grading was seen 33.3% of the time for PGm. When using AVA to grade AS, pre-CPB TEE produced an AS severity 1 grade higher, 1 grade lower, or 2 grades lower than preoperative TTE, 17.9%, 14.6%, and 0.8% of the time, respectively. Any discrepancy in grading was seen 33.3% of the time for AVA.
All 4 metrics’ agreement between pre-CPB TEE and preoperative TTE were significantly greater than would be expected by chance (p < 0.001 for all). Values for the kappa statistic ranged from 0.51 to 0.70, suggesting weak to moderate agreement. PGm exhibited the largest level of agreement with k of 0.70 (95% CI: 0.61-0.80), while DI had the weakest agreement (k = 0.51, 95% CI: 0.39, 0.64). The 95% CIs for k values of PGm and DI did not overlap, indicating that agreement between severity estimated from preoperative TTE and preCPB TEE using PGm was significantly better than the agreement between preoperative TTE and pre-CPB TEE for DI. There were no other notable differences in agreement between the 4 measures. Weighted kappa statistics with 95% CIs also are provided in Table 3. Stratification by Left Ventricular Function As a secondary analysis, the authors also examined agreement between TTE and TEE stratified by left ventricular function, grouping participants according to those with LVEF <50% and those with LVEF 50%. There were 17 participants with LVEF <50% and 106 with LVEF 50%. The impact of pre-CPB TEE for each grading parameter on AS severity when compared with preoperative TTE stratified by left ventricular function is summarized in Table 4. For Vp and PGm, the percentage of participants showing any change in severity between those with LVEF <50% versus >50% was similar. In contrast, for AVA and DI, the percentage of participants with no change in severity was 18.2% and 12.3% lower in those with LVEF 50. All 4 metrics’ agreement between TEE and TTE were significantly greater than would be expected by chance (p < 0.001 for all). Values for the kappa statistic by LVEF status ranged from 0.50 to 0.82 and were similar between patients with LVEF <50 versus 50 for Vp and PGm. Using AVA to determine severity, the kappa value was higher for patients with LVEF <50 compared with LVEF 50 (k = 0.82 v k = 0.58). Of note, none of the patients with LVEF <50 had a severity of 1 grade lower using DI and TEE; however, the agreement between TEE and TTE measures of severity based on DI were similar for those with good or poor ventricular function. Weighted kappa statistics with 95% CIs by left ventricular function also are provided in Table 4. Discussion The authors could not demonstrate that DI was better than Vp, PGm, or AVA at matching AS grades between intraoperative pre-CPB TEE and preoperative TTE. The results showed that DI, Vp, PGm, and AVA all had similar rates of overall disagreement between pre-CPB and TTE, but DI showed less agreement compared with PGm between TEE and TTE when accounting for where dissimilarities occur (Table 3). For instance, Vp and PGm illustrated a predictable tendency to underestimate AS severity during pre-CPB TEE versus TTE. Alternatively, AVA and DI illustrated a less predictable tendency to overestimate AS severity during pre-CPB TEE versus
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
2381
Table 3 Change in AS Severity Based on Imaging Context and Grading Parameter AS Grading Parameter by Preoperative TTE (n = 123)
Impact of pre-CPB TEE on AS grade
1 grade higher No change* 1 grade lower 2 grades lower Any change Kappa statistic (95% CI)
Vp
PGm
AVA
DI
5 (4.1) 74 (60.1) 38 (30.9) 6 (4.9) 49 (39.8) 0.62 (0.51-0.73)
3 (2.4) 82 (66.7) 36 (29.3) 2 (1.6) 41 (33.3) 0.70 (0.61-0.80)
22 (17.9) 82 (66.7) 18 (14.6) 1 (0.8) 41 (33.3) 0.62 (0.49-0.74)
26 (21.1) 81 (65.9) 16 (13.0) 0 (0) 42 (34.1) 0.51 (0.39-0.64)
NOTE. Data expressed as number (percentage) except for entries for the kappa statistic, which are reported as k (95% CI). Abbreviations: AS, aortic stenosis; AVA, aortic valve area; CPB, cardiopulmonary bypass; DI, dimensionless index; Vp, peak velocity; PGm, mean pressure gradient; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography. * The row reporting “No change” is equivalent to the number (percentage) that agree between TEE and TTE.
TTE. A graphical comparison for each parameter is represented in Fig 4. There are physiological reasons, decreased preload and afterload caused by positive-pressure ventilation and general anesthesia, that can explain the tendency for Vp and PGm to underestimate AS during pre-CPB TEE. Additionally, these findings mirror the results from a prior study comparing preoperative TTE with pre-CPB TEE, which showed an average drop of 6.6 mmHg when using TEE.5 This current study showed a similar average decrease in PGm of 7.0 mmHg during pre-CPB TEE. However, it is surprising that DI would not be more consistently matched between pre-CPB TEE and TTE. DI is a theoretical solution to grading disagreement problems based on the anesthetized state. Even if there are changes in loading conditions between imaging contexts, both the numerator and the denominator of the DI equation should move proportionally in the same direction and remain constant in the same patients. The DI also accounts for differences imparted by the imaging
techniques. For example, if differences between TTE and TEE cause changes in the Doppler angle of interrogation with LVOT and aortic outflow, these changes should affect the numerator and denominator of the DI equation proportionally, thereby rendering DI the same. In the results, the AVA and DI appear to track one another, which makes sense given that they both represent the same proportion (VTILVOT/VTIAV), except that AVA takes LVOT diameter into account. The average LVOT diameter for the study patients was slightly higher during TEE (2.2 cm2 v 2.1 cm2), meaning that the decrease in average AVA and increase in severity during TEE is not due to LVOT diameter discrepancies. Therefore, DI’s inability to match severity lies in the fact that VTI measurement differences existed between TTE and pre-CPB TEE. Moreover, a slight tendency for DI to overestimate severity means that, on average, VTIAV increased relative to VTILVOT or VTILVOT decreased relative to VTIAV during TEE. Table 2 shows that the latter scenario occurred in the study. Both average VTILVOT and VTIAV values decreased
Table 4 Change in AS Severity Stratified on Left Ventricular Function Based on Imaging Context and Grading Parameter AS Grading Parameters by Intraoperative TEE as Compared With Preoperative TTE for Participants With LVEF <50 (n = 17) and LVEF 50 (n =106) PGm
Vp
1 grade higher No change 1 grade lower 2 grades lower Any change % agreement (95% CI) Kappa statistic (95% CI)
AVA
DI
LVEF <50
LVEF 50
LVEF <50
LVEF 50
LVEF <50
LVEF 50
LVEF <50
LVEF 50
2 (11.8) 10 (58.5) 5 (29.4) 0 (0.00) 7 (41.2) 58.5 (35.4-82.2) 0.58 (0.16-0.90)
3 (2.83) 65 (61.3) 32 (30.2) 6 (5.70) 41 (38.7) 61.3 (50.3-70.6) 0.59 (0.46-0.72)
1 (5.88) 11 (64.7) 5 (29.4) 0 (0.00) 6 (35.3) 64.7 (42.0-87.4) 0.64 (0.38-0.92)
2 (1.89) 71 (67.0) 31 (29.4) 2 (1.90) 36 (32.1) 67 (56.1-75.9) 0.69 (0.58-0.79)
1 (5.88) 14 (82.4) 2 (11.8) 0 (0.00) 3 (17.6) 82.4 (64.2-1.00) 0.82 (0.63-1.00)
21 (19.8) 68 (64.2) 16 (15.1) 1 (0.9) 41 (36.6) 64.2 (53.2-73.3) 0.58 (0.43-0.72)
4 (23.5) 13 (76.5) 0 (0.00) 0 (0.00) 4 (23.5) 76.5 (56.3-96.6) 0.79 (0.55-1.00)
22 (20.8) 68 (64.2) 16 (15.1) 0 (0.00) 40 (35.7) 64.2 (53.2-43.3) 0.5 (0.36-0.63)
NOTE. Data expressed as number (percentage) except for entries for percent agreement and for the kappa statistic, which are reported as value (95% CI). Abbreviations: AS, aortic stenosis; AVA, aortic valve area; DI, dimensionless index; LVEF, left ventricular ejection fraction; PGm, mean pressure gradient; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; Vp, peak velocity.
