Abstract
Objectives We proposed several new echo-parameters (augmented blood pressure) derived from blood pressure and aortic valve gradient measurements and hypothesized that they can accurately predict 1-year mortality post-TAVR.
Background In the era of transcatheter aortic valve replacement (TAVR), predicting post-procedural outcome of patients is one of the most important goals in structural heart research.
Methods Patients in the Mayo Clinic National Cardiovascular Diseases Registry (NCDR)-TAVR database who underwent TAVR between January 1, 2012 and June 30, 2017 were identified to retrieve baseline demographics, STS risk score, ECG, cardiac computed tomography, echocardiographic and mortality data. Augmented blood pressure parameters and valvulo-arterial impedance were evaluated by Cox regression. After logistic model generation, receiver operating curve analysis was used to assess the model performance against STS risk score.
Results The final cohort contained 883 patients with the mean age of 81.3±8.5 years old and 58.2% were male. The mean STS risk score was 8.1±5.1. The median follow-up duration was 353 days and one-year all-cause mortality rates was 13.3%. Multivariate Cox regression showed that augmented SBP and augmented MAP parameters were independent predictors of 1-year mortality (all p<0.0001). A single-parameter model based on augmented MAP1 supersedes STS risk score in predicting 1-year mortality (AUC 0.697 vs. 0.591, p=0.0055).
Conclusion Augmented mean arterial pressure provides a simple but effective approach for clinicians to quickly estimate the clinical outcome of TAVR patients and should be incorporated in the assessment of TAVR candidacy.
Condensed Abstract We proposed and validated a new echo-parameter, augmented mean arterial pressure, in a cohort contains 883 patients who underwent TAVR between January 1, 2012 and June 30, 2017. Multivariate Cox regression showed that augmented MAP parameters was independent predictors of 1-year mortality (all p<0.0001). A single-parameter model based on augmented MAP supersedes STS risk score in predicting 1-year mortality (AUC 0.697 vs. 0.591, p=0.0055). Augmented MAP provides a simple but effective approach for clinicians to quickly estimate the clinical outcome of TAVR patients and should be incorporated in the assessment of TAVR candidacy.
Background/Introduction
The success of transcatheter aortic valve replacement (TAVR) has substantially changed the landscape of management of aortic valve disease(1–3). With the expansion of TAVR indications, it is anticipated that more TAVR procedures will be performed in the foreseeable future (4). Given the underlying comorbidities, the clinical outcomes after TAVR procedure have gained significant attention that involves both conventional and machine learning research approaches (5–12). There are newer studies focused on the relationship of blood pressure and valvulo-arterial impedance (Zva), which estimates the overall left ventricular afterload, and its effect on TAVR prognosis (8–10,13). Elevated Zva has been reported to be associated with worse quality of life and exercise performance at one year (10), while there are inconsistent results in predicting long-term mortality (14–16). Our group previously reported that cardiac power index and gradient-adjusted cardiac power index are good predictors of 1-year mortality after TAVR (6,7). In the calculation of gradient-adjusted cardiac power index, transvalvular gradient (mean transvalvular gradient or instantaneous peak transvalvular gradient) was added to the systolic blood pressure as augmented systolic blood pressure to calculate augmented mean arterial pressure (MAP). The augmented MAP component is conceptually close to the summation of valvular load and arterial load, which is the numerator of the Zva formula (1–3). Nagura et al. suggested that Zva is more sensitive to the change of stroke value index but not the arterial load, and the potential measuring error from stroke value index can be magnified in Zva calculation(15). Along this line, we considered simplifying Zva by removing the stroke volume index and see whether the augmented blood pressure can predict clinical outcomes in TAVR patients. We hypothesized that augmented systolic blood pressure and augmented mean arterial pressure can accurately predict the 1-year mortality in patients who underwent a TAVR procedure.
Materials and Methods
Study population, baseline demographics and clinical data
A chart review was conducted on patients included in the Mayo Clinic National Cardiovascular Diseases Registry (NCDR)-TAVR database which included patients from three major academic medical centers located in Rochester, MN, Phoenix, AZ, and Jacksonville, FL. We identified all patients aged ≥18 years who underwent TAVR between January 1, 2012 and June 30, 2017. Baseline demographics, lab data, device data, STS-PROM risk score, and follow-up data were directly extracted from the database. Patients with missing values in augmented blood pressure and valvulo-arterial impedance were excluded to ensure the consistency of model development.
