The Incremental Value of Right Ventricular Size and Strain in the Risk Assessment of Right Heart Failure Post - Left Ventricular Assist Device Implantation

Accepted Manuscript

The Incremental Value of Right Ventricular Size and Strain in the Risk Assessment of Right Heart Failure post-Left Ventricular Assist Device Implantation Marie Aymami MD MS , Myriam Amsallem MD MS , Jackson Adams , Karim Sallam MD , Kegan Moneghetti MBBS (hons) , Matthew Wheeler MD PhD , William Hiesinger MD , Jeffrey Teuteberg MD PhD , Dana Weisshaar MD , Jean-Philippe Verhoye MD PhD , Y. Joseph Woo MD , Richard Ha MD , Franc¸ois Haddad MD , Dipanjan Banerjee MD PII: DOI: Reference:

S1071-9164(18)31152-7 https://doi.org/10.1016/j.cardfail.2018.10.012 YJCAF 4224

To appear in:

Journal of Cardiac Failure

Received date: Revised date: Accepted date:

22 May 2018 11 September 2018 24 October 2018

Please cite this article as: Marie Aymami MD MS , Myriam Amsallem MD MS , Jackson Adams , Karim Sallam MD , Kegan Moneghetti MBBS (hons) , Matthew Wheeler MD PhD , William Hiesinger MD , Jeffrey Teuteberg MD PhD , Dana Weisshaar MD , Jean-Philippe Verhoye MD PhD , Y. Joseph Woo MD , Richard Ha MD , Franc¸ois Haddad MD , Dipanjan Banerjee MD , The Incremental Value of Right Ventricular Size and Strain in the Risk Assessment of Right Heart Failure post-Left Ventricular Assist Device Implantation, Journal of Cardiac Failure (2018), doi: https://doi.org/10.1016/j.cardfail.2018.10.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Highlights Predicting right heart failure (RHF) post-LVAD implantation remains challenging.



RV end-diastolic area and strain are complementary prognostic markers.



They provide incremental risk stratification for RHF to validated risk scores.

AC

CE

PT

ED

M

AN US

CR IP T



1

ACCEPTED MANUSCRIPT

The Incremental Value of Right Ventricular Size and Strain in the Risk Assessment of Right Heart Failure post-Left Ventricular Assist Device Implantation

CR IP T

Short title: Right heart failure post-LVAD

Authors: Marie Aymami MD MSa,b,†, Myriam Amsallem MD MSa,c,†*, Jackson Adams1, Karim Sallam MDa, Kegan Moneghetti MBBS (hons)a,d, Matthew Wheeler MD PhDa, William Hiesinger MDe, Jeffrey

AN US

Teuteberg MD PhDa, Dana Weisshaar MDf, Jean-Philippe Verhoye MD PhDa, Y. Joseph Woo MDe, Richard Ha MDg, François Haddad MDa††, Dipanjan Banerjee MDa†† Institutions: aDivision

of Cardiovascular Medicine, Stanford University School of Medicine and Stanford

of Cardiac, Thoracic and Vascular Surgery, University Hospital of Rennes, Rennes, France

cResearch

and Innovation Unit, INSERM U999, DHU Torino, Paris Sud University, Marie Lannelongue

ED

bDivision

M

Cardiovascular Institute, Palo Alto, California

Hospital, Le Plessis Robinson, France

of Medicine, St Vincent’s Hospital, University of Melbourne, Australia

eDepartment

of Cardiothoracic Surgery, Stanford University School of Medicine, Stanford, California

CE

fKaiser

PT

dDepartment

Permanente Northern California Advanced Heart Failure Program, Santa Clara, California of Cardiothoracic Surgery, Kaiser Permanente, California

AC

gDivision

† and †† all

authors contributed equally to the study.

*Corresponding author: Myriam Amsallem MD MS, Stanford University School of Medicine, 300 Pasteur Drive, Stanford, CA 94304, phone: (650) 497 9460; fax: (650) 725-1599; email: [email protected]

2

ACCEPTED MANUSCRIPT

Word count: abstract (200), text (3,532 excluding references, tables and figure legends), 26 references, 4 figures, 2 tables and 6 supplementary materials (2 figure and 4 tables).

PT

ABSTRACT

ED

M

AN US

CR IP T

Graphical abstract

CE

Background: Right heart failure (RHF) post-left ventricular assist device (LVAD) implantation is

AC

associated with high morbi-mortality. Existing risk scores includes semi-quantitative evaluation of right ventricular (RV) dysfunction. This study aims to determine whether quantitative evaluation of both RV size and function improve risk stratification for RHF post-LVAD implantation beyond validated scores. Methods and results: From 2009 to 2015, 158 patients who underwent implantation of continuous-flow device with complete echocardiographic and hemodynamic data were included. Quantitative RV parameters included RV end-diastolic (RVEDAI) and end-systolic area index, RV free-wall longitudinal

3

ACCEPTED MANUSCRIPT

strain (RVLS), fractional area change, tricuspid annular plane systolic excursion, right atrial area and pressure. Independent correlates of early RHF (<30 days) were determined using logistic regression analysis. Mean age was 56 ±13 years, with 79% males; 49% were INTERMACS ≤ 2. RHF occurred in 60 patients (38%) with 20 (13%) requiring right ventricular assist device. On multivariate analysis,

CR IP T

INTERMACS profiles (adjusted odds ratio 2.38 [1.47-3.85]), RVEDAI (1.61 [1.08-2.32]) and RVLS (2.72 [1.65-4.51]) were independent correlates of RHF (all p<0.05). Both RVLS and RVEDAI were incremental to validated risk scores (including the EUROMACS score) for early RHF post-LVAD (all p<0.01).

AN US

Conclusion: RV end-diastolic and strain are complementary prognostic markers of RHF post-LVAD.

AC

CE

PT

ED

M

Keywords: Echocardiography; Imaging; Left Ventricular Assist Device; Outcomes; Right Ventricle.

