Intraoperative Hemodynamic and Echocardiographic Measurements Associated With Severe Right Ventricular Failure After Left Ventricular Assist Device Implantation

Abstract

BACKGROUND:

Severe right ventricular failure (RVF) after left ventricular assist device (LVAD) implantation increases morbidity and mortality. We investigated the association between intraoperative right heart hemodynamic data, echocardiographic parameters, and severe versus nonsevere RVF.

METHODS:

A review of LVAD patients between March 2013 and March 2016 was performed. Severe RVF was defined by the need for a right ventricular mechanical support device, inotropic, and/or inhaled pulmonary vasodilator requirements for >14 days. From a chart review, the right ventricular failure risk score was calculated and right heart hemodynamic data were collected. Pulmonary artery pulsatility index (PAPi) [(pulmonary artery systolic pressure – pulmonary artery diastolic pressure)/central venous pressure (CVP)] was calculated for 2 periods: (1) 30 minutes before cardiopulmonary bypass (CPB) and (2) after chest closure. Echocardiographic data were recorded pre-CPB and post-CPB by a blinded reviewer. Univariate logistic regression models were used to examine the performance of hemodynamic and echocardiographic metrics.

RESULTS:

A total of 110 LVAD patients were identified. Twenty-five did not meet criteria for RVF. Of the remaining 85 patients, 28 (33%) met criteria for severe RVF. Hemodynamic factors associated with severe RVF included: higher CVP values after chest closure (18 ± 9 vs 13 ± 5 mm Hg; P = .0008) in addition to lower PAPi pre-CPB (1.2 ± 0.6 vs 1.7 ± 1.0; P = .04) and after chest closure (0.9 ± 0.5 vs 1.5 ± 0.8; P = .0008). Post-CPB echocardiographic findings associated with severe RVF included: larger right atrial diameter major axis (5.4 ± 0.9 vs 4.9 ± 1.0 cm; P = .03), larger right ventricle end-systolic area (22.6 ± 8.4 vs 18.5 ± 7.9 cm²; P = .03), lower fractional area of change (20.2 ± 10.8 vs 25.9 ± 12.6; P = .04), and lower tricuspid annular plane systolic excursion (0.9 ± 0.2 vs 1.1 ± 0.3 cm; P = .008). Right ventricular failure risk score was not a significant predictor of severe RVF. Post-chest closure CVP and post-chest closure PAPi discriminated severe from nonsevere RVF better than other variables measured, each with an area under the curve of 0.75 (95% CI, 0.64–0.86).

CONCLUSIONS:

Post-chest closure values of CVP and PAPi were significantly associated with severe RVF. Echocardiographic assessment of RV function post-CPB was weakly associated with severe RVF.

Left ventricular assist device (LVAD) implantation is commonly complicated by severe right ventricular failure (RVF), as defined by the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) in 20%–44% of cases.¹ Among patients with severe RVF post-LVAD, delay in placing a right-sided mechanical assist device has been associated with higher mortality when compared to a planned biventricular mechanical support.^2–4 Many centers implanting LVAD devices use published preoperative risk stratifying tools to select optimal recipients for the procedure. However, severe RVF remains a commonly encountered problem and no specific risk calculator has been widely adopted. The early intraoperative identification of patients at risk of severe RVF could be important when considering treatment options including temporary right-sided mechanical support to prevent end-organ damage.

Combining hemodynamic data together with echocardiographic metrics may also have an important role for early detection of severe RVF as defined by INTERMACS in LVAD patients.⁵ The goal of this preliminary study was to examine whether: (1) intraoperative central venous pressure (CVP) and pulmonary artery pulsatility index (PAPi) values and (2) quantitative intraoperative transesophageal echocardiography (TEE) measurements differ in subjects with severe compared to those with nonsevere RVF after LVAD implantation.