Indications for TAVR were based on the standard FDA criteria at the time of the procedure, and primarily included patients deemed at increased risk for surgical aortic valve replacement (SAVR). The Institutional Review Board at Mayo Clinic approved the study protocol and research authorization was provided by all the patients to utilize their medical information.
Transthoracic echocardiography with 2-dimension imaging and Doppler were performed pre-procedure using commercially available ultra-sound scanners (Philips iE33; Phillips Medical Systems, Andover, MA, USA; GE Vivid E9, GE Healthcare, Milwaukee, WI, USA). All echocardiograms were interpreted by experienced echocardiographers who had achieved COCATS level III training, according to the American College of Cardiology, European Association of Echocardiography (EAE) and American Society of Echocardiography (ASE) guidelines (17–19). Offline measurements of the images were obtained using ProSolv Cardiovascular Analyzer 3.0 (ProSolv Cardiovascular Inc., Indianapolis, IN, USA).
Electrocardiography was performed prior to and after the procedure according to the guidelines (20), and the patients were carefully monitored for the occurrence of conduction abnormalities post-procedure. Multi-Detector Computed Tomography (MDCT) performed within a month of the procedure, and pre-procedural sizing of the annulus was performed (21,22). 3mensio Valves software version 7.0 (Pie Medical Imaging, Maastricht, the Netherlands) was used to analyze the Digital Imaging and Communications in Medicine (DICOM) data and to estimate the point in the cardiac cycle when the aortic valve was maximally open using the live movie feature (23). All-cause mortality data were extracted from the NCDR-TAVR database.
Calculation of augmented systolic blood pressure, mean arterial pressure and valvulo-arterial impedance
The augmented blood pressure calculation formulas are stated as below:
Augmented SBP1(AugSBP1): Mean aortic valve gradient (mean AVG) was added to systolic blood pressure (Equation 1), and augmented MAP1(AugMAP1) was calculated by replacing the SBP with augmented SBP1 in the MAP formula (Equation 3);
Augmented SBP2(AugSBP2): Aortic valve maximal instantaneous gradient was added to systolic blood pressure (Equation 2), and augmented MAP2(AugMAP2) was calculated by replacing the SBP with augmented SBP2(Equation 4); and
Augmented MAP3(AugMAP3): Aortic valve mean gradient was added to mean arterial pressure (Equation 5) (6). Valvulo-arterial impedance (Zva) was calculated according to the standard formula by dividing the sum of the systolic blood pressure and mean transvalvular gradient by stroke volume index (SVI)(15).
Statistical analysis
Patients were grouped into alive and deceased groups and analyzed accordingly. Continuous variables were summarized as mean ± standard deviation, and the differences among groups were evaluated with Mann-Whitney U test. Categorical variables were expressed as counts and percentages, and differences among groups were evaluated with Chi-square test. All the two group comparisons were summarized as alive group versus deceased group if not otherwise specified. Kaplan-Meier curve and Cox regression were used for survival analysis; the median of AugSBP1 and AugMAP1 were used as the cutoff to group the patients for Kaplan-Meier analysis. Each augmented blood pressure measurements and STS-PROM risk score were used to develop logistic regression models separately. Receiver operating (ROC) curve analysis with area under the curve (AUC) was used to assess the accuracy of logistic regression models against STS-PROM risk score, and DeLong’s test was used to assess the difference of model accuracy. For all the hypothesis, a p value of less than 0.05 was used as the cutoff of statistical significance. All the analyses were performed in Python version 3.7.10.
Results
Study Population and Baseline Demographics
A total of 883 patients were included for the final analysis after excluding 188 patients in whom values in any augmented blood pressure parameters or Zva were missing. The mean age was 81.3±8.5 years old, 58.2% were male (n=514), and 97.0% were white. The mean STS risk score was 8.1±5.1. The median follow-up duration was 353 days and one-year all-cause mortality rates was 13.3% (n=118). Systolic blood pressure (alive versus deceased: 130.6±20.8 mmHg vs. 117.6±17.4 mmHg, p=0.0001), diastolic blood pressure (68.4±12.8 mmHg vs. 61.4±11.8 mmHg, p<0.0001) and mean arterial pressure (89.1±13.2 mmHg vs. 80.2±11.9 mmHg, p<0.0001) were significantly higher in the alive group. Detailed data are summarized in Table 1 and Supplement Table 1.