2382
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
Fig 4. Scatter plots of TTE (y-axis) by TEE (x-axis) for each metric. The 45o line represents perfect agreement with TEE. Points above the line represent participants for whom values during TTE > TEE and points below the line represent participants for whom values during TTE < TEE. The Pearson’s correlation between TEE and TTE also is provided on each plot. All correlations were significant at p < 0.001. TEE, transesophageal echocardiography; TTE, transthoracic echocardiograph.
when moving from TTE to TEE, which may be attributable to physiological changes from general anesthesia as discussed previously. However, the average VTILVOT decreased more relative to the average VTIAV. There are a few reasons why severity by DI may have changed this way. First, the trend toward worsening AS as measured by DI during pre-CPB TEE may reflect worsening AS. In this case, VTIAV for a given patient would decrease less, relative to VTILVOT. The AVA worsens by 0.1 cm2/y; nevertheless, this is an average, and some patients worsen by up to 0.5 to 0.72 cm2/y.16,17 The authors chose an absolute cutoff of 5 months between preoperative TTE and pre-CPB TEE to reduce the impact of worsening AS between examinations while allowing for a reasonable sample size. Overall, the mean time between examinations was only 29 days, therefore the risk of significantly worsening AS would be small but not eliminated (Table 2).
Secondly, there is known interoperator variability for DI measurements when using VTIs. For example, Sacchi et al.18 demonstrated in a study with 25 operators that the interoperator coefficient of variation was 9.3%, while the intraoperator coefficient of variation was 13.9% for measuring DI. Sixtytwo patients in our study were within §0.05 of a DI severity cutoff on preoperative TTE. Therefore, small measurement differences between operators, amounting to clinically insignificant discrepancies, may have crossed grading cutoffs often enough to lead to multiple reclassifications of severity. Lastly, a mix of LVOT VTI measurements was done with a double envelope technique while others were accomplished with separate pulsed-wave Doppler acquisitions. This difference in acquisition may have had an impact on DI calculations. However, it is difficult to predict in which direction this may have pushed DI when moving from TTE to TEE. Previous work has shown good correlation between a double envelope
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
technique during TEE compared with separate envelope acquisition during TTE, so it is unclear if this would have had any appreciable impact.19 One other study, by Uda et al.,20 attempted a similar comparison between preoperative TTE and pre-CPB TEE. Their analysis showed that PGm and Vp underestimate AS during pre-CPB compared with TTE, but they found fewer patients reclassified to a different AS grade when using DI.20 However, the current study’s patients were taken from a pool consisting of all grades of AS, whereas most of Uda et al.’s patients had severe AS. For example, 41% of the patient population had preoperative TTE DI values <0.25, whereas 74% of their patient population had preoperative TTE DI values <0.25. If DI during pre-CPB TEE slightly overestimates AS severity compared with preoperative TTE, then more agreement is likely when mostly severe AS patients are being evaluated, because one cannot ascribe a grade more severe than severe. Thus, the results may be compatible. Limitations There were several limitations to this study. Given its retrospective nature, selection bias is possible. However, the authors screened a large group of patients over an extended period across multiple National Board of Echocardiographycertified echocardiographers in dedicated cardiology and cardiothoracic anesthesiology fellowship programs. Additionally, all exams were reviewed again by a certified investigator to ensure quality. As mentioned previously, the time between preoperative TTE and intraoperative pre-CPB examinations was a limitation. An absolute cutoff of 5 months between exams was used, and the average time between exams was approximately 29 days. Thus, for some patients, enough time may have elapsed between examinations so that their AS had become worse. The largest limitation resides in the inability to control for hemodynamics. Although the authors’ institutional protocols recommend obtaining pre-CPB AS assessments at a patient’s baseline blood pressure, defined as within 20% of a patient’s preoperative outpatient blood pressure, there was no way to double-check that this always occurred during pre-CPB because the intraoperative anesthetic record time stamps did not always match the TEE imaging time stamps. Likewise, although the authors ensured similar rhythms, the authors did not ensure matching heart rates between examinations. To achieve a coherent comparison between TTE and pre-CPB, the hemodynamic states would need to be matched. If the outpatient TTE blood pressure was not at the patient’s baseline, for example, the pre-CPB blood pressure target would not be the same as during TTE. Additionally, it is unclear if blood pressure or some other cardiac output measure best would harmonize the conditions for TTE and pre-CPB TEE exams. Finally, given the small number of participants with poor left ventricular function (n = 17), care should be taken in interpreting any difference in observed disagreement based on this group of patients. This stratified analysis was exploratory and was meant to provide preliminary information about differences in grading based on left ventricular function.