4

ACCEPTED MANUSCRIPT

INTRODUCTION Right heart failure (RHF) represents a major cause of mortality and morbidity following left ventricular assist device (LVAD) implantation.1,2 In the last few years, several predictive scores for post-LVAD RHF were developed including the CRITT (Central venous pressure, severe Right ventricular dysfunction,

CR IP T

preoperative Intubation, severe Tricuspid regurgitation and Tachycardia), EUROMACS (European Registry for Patients with Mechanical Circulatory Support) and Michigan scores.3–7 Among them, the CRITT score has shown reasonable external validation with c-statistics in the range of 0.65 to 0.70.7 The recently published EUROMACS study derived a score based on a cohort of 2000 patients, including the

AN US

Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profile, hemoglobin, ratio of right atrial/pulmonary capillary wedge pressure (RAP/PAWP), use of multiple inotropes and severe right ventricular (RV) dysfunction assessed semi-quantitatively.8 This score had a c-statistics of 0.67 in their validation cohort, but is still pending external validation.

M

Previous risk scores have included a variety of different preoperative parameters that can be divided into three categories, i.e. right heart parameters (RV systolic or diastolic function, tricuspid

ED

regurgitation TR), heart failure syndrome severity (INTERMACS profile, vasopressor requirements) or end-organ involvement (anemia, hepatic or renal impairment, intubation), as summarized in

PT

supplementary Table S1. However, right heart assessment has been limited by the use of semiquantitative estimation of right heart function in validated risk scores. While preoperative RV free-wall

CE

longitudinal strain (RVLS) has recently emerged as predictor of RHF in several small cohorts,9–11 very few studies have explored the value of RV size12 and No study to date has assessed the complementary

AC

prognostic value of quantitative metrics of RV size and function to predict RHF post-LVAD. For this study, we hypothesized that considering both quantitative RV size and deformation

imaging based RVLS are incremental to validated risk scores in assessing the risk of RHF post-LVAD implantation. The first objective of our study was to define the prevalence of right heart dysfunction and enlargement based on quantitative echocardiography in patients referred for continuous flow-LVAD implantation. The second objective was to determine whether quantitative evaluation of both RV size and

5

ACCEPTED MANUSCRIPT

RVLS are complementary in assessing the risk of RHF. Finally, as semi-quantitative evaluation of RV parameters is incorporated in previous risk scores, we wanted to evaluate the concordance between semiquantitative and quantitative evaluation of RV function and size.

CR IP T

MATERIAL and METHODS Study population

This study is a retrospective study based on Stanford prospective registry of patients referred for LVAD implantation. Between February 2009 and August 2015, 191 consecutive patients (older than 18 years

AN US

old) underwent isolated LVAD implantation at Stanford University Medical Center (CA, USA). Patients with left heart failure requiring continuous-flow LVAD implantation were included if total bilirubin levels were available within two weeks of echocardiography. Patients were excluded if comprehensive echocardiographic study (n=18) allowing evaluation of the right heart or complete right heart

M

catheterization data (n=15) were not available. Isolated LVAD implantations were done under cardiopulmonary bypass on a beating heart. Stanford University Institutional Review Board approved the

ED

study (#25673), which was conducted in agreement with the Helsinki-II declaration. Informed consent was obtained from all patients.

PT

Echocardiography

Digitized studies were acquired using Philips IE 33 ultrasound systems (Philips, Amsterdam, The

CE

Netherlands). All measures of dimensions and function were averaged over 3 cycles and analysed according to the latest guidelines by two blinded certified readers (My.A. and M.A.). RV end-diastolic

AC

and end-systolic area were measured from an RV focused apical 4-chamber view and indexed to body surface area (RVEDAI and RVESAI respectively). RA area indexed on body surface area was measured the RV focused apical 4-chamber view. RV function was quantified using free-wall Lagrangian longitudinal strain (RVLS), RV fractional area change (RVFAC) and tricuspid annular plane systolic excursion (TAPSE). RVLS was measured from mid-endocardial end-diastolic and end-systolic manually traced lengths of the RV free-wall, calculated as (end-systolic length−end-diastolic length)/end-diastolic

6

ACCEPTED MANUSCRIPT

length and expressed as absolute values, as previously validated.13 In the absence of evidence-based consensus on thresholds of RV enlargement or dysfunction severity, we predefined them according to previous published healthy cohorts (as 30% of the upper limit for areas and 40% of the lower limit for function). RV enlargement was classified as severe if RVEDAI > 16cm2/m2 in men and > 15cm2/m2 in

CR IP T

women,14 and RV dysfunction as severe if RVLS < 12%.15 RAP was estimated from the inferior vena cava size and collapse as previously validated and classified as significantly increased (≥ 15mmHg) if the inferior vena cava > 2 cm and the collapse index was < 50%. Severe TR was defined in the presence of a vena contracta ≥ 7mm, reversed systolic hepatic vein flow, proximal isovelocity surface area radius >

Clinical, Hemodynamic and Laboratory data

AN US

9mm, and very large central jet or moderate eccentric wall impinging jet.16

Clinical and laboratory preoperative data (including serum creatinine and total bilirubin) available within a month of echocardiography were also included. The Michigan score’s total bilirubin threshold of ≥ 2.0

M

mg/dL was chosen.3 Right heart catheterization within 3 months of inclusion (available within a week for 63% of patients) was performed through the internal jugular or right femoral vein. RAP, systolic, mean

ED

and diastolic pulmonary arterial pressures, PAWP, pulmonary vascular resistance and pulmonary vascular resistance index (measured as transpulmonary gradient divided by cardiac index) and cardiac index (using

PT

the thermodilution method) were measured. Preoperative catheterization or echocardiograms were not routinely repeated the day before surgery if the patient’s clinical status was deemed unchanged compared

CE

to baseline studies.

AC

Endpoints

Patients were prospectively followed after enrolment in the registry; follow-up was concluded on October 2016. The primary endpoint was early (<30 days) RHF predefined according to the INTERMACS definition as sustained elevation of central venous pressure >16 mmHg and the need for prolonged inotropes beyond 7 days, or the unplanned need for RV assist device implantation. Two physicians independently verified RHF adjudication (MA and DB); consensus had to be reached prior to data

7

ACCEPTED MANUSCRIPT

analysis. As inhaled nitric oxide was often prophylactically used at our institution during the study period, it could not be used to define RHF. The secondary endpoint was unplanned need for RVAD post-LVAD implantation. The main criteria for RVAD placement was progressive RV dysfunction as measured by echocardiography and hemodynamic indices (high central venous pressure >16mmHg) leading to

CR IP T

increasing doses of inotropes with evidence of end organ malperfusion (e.g. rising serum lactate levels, liver function tests and/or serum creatinine levels). Death was additionally verified through the National Social Security Death Index, whereas heart transplantation and hospitalization for heart failure (defined by >24 hours of hospitalization) were verified through chart review and follow-up.