Augmented blood pressure, echocardiography and Zva measurements
There were no significant difference in LV ejection fraction (alive versus deceased: 57.3±12.8 vs. 56.4±13.5, p= 0.3959), mean aortic valve gradient (43.4±13.5 vs. 42.3±13.0, p= 0.2739) and maximal aortic valve instantaneous gradient (71.5±21.1 vs. 69.7±20.8, p=0.1760) between the 2 groups. However, stroke volume index was significantly higher in the alive group (44.56±10.17 vs. 42.45±9.56, p= 0.0083). Regarding augmented blood pressure measurements, both AugSBP1 (173.9±25.5 vs. 159.94±22.03, p<0.0001) and AugSBP2 (202.1±30.8 vs. 186.7±28.9, p<0.0001) were significantly higher in the alive group. There were similar findings for AugMAP1 (103.6±14.3 vs. 94.3±12.8, p<0.0001), AugMAP2 (113.0±15.5 vs. 103.2±14.4, p<0.0001) and AugMAP3 (132.5±19.7 vs. 122.5±17.9, p<0.0001). Box plots are used to visualize the augmented blood pressure data (Figure 1). There was no significant difference, but a trend of higher Zva in the alive group (4.1± 1.1 vs. 3.9± 0.9, p=0.07).
Kaplan-Meier analysis and Cox regression
The median of AugSBP1 was 170 mmHg, and the median of AugMAP1 is 102 mmHg. In Kaplan-Meier analysis, both parameters demonstrated significant survival difference when comparing patients with ≥ cutoff vs. patients < cutoff (all p<0.0001) (Figure 2A and 2B). In univariate Cox regression, AugSBP1, AugSBP2, AugMAP1, AugMAP2 and AugMAP3 were independently associated with 1-year all-cause mortality, and these associations remained significant after adjusting for age, sex and STS-PROM risk score in multivariate Cox regression (all p< 0.0001). In contrast, Zva was not independently associated with 1-year mortality in either univariate or multivariate Cox regression. Table 2 summarizes the Cox regression data.
Model development and evaluation
STS score mode had an AUC performance of 0.591 (95%CI 0.532 - 0.649) in predicting 1-year mortality and was used as the reference model to evaluate other models’ performance. The p-values below compared the AUC difference for each model against the STS score model. Among all the logistic regression models developed to predict 1-year post TAVR mortality, AugMAP1(AUC 0.697, 95%CI 0.648 - 0.747, p=0.0055) and AugMAP2(AUC 0.687, 95%CI 0.638 - 0.738, p=0.012) significantly outperformed the STS score. AugSBP1(AUC 0.661, 95%CI 0.611 - 0.715, p=0.069), AugSBP2(AUC 0.645, 95%CI 0.595 - 0.7, p=0.159) and AugMAP3(AUC 0.650, 95%CI 0.600 - 0.705, p=0.115) were comparable/non-inferior in predicting mortality when compared to STS score. The AUC for Zva (0.542, 95%CI 0.490 - 0.597, p=0.24) was lower than the STS score, however the difference was not statistically significant. Figure 3. demonstrates all the ROC curves.
Discussion
In this retrospective study, we were able to demonstrate that baseline augmented mean arterial pressure provided significantly better accuracy in predicting 1-year all-cause mortality in TAVR patients when compared to STS-PROM risk score. Our findings suggest that augmented mean arterial pressure, as a surrogate marker of cardiac contractile reserve, is closely related to clinical outcome in TAVR patients. To the best of our knowledge, this is the first report of using single-parameter models based on augmented blood pressure measurements to predict mortality in TAVR patients and was found to have significantly superior accuracy against STS-PROM risk score.
The physiology meaning of augmented blood pressure
When calculating the augmented blood pressure, we assumed that adding either mean or maximal instantaneous gradient to the systemic systolic blood pressure can reflect the true systolic pressure generated by the left ventricle (6). While non-invasive MAP is practically calculated as the summation of diastolic blood pressure and 1/3 pulse pressure, MAP also equals to the product of cardiac output and systemic vascular resistance (SVR) (assuming a negligible central venous pressure level). Therefore, MAP can be considered as the capability of the left ventricle to generate cardiac output based on a given systemic vascular resistance, as MAP is proportional to cardiac output when systemic vascular resistance is a constant. In the setting of aortic stenosis, higher augmented MAP indicates better cardiac contractile reserve to generate higher blood pressure against the afterload (stenotic aortic valve and SVR). This may also explain the strong hazard ratio (HR 0.95 per mmHg increase, p<0.0001) of AugMAP1 in Cox regression, which is simply reflecting that better contractile reserve should lead to better clinical outcome. Comparing the performance of AugMAP1, AugMAP2, and AugMAP3, the improvement of accuracy (AUC) in predicting mortality suggests that the closer a parameter is to the real mean arterial pressure of the left ventricle (i.e. the contractile reserve of the left ventricle), the better it will be in outcome prediction. As for AugSBP1 and AugSBP2, while the AUC is still better than STS-PROM in predicting 1-year mortality, the p-values were not as significant as AugMAP1 and AugMAP2. This can be explained by the same theory regarding real MAP: since the diastolic blood pressure component was removed, the AugSBP measurements are deviated to the end of systolic blood pressure, and not truly reflecting the real MAP/contractile reserve.