2383
Clinical Relevance This study reaffirms that PGm and Vp measured by intraoperative pre-CPB TEE underestimates AS severity compared with preoperative TTE. This work suggests that this underestimation holds true across multiple grades of AS as well as in patients with low systolic function. Thus, the intraoperative echocardiographer should apply PGm or Vp grading cutoffs with these limitations in mind when determining AS severity. However, if one obtains a severe AS grade via PGm or Vp during intraoperative pre-CPB TEE, it strongly suggests a severe AS grade by preoperative TTE. Thus, PGm and Vp may serve as a useful screening tool for identifying incidental severe AS. Using DI avoids the pitfall of predictably underestimating AS during pre-CPB; however, this study was unable to show that it was any more reliable than AVA calculation for matching overall grading severity between pre-CPB TEE and TTE. Prospective studies controlling for hemodynamics and standardizing VTI acquisition are needed to better compare DI measurements during TTE versus TEE. Conclusions A stepwise, practical approach weighing all grading parameters and their respective limitations should be used to handle AS grading challenges. The Vp and PGm values during preCPB TEE reliably underestimate AS compared with preoperative TTE. Both AVA and DI have similar grading discrepancy rates when moving from preoperative TTE to pre-CPB TEE. No one grading parameter in isolation can deliver a completely reliable overall grade of AS during pre-CPB TEE. More work is needed to refine a consistent, practical approach to assigning an AS grade in this clinical context. Conflicts of Interests The authors have no conflicts of interest to disclose. References 1 Eltzschig HK, Rosenberger P, Loffler M, et al. Impact of intraoperative transesophageal echocardiography on surgical decisions in 12,566 patients undergoing cardiac surgery. Ann Thorac Surg 2008;85:845–52. 2 Karagounis A, Valencia O, Chandrasekaran V, et al. Management of patients undergoing coronary artery bypass graft surgery with mild to moderate aortic stenosis. J Heart Valve Dis 2004;13:369–73. 3 Nishimura RA, Otto CM, Bonow RO, et al. 2014 AHA/ACC Guideline for the Management of Patients With Valvular Heart Disease: Executive Summary: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2014;63:2438–88. 4 Whitener G, McKenzie J, Akushevich I, et al. Discordance in grading methods of aortic stenosis by pre-cardiopulmonary bypass transesophageal echocardiography. Anesth Analg 2016;122:953–8. 5 Whitener G, Sivak J, Akushevich I, et al. Grading aortic stenosis with mean gradient and aortic valve area: A comparison between preoperative transthoracic and precardiopulmonary bypass transesophageal echocardiography. J Cardiothorac Vasc Anesth 2016;30:1254–9.
2384
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
6 Otto CM, Pearlman AS, Comess KA, et al. Determination of the stenotic aortic valve area in adults using Doppler echocardiography. J Am Coll Cardiol 1986;7:509–17;1986. 7 Oh JK, Taliercio CP, Holmes DR Jr, et al. Prediction of the severity of aortic stenosis by Doppler aortic valve area determination: Prospective Doppler-catheterization correlation in 100 patients. J Am Coll Cardiol 1988;11:1227–34. 8 Galan A, Zoghbi WA, Quinones MA. Determination of severity of valvular aortic stenosis by Doppler echocardiography and relation of findings to clinical outcome and agreement with hemodynamic measurements determined at cardiac catheterization. Am J Cardiol 1991;67:1007–12. 9 Otto CM, Burwash IG, Legget ME, et al. Prospective study of asymptomatic valvular aortic stenosis. Clinical, echocardiographic, and exercise predictors of outcome. Circulation 1997;95:2262–70. 10 Rusinaru D, Malaquin D, Marechaux S, et al. Relation of dimensionless index to long-term outcome in aortic stenosis with preserved LVEF. JACC. Cardiovasc Imaging 2015;8:766–75. 11 Baumgartner H, Hung J, Bermejo J, et al. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. J Am Soc Echocardiogr 2009;22:1–23; quiz 101-2. 12 Baumgartner H, Hung J, Bermejo J, et al. Recommendations on the echocardiographic assessment of aortic valve stenosis: A focused update from the European Association of Cardiovascular Imaging and
13
14 15 16
17 18
19
20
the American Society of Echocardiography. J Am Soc Echocardiogr 2017;30:372–92. Zoghbi WA, Farmer KL, Soto JG, et al. Accurate noninvasive quantification of stenotic aortic valve area by Doppler echocardiography. Circulation 1986;73:452–9. Fleiss JL, Cohen J, Everitt BS. Large sample standard errors of kappa and weighted kappa. Psychol Bull 1969;72:323–7. McHugh ML. Interrater reliability: The kappa statistic. Biochem Med 2012;22:276–82. Otto CM, Pearlman AS, Gardner CL. Hemodynamic progression of aortic stenosis in adults assessed by Doppler echocardiography. J Am Coll Cardiol 1989;13:545–50. Faggiano P, Ghizzoni G, Sorgato A, et al. Rate of progression of valvular aortic stenosis in adults. Am J Cardiol 1992;70:229–33. Sacchi S, Dhutia NM, Shun-Shin MJ, et al. Doppler assessment of aortic stenosis: a 25-operator study demonstrating why reading the peak velocity is superior to velocity time integral. Eur Heart J Cardiovasc Imaging 2018; 19:1380–9. Maslow AD, Mashikian J, Haering JM, et al. Transesophageal echocardiographic evaluation of native aortic valve area: Utility of the double-envelope technique. J Cardiothorac Vasc Anesth 2001;15:293–9. Uda Y, Cowie B, Kluger R. Comparison of preoperative and intraoperative assessment of aortic stenosis severity by echocardiography. Br J Anaesth 2017;118:699–704.