METHODS

Study Design

This was a retrospective cohort study of patients who received an LVAD between March 2013 and March 2016 at Tufts Medical Center in Boston, MA. All data were retrieved from chart reviews. The study was approved by the hospital’s institutional review board and the requirement for written informed consent was waived by the institutional review board.

Patient Cohort

Patients were eligible to be included in the study if they received either a HeartMate II (Thoratec, Pleasanton, CA) or a HeartWare HVAD (HeartWare, Oakville CA) LVAD. Patients were excluded if: (1) hemodynamic data needed to calculate PAPi or comprehensive echocardiographic images were missing, (2) additional procedure that may have affected the cardiac geometry and/or hemodynamic measurements, such as tricuspid, mitral, and/or aortic valve repair/replacement, were performed at the time of LVAD implantation; (3) left the operating room with an open chest, which makes comparing open and closed chest patients untenable; or (4) had RV mechanical support introduced concurrently with LVAD implantation. Patients on preoperative inotropic support were included in the analysis. While subjects proceeding with LVAD implantation were clinically risk stratified by a multidisciplinary team, no published risk stratification tool was used. All patients received general endotracheal anesthesia, a pulmonary artery catheter, and a comprehensive transesophageal (TEE) study. Intraoperative anesthetic, ventilator, hemodynamic, and fluid management was determined by the provider and was not protocol driven. RV mechanical support options included: RV Impella device (Abiomed, Danvers, MA), CentriMag device (Thoratec, Pleasanton, CA), or TandemHeart percutaneous VAD (CardiacAssist, Inc, Pittsburgh, PA). All RV mechanical support was implemented after the LVAD procedure; device and timing were not protocol driven and determined by the collaborative clinical judgment of the heart failure team that included cardiology, surgery, and critical care services.

Data Collection

Preoperative clinical and laboratory data, intraoperative hemodynamic measurements, echocardiographic data, and postoperative clinical outcome data were retrieved from chart review. We used the standard clinical definition of severe RVF provided by INTERMACS, defined by meeting one or more of the following criteria: the need for a RV mechanical support device (right ventricle assist device [RVAD]) or need for inotropic/inhaled pulmonary vasodilator for >14 days postoperatively.⁵ Two groups of patients were created from this definition: severe RVF and nonsevere RVF.

Hemodynamic Parameters

The right ventricular failure risk score (RVFRS) was calculated by the summation and weighting of 4 preoperative variables: vasopressors (4 points), aspartate aminotransferase ≥80 IU/L (2 points), bilirubin ≥2.0 mg/dL (2.5 points), creatinine ≥2.3 mg/dL (3 points).⁶ CVP, PA systolic pressure (PASP), and PA diastolic pressures (PADP) were recorded continuously with Anesthesia Touch (Plexus Technology Group, LLC, Jackson, MI) and PAPi [(PASP − PADP)/CVP] was calculated by averaging data over 15-minute time intervals at 2 periods: (1) 30 minutes before initiation of cardiopulmonary bypass (CPB), and (2) after chest closure immediately before leaving the operating room.

Echocardiographic Parameters

Two-dimensional color flow and Doppler intraoperative TEE were performed using a Phillips iE33 Ultrasound System (Phillips Medical Systems, Andover, MA). The images were obtained by, or under the direct supervision of, a National Board of Echocardiography-certified advanced perioperative echocardiographer according to the Tufts Medical Center intraoperative image acquisition protocol for LVADs. Two comprehensive intraoperative examinations were completed, 1 before LVAD insertion (pre-CPB) and 1 after (post-CPB) (Supplemental Digital Content, Document 1, http://links.lww.com/AA/C434). TEE images were stored and analyzed using XCelera PACS (Philips, Eindhoven, the Netherlands). Echocardiographic variables of particular interest were chosen based on our institutional experience and published preoperative associations with severe RVF (PAPi, CVP, fractional area of change [FAC], right atrial diameter major [RAdMajor], and tricuspid annular plane systolic excursion [TAPSE]).^6–10 The collected echocardiographic metrics included: RV end-diastolic area, RV end-systolic area, RVFAC, right atrial (RA) end-systolic area (ESRAA), right atrial diameter major (RAdMajor), and minor, RV diastolic basal diameter, RV/LV end-diastolic basal diameter ratios, and TAPSE (Figure 1). TAPSE was measured by manually tracking the excursion of the tricuspid annulus from end-diastole to end-systole. Tricuspid and mitral valve regurgitation were graded according to the American Society of Echocardiographic guidelines by 1 blinded reviewer.¹¹ If specific metrics were not obtained or echocardiographic windows were suboptimal to perform the evaluation, the specific metric was excluded in the final analysis.

