Journal Pre-proof
Timing and trends of right atrial pressure and risk of right heart failure after left ventricular assist device implantation Gaurav Gulati MD , Nilay Sutaria MD , Amanda R Vest MBBS, MPH , David D DeNofrio MD , Masashi Kawabori MD , Gregory Couper MD , Michael Kiernan MD, MS PII: DOI: Reference:
S1071-9164(19)31486-1 https://doi.org/10.1016/j.cardfail.2020.01.013 YJCAF 4480
To appear in:
Journal of Cardiac Failure
Received date: Revised date: Accepted date:
3 September 2019 30 December 2019 17 January 2020
Please cite this article as: Gaurav Gulati MD , Nilay Sutaria MD , Amanda R Vest MBBS, MPH , David D DeNofrio MD , Masashi Kawabori MD , Gregory Couper MD , Michael Kiernan MD, MS , Timing and trends of right atrial pressure and risk of right heart failure after left ventricular assist device implantation, Journal of Cardiac Failure (2020), doi: https://doi.org/10.1016/j.cardfail.2020.01.013
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2020 Published by Elsevier Inc.
Timing and trends of right atrial pressure and risk of right heart failure after left ventricular assist device implantation Gaurav Gulati1 MD; Nilay Sutaria1 MD; Amanda R Vest1 MBBS, MPH; David D DeNofrio1 MD; Masashi Kawabori1 MD; Gregory Couper1 MD; and Michael Kiernan1 MD, MS
Affiliations: 1. Cardiovascular Center, Division of Cardiology, Tufts Medical Center, Boston MA
Corresponding author: Michael Kiernan South Bldg, 6th floor 800 Washington St., Box 5931 Boston, MA 02111 Phone: 617-636-8068 Fax: 617-636-6030
[email protected]
Word count: 3848
Abstract Background Elevated right atrial pressure (RAP) is associated with poor outcomes after LVAD implantation. However, the optimal time for RAP measurement and the importance of resolution of right heart congestion prior to LVAD implantation remain unclear.
Methods and Results We performed a retrospective cohort study of 134 consecutive LVAD recipients from our institution. Congestion was defined as RAP ≥ 14 mmHg and was assessed at hospital admission and implant. The primary outcome was death or RVAD implantation. When stratified by congestion status at admission, congested and non-congested patients had similar event-free survival rates (HR 1.2, 95% CI 0.6–2.6). However, when stratified at implant, congested patients had a higher rate death or RVAD implantation (HR 2.5, 95% CI 1.1–5.6). Patients were then divided into 4 groups based on their trajectory of congestion status: no congestion, resolved congestion, new congestion, or persistent congestion. Patients with no congestion and resolved congestion had similar outcomes, while patients with persistent congestion had a markedly increased rate of death or RVAD implantation (HR 3.1, 95% CI 1.3–7.6).
Conclusion
RAP at LVAD implantation is more strongly associated with postoperative outcomes than admission RAP. Patients not responsive to decongestive therapies, with persistently elevated RAP, represent a high risk cohort for adverse outcomes following LVAD implantation. Keywords: heart failure, LVAD, hemodynamics
Introduction Elevated right atrial pressure (RAP) has been associated with a number of adverse outcomes after LVAD implantation, including need for renal replacement therapy,1 bleeding,2 right heart failure (RHF), and death.3–6 A number of studies have developed risk prediction models to better identify patients at high risk for poor postoperative outcomes after LVAD implantation. Elevated right atrial pressure (RAP) has been identified as an important risk marker in multiple models.3–6 Risk models that do not include RAP may include other variables that reflect the sequelae of right heart congestion, including elevated serum of bilirubin or creatinine levels.7,8
These risk prediction models have primarily been derived from retrospective analyses and the timing of preoperative RAP assessment relative to LVAD implant date has not generally been reported. Hemodynamic variables, including RAP, are dynamic, changing continuously in response to therapeutic interventions that impact physiologic changes in volume status as well as changes in vascular and biventricular compliance and function. Rather than evaluating a ‘snapshot’ in time, we sought to investigate the importance of trending hemodynamic data as the optimal time to assess RAP preoperatively for the purposes of risk stratification is unknown. Furthermore, data regarding the clinical impact of preoperative hemodynamic optimization, commonly targeting a reduction in RAP, on postoperative outcomes are limited. It remains uncertain whether elevated RAP is simply a marker of poor RV function, or whether it is a modifiable risk factor and a potential target for therapeutic interventions.