Contents lists available at ScienceDirect
Journal of Cardiothoracic and Vascular Anesthesia journal homepage: www.jcvaonline.com
Original Article
Grading Aortic Valve Stenosis With Dimensionless Index During Pre-cardiopulmonary Bypass Transesophageal Echocardiography: A Comparison With Transthoracic Echocardiography 1 George B. Whitener, MD, FASE*, , Paul C. Shanahany, Bethany J. Wolfz, Alan C. Finley, MD, FASE* *
Department of Anesthesia and Perioperative Medicine, Medical University of South Carolina, Charleston, SC y University of Rochester School of Medicine and Dentistry, Rochester, NY z Department of Public Health Sciences, Medical University of South Carolina, Charleston, SC
Objective: The authors hypothesized that grading valvular aortic stenosis (AS) with dimensionless index (DI) during intraoperative pre-cardiopulmonary bypass (pre-CPB) transesophageal echocardiography (TEE) would match the grade of AS during preoperative transthoracic echocardiography (TTE) for the same patients more often than when using peak velocity (Vp), mean pressure gradient (PGm), or aortic valve area (AVA). Design: Retrospective, observational. Setting: Single university hospital. Participants: The participants in this study included 123 cardiac surgical patients with any degree of AS, who underwent open cardiac surgery between 2010 and 2016 at the Medical University of South Carolina and had Vp, PGm, AVA, and DI values available from reporting databases or archived imaging. Interventions: None. Measurements and Main Results: When using DI, pre-CPB TEE grading of AS severity was 1 grade higher 21.1% of the time and 1 grade lower 13.0% of the time compared with TTE, for an overall disagreement rate of 34.1%. The overall disagreement rates between pre-CPB TEE and TTE for Vp, PGm, and AVA were 39.8%, 33.3%, and 33.3%, respectively. Conclusions: The authors could not demonstrate that DI was better than Vp, PGm, or AVA at matching AS grades between intraoperative preCPB TEE and preoperative TTE. When DI was used, pre-CPB TEE was more likely to overestimate than underestimate the severity of AS compared with TTE. However, when Vp or PGm was used, pre-CPB TEE was more likely to underestimate the severity of AS compared with TTE. A comprehensive approach without overemphasis on 1 parameter should be used for AS assessment by intraoperative TEE. Ó 2019 Elsevier Inc. All rights reserved. Key Words: aortic stenosis; dimensionless index; grading; pre-cardiopulmonary bypass transesophageal echocardiography; intraoperative transesophageal echocardiography
G. Whitener helped design the study, conduct the study, analyze the data, and write the manuscript. P. Shanahan helped conduct the study, acquire the data, and write the manuscript. B. Wolf helped design the study, analyze the data, provide statistical analysis, and write the manuscript. A. Finley helped design the study and edit the manuscript. 1 Address reprint requests to George Whitener, MD, FASE, Department of Anesthesia and Perioperative Medicine, Medical University of South Carolina, 25 Courtenay Dr, 4210 Ashley River Tower, Charleston, SC 29425. E-mail address: [email protected] (G.B. Whitener). https://doi.org/10.1053/j.jvca.2019.03.046 1053-0770/Ó 2019 Elsevier Inc. All rights reserved.
ACCURATELY GRADING VALVULAR AORTIC stenosis (AS) during pre-cardiopulmonary bypass (pre-CPB) transesophageal echocardiography (TEE) is important, especially in the clinical context of incidental AS, when a plan has not been made prior to coming to the operating room to address a pathological valve. The rate of incidental AS is low, but not insignificant. In a cohort of 3,835 patients undergoing coronary artery bypass grafting, 3.3% of patients had an additional mitral or
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
aortic valve procedure added to the operation due to pre-CPB TEE findings.1 In another large retrospective study, the rate of incidental mild or moderate AS in 8,202 patients undergoing coronary artery bypass grafting was reported to be 2.4%.2 The ability to delineate mild, moderate, and severe AS in the setting of another cardiac surgical procedure is important because, in this context, the grade of AS affects surgical decision-making.3 Even under ideal conditions, inconsistencies between AS grading parameters complicate overall severity assignment. Additionally, the presence of general anesthesia complicates intraoperative pre-CPB TEE because of its impact on these grading parameters. For instance, the discordance rate between aortic valve area (AVA) and mean pressure gradiant (PGm) to grade AS is worse during pre-CPB TEE compared with transthoracic echocardiography (TTE).4 Moreover, PGm preferentially underestimates the degree of AS severity during pre-CPB TEE compared with TTE.5 Aortic stenosis grading cutoffs were derived from patients without general anesthesia or positive pressure ventilation.3,69 Nevertheless, these cutoffs are the only grading standard available for the intraoperative echocardiographer. Thus, an
2377
ideal grading parameter for assessing incidental AS during pre-CPB would be affected little by hemodynamic loading conditions, would be reproducible while limiting intrinsic measurement errors, and would have a high correlation between pre-CPB TEE and TTE to allow translation of cutoff values. The dimensionless index (DI) is a potentially ideal AS grading parameter because it does not require measuring the left ventricular outflow tract, as AVA does, and, unlike peak velocity (Vp) and PGm, DI maintains stability across varying loading conditions and angles of Doppler interrogation.10 Therefore, the authors hypothesized that grading AS with DI during pre-CPB TEE would match the grade of AS during TTE more often than when using Vp, PGm, or AVA. Methods Patient Selection After submitting a waiver of consent for a retrospective review and obtaining approval from the Medical University of South Carolina institutional review board, the authors obtained
Fig 1. Flow diagram illustrating selection of study population. AI, aortic insufficiency; AS, aortic stenosis; AV, aortic valve; MR, mitral regurgitation; RV, right ventricle; TEE, transesophageal echocardiography; TR, tricuspid regurgitation; TTE, transthoracic echocardiography.
2378
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
the records for all patients undergoing any cardiac surgical procedure from January 1, 2010 to August 11, 2016, who had any degree (mild, moderate, or severe) of AS listed on the institution’s intraoperative TEE report. The authors simultaneously selected for patients with normal right ventricular function, mild or less aortic regurgitation, mild or less mitral regurgitation, and mild or less tricuspid regurgitation. Once these patients were identified, the following AS grading parameters were obtained from the report: Vp, PGm, AVA, and DI. Of these patients, the authors cross-referenced their respective preoperative TTE reports to compare the aforementioned grading parameters. The authors excluded any patients with TTE reports more than 5 months from the time of surgery. TTE reports closest to the time of surgery were used in situations where multiple studies were available. Subsequently, all matching pre-CPB TEE and TTE examinations were reviewed by an echocardiographer certified by the National Board of Echocardiography in advanced perioperative echocardiography to obtain any missing values and corroborate the reports with imaging. Patients with missing, poor imaging, or inadequate Doppler-derived waveforms were excluded. During review, the authors also found and excluded patients who had a bioprosthetic aortic valve, pericardial effusion, abnormal right ventricular function by TTE, or greater than mild valvular regurgitation by TTE (Fig 1). Echocardiographic Variables and Acquisition The Vp, PGm, and AVA values were recorded routinely and available in both reporting systems. Examinations were reviewed to ensure that measurements were obtained according to guidelines by the American Society of Echocardiography.11,12 The Vp and PGm values were obtained by integrating
a continuous-wave Doppler signal of flow across the aortic valve. For pre-CPB TEE, Vp and PGm were obtained from a deep transgastric long-axis view with the sample gate aligned as parallel to blood flow as possible. If unavailable, a transgastric long-axis view was used. For TTE, aortic valve flow was interrogated from one of multiple viewing angles, including the apical, right parasternal, suprasternal, subcostal, or supraclavicular viewing windows, and the view yielding the highest-velocity jet was used. The AVA was calculated by using the continuity equation6,12,13 for both TTE and pre-CPB TEE: ðCSALVOT Þ ðVTILVOT Þ AVA cm2 ¼ ; VTIAV where CSALVOT represents the cross-sectional area (CSA) of the left ventricular outflow track (LVOT), VTILVOT is the velocity-time integral (VTI) of flow through the LVOT, while VTIAV is the VTI of flow through the aortic valve. To calculate the CSALVOT, an LVOT diameter was obtained in the midesophageal long-axis view in pre-CPB TEE and by the parasternal long-axis view during preoperative TTE. The LVOT diameters were measured 0.5 to 1.0 cm to the ventricular side of the valve annulus at the site of LVOT VTI measurement as was previously recommended by the European Association of Echocardiography and the American Society of Echocardiography.11 The DI was calculated by the following equation12: DI ¼
ðVTILVOT Þ ðVTIAV Þ
The DI routinely was not reported; therefore, the TTE images were reviewed for each patient by a certified echocardiographer to confirm the waveform values for LVOT
Fig 2. TTE images showing (A) LVOT diameter measurement in a parasternal long-axis aortic valve view, (B) LVOT-VTI by pulsed-wave Doppler in an apical view, and (C) AV-VTI by continuous-wave Doppler in an apical view. AV, aortic valve; LVOT, left ventricular outflow tract; TTE, transthoracic echocardiography; VTI, velocity-time integral.