Figure 1.

A–C, Illustrations and actual images of the 4-chamber view with the probe turned toward the right ventricle (RV). A and D, The end of diastole. B and C, The end of systole. The red area is indicating the measurements for end-systolic right atrial area (ESRAA), as well as RV end-systolic area (RVESA) and RV end-diastolic area (RVEDA), also used in the calculation of RV fractional area of change. C, The measurements (at the end of systole) for right atrial diameter major and right atrial diameter minor. D, An illustration and actual image of the 4-chamber view (at the end of diastole). The red lines are indicating measurements of RV and left ventricle (LV) diameters and the corresponding RV/LV diameter ratios (at the end of diastole).

Statistical Analysis

Differences in demographic and clinical characteristics, and hemodynamic and echocardiographic variables, between the severe RVF and nonsevere RVF groups were tested using χ², t test, and Mann-Whitney U tests, as appropriate. Univariate logistic regression models were used to test each variable’s association with postoperative severe RVF. Odds ratios and area under the receiver operating characteristic curve (AUC) values were calculated for each of the variables of interest, along with 95% CI. Statistical analysis was performed with SAS v9.4 (Cary, NC) with 2-sided tests and α = .05.

Sample Size Calculation.

It was estimated that approximately 100 patients from our center would have available data and be eligible for inclusion in this analysis and that approximately one-third (ie, 33) would meet criteria for severe RVF. Assuming estimated AUCs ranging from 0.60 to 0.75, we calculated this sample size would provide sufficient precision to construct 2-sided 95% CIs around these estimates with widths of 0.24 to 0.21, respectively (PASS 14 Power Analysis and Sample Size Software. NCSS, Kaysville, UT).

RESULTS

Patient Characteristics

Between March 2013 and March 2016, a total of 110 patients at Tufts Medical Center received an LVAD. Twenty-five were missing key data and were excluded from the study, leaving a cohort of 85 patients. Of those, 57 (67%) of patients were classified as nonsevere RVF, while 28 (33%) patients met criteria for severe RVF (any inotrope and/or pulmonary vasodilator >14 days, or RV mechanical support). Twenty-eight patients received a HeartMate LVAD and 57 received a HeartWare LVAD (P > .9 between RVF groups). In the severe RVF group, 8 patients met the definition based on the need for inhaled pulmonary vasodilators and 14 for intravenous inotropic agents (2 required dobutamine and 12 required milrinone). Six patients from the severe RVF group received a RVAD; 5 patients received RV Impella devices (Abiomed), and 1 patient received a TandemHeart percutaneous VAD (CardiacAssist, Inc) as a separate, postoperative procedure, on the same date as the LVAD. The RVFRS was not statistically different between the severe and nonsevere RVF groups (P = .09; Table 1). Patients who received a greater volume of resuscitation fluids appeared to develop severe RVF more frequently (P = .03; Table 2).

Table 1.