An overlooked outcome predictor: augmented blood pressure
In both univariate and multivariate cox regression analysis, all the augmented blood pressure parameters were shown to be independent predictors of 1-year mortality, and AugMAP1 had the best performance among them (HR 0.96 per mmHg increase, 95% CI: 0.94-0.97, p<0.0001). AugMAP1 also superseded STS risk score in predicting 1-year mortality in head-to-head ROC curve comparison (p=0.0055). As the standard approach of TAVR assessment, STS-PROM score is considered as a good predictor for both short-and long-term post TAVR mortality(24– 26). Hemmann et al. reported that STS-PROM score an AUC of 0.679 (95% CI: 0.610-0.748) in predicting 1-year post TAVR mortality based on a cohort with 426 patients and is superior to EuroSCORE and EuroSCORE2(25). In another study with 3491 TAVR patients, the STS score had an AUC of 0.61 (95% CI: 0.56-0.67) in predicting 30-day post-TAVR mortality (26). Overall, the AUC performance of STS score based on the prior studies is in the range of 0.6-0.7, which is similar with the findings in our patient cohort (AUC 0.591, 95%CI 0.532 - 0.649). Although our cohort size in the current study is slightly smaller due to excluding patients with missing values, the accuracy of this single parameter model using AugMAP1 almost approaches the same level of our previously published GBM machine learning model developed from the same database (5). Actually, among all patients with available AugMAP1 measurements in our TAVR database (n=971), the AUC of AugMAP1 was 0.703 (95%CI: 0.653 - 0.743) in contrast to 0.611(95%CI: 0.554 - 0.665) for STS risk score (p=0.009). Likewise, AugSBP1 had a very nice performance, which was at least non-inferior to STS risk score (AUC 0.661 vs. 0.591, p=0.069). Based on the Kaplan-Meier survival analysis, using the median cutoff values of AugSBP1 (170 mmHg) and AugMAP1 (102 mmHg), we were able to demonstrate excellent performance in differentiating patient mortality (Figure 1A and 1B). The two parameters are simple, easy to calculate, and surprisingly accurate with only few variables as their input. (24– 26).
A possible concern for this approach is that the parameter mostly relies on blood pressure measurements, which varies from time to time, and suboptimal blood pressure control may cause higher blood pressure readings. In our cohort, we observed that all SBP, DBP and MAP measurements were significantly higher in the alive group (all p< 0.0001), however the mean SBP in the alive group was reasonably well controlled (130.6±20.8 mmHg). There was no significant difference between the alive and deceased groups regarding antihypertensive medications. However, considering augmented blood pressure as the surrogate marker of cardiac contractile reserve, we would argue that higher (but under controlled) baseline blood pressure stands for better cardiac contractile reserve in these patients, and can potentially be related to better outcome. This concept is supported by earlier studies that patients with higher blood pressure/ hypertension after TAVR procedure are associated with better outcomes(27), in contrast to lower blood pressure(9). Furthermore, patients who developed hypertension post TAVR procedure were found to have significantly improved cardiac output and stroke volume(27), which suggests these patients had better cardiac contractile reserve and therefore were able to generate higher blood pressure after relieving stenosis at the level of valve, in contrast to those with lower blood pressure(9). Figure 4 demonstrates two representative cases with higher versus lower AugSBP/AugMAP.