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
VTI by pulsed-wave Doppler and aortic valve VTI by continuous-wave Doppler. Patients with missing or poor imaging were excluded. Of note, the authors assured that the VTIs used to calculate the reported AVA by continuity equation were the same waveforms used to calculate DI. Examples of VTI acquisition for DI calculation along with LVOT measurement for both TTE and TEE are shown in Fig. 2 and 3. Statistical Analysis Descriptive statistics were calculated for all patient characteristics and for the 4 TTE and TEE measures described in the previous section. Patient severity was determined using the grading severity cutoffs taken from current American Heart Association, American College of Cardiology, and American Society of Echocardiography guideline recommendations and applied to each grading parameter for each patient for both pre-CPB TEE and TTE.3,12 The used cutoffs are summarized in Table 1. The authors examined the agreement in severity grading for each metric as determined by TTE and pre-CPB TEE across all patients and performed additional stratification of participants by left ventricular ejection fraction <50% versus 50%. Agreement for each measure was evaluated using the weighted kappa statistic and tested whether the kappa statistic took a value of 0 (indicating no agreement). The weighted kappa statistic ranges between 1 (perfect agreement) and ¡1 (completely contradictory) and accounts for the magnitude of the difference between the 2 methods in estimating agreement.14 Guidelines for interpreting the kappa statistic indicate that values between 0.4 and 0.59 suggest weak agreement, values between 0.6 to 0.79 suggest moderate agreement, and values greater than 0.8 suggest strong agreement.15 The authors also calculated 95% confidence
2379
Table 1 Aortic Stenosis Grading Limits Measurement
Mild
Moderate
Severe
Peak velocity (m/s) Mean gradient (mmHg) Aortic valve area (cm2) Indexed valve area (cm2/m2) Dimensionless index
<3.0 <20 >1.5 >0.85 >0.50
3.0-4.0 20-40 1.0-1.5 0.60-0.85 0.25-0.50
4.0 40 <1.0 <0.6 <0.25
Adapted from Baumgartner et al.11 and Nishimura et al.3
intervals (CIs) for the weighted kappa statistics to compare the level of agreement between TTE and pre-CPB TEE for DI with the agreement for Vp, PGm, and AVA, where nonoverlapping CIs for kappa statistics indicate that the metric with higher agreement between TTE and pre-CPB TEE is significantly better. The authors also estimated the kappa statistic stratified by Simpson’s LVEF of <50% or 50%. As a secondary measure of agreement, the authors looked at percent agreement with 95% CIs for each metric defined as the percent of participants for whom the TTE and TEE severity for each metric was equivalent. Finally, the authors examined the correlation between TTE and TEE measures for each metric using Pearson’s correlation. All analyses were conducted in SAS, version 9.4 (SAS Institute, Cary, NC). Results A total of 319 patients met initial inclusion criteria. Once all pre-CPB and TTE examinations were reviewed, 123 patients were left to compare AS grading parameters between pre-CPB
Fig 3. TEE images in the same patient as Fig 2 showing (A) LVOT diameter measurement in a midesophageal long-axis view, (B) LVOT-VTI by pulsed-wave Doppler in a deep transgastric long-axis view, and (C) AV-VTI by continuous-wave Doppler in a deep transgastric long-axis view. Note the lower Vp (labeled AV Vmax on the image), PGm, LVOT-VTI, and AV-VTI values compared with transthoracic echocardiography in Fig 2. AV, aortic valve; LVOT, left ventricular outflow tract; PGm, mean pressure gradient; TEE, transesophageal echocardiography; Vp, peak velocity; VTI, velocity-time integral.
2380
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
Table 2 Patient Demographics Characteristic Age at surgery, y* Male sex, n (%)y Body surface area, m2* Mean time from TTE to TEE, d* Surgical procedure, n (%)y AVR CABG + AVR CABG TAVR AVR + ascending aortic repair Heart transplant Echo parameter Average LVOT diameter by TTE, cm* Average LVOT diameter by TEE, cm* Average SVI by TTE, mL/m2* Average SVI by TEE, mL/m2*
Average mean gradient by TTE, mmHg* Average mean gradient by TEE, mmHg* Average peak velocity by TTE, m/s* Average peak velocity by TEE, m/s* Average AVA by TTE, cm2* Average AVA by TEE, cm2* Average VTILVOT by TTE, cm* Average VTILVOT by TEE, cm* Average VTIAV by TTE, cm* Average VTIAV by TEE, cm* Average DI by TTE* Average DI by TEE*
n = 123 71.3 § 10.9 84 (68) 1.98 § 0.25 28.8 § 34.6 40 (32.5) 30 (24.4) 23 (18.7) 21 (17.1) 8 (6.5) 1 (0.8) n = 123 2.1 § 0.21 2.2 § 0.24 44.0 § 11 42.6 § 12 35.1 § 18.5 28.1 § 16.7 3.74 § 1.00 3.33 § 0.92 1.08 § 0.44 1.06 § 0.37 24.4 § 5.80 22.2 § 6.40 88.9 § 29.6 86.2 § 29.5 0.30 § 0.12 0.28 § 0.10
Abbreviations: AVA, aortic valve area; AVR, aortic valve replacement; CABG, coronary artery bypass grafting; CPB, cardiopulmonary bypass; DI, dimensionless index; SVI, stroke volume index; TAVR, transcutaneous aortic valve replacement; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; VTI, velocity-time integral. * Data presented as mean § standard deviation. y Data presented as frequency in the population.