Preoperative Baseline Characteristics

Demographics and Preoperative Characteristics	Nonsevere RVF (n = 57)	Severe RVF (n = 28)	P Value
Age	55 ± 12	56 ± 11	.58
Male sex: N (%)	44 (78.57)	24 (85.71)	.22
Body mass index	28.16 ± 6.03	28.08 ± 6.35	.96
Body surface area	1.95 ± 0.23	2.01 ± 0.34	.32
Ischemic CM (versus nonischemic)	19 (33.93)	7 (25.00)	.45
Re-do sternotomy	11 (19.64)	4 (14.29)	.76
Preoperative mechanical ventilation: N (%)	0 (0)	1 (3.70)	.33
LVEF <20%: N (%)	52 (92.86)	27 (96.43)	.66
CRF (Cr > 2.3 mg/dL): N (%)	3 (5.00)	2 (8.00)	.74
	Nonsevere RVF	Severe RVF	P Value
Laboratory variables
Creatinine (mg/dL)	1.22 ± 0.52	1.43 ± 0.72	.13
BUN (mg/dL)	23.57 ± 14.29	33.07 ± 14.89	.01
AST (IU/L)	28.21 ± 19.83	32.54 ± 14.89	.31
Total bilirubin (mg/dL)	0.97 ± 0.75	1.43 ± 1.02	.02
Support device variables
HeartWare device: N (%)	36 (64.29)	18 (64.29)	>.99
HeartMate device: N (%)	20 (35.71)	10 (35.71)	>.99
Bridge to transplant: N (%)	37 (66.07)	20 (71.43)	>.99
Scoring variables
INTERMACS score	2.35 ± 0.85	2.81 ± 0.72	.01
RVFRS: median (interquartile range)	0 (0, 0)	0 (0, 2.5)	.09a

All data are presented as mean ± standard deviation or N (%) except where noted. Comparisons are made with 2-tailed t test and χ² tests.

Abbreviations: AST, aspartate aminotransferase; BUN, blood urea nitrogen; CM, cardiomyopathy; CRF, chronic renal failure; HbA1c, Hemoglobin A1c; INTERMACS, interagency registry for mechanically assisted circulatory support; LVEF, left ventricular ejection fraction; RVF, right ventricular failure; RVFRS, right ventricular failure risk score.

^aMann-Whitney U test.

Table 2.

Intraoperative Variables After Chest Closure

Medications	Nonsevere RVF (n = 57)	Severe RVF (n = 28)	P Value
Epinephrine dose (μg·kg−1·minute−1)	0.01 ± 0.02	0.02 ± 0.02	.08
Norepinephrine dose (μg/min) median (interquartile range)	5 (2, 9)	2 (0. 4)	.005a
Phenylephrine dose (μg/min) median (interquartile range)	0 (0, 40)	0 (0, 45)	.70a
Vasopressin dose (units/min) median (interquartile range)	0 (0, 0.04)	0.02 (0.005, 0.06)	.04a
Milrinone dose (μg·kg−1·minute−1) median (interquartile range)	0.5 (0.375, 0.5)	0.5 (0.375, 0.5)	.48a
Dobutamine dose (μg·kg−1·minute−1) median (interquartile range)	0 (0, 0)	0 (0, 5)	.03a
Dopamine dose (μg·kg−1·minute−1) median (interquartile range)	0 (0, 0)	0 (0, 0)	.20a
	Nonsevere RVF	Severe RVF	P Value
Fluids and blood products (mL)
Packed red blood cells	318 ± 458	700 ± 931	.02
Fresh frozen plasma	472 ± 422	572 ± 552	.41
Platelets	182 ± 164	266 ± 222	.055
Cryoprecipitate	71 ± 71	93 ± 83	.22
Cell Saver (mL)	608 ± 410	920 ± 1044	.06
Crystalloid (mL)	1888 ± 1172	2223 ± 1010	.21
Total fluid volumeb (mL)	3729 ± 2011	4993 ± 1810	.03

All data are presented as mean ± standard deviation or N (%) except where noted. Comparisons are made with 2-tailed t test and χ² tests.

Abbreviation: RVF, right ventricular failure.

^aMann-Whitney U test.

^bTotal of all crystalloid, packed red blood cells, fresh frozen plasma, platelets, and cryoprecipitate.