AN US

Risk scores

The following scores (Table S1) were calculated: Michigan score, Penn RVAD risk score, Utah RV risk score, CRITT score, score derived from the Kormos model’s variables, EUROMACS score.3–8 Statistical analysis

M

Data are summarized as mean ±standard deviation (SD) for continuous variables according to the central limit theorem, and number (%) for categorical variables. Comparisons between two groups were

ED

performed using Student t-tests for continuous variables and Chi square tests for categorical data; comparisons between multiple groups were performed using one-way variance analysis (ANOVA).

PT

Correlations between right heart variables were expressed using Pearson’s correlation coefficients (r) and their p-values. Transplant-free survival estimates were based on the Kaplan-Meier method and compared

CE

by log-rank statistics. Logistic regression models were used to define association with outcome (as all events considered occurred in the postoperative 30-day period). First, a model of right heart parameters

AC

was built to determine which independent variables (with p<0.10) to add in the hierarchical model (backward). Receiver operating characteristic curves were constructed to estimate the c-statistics of risk scores with 95% confidence interval using early RHF as the binary outcome, and to determine criterion of right heart metrics for diagnostic accuracy of RHF. The incremental value of right heart quantitative parameters to INTERMACS profiles and risk scores (excluding variables already included in scores) was assessed using multivariable logistic regression, and expressed using Chi square values. The

8

ACCEPTED MANUSCRIPT

INTERMACS profile was entered as categorical: 1, 2 and ≥3, using the latter (≥3) as the reference group for risk assessment. Interobserver reproducibility for right heart assessment was estimated in 20% of the cohort interpreted by a second certified cardiologist (MA) blinded to the first interpretation, and expressed using intraclass correlation coefficients and their 95% confidence interval. P-values <0.05 were

CR IP T

considered statistically significant. All statistical analyses were performed using SPSS version 23.0 (SPSS, Chicago, IL).

RESULTS

AN US

Study population

In total, 158 patients were included in the study: 94 underwent Heartmate II (manufactured by Thoratec Corp., now Abbott Laboratory, Pleasanton, CA) placement, 57 received the HeartWare HVAD device (manufactured by HeartWare Corp., now Medtronic, Framingham, MA) and 7 underwent Jarvik 2000

M

(Jarvik Heart Inc., New York, NY) implantation. Table 1 summarizes the baseline clinical, hemodynamic and echocardiographic characteristics of the included population (n=158). Mean age was 56 ±13 years

ED

with 79% of males and 49% were INTERMACS 1 or 2. Nine patients underwent concomitant valvular surgery, including 4 tricuspid valve repairs. When comparing the final cohort (n=158) to the initial cohort

PT

(n=191), there was no difference in age (p=0.65), sex (p=0.79), INTERMACS profiles (p=0.93) or risk of RHF (39.7% before and 38.0%, p=0.74). Supplementary Table S2 presents the characteristics of the 33

CE

patients excluded from the study, who had more frequently ischemic etiology of heart failure (49%), but no difference in terms of baseline clinical demographics or incidence of RHF or RVAD need.

AC

Outcomes

RHF occurred in 60 of the 158 patients (38.0%), including 20 patients (12.7% of the total cohort) requiring unplanned RVAD implantation (with a median time from LVAD to RVAD of 1 day). The incidental rates of RHF and RVAD need according to the year of inclusion are presented in supplementary Table S3. There was no significant difference in incidence of RHF according to the type of device implanted (37.1% for Heartmate II and HVAD versus 57.1% for Jarvik devices, p=0.29). Early

9

ACCEPTED MANUSCRIPT

post-operative complications and mortality was higher in patients with RHF (Figure 1). One-year transplant-free survival (±standard error) was significantly better in patients without RHF (66.1 ±4.8%) than those with RHF who did not required RVAD (37.0 ±7.7%) and those with RHF requiring an RVAD (15.0 ±8.0%). Fourteen patients underwent cardiac transplant because of RHF onset, including 4 patients

CR IP T

who required RVAD. One patient initially implanted with LVAD as destination therapy underwent cardiac transplant. Transplant-free survival curves of survivors at 30 days are presented in supplementary Figure S1; while that most of the risk occurs during the 30 days, the survival curves significantly differed according to RHF onset (log rank test p<0.001).

AN US

Prevalence of abnormal right heart characteristics

Figure 2A presents the correlations between right heart imaging metrics in the population. Correlation between RVLS and RVEDAI (r=-0.26) was weaker than the correlation between RVLS and RVESAI (r=0.45), although borderline for statistical significance (p=0.05). Severe RV end-diastolic enlargement

M

(RVEDAI >15cm2/m2 in women and >16 in men) was observed in 53.7%, severe RV dysfunction (RVLS <12%) in 27.2% of patients, severe TR in 22.2%, RAP ≥15mmHg in 67.7% and increased bilirubin

ED

(≥2.0mg/dL) in 18.4% (Figure 2B-D). Among the 85 patients with severe RV enlargement, 26 (30.6%) had severe TR, including one patient who underwent tricuspid valve repair concomitant to LVAD

PT

implantation. Figure 2E illustrates using a Venn diagram the overlap between abnormal features based on RVLS, RVEDAI and RAP.