and TTE (Fig 1). Baseline patient and echocardiographic characteristics for the study population are shown in Table 2.The impact of pre-CPB TEE for each grading parameter on AS severity compared with preoperative TTE is summarized in Table 3. When using DI to grade AS, pre-CPB TEE produced an AS severity 1 grade higher or 1 grade lower than preoperative TTE, 21.1% and 13.0% of the time, respectively. Any discrepancy in severity was seen 34.1% of the time for DI. When using Vp, pre-CPB revealed an AS severity 1 grade higher, 1 grade lower, or 2 grades lower than preoperative TTE, 4.1%, 30.9%, and 4.9% of the time, respectively. Any discrepancy in grading was seen 39.8% of the time for Vp. When using PGm to grade AS, pre-CPB TEE produced an AS severity 1 grade higher, 1 grade lower, or 2 grades lower than preoperative TTE 2.4%, 29.3%, and 1.6% of the time, respectively. Any discrepancy in grading was seen 33.3% of the time for PGm. When using AVA to grade AS, pre-CPB TEE produced an AS severity 1 grade higher, 1 grade lower, or 2 grades lower than preoperative TTE, 17.9%, 14.6%, and 0.8% of the time, respectively. Any discrepancy in grading was seen 33.3% of the time for AVA.
All 4 metrics’ agreement between pre-CPB TEE and preoperative TTE were significantly greater than would be expected by chance (p < 0.001 for all). Values for the kappa statistic ranged from 0.51 to 0.70, suggesting weak to moderate agreement. PGm exhibited the largest level of agreement with k of 0.70 (95% CI: 0.61-0.80), while DI had the weakest agreement (k = 0.51, 95% CI: 0.39, 0.64). The 95% CIs for k values of PGm and DI did not overlap, indicating that agreement between severity estimated from preoperative TTE and preCPB TEE using PGm was significantly better than the agreement between preoperative TTE and pre-CPB TEE for DI. There were no other notable differences in agreement between the 4 measures. Weighted kappa statistics with 95% CIs also are provided in Table 3. Stratification by Left Ventricular Function As a secondary analysis, the authors also examined agreement between TTE and TEE stratified by left ventricular function, grouping participants according to those with LVEF <50% and those with LVEF 50%. There were 17 participants with LVEF <50% and 106 with LVEF 50%. The impact of pre-CPB TEE for each grading parameter on AS severity when compared with preoperative TTE stratified by left ventricular function is summarized in Table 4. For Vp and PGm, the percentage of participants showing any change in severity between those with LVEF <50% versus >50% was similar. In contrast, for AVA and DI, the percentage of participants with no change in severity was 18.2% and 12.3% lower in those with LVEF 50. All 4 metrics’ agreement between TEE and TTE were significantly greater than would be expected by chance (p < 0.001 for all). Values for the kappa statistic by LVEF status ranged from 0.50 to 0.82 and were similar between patients with LVEF <50 versus 50 for Vp and PGm. Using AVA to determine severity, the kappa value was higher for patients with LVEF <50 compared with LVEF 50 (k = 0.82 v k = 0.58). Of note, none of the patients with LVEF <50 had a severity of 1 grade lower using DI and TEE; however, the agreement between TEE and TTE measures of severity based on DI were similar for those with good or poor ventricular function. Weighted kappa statistics with 95% CIs by left ventricular function also are provided in Table 4. Discussion The authors could not demonstrate that DI was better than Vp, PGm, or AVA at matching AS grades between intraoperative pre-CPB TEE and preoperative TTE. The results showed that DI, Vp, PGm, and AVA all had similar rates of overall disagreement between pre-CPB and TTE, but DI showed less agreement compared with PGm between TEE and TTE when accounting for where dissimilarities occur (Table 3). For instance, Vp and PGm illustrated a predictable tendency to underestimate AS severity during pre-CPB TEE versus TTE. Alternatively, AVA and DI illustrated a less predictable tendency to overestimate AS severity during pre-CPB TEE versus
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
2381
Table 3 Change in AS Severity Based on Imaging Context and Grading Parameter AS Grading Parameter by Preoperative TTE (n = 123)
Impact of pre-CPB TEE on AS grade
1 grade higher No change* 1 grade lower 2 grades lower Any change Kappa statistic (95% CI)
Vp
PGm
AVA
DI
5 (4.1) 74 (60.1) 38 (30.9) 6 (4.9) 49 (39.8) 0.62 (0.51-0.73)
3 (2.4) 82 (66.7) 36 (29.3) 2 (1.6) 41 (33.3) 0.70 (0.61-0.80)
22 (17.9) 82 (66.7) 18 (14.6) 1 (0.8) 41 (33.3) 0.62 (0.49-0.74)
26 (21.1) 81 (65.9) 16 (13.0) 0 (0) 42 (34.1) 0.51 (0.39-0.64)
NOTE. Data expressed as number (percentage) except for entries for the kappa statistic, which are reported as k (95% CI). Abbreviations: AS, aortic stenosis; AVA, aortic valve area; CPB, cardiopulmonary bypass; DI, dimensionless index; Vp, peak velocity; PGm, mean pressure gradient; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography. * The row reporting “No change” is equivalent to the number (percentage) that agree between TEE and TTE.