Valvulo-arterial impedance: a close shot
Back in 2009, the concept of valvulo-arterial impedance was introduced by Hachicha et al., and this parameter was shown to be independently associated with mortality in aortic stenosis patients who underwent either surgical aortic valve replacement or medical therapy(28). After that, valvulo-arterial impedance was evaluated in multiple studies as a post-TAVR outcome predictor (10,14–16,29,30). A cohort study containing 202 TAVR patients reported that patients who died within 6 months had higher baseline Zva and less improvement in post-TAVR Zva. While the authors reported that baseline Zva was independently associated with 6-months all-cause mortality, this conclusion was only based on univariate logistic regression and was not adjusted for other covariates(16). In terms of longer-term mortality, Katsanos et al. showed that higher baseline Zva was independently associated with mortality at 2 years(14). However, their study also had a smaller patient cohort (n=116), and the mortality endpoint only happened in 21 (18%) patients. Another study based on the OCEAN-TAVR registry included 1004 patients but reported that post-TAVR Zva was not associated with two-year all-cause mortality in multivariate analysis; the correlation between baseline Zva and mortality was not assessed in this study (15). Importantly, none of the above studies had used Zva to generate a single-parameter model and compare the performance with STS risk score. In contrast to above studies (14–16), we did not observe significant difference of Zva between the alive and deceased group, this is likely due to the higher stroke volume index in the alive group, which balanced out the effect of the numerator (AugSBP1). In our cohort, baseline Zva was not an independent predictor of 1-year post-TAVR mortality in either univariate or multivariate Cox regression analysis, and the AUC of Zva logistic regression model was smaller than STS-PROM risk score (0.542 vs. 0.591, p=0.24), although not statistically significant. Compared to the performance of augmented blood pressure measurements, the component of total load (SBP + mean AVG, equals to AugSBP1) is likely the key portion for Zva to be associated with mortality in prior works(14–16). Introducing stroke volume index as the denominator substantially cancelled the contribution of cardiac output from the total load term and can bring additional measuring error to this parameter(15). Our cardiac contractile reserve theory anticipates that the best accuracy Zva can achieve is at the level of AugSBP1 concerning this is the numerator of the formula. However, with the contribution and potential measurement error from stroke volume index, Zva deviates more from the augmented MAP thus had worse performance compared to AugSBP1. Our results comparing AugMAP and Zva support the concept that cardiac contractile reserve is more important than the vascular load in determining the outcome of TAVR patients.
Incorporating augmented MAP in the assessment of TAVR patients
Compared to STS risk score, which requires input of more than 70 variables and relies on an online calculator after completing extensive workup(31), our models with augmented MAP and SBP provide a simple and effective way that only requires 2-3 readily available variables (SBP, DBP and mean AVG) to make a superior prediction. The median cutoffs (AugSBP1: 170 mmHg and AugMAP1: 102 mmHg) can be easily calculated at bedside to provide real-time assessment, which may facilitate the process for workup and decision-making and can potentially change the assessment of TAVR candidacy.
Conclusion
As a single parameter, augmented mean arterial pressure (AugMAP1) supersedes the STS risk score in predicting 1-year mortality in TAVR patients. It provides a simple but effective approach for clinicians to quickly estimate the clinical outcome of TAVR patients and should be incorporated in the assessment of TAVR candidacy.
Limitations
This study is limited by its retrospective nature, and external validation is not available at the time of the model development. To maintain a sufficient sample size, we did not further divide our patient cohort into subgroups for ROC analysis. Concerning the period of TAVR procedures, most of the patients in this cohort were considered intermediate-to-high risk, so low-risk patients may not be well represented in this development cohort. The time from the baseline echocardiography/ blood pressure measurement to the TAVR procedure varied among patients, and the potential changes in between were not considered in our model. While our approach eliminated the contribution of stroke volume index, the component of systolic blood pressure and transvalvular gradient in this parameter still inherited the intrinsic limitation of Zva(32). Also, invasive hemodynamic measurements were not available to validate the correlations between non-invasively measured augmented MAP and the invasive MAP, which is a potential direction for future studies.
Perspectives
Augmented mean arterial blood pressure provides a simple but effective approach for clinicians to quickly estimate the clinical outcome of TAVR patients, and should be incorporated in the assessment of TAVR candidacy.
Data Availability
NA
Legends
Footnotes
Disclosure: None.
Abbreviations
- AVG
- aortic valve gradient
- AugSBP
- augmented systolic blood pressure
- AugMAP
- augmented mean arterial pressure
- CO
- cardiac output
- DBP
- diastolic blood pressure
- MAP
- mean arterial pressure
- SBP
- systolic blood pressure
- SVI
- stroke volume index
- SAVR
- surgical aortic valve replacement
- TAVR
- transcatheter aortic valve replacement
- Zva
- valvulo-arterial impedance