Hemodynamic and Echocardiographic Discriminators of Severe RVF

Hemodynamic factors with significant differences in the severe versus nonsevere RVF groups included: lower pre-CPB PAPi (P = .045); higher post-chest closure PADP (P = .01); higher post-chest closure CVP (P < .01); and lower post-chest closure PAPi (P < .01; Tables 3 and 4). Echocardiographic findings associated with severe RVF included: larger post-CPB RV end-systolic area (P = .034); lower post-CPB RVFAC (P = .041); lower post-CPB TAPSE (P < .01); higher post-CPB ESRAA (P < .05); and higher post-CPB RAdMajor (P = .026; Table 4).

Table 3.

Pre-LVAD Intraoperative Echocardiographic and Hemodynamic Predictors of Severe RVF

Variables Pre-CPB	Nonsevere RVF (n= 57)	Severe RVF (n = 28)	P Value
Echocardiographic variables
RVEDA (cm2)	28.8 ± 9.4	29.2 ± 8.2	.87
RVESA (cm2)	23.2 ± 8.9	23.8 ± 8.1	.77
RVFAC	20.8 ± 10.8	19.4 ± 9.4	.54
ESRAA (cm2)	23.6 ± 7.9	25.3 ± 7.2	.35
TAPSE (cm)	1.5 ± 0.5	1.3 ± 0.4	.06
TVR grade
Median (interquartile range)	2 (1, 2)	2 (1, 2)	.92a
MVR grade
Median (interquartile range)	2 (1, 3)	2 (2, 4)	.31a
RAdMinor (cm)	4.9 ± 0.9	4.8 ± .1.0	.79
RAdMajor (cm)	5.4 ± 0.9	5.7 ± .0.8	.11
RVEDD (cm)	4.1 ± 0.9	4.4 ± .0.8	.21
LVEDD (cm)	6.6 ± 1.0	6.5 ± .1.0	.88
RV/LV diameter ratio	0.6 ± 0.2	0.7 ± 0.1	.22
Hemodynamic variables
PASP (mm Hg)	43.9 ± 14.5	45.8 ± 14.4	.59
PADP (mm Hg)	24.9 ± 9.9	26.2 ± 7.9	.57
PA pulse pressure (mm Hg)	19.0 ± 8.3	19.6 ± 8.4	.76
CVP (mm Hg)	14.3 ± 6.9	18.6 ± 11.8	.053
PAPi (mm Hg)	1.7 ± 1.0	1.2 ± 0.6	.045

All data are presented as mean ± standard deviation or N (%) except where noted. Comparisons are made with 2-tailed t test and χ² tests.

Abbreviations: CPB, cardiopulmonary bypass; CVP, central venous pressure; ESRAA, end-systolic right atrial area; LVAD, left ventricular assist device; LVEDD, left ventricle end-diastolic diameter; MR, mitral regurgitation; PA, pulmonary artery; PADP, pulmonary artery diastolic pressure; PAPi, pulmonary artery pulsatility index; PASP, pulmonary artery systolic pressure; RAdMajor, right atrial diameter major; RAdMinor, right atrial diameter minor; RVEDA, right ventricular end-diastolic area; RVEDD, right ventricle end-diastolic diameter; RVESA, right ventricular end-systolic area; RVF, right ventricular failure; RVFAC, right ventricle fractional area of change; RV/LV, right ventricle/left ventricle; TAPSE, tricuspid annular plane systolic excursion; TR, tricuspid regurgitation.

^aMann-Whitney U test.

Table 4.