CE

Univariable and multivariable analysis Among right heart metrics, RVLS, RVEDAI, RVESAI, RAP, RAP/PAWP, total bilirubin and

AC

INTERMACS profiles were strongly associated with an increased risk of RHF on univariable analysis (Table 2). Severe RV dysfunction (based on RVLS) and RV end-diastolic enlargement were associated with a three-fold increased risk of RHF (p<0.01). To minimize overfitting the multivariable model, only 6 variables were included in addition to age and sex: INTERMACS profile, RVLS, RVEDAI, RAP/PAWP ratio, total bilirubin and severe TR. The choice of variables was based on the following rationale: (1) INTERMACS categories (1, 2 and ≥3)

10

ACCEPTED MANUSCRIPT

as previously reported,2 (2) RVLS emerges as a strong correlate of outcome in a recent meta-analysis,11 (3) RVEDAI as it was less correlated to RVLS in Figure 2A, suggesting more complementarity, (4) RAP/PAWP ratio emerging as a strong predictor of outcome in the recent EUROMACS cohorts,8 (5) total bilirubin as part in the Michigan’s score,3 and (6) severe TR according to the CRITT definition, as it can

CR IP T

influence estimation of RV dimensions. On multivariable analysis, INTERMACS profile (adjusted OR = 2.38 [1.47-3.85], p<0.001), RVLS (OR per standard deviation worsening = 1.57 [1.04-2.31], p=0.03) and RVEDAI (OR per standard deviation increase = 1.61 [1.08-2.32], p=0.02) were retained in the model for RHF (χ2 = 36.6, p<0.001). When performing a similar multivariable model replacing RVEDAI by

AN US

RVESAI, only INTERMACS profile (adjusted OR = 1.79 [1.33-3.57], p<0.01) and RVESAI (OR per standard deviation increase = 1.69 [1.15-2.52], p<0.01) were retained in the model (χ2 = 29.8, p<0.001). Using post-hoc receiver operating characteristic analysis, the criterion of RVLS and RVEDAI for balanced sensitivity (Se) and specificity (Spe) were: RVLS <12.7% (Se 60%, Spe 73%) and RVEDAI

M

>15.3cm2/m2 in women (Se 83%, Spe 67%) and >16.1cm2/m2 in men (Se 65%, Spe 52%). Incremental value of RVLS and RVEDAI to risk scores

ED

The risk of RHF was higher in presence of both preoperative RV dysfunction and RV enlargement than in presence of only one abnormality (p<0.03, Figure 3A). Patients with both RV

PT

dysfunction and enlargement had similar risk of RHF (about 70%) than patients in INTERMACS class 1 (Figure 3B). Most of patients with RV dysfunction and enlargement were in INTERMACS class 2 (39%)

CE

or ≥3 (43%), suggesting additive value of having both abnormal phenotypes to the INTERMACS classification for prediction of early RHF post-LVAD implantation.

AC

The performance of previously published scores for RHF prediction was moderate (c-statistic

ranging from 0.51 to 0.67) with the best nominal discrimination being for the CRITT and EUROMACS score, as shown in Figure 3C. Quantitative metrics (RVLS and RVEDAI) improved the discrimination of previously validated scores (Figure 3D). For the Michigan score, adding RVLS and RVEDAI (as continuous variables) to the model improved discrimination for RHF (c-statistic of 0.70 [0.62 – 0.77]). For the Kormos model, adding RVLS and RVEDAI also significantly improved discrimination (0.71

11

ACCEPTED MANUSCRIPT

[0.64 – 0.78]). For the EUROMACS score, adding RVEDAI (as RV dysfunction is already included) also improved discrimination for RHF (0.69 [0.61 – 0.76]). For the INTERMACS profiles, adding RVLAS and RVEDAI also improved discrimination for RHF (0.75 [0.67 – 0.81]). RVAD outcome analysis

CR IP T

Univariate analysis for RVAD need (secondary endpoint) is presented in supplementary Table S4. While INTERMACS profiles were significantly associated with the need of need for RVAD placement (OR=0.50 [0.26-0.95], p=0.03), neither RVLS nor RVEDAI were reaching statistical significance. Adding RVLS and RVEDAI (as continuous variables) did not significantly improved the risk

AN US

stratification for RVAD need of risk scores (Figure S2). Methodology of RV assessment

The association between semi-quantitative assessment of RV size and function reported by the Stanford Hospital Echocardiography Laboratory and quantitative metrics of RVLS or RVEDAI are presented in

M

Figure 4A. The semi-quantitative distinction between moderate and severe alterations demonstrated significant overlap. Classification of severe RV dysfunction also demonstrated significant variability

ED

according to the metrics used (Figure 4B). Quantitative assessment of right heart size and function demonstrated excellent reproducibility; the intraclass correlation coefficient was 0.96 [0.90-0.98] for

PT

RVEDAI, 0.98 [0.95-0.99] for RVESAI, 0.92 [0.83-0.97] for RVLS, 0.90 [0.77-0.96] for RVFAC, 0.72

CE

[0.37-0.88] for TAPSE and 0.97 [0.94-0.99] for RA area index.

DISCUSSION

AC

The main finding of this study is that preoperative quantitative measurement of RV size is complementary to strain to improve risk stratification for RHF in patients undergoing implantation for continuous-flow LVAD.

Early moderate to severe RHF is an important cause of mortality and morbidity following LVAD implantation, as confirmed in our study. In the era of continuous-flow LVAD, RHF occurs in 9 to 42% of cases, depending on the diagnostic criteria and severity used.17–19 In an effort to better estimate surgical

12

ACCEPTED MANUSCRIPT

risk and for plan preemptive RVAD, several pre-LVAD scores have been developed. These scores have often considered different factors including right heart parameters (semi-quantitative RV systolic function, TR, RV stroke work), acuity and severity of heart failure syndrome (INTERMACS profile, vasopressor requirements), end-organ involvement (liver or renal impairment, intubation) or medication

CR IP T

profile.3–6,8 Consistent with previous studies, the CRITT score demonstrated moderate discrimination in our cohort with a c-statistic around 0.67.20 Our study is also the first to externally validate the recently described EUROMACS showing a similar discrimination to the CRITT score (c-statistic of 0.66).8 In contrast, RAP/PAWP or hemoglobin were not as strongly associated with outcome in our cohort.

AN US

Our findings also highlight the high prevalence of preoperative RV severe dysfunction, and its variability according to the quantitative metrics used. In our cohort, the prevalence of severe RV dysfunction was ranging from 27% when using the 2005 ASE guidelines thresholds for RVFAC (<17%) 21 to 37% when using RVLS (<12%) or TAPSE (<12mm). It should be acknowledged that these different

M

metrics of RV dysfunction are not equivalent, since they provide information on different components of RV systolic function (i.e. radial versus longitudinal myocardial fiber shortening).22 In addition, the

ED

thresholds of dysfunction severity remain not consensual and require further investigations, specifically across different etiology of right heart remodeling.