TTE. A graphical comparison for each parameter is represented in Fig 4. There are physiological reasons, decreased preload and afterload caused by positive-pressure ventilation and general anesthesia, that can explain the tendency for Vp and PGm to underestimate AS during pre-CPB TEE. Additionally, these findings mirror the results from a prior study comparing preoperative TTE with pre-CPB TEE, which showed an average drop of 6.6 mmHg when using TEE.5 This current study showed a similar average decrease in PGm of 7.0 mmHg during pre-CPB TEE. However, it is surprising that DI would not be more consistently matched between pre-CPB TEE and TTE. DI is a theoretical solution to grading disagreement problems based on the anesthetized state. Even if there are changes in loading conditions between imaging contexts, both the numerator and the denominator of the DI equation should move proportionally in the same direction and remain constant in the same patients. The DI also accounts for differences imparted by the imaging
techniques. For example, if differences between TTE and TEE cause changes in the Doppler angle of interrogation with LVOT and aortic outflow, these changes should affect the numerator and denominator of the DI equation proportionally, thereby rendering DI the same. In the results, the AVA and DI appear to track one another, which makes sense given that they both represent the same proportion (VTILVOT/VTIAV), except that AVA takes LVOT diameter into account. The average LVOT diameter for the study patients was slightly higher during TEE (2.2 cm2 v 2.1 cm2), meaning that the decrease in average AVA and increase in severity during TEE is not due to LVOT diameter discrepancies. Therefore, DI’s inability to match severity lies in the fact that VTI measurement differences existed between TTE and pre-CPB TEE. Moreover, a slight tendency for DI to overestimate severity means that, on average, VTIAV increased relative to VTILVOT or VTILVOT decreased relative to VTIAV during TEE. Table 2 shows that the latter scenario occurred in the study. Both average VTILVOT and VTIAV values decreased
Table 4 Change in AS Severity Stratified on Left Ventricular Function Based on Imaging Context and Grading Parameter AS Grading Parameters by Intraoperative TEE as Compared With Preoperative TTE for Participants With LVEF <50 (n = 17) and LVEF 50 (n =106) PGm
Vp
1 grade higher No change 1 grade lower 2 grades lower Any change % agreement (95% CI) Kappa statistic (95% CI)
AVA
DI
LVEF <50
LVEF 50
LVEF <50
LVEF 50
LVEF <50
LVEF 50
LVEF <50
LVEF 50
2 (11.8) 10 (58.5) 5 (29.4) 0 (0.00) 7 (41.2) 58.5 (35.4-82.2) 0.58 (0.16-0.90)
3 (2.83) 65 (61.3) 32 (30.2) 6 (5.70) 41 (38.7) 61.3 (50.3-70.6) 0.59 (0.46-0.72)
1 (5.88) 11 (64.7) 5 (29.4) 0 (0.00) 6 (35.3) 64.7 (42.0-87.4) 0.64 (0.38-0.92)
2 (1.89) 71 (67.0) 31 (29.4) 2 (1.90) 36 (32.1) 67 (56.1-75.9) 0.69 (0.58-0.79)
1 (5.88) 14 (82.4) 2 (11.8) 0 (0.00) 3 (17.6) 82.4 (64.2-1.00) 0.82 (0.63-1.00)
21 (19.8) 68 (64.2) 16 (15.1) 1 (0.9) 41 (36.6) 64.2 (53.2-73.3) 0.58 (0.43-0.72)
4 (23.5) 13 (76.5) 0 (0.00) 0 (0.00) 4 (23.5) 76.5 (56.3-96.6) 0.79 (0.55-1.00)
22 (20.8) 68 (64.2) 16 (15.1) 0 (0.00) 40 (35.7) 64.2 (53.2-43.3) 0.5 (0.36-0.63)
NOTE. Data expressed as number (percentage) except for entries for percent agreement and for the kappa statistic, which are reported as value (95% CI). Abbreviations: AS, aortic stenosis; AVA, aortic valve area; DI, dimensionless index; LVEF, left ventricular ejection fraction; PGm, mean pressure gradient; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; Vp, peak velocity.
2382
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
Fig 4. Scatter plots of TTE (y-axis) by TEE (x-axis) for each metric. The 45o line represents perfect agreement with TEE. Points above the line represent participants for whom values during TTE > TEE and points below the line represent participants for whom values during TTE < TEE. The Pearson’s correlation between TEE and TTE also is provided on each plot. All correlations were significant at p < 0.001. TEE, transesophageal echocardiography; TTE, transthoracic echocardiograph.
when moving from TTE to TEE, which may be attributable to physiological changes from general anesthesia as discussed previously. However, the average VTILVOT decreased more relative to the average VTIAV. There are a few reasons why severity by DI may have changed this way. First, the trend toward worsening AS as measured by DI during pre-CPB TEE may reflect worsening AS. In this case, VTIAV for a given patient would decrease less, relative to VTILVOT. The AVA worsens by 0.1 cm2/y; nevertheless, this is an average, and some patients worsen by up to 0.5 to 0.72 cm2/y.16,17 The authors chose an absolute cutoff of 5 months between preoperative TTE and pre-CPB TEE to reduce the impact of worsening AS between examinations while allowing for a reasonable sample size. Overall, the mean time between examinations was only 29 days, therefore the risk of significantly worsening AS would be small but not eliminated (Table 2).
Secondly, there is known interoperator variability for DI measurements when using VTIs. For example, Sacchi et al.18 demonstrated in a study with 25 operators that the interoperator coefficient of variation was 9.3%, while the intraoperator coefficient of variation was 13.9% for measuring DI. Sixtytwo patients in our study were within §0.05 of a DI severity cutoff on preoperative TTE. Therefore, small measurement differences between operators, amounting to clinically insignificant discrepancies, may have crossed grading cutoffs often enough to lead to multiple reclassifications of severity. Lastly, a mix of LVOT VTI measurements was done with a double envelope technique while others were accomplished with separate pulsed-wave Doppler acquisitions. This difference in acquisition may have had an impact on DI calculations. However, it is difficult to predict in which direction this may have pushed DI when moving from TTE to TEE. Previous work has shown good correlation between a double envelope
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
technique during TEE compared with separate envelope acquisition during TTE, so it is unclear if this would have had any appreciable impact.19 One other study, by Uda et al.,20 attempted a similar comparison between preoperative TTE and pre-CPB TEE. Their analysis showed that PGm and Vp underestimate AS during pre-CPB compared with TTE, but they found fewer patients reclassified to a different AS grade when using DI.20 However, the current study’s patients were taken from a pool consisting of all grades of AS, whereas most of Uda et al.’s patients had severe AS. For example, 41% of the patient population had preoperative TTE DI values <0.25, whereas 74% of their patient population had preoperative TTE DI values <0.25. If DI during pre-CPB TEE slightly overestimates AS severity compared with preoperative TTE, then more agreement is likely when mostly severe AS patients are being evaluated, because one cannot ascribe a grade more severe than severe. Thus, the results may be compatible. Limitations There were several limitations to this study. Given its retrospective nature, selection bias is possible. However, the authors screened a large group of patients over an extended period across multiple National Board of Echocardiographycertified echocardiographers in dedicated cardiology and cardiothoracic anesthesiology fellowship programs. Additionally, all exams were reviewed again by a certified investigator to ensure quality. As mentioned previously, the time between preoperative TTE and intraoperative pre-CPB examinations was a limitation. An absolute cutoff of 5 months between exams was used, and the average time between exams was approximately 29 days. Thus, for some patients, enough time may have elapsed between examinations so that their AS had become worse. The largest limitation resides in the inability to control for hemodynamics. Although the authors’ institutional protocols recommend obtaining pre-CPB AS assessments at a patient’s baseline blood pressure, defined as within 20% of a patient’s preoperative outpatient blood pressure, there was no way to double-check that this always occurred during pre-CPB because the intraoperative anesthetic record time stamps did not always match the TEE imaging time stamps. Likewise, although the authors ensured similar rhythms, the authors did not ensure matching heart rates between examinations. To achieve a coherent comparison between TTE and pre-CPB, the hemodynamic states would need to be matched. If the outpatient TTE blood pressure was not at the patient’s baseline, for example, the pre-CPB blood pressure target would not be the same as during TTE. Additionally, it is unclear if blood pressure or some other cardiac output measure best would harmonize the conditions for TTE and pre-CPB TEE exams. Finally, given the small number of participants with poor left ventricular function (n = 17), care should be taken in interpreting any difference in observed disagreement based on this group of patients. This stratified analysis was exploratory and was meant to provide preliminary information about differences in grading based on left ventricular function.