Post-LVAD Post-Chest Closure Intraoperative Echocardiographic and Hemodynamic Predictors of Severe RVF

Variables Post-CPB	Nonsevere RVF (n = 57)	Severe RVF (n = 28)	P Value
Echocardiographic variables
RVEDA (cm2)	24.5 ±48.1	28.00 ± 8.72	.07
RVESA (cm2)	18.5 ±87.9	22.56 ± 8.40	.034
RVFAC	26.0 6.12.6	20.19 ± 10.84	.041
ESRAA (cm2)	19.7 ±97.2	23.05 ± 6.75	0.045
TAPSE (cm)	1.1 ±.0.3	0.94 ± 0.23	.008
TVR grade: median (interquartile range)	1(1, 2)	2 (1, 2)	.16a
MVR grade: median (interquartile range)	1 (1, 2)	1 (1, 2)	.25a
RAdMinor (cm)	4.6 ± 1.0	5.1 ± 1.0	.09
RAdMajor (cm)	4.9 ± 1.0	5.4 ± 0.9	.026
RVEDD (cm)	4.3 ± 0.9	4.6 ± 0.9	.10
LVEDD (cm)	5.4 ± 1.2	5.4 ± 1.2	.97
RV/LV diameter ratio	0.8 ± 0.2	0.9 ± 0.2	.23
Variables After Chest Closure	Nonsevere RVF	Severe RVF	P Value
Hemodynamic variables
PASP (mm Hg)	37.5 ± 9.3	38.8 ± 9.2	.54
PADP (mm Hg)	20.0 ± 6.0	23.8 ± 5.1	.007
PA pulse pressure (mm Hg)	17.5 ± 5.7	15.0 ± 6.0	.08
CVP (mm Hg)	12.9 ± 4.8	18.1 ± 8.6	.0008
PAPi (mm Hg)	1.5 ± 0.8	0.9 ± 0.5	.0008
Univariate Logistic Regression Results for Echocardiographie and Hemodynamic Variables of Interest
Measurement (n = 81)	Odds Ratio (95% CI)	P Value	AUC (95% CI)
Post-CPB TAPSE	0.08 (0.01, 0.51)	.008	0.68 (0.56–0.80)
Post-CPB RAdMajor	1.70 (1.05, 2.76)	.03	0.66 (0.54–0.78)
Post-CPB RVFAC	0.99 (0.94, 1.03)	.6	0.53 (0.40–0.67)
RVFRS	1.21 (0.91, 1.61)	.2	0.58 (0.48–0.68)
Post-chest closure PAPi	0.20 (0.07, 0.55)	.002	0.75 (0.64–0.86)
Post-chest closure CVP	1.21 (1.08, 1.36)	.001	0.75 (0.64–0.86)

All data are presented as mean ± standard deviation or N (%) except where noted. Comparisons are made with 2-tailed t test and χ² tests.

Abbreviations: AUC, area under the curve; CI, confidence interval; CPB, cardiopulmonary bypass; CVP, central venous pressure; ESRAA, end-systolic right atrial area; LVAD, left ventricular assist device; LVEDD, left ventricle end-diastolic diameter; MR, mitral regurgitation; PA, pulmonary artery; PADP pulmonary artery diastolic pressure; PAPi, pulmonary artery pulsatility index; PASP, pulmonary artery systolic pressure; RAdMajor, right atrial diameter major; RAdMinor, right atrial diameter minor; RVEDA, right ventricular end-diastolic area; RVEDD, right ventricle end-diastolic diameter; RVFRS, right ventricular failure risk score; RVESA, right ventricular end-systolic area; RVF, right ventricular failure; RVFAC, right ventricle fractional area of change; RV/LV, right ventricle/left ventricle; TAPSE, tricuspid annular plane systolic excursion; TR, tricuspid regurgitation.

^aMann-Whitney U test.

The results of the univariate logistic regression analyses for the 81 subjects with complete data on all variables of interest are shown in Table 4. Both post-closure PAPi and CVP discriminated severe versus nonsevere RVF, each with an AUC of 0.75 (95% CI, 0.64–0.86). Echocardiographic variables such as TAPSE, RVFAC, and RAdMajor had AUCs ranging from 0.53 (95% CI, 0.40–0.67) to 0.68 (95% CI, 0.56–0.80), while RVFRS had an AUC of 0.58 (95% CI, 0.48–0.68) (Table 4, Figure 2).