PT

In patients referred for LVAD implantation, RVLS has been the most promising prognostic imaging marker for stratifying patients at risk of early RHF post-implantation. Three studies have shown

CE

the value of two-dimensional speckle-tracking based RV free-wall longitudinal strain to predict RHF.9,11,23 The largest study by Kato et al. consisted of 117 patients, finding a cut-off value of -9.6% as significantly

AC

increasing the Michigan score c-statistics from 0.66 to 0.77 (p<0.01).23 One important caveat of RVLS is the need for good quality acoustic windows, which may be challenging in this particular population. In our study, we used the manual method to overcome this issue. The threshold identified by our method was slightly higher in absolute value (-12.7%), highlighting the importance of methodological consideration when using novel metrics.

13

ACCEPTED MANUSCRIPT

While strain has been the focus of several publications, there is less evidence on the value of quantitative RV size for risk prediction of RHF post-LVAD. To our knowledge, only one study using three-dimensional echocardiography by Kiernan et al. has shown the value of RV end-systolic volume as prognostic markers of RHF.12 The original contribution of our study, using two-dimensional

CR IP T

echocardiography, is to demonstrate the complementarity of quantitative RV end-diastolic size to strain for risk prediction of RHF post-LVAD. As RVEDAI was less correlated to RVLS than RVESAI, it demonstrated a better complementarity than RVESAI, which integrates both RV remodeling and functional features of the right ventricle.24 From a pathophysiological perspective, the presence of

AN US

preoperative RV enlargement is believed to impair the ability of the right heart to adapt to septal shifts or change in geometry that can occur following LVAD implantation; hence being a significant contributor to RHF post-LVAD.25 The main methodological challenge of measuring RV areas is due to the crescent shape of the RV and the need to ensure a focused RV view and absence of foreshortening. This

M

measurement is best achieved using the RV focused apical 4-chamber view, as recommended in the latest guidelines and as routinely performed at our institution, ensuring excellent inter-observer

Clinical implications

ED

reproducibility.14,26

PT

The main clinical implication of our study is to highlight the complementarity of quantitative measurements of RV end-diastolic area index and RVLS for risk prediction of RHF post-LVAD and their

CE

incremental value to previous risk scores, including the recent EUROMACS score. The other major finding of our cohort points out to the variability of semi-quantitative right heart assessment using

AC

echocardiography. Within our lab, there was an acceptable overall correlation between actual measures and the semi-quantitative assessment. However there was still significant overlap between moderate and severe classifications. As right heart evaluation still relies on semi-quantitative assessment in routine clinical practice exposing to inter-lab variability, it may partially contribute to the moderate external validation of previous risk scores. This could be an argument to favour quantitative analysis and integrating multiple complementary metrics for right heart assessment pre-LVAD implantation.

14

ACCEPTED MANUSCRIPT

Study limitations This study has some limitations. The first limitation comes from the single-center study design. However, it was primarily intended as a validation of previous studies and exploration of the incremental value of RV size and strain for risk prediction of RHF post-LVAD. This is, to date, the first external

CR IP T

validation of the EUROMACS score. The second limitation may come from the exclusion of 17% of the initial population. However, the included cohort remained representative of the overall continuous-flow cohort and the rate of RHF did not change after exclusions. The third limitation of using RV areas is their load dependency, which has not been explored in this study. The fourth limitation is that the study is

AN US

likely underpowered to demonstrate the value of quantitative right heart metrics for risk prediction of the secondary endpoint (i.e. RVAD need post-LVAD). Finally, this study has not prospectively collected data on perioperative complications and aggravating factors that affect the right ventricle (such as the number

M

of blood transfusion).

CONCLUSIONS

ED

The prediction of right heart failure after LVAD implantation remains challenging. This study highlights the complementarity of RV end-systolic dimension and free-wall longitudinal strain for risk

AC

CE

PT

prediction of RHF post-LVAD implantation.

15

ACCEPTED MANUSCRIPT

ACKNOWLEDGMENTS The authors would like to thank Stanford Cardiovascular Insitute, the Pai Chan Lee Research Fund and the Orathi Foundation for their support.

CR IP T

FUNDING: M. Aymami received a grant from the Federation Francaise de Cardiologie (France). M. Amsallem received a 2016 Young Investigator Seed Grant from the Vera Moulton Wall Center at Stanford (USA) and is supported by a public grant overseen by the French National Research Agency as part of the second Investissements d’Avenir program (ANR-15-RHUS-0002). None of the funding source

AN US

has been involved in the study design, collection, analysis and interpretation of data, writing of the report; or in the decision to submit the article for publication.

DISCLOSURE: None of the authors have any conflict of interest relative to this study. J. Teuteberg has

M

received financial support from Medtronic, Abiomed and CareDx (as a speaker and member of advisory boards) and from Abbott (as member of the HeartMate3 clinical events committee). D. Banerjee has

ED

received financial research support from Abbott and Medtronic. All authors have approved the final

AC

CE

PT

article.

16

ACCEPTED MANUSCRIPT

REFERENCES

1. Patlolla B, Beygui R, Haddad F. Right-ventricular failure following left ventricle assist device implantation. Curr Opin Cardiol 2013;28:223–233.

CR IP T

2. Kirklin JK, Pagani FD, Kormos RL, Stevenson LW, Blume ED, Myers SL, Miller MA, Baldwin JT, Young JB, Naftel DC. Eighth annual INTERMACS report: Special focus on framing the impact of adverse events. J Heart Lung Transplant 2017;36:1080–1086. 3. Matthews JC, Koelling TM, Pagani FD, Aaronson KD. The right ventricular failure risk score a pre-operative tool for assessing the risk of right ventricular failure in left ventricular assist device candidates. J Am Coll Cardiol 2008;51:2163–2172.

AN US

4. Kormos RL, Teuteberg JJ, Pagani FD, Russell SD, John R, Miller LW, Massey T, Milano CA, Moazami N, Sundareswaran KS, Farrar DJ, HeartMate II Clinical Investigators. Right ventricular failure in patients with the HeartMate II continuous-flow left ventricular assist device: incidence, risk factors, and effect on outcomes. J Thorac Cardiovasc Surg 2010;139:1316–1324.