2383
Clinical Relevance This study reaffirms that PGm and Vp measured by intraoperative pre-CPB TEE underestimates AS severity compared with preoperative TTE. This work suggests that this underestimation holds true across multiple grades of AS as well as in patients with low systolic function. Thus, the intraoperative echocardiographer should apply PGm or Vp grading cutoffs with these limitations in mind when determining AS severity. However, if one obtains a severe AS grade via PGm or Vp during intraoperative pre-CPB TEE, it strongly suggests a severe AS grade by preoperative TTE. Thus, PGm and Vp may serve as a useful screening tool for identifying incidental severe AS. Using DI avoids the pitfall of predictably underestimating AS during pre-CPB; however, this study was unable to show that it was any more reliable than AVA calculation for matching overall grading severity between pre-CPB TEE and TTE. Prospective studies controlling for hemodynamics and standardizing VTI acquisition are needed to better compare DI measurements during TTE versus TEE. Conclusions A stepwise, practical approach weighing all grading parameters and their respective limitations should be used to handle AS grading challenges. The Vp and PGm values during preCPB TEE reliably underestimate AS compared with preoperative TTE. Both AVA and DI have similar grading discrepancy rates when moving from preoperative TTE to pre-CPB TEE. No one grading parameter in isolation can deliver a completely reliable overall grade of AS during pre-CPB TEE. More work is needed to refine a consistent, practical approach to assigning an AS grade in this clinical context. Conflicts of Interests The authors have no conflicts of interest to disclose. References 1 Eltzschig HK, Rosenberger P, Loffler M, et al. Impact of intraoperative transesophageal echocardiography on surgical decisions in 12,566 patients undergoing cardiac surgery. Ann Thorac Surg 2008;85:845–52. 2 Karagounis A, Valencia O, Chandrasekaran V, et al. Management of patients undergoing coronary artery bypass graft surgery with mild to moderate aortic stenosis. J Heart Valve Dis 2004;13:369–73. 3 Nishimura RA, Otto CM, Bonow RO, et al. 2014 AHA/ACC Guideline for the Management of Patients With Valvular Heart Disease: Executive Summary: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2014;63:2438–88. 4 Whitener G, McKenzie J, Akushevich I, et al. Discordance in grading methods of aortic stenosis by pre-cardiopulmonary bypass transesophageal echocardiography. Anesth Analg 2016;122:953–8. 5 Whitener G, Sivak J, Akushevich I, et al. Grading aortic stenosis with mean gradient and aortic valve area: A comparison between preoperative transthoracic and precardiopulmonary bypass transesophageal echocardiography. J Cardiothorac Vasc Anesth 2016;30:1254–9.
2384
G.B. Whitener et al. / Journal of Cardiothoracic and Vascular Anesthesia 33 (2019) 23762384
6 Otto CM, Pearlman AS, Comess KA, et al. Determination of the stenotic aortic valve area in adults using Doppler echocardiography. J Am Coll Cardiol 1986;7:509–17;1986. 7 Oh JK, Taliercio CP, Holmes DR Jr, et al. Prediction of the severity of aortic stenosis by Doppler aortic valve area determination: Prospective Doppler-catheterization correlation in 100 patients. J Am Coll Cardiol 1988;11:1227–34. 8 Galan A, Zoghbi WA, Quinones MA. Determination of severity of valvular aortic stenosis by Doppler echocardiography and relation of findings to clinical outcome and agreement with hemodynamic measurements determined at cardiac catheterization. Am J Cardiol 1991;67:1007–12. 9 Otto CM, Burwash IG, Legget ME, et al. Prospective study of asymptomatic valvular aortic stenosis. Clinical, echocardiographic, and exercise predictors of outcome. Circulation 1997;95:2262–70. 10 Rusinaru D, Malaquin D, Marechaux S, et al. Relation of dimensionless index to long-term outcome in aortic stenosis with preserved LVEF. JACC. Cardiovasc Imaging 2015;8:766–75. 11 Baumgartner H, Hung J, Bermejo J, et al. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. J Am Soc Echocardiogr 2009;22:1–23; quiz 101-2. 12 Baumgartner H, Hung J, Bermejo J, et al. Recommendations on the echocardiographic assessment of aortic valve stenosis: A focused update from the European Association of Cardiovascular Imaging and
13
14 15 16
17 18
19
20
the American Society of Echocardiography. J Am Soc Echocardiogr 2017;30:372–92. Zoghbi WA, Farmer KL, Soto JG, et al. Accurate noninvasive quantification of stenotic aortic valve area by Doppler echocardiography. Circulation 1986;73:452–9. Fleiss JL, Cohen J, Everitt BS. Large sample standard errors of kappa and weighted kappa. Psychol Bull 1969;72:323–7. McHugh ML. Interrater reliability: The kappa statistic. Biochem Med 2012;22:276–82. Otto CM, Pearlman AS, Gardner CL. Hemodynamic progression of aortic stenosis in adults assessed by Doppler echocardiography. J Am Coll Cardiol 1989;13:545–50. Faggiano P, Ghizzoni G, Sorgato A, et al. Rate of progression of valvular aortic stenosis in adults. Am J Cardiol 1992;70:229–33. Sacchi S, Dhutia NM, Shun-Shin MJ, et al. Doppler assessment of aortic stenosis: a 25-operator study demonstrating why reading the peak velocity is superior to velocity time integral. Eur Heart J Cardiovasc Imaging 2018; 19:1380–9. Maslow AD, Mashikian J, Haering JM, et al. Transesophageal echocardiographic evaluation of native aortic valve area: Utility of the double-envelope technique. J Cardiothorac Vasc Anesth 2001;15:293–9. Uda Y, Cowie B, Kluger R. Comparison of preoperative and intraoperative assessment of aortic stenosis severity by echocardiography. Br J Anaesth 2017;118:699–704.