Figure 2.

Receiver operator curves (ROCs) representing univariate analysis of intraoperative hemodyanmic and echocardiographic measures as discriminators of severe right ventricular failure (RVF) from not-severe RVF. Odds ratios and area under the receiver operating characteristic curve (AUC) values were calculated for each of the variables of interest, along with 95% CIs. Statistical analysis was performed with SAS v9.4 (Cary, NC) with 2-sided tests and α = .05. A, ROCs for post-chest closure central venous pressure (CVP), post-chest closure pulmonary artery pulsatility index (PAPi), and RVF risk score (RVFRS). B, ROCs for post-cardiopulmonary bypass (CPB) tricuspid annular plane systolic excursion (TAPSE), post-CPB right atrial dimension major (RAdMajor), and post-CPB right ventricle fractional area of change (RVFAC).

DISCUSSION

The intraoperative measurement of CVP and PAPi after chest closure discriminated subjects with severe RVF from those with nonsevere RVF after LVAD implantation. RVFRS was not discriminatory. Echocardiographic variables of RV size and function and RA size showed a weak association with severe RVF.

Pathophysiology of RVF Post-LVAD

Implantation of an LVAD unloads the LV and increases RV preload.¹² There is simultaneous reduction of left atrial pressures and pulmonary vascular resistance although in some patients reduction in pulmonary vascular resistance may be less pronounced due to vascular remodeling.^13–16 Increases in RV preload due to increased venous return along with the perioperative administration of blood products and intravenous fluid can cause RV distention, exacerbate wall stress, and worsen tricuspid regurgitation.¹⁷ Of potential relevance, we did observe an association between total intraoperative fluid administration with severe RVF (Table 2). By reducing LV intracavity pressure, LVAD can cause a leftward shift of the intraventricular septum, distort the RV’s crescent geometry, and impair its systolic function. Indeed, persistent septal deviation has been associated with worse outcomes.^18–20

Existing Preoperative Risk Models for Severe RVF in the Setting of LVADs

Numerous preoperative clinical scoring systems purported to predict RVF in patients undergoing LVAD insertion have been published, utilizing a plethora of clinical parameters.^{8–10,21–23} Reproduction of many of these clinical scoring models has produced moderate results at best.²⁴ This problem is compounded by the variable use of different definitions of RVF among existing studies. As most patients with LVADs have at least some degree of RVF, the focus should be on the severity of RVF. Preoperative risk assessment tools are also limited by perioperative events such as transfusion-associated circulatory overload that may further impair the RV.

Using the preoperative need for vasopressors, aspartate aminotransferase, bilirubin, and creatinine levels, Matthews et al⁶ developed RVFRS and reported AUC of 0.73 for RVF after LVAD placement.⁵ While RVFRS has been widely referenced in the literature as a benchmark for subsequent studies, its performance in predicting RVF has been inconsistent.^{8–10,21–23} Matthews et al⁶ also investigated the application of preoperative transthoracic echocardiography but used subjective visual assessment to characterize RV function and severity of tricuspid regurgitation. Importantly, the authors defined severe RVF according to INTERMACS, that is, unexpected need for RVAD and need for inotropic agents for >14 days. This allows for a reasonable comparison with our study. Similar to other published studies, we report that RVFRS did not predict RV failure.^{8–10,21–23}

The PAPi

PAPi is a novel hemodynamic marker associated with severe RV dysfunction after an acute inferior wall infarct and after LVAD implantation.^25–27 Utilizing pulmonary artery pulse pressure as a surrogate of RV stroke volume and CVP as an estimate of RV filling pressures, PAPi may provide an insight to RV function. PAPi has been compared against other hemodynamic of RV function indices, such as CVP, CVP/pulmonary capillary wedge pressure, and RV stroke work index, which has shown PAPi to be a superior metric of RV function.^25,26 Our study also suggests that PAPi after chest closure is different between the severe and nonsevere RVF groups in the intraoperative setting; however, the diagnostic performance of PAPi was very similar to CVP. Of note, PA pulse pressure (the numerator in the PAPi calculation) compared to CVP (the denominator) did not have an independent discriminatory value (Table 4). Therefore, based on this dataset, we cannot conclude whether PAPi has a better discriminatory power over CVP alone.