M

5. Fitzpatrick JR, Frederick JR, Hsu VM, Kozin ED, O’Hara ML, Howell E, Dougherty D, McCormick RC, Laporte CA, Cohen JE, Southerland KW, Howard JL, Jessup ML, Morris RJ, Acker MA, Woo YJ. Risk score derived from pre-operative data analysis predicts the need for biventricular mechanical circulatory support. J Heart Lung Transplant 2008;27:1286–1292.

ED

6. Drakos SG, Janicki L, Horne BD, Kfoury AG, Reid BB, Clayson S, Horton K, Haddad F, Li DY, Renlund DG, Fisher PW. Risk factors predictive of right ventricular failure after left ventricular assist device implantation. Am J Cardiol 2010;105:1030–1035.

PT

7. Atluri P, Goldstone AB, Fairman AS, MacArthur JW, Shudo Y, Cohen JE, Acker AL, Hiesinger W, Howard JL, Acker MA, Woo YJ. Predicting right ventricular failure in the modern, continuous flow left ventricular assist device era. Ann Thorac Surg 2013;96:857– 863.

AC

CE

8. Soliman OI, Akin S, Muslem R, Boersma E, Manintveld OC, Krabatsch T, Gummert JF, By TMMH de, Bogers AJJC, Zijlstra F, Mohacsi P, Caliskan K, EUROMACS investigators. Derivation and Validation of a Novel Right-Sided Heart Failure Model After Implantation of Continuous Flow Left Ventricular Assist Devices:The EUROMACS (European Registry for Patients with Mechanical Circulatory Support) Right-Sided Heart Failure Risk Score. Circulation 2018;137:891-906. 9. Cameli M, Righini FM, Lisi M, Bennati E, Navarri R, Lunghetti S, Padeletti M, Cameli P, Tsioulpas C, Bernazzali S, Maccherini M, Sani G, Henein M, Mondillo S. Comparison of right versus left ventricular strain analysis as a predictor of outcome in patients with systolic heart failure referred for heart transplantation. Am J Cardiol 2013;112:1778–1784. 10. Boegershausen N, Zayat R, Aljalloud A, Musetti G, Goetzenich A, Tewarie L, Moza A, Amerini A, Autschbach R, Hatam N. Risk factors for the development of right ventricular 17

ACCEPTED MANUSCRIPT

failure after left ventricular assist device implantation-a single-centre retrospective with focus on deformation imaging. Eur J Cardio-Thorac Surg 2017;52:1069-1076. 11. Bellavia D, Iacovoni A, Scardulla C, Moja L, Pilato M, Kushwaha SS, Senni M, Clemenza F, Agnese V, Falletta C, Romano G, Maalouf J, Dandel M. Prediction of right ventricular failure after ventricular assist device implant: systematic review and meta-analysis of observational studies. Eur J Heart Fail 2017;19:926-946.

CR IP T

12. Kiernan MS, French AL, DeNofrio D, Parmar YJ, Pham DT, Kapur NK, Pandian NG, Patel AR. Preoperative three-dimensional echocardiography to assess risk of right ventricular failure after left ventricular assist device surgery. J Card Fail 2015;21:189–197.

AN US

13. Amsallem M, Sweatt AJ, Aymami MC, Kuznetsova T, Selej M, Lu H, Mercier O, Fadel E, Schnittger I, McConnell MV, Rabinovitch M, Zamanian RT, Haddad F. Right Heart EndSystolic Remodeling Index Strongly Predicts Outcomes in Pulmonary Arterial Hypertension: Comparison With Validated Models. Circ Cardiovasc Imaging 2017;10.pii:e005771.

M

14. Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, Flachskampf FA, Foster E, Goldstein SA, Kuznetsova T, Lancellotti P, Muraru D, Picard MH, Rietzschel ER, Rudski L, Spencer KT, Tsang W, Voigt J-U. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2015;28:1-39.e14.

ED

15. Amsallem M, Boulate D, Kooreman Z, Zamanian RT, Fadel G, Schnittger I, Fadel E, McConnell MV, Dhillon G, Mercier O, Haddad F. Investigating the value of right heart echocardiographic metrics for detection of pulmonary hypertension in patients with advanced lung disease. Int J Cardiovasc Imaging 2017;33:825–835.

CE

PT

16. Amsallem M, Sternbach JM, Adigopula S, Kobayashi Y, Vu TA, Zamanian R, Liang D, Dhillon G, Schnittger I, McConnell MV, Haddad F. Addressing the Controversy of Estimating Pulmonary Arterial Pressure by Echocardiography. J Am Soc Echocardiogr 2016;29:93-102.

AC

17. Ochiai Y, McCarthy PM, Smedira NG, Banbury MK, Navia JL, Feng J, Hsu AP, Yeager ML, Buda T, Hoercher KJ, Howard MW, Takagaki M, Doi K, Fukamachi K. Predictors of severe right ventricular failure after implantable left ventricular assist device insertion: analysis of 245 patients. Circulation 2002;106:I198-202. 18. Dang NC, Topkara VK, Mercando M, Kay J, Kruger KH, Aboodi MS, Oz MC, Naka Y. Right heart failure after left ventricular assist device implantation in patients with chronic congestive heart failure. J Heart Lung Transplant 2006;25:1–6. 19. Grant ADM, Smedira NG, Starling RC, Marwick TH. Independent and incremental role of quantitative right ventricular evaluation for the prediction of right ventricular failure after left ventricular assist device implantation. J Am Coll Cardiol 2012;60:521–528.

18

ACCEPTED MANUSCRIPT

20. Kalogeropoulos AP, Kelkar A, Weinberger JF, Morris AA, Georgiopoulou VV, Markham DW, Butler J, Vega JD, Smith AL. Validation of clinical scores for right ventricular failure prediction after implantation of continuous-flow left ventricular assist devices. J Heart Lung Transplant 2015;34:1595–1603.

CR IP T

21. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, Solomon SD, Spencer KT, Sutton MSJ, Stewart WJ, Chamber Quantification Writing Group, American Society of Echocardiography’s Guidelines and Standards Committee, European Association of Echocardiography. Recommendations for chamber quantification: a report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440–1463.