Echocardiography

There have been multiple studies investigating the predictive value of echocardiography for severe RVF after LVAD implantation.^28–35 A major limitation of this study is the inconsistent or unreported acquisition time of the echocardiogram in relation to the LVAD procedure. Exact TEE acquisition times were a limitation of our study. Nonetheless, there has been some consistency with measurement of TAPSE.^28–35 We found lower TAPSE and RVFAC in the severe RVF group after chest closure. Both metrics of right atrial dilation (RAdMajor and ESRAA) were increased in the severe RVF group. Other authors have suggested that RAdMajor is closely related to RA pressure.³⁶ While these significant statistical associations of echocardiographic metrics were present, the actual differences were small and clinically insignificant. Notably, we did not observe a difference in intraoperative RV/LV basal diameter ratios in our study probably due to the concurrent adjustments of LVAD speeds and titration of inotropic support to avoid leftward interventricular septal deviation.

STUDY LIMITATIONS

Similar to other published risk models discerning between severe and nonsevere RVF after LVAD implantation, our single institution study is limited by its small sample size and retrospective design.^{6,8–10,21–23} The statistical associations of hemodynamic and echocardiographic data with severe RVF have unproven clinical applicability, and caution should be exercised in interpreting the results of the many tests of association with the outcome as no adjustment for multiple comparisons were made in this preliminary study. Due to the small sample size of the study, we were also likely underpowered to detect minimally important clinical differences, particularly for skewed variables like RVFRS. As with any retrospective chart review, limitations include possible bias due to missing data and inconsistent clinical practice over time. These could be overcome in a prospective study. We were unable to investigate potentially important hemodynamic metrics such as pulmonary vascular resistance, CVP/pulmonary capillary wedge pressure, and RV stroke work index. Furthermore, continuously adjusted LVAD speed, to control the patient’s hemodynamic course, could influence LV decompression, interventricular septal position, and RV function. Our ability to include intraoperative LVAD settings in this study was limited by the intermittent charting in perfusionists’ records.

We were also limited in the number of echocardiographic metrics we could review. For example, other studies have suggested tricuspid annular systolic velocity, early tricuspid inflow velocities/tricuspid annular early diastolic velocity, and RV strain to be associated with severe RVF.^28,32,33 Additionally, there is some evidence that 3 dimensional RV end-diastolic, end-systolic volumes and ejection fraction may be predictors of severe RVF.³⁷ Because TEE measurements and hemodynamic measurements in our retrospective study were not consistently recorded at the same time point, we are unable to report whether the combination of hemodynamic with TEE anatomic data provides additional predictive value.

CONCLUSIONS

This study provides evidence that CVP and PAPi after chest closure may be associated with severe RVF after LVAD implantation. Quantitative echocardiographic metrics of right heart geometry and function acquired post-CPB were weakly correlated with severe RVF. Our data suggest that future prospective analysis is warranted; however, it is needed before any clinical application can be suggested.

KEY POINTS.

• Question: Can intraoperative hemodynamic and echocardiographic variables discriminate between severe and nonsevere right ventricular failure after left ventricular assist device implantation?

• Findings: Pulmonary artery pulsatility index and central venous pressure after chest closure discriminate between nonsevere and severe right ventricular failure. Echocardiographic metrics of right ventricular function and right atrial size obtained after separation from cardiopulmonary bypass appear weakly associated with severe right ventricular failure.

• Meaning: Post–chest closure values of central venous pressure and pulmonary artery pulsatility index may be indicators of progression to severe right ventricular failure.