AN US

22. Haddad F, Hunt SA, Rosenthal DN, Murphy DJ. Right ventricular function in cardiovascular disease, part I: Anatomy, physiology, aging, and functional assessment of the right ventricle. Circulation 2008;117:1436–1448. 23. Kato TS, Jiang J, Schulze PC, Jorde U, Uriel N, Kitada S, Takayama H, Naka Y, Mancini D, Gillam L, Homma S, Farr M. Serial echocardiography using tissue Doppler and speckle tracking imaging to monitor right ventricular failure before and after left ventricular assist device surgery. JACC Heart Fail 2013;1:216–222.

M

24. Carabello BA, Spann JF. The uses and limitations of end-systolic indexes of left ventricular function. Circulation 1984;69:1058–1064.

ED

25. Houston BA, Shah KB, Mehra MR, Tedford RJ. A new ‘twist’ on right heart failure with left ventricular assist systems. J Heart Lung Transplant 2017;36:701–707.

AC

CE

PT

26. Amsallem M, Lu H, Tang X, Do Couto Francisco N, Kobayashi Y, Moneghetti K, Shiran H, Kim P, Sibhatu S, Rogers I, Schnittger I, Liang D, Haddad F. Optimizing Right Ventricular Focused Four-Chamber Views using Three-Dimensional Imaging, a comparative Magnetic Resonance based study. Int J Cardiovasc Imaging 2018.doi:10.1007/s10554-018-1356-7.

19

ACCEPTED MANUSCRIPT

TABLES Table 1. Baseline characteristics of the study population. n=158 Demographics 56.0 ±12.6

Male sex (%)

124 (78.5)

CR IP T

Age (years) Body mass index, kg/m2

27.5 ±6.2

Ischemic etiology (%)

48 (30.4)

Destination Therapy (%)

58 (36.7)

Chronic kidney disease (%)

76 (48.1) 40 (25.3)

History of atrial fibrillation or flutter (%) Clinical conditions at LVAD implantation Heart rate (bpm)

69 (43.7)

AN US

Prior cardiac surgery (%)

85.8 ±18.1

Systolic blood pressure (mmHg)

101.4 ±12.6 30 (19.0)

Inotropes (%)†

116 (73.4)

M

Preoperative mechanical circulatory support (%)*

INTERMACS class (%) 1 2

PT

≥3

ED

Vasopressors (%)†

24 (15.2) 23 (14.6) 54 (34.2) 81 (51.2)

Preoperative laboratory data Hemoglobin (g/dL)

11.6 ±3.0

CE

Hemoglobin ≤10.0 g/dL

43 (27.2) 2

70.8 ±36.4

Blood urea nitrogen (mg/dL)

29.2 ±15.7

Aspartate aminotransferase (U/L)

50.2 ±64.7

AC

MDRD estimated GFR (ml/min/1.73m )

Total bilirubin (mg/dL)

1.3 ±1.1

Preoperative invasive hemodynamics Mean right atrial pressure RAP (mmHg)

12.4 ±7.2

Mean pulmonary arterial pressure (mmHg)

35.7 ±10.9

Pulmonary arterial wedge pressure PAWP (mmHg)

24.5 ±8.7

20

ACCEPTED MANUSCRIPT

RAP/PAWP

0.50 ±0.22 2

Cardiac index (L/min/m )

1.9 ±0.5

Pulmonary vascular resistance (WU)

3.4 ±2.2

Pulmonary vascular resistance >3WU

75 (47.5)

Preoperative echocardiography 19.9 ±5.6

Left ventricular internal end-diastolic diameter (mm) 2

2

Right ventricular end-diastolic area index (cm /m ) Right ventricular end-systolic area index (cm2/m2) 2

2

Right atrial area index (cm /m )

70.5 ±11.2

CR IP T

Left ventricular ejection fraction (%)

17.3 ±3.8 12.4 ±3.5 11.5 ±3.6

Tricuspid annular plane systolic excursion (mm)

23.1 ±6.7

AN US

Right ventricular fractional area change (%)

14.2 ±4.1

Right ventricular longitudinal strain (absolute, %)

14.1 ±3.5

Severe tricuspid regurgitation (%)

35 (22.2)

Right atrial pressure (mmHg)

15.3 ±5.6

Data is presented as mean ±standard deviation or number (percentage). Chronic kidney disease was defined by glomerular filtration rate (GFR) <60mL/min using the MDRD formula. *Included intra-aortic

M

ballon pumps and miniaturized LVADs (Tandem Heart, Impella). †Inotropes included milrinone,

AC

CE

PT

ED

dobutamine, and dopamine. Vasopressors included norepinephrine, phenylephrine, and vasopressin.

21

ACCEPTED MANUSCRIPT

Table 2. Univariable analysis logistic regression analysis of correlates of right heart failure after LVAD implantation. OR*

95% CI

p-value

Age (years)

0.68

0.46 – 0.94

0.02

Male sex

3.42

1.32 – 8.87

0.01

Body surface area (m2)

1.15

0.83 – 1.60

0.39

Ischemic etiology

0.63

0.31 – 1.29

0.20

Destination Therapy

0.49

0.24 – 0.98

0.04

Chronic kidney disease

0.85

0.45 – 1.62

0.63

Prior cardiac surgery

0.32

0.14 – 0.75

<0.01

History of atrial fibrillation or flutter

1.10

0.30 – 4.05

0.89

1.39

0.94 – 1.92

0.12

0.98

0.68 – 1.28

0.88

2.35

1.28 – 3.73

<0.01

4.59

2.03 – 10.38

<0.001

2.88

1.26 – 6.55

0.01

2.41

1.01 – 5.73

0.05

Heart rate (bpm) Systolic blood pressure (mmHg) Mechanical ventilation duration (days)

M

Preoperative mechanical circulatory support†

AN US

Clinical condition at LVAD implantation

Inotropes†

INTERMACS class

ED

Vasopressors†

CR IP T

Demographics

<0.001 6.42

2.38 – 17.33

<0.001

2.77

0.96 – 4.37

<0.001

1

-

-

Cardiopulmonary bypass duration (min)

1.59

1.00 – 1.92

0.06

Open-heart combined surgery

3.62

1.04 – 12.58

0.04

Hemoglobin (g/dL)

0.94

0.68 – 1.32

0.94

Hemoglobin ≤10.0 g/dL

1.10

0.53 – 2.25

0.81

Platelet count (G/L)

1.00

0.84 – 1.52

0.47

White blood cell count (G/L)

1.37

0.97 – 1.86

0.07

Creatinine (mg/dL)

1.18

0.84 – 1.67

0.34

MDRD estimated GFR (ml/min/1.73m2)

1.44

0.69 – 2.06

0.76

1

PT

2

≥3 (reference group)

CE

Surgical procedure

AC

Preoperative laboratory data

22

ACCEPTED MANUSCRIPT

Blood urea nitrogen (mg/dL)

1.17

0.85 – 1.59

0.34

Albumin (g/dL)

0.65

0.46 – 0.92

0.01

Aspartate aminotransferase (U/L)

1.00

0.72 – 1.38

0.89

Total bilirubin (mg/dL)

1.81

1.23 – 2.67

<0.001

Right atrial pressure RAP (mmHg)

1.42

1.01 – 1.98

0.03

Mean pulmonary arterial pressure (mmHg)

0.97

0.72 – 1.37

0.86

Pulmonary arterial wedge pressure PAWP (mmHg)

0.96

0.71 – 1.28

0.99

RAP/PAWP

1.68

1.20 – 2.36

<0.01

0.86 – 1.64

0.29

0.64 – 1.24

0.48

0.29 – 1.06

0.07

2

1.19

Pulmonary vascular resistance (WU)

0.89

Pulmonary vascular resistance > 3 WU

0.55

AN US

Cardiac index (L/min/m )

CR IP T

Preoperative invasive hemodynamics

Preoperative echocardiography

0.71

0.52 – 1.02

0.06

1.03

0.80 – 1.39

0.83

1.03

1.02 – 1.05

<0.001

1.73

1.26 – 2.47

<0.001

2.65

1.35 – 5.20

<0.01

1.84

1.31 – 2.66

0.001

Tricuspid annular plane systolic excursion (mm)

0.67

0.48 – 0.96

0.03

RV fractional area change (%)

0.61

0.42 – 0.87

<0.001

RV longitudinal strain (absolute value, %)

0.56

0.40 – 0.80

<0.001

RV longitudinal strain < 12%

3.18

1.54 – 6.55

<0.001

Severe tricuspid regurgitation

2.04

0.96 – 4.37

0.07

Right atrial pressure

1.46

1.02 – 2.09

0.04

Right atrial pressure ≥15mmHg

1.89

0.87 – 4.08

0.11

Left ventricular ejection fraction (%)

Left ventricular internal end-diastolic diameter (mm) 2

2

RA area index (cm /m ) RV end-diastolic area index (cm2/m2) >16cm2/m2 (men)

AC

CE

PT

ED

RV end-systolic area index (cm2/m2)

M

RV end-diastolic area index > 15 (women),

Data is presented as mean ±standard deviation or number (percentage). *Odds ratio is adjusted by the standard deviation of the variable considered. †Included intra-aortic ballon pumps and miniaturized LVADs (Tandem Heart, Impella). †Inotropes included milrinone, dobutamine, and dopamine. Vasopressors included norepinephrine, phenylephrine, and vasopressin.

23

ACCEPTED MANUSCRIPT

AN US

CR IP T

FIGURE LEGENDS

Figure 1. (A) Two-year Kaplan Meier cumulative transplant-free survival curve of patients who underwent LVAD implantation and developed or not right heart failure (RHF) with or without right ventricular assist device (RVAD) implantation. (B) Comparative early outcomes in patients with or

M

without RHF post-LVAD implantation. Data is presented as mean ±standard deviation or number

AC

CE

PT

ED

(percentage).

24

AN US

CR IP T

ACCEPTED MANUSCRIPT

Figure 2. Prevalence of preoperative right heart abnormalities. (A) Correlation heatmap of

M

echocardiographic right heart metrics. Significant correlations (with p values <0.005 corrected for multiple analysis) are presented using Pearson’s correlation coefficients; non-significant correlations are

ED

left blank. Invasive pulmonary vascular resistance index available within a year of echocardiography is displayed, although limited by the time delay. (B-D) Distribution of right ventricular end-diastolic area

PT

index (RVEDAI), free-wall longitudinal strain (RVLS), tricuspid regurgitation (TR) severity classified as

CE

mild, moderate (mod.) or severe, non-invasive right atrial pressure (RAP) estimate and total bilirubin levels. (E) Venn diagram demonstrating the prevalence of RV end-diastolic enlargement (based on RV

AC

end-diastolic area index RVEDAI predefined sex-specific thresholds), severe RV dysfunction (based on RV longitudinal strain RVLS) and elevated right atrial pressure (RAP).

25

M

AN US

CR IP T

ACCEPTED MANUSCRIPT

Figure 3. (A-B) Incidence of right heart failure (RHF) according to the presence and number of abnormal

ED

findings or INTERMACS profile (compared using one-way ANOVA). RVEDAI: right ventricular enddiastolic area (>15cm2/m2 for women, >16cm2/m2 for men); RVLS: right ventricular longitudinal strain

PT

(<12% in absolute value). (C) Receiving operating curve for right heart failure for the CRITT, EUROMACS, Michigan, Kormos et al.’s, Utah and Penn scores. (D) Incremental value of RVLS and

CE

RVEDAI (as continuous values) to risk scores and INTERMACS profiles for prediction of RHF post-

AC

LVAD implantation.

26

CR IP T

ACCEPTED MANUSCRIPT

Figure 4. (A) Correspondence between semi-quantitative and quantitative assessment of right ventricular

AN US

(RV) function and end-diastolic size. The semi-quantitative assessment in normal/mild, moderate and severe alterations was clinically reported by the clinical echocardiography laboratory. (B) Venn diagram demonstrating the prevalence of severe RV dysfunction using tricuspid annular plane systolic excursion

AC

CE

PT

ED

M

(TAPSE), RV longitudinal strain (RVLS) or fractional area change (RVFAC).

27