The primary objectives of this study were to evaluate: 1) the risk of death or RVAD utilization associated with baseline RAP (measured at time of initial hemodynamic assessment) versus RAP recorded within 24 hours of LVAD implantation, and 2) differences in outcomes between patients who were able to be successfully decongested at the time of LVAD implantation compared with those with persistent congestion. We hypothesized that elevated RAP at the time of implantation would be more predictive of poor postoperative outcomes, and that the ability to decongest (achieve a RAP < 14 mmHg) would be associated with a more favorable postoperative course.
Methods Study population and data collection We performed a retrospective cohort study of all adult patients that received a durable LVAD at Tufts Medical Center, Boston, MA, from October 2014 to February 2018. Patients were excluded if they received extracorporeal membrane oxygenation (ECMO) support prior to LVAD implantation, or if hemodynamic data were unavailable. At our institution, patients with advanced heart failure are managed with pulmonary artery catheter-guided use of diuretics, vasodilators, and inotropes. While specific hemodynamic targets are not protocolized, attempts are made to decongest patients to clinical euvolemia with RAP less than 10 mmHg when clinically feasible, while maintaining/improving end-organ function prior to LVAD implantation. Hemodynamic data were collected from catheterization laboratory reports and progress notes during each patient’s admission during which the LVAD implantation was performed. Admission right atrial pressure (RAPadmit) was defined as the RAP on the first day of the patient’s
hospitalization with a pulmonary artery catheter and implant RAP (RAPimplant) was defined as the RAP measured closest to LVAD implantation. Echocardiographic data were collected from the most recent echocardiogram prior to LVAD implantation. Laboratory values were collected on the day of LVAD implantation. Patients were followed until death, heart transplantation, last follow up, or until July 2018. This study was approved by the Tufts Medical Center Institutional Review Board and the requirement for informed consent was waived.
Outcomes The primary study outcome was the time to death or RVAD implantation, with patients censored at the time of transplantation. Secondary outcomes included severe early RHF (the composite of death within 30 days, RVAD within 30 days, or need for postoperative inotropes beyond 14 days); and intensive care unit (ICU) and total post-implant lengths of stay (LOS).
Statistical analysis Congestion at admission or implant was defined as RAP ≥ 14 mmHg. This cutoff was chosen a priori based on prior literature reporting critical RAP values between 13 mmHg and 17 mmHg, and prior work from our institution using a threshold of 14 mmHg.4,5,9 As congestion status can change between admission and implant, the study population was divided into four congestion groups by their congestion status at admission and implant. The four congestion groups were defined as: 1) no congestion (RAPadmit and RAPimplant < 14 mmHg), 2) resolved congestion (RAPadmit ≥ 14 mmHg and RAPimplant < 14 mmHg), 3) new congestion (RAPadmit < 14 mmHg and RAPimplant ≥ 14 mmHg), and 4) persistent congestion (RAPadmit and RAPimplant ≥ 14 mmHg).
Baseline characteristics were compared between congestion groups using ANOVA for normally distributed continuous variables, Kruskal-Wallis test for non-normally distributed continuous variables, and Fisher’s exact test for categorical variables.
The primary outcome, time to death or RVAD implantation, was analyzed using the KaplanMeier method and Cox proportional hazards regression modeling. To assess the relative importance of the time of RAP measurement, the associations between the primary outcome and congestion at admission (resolved and persistent congestion groups vs no and new congestion groups) and congestion at implant (new and persistent congestion vs no and resolved congestion) were analyzed separately in a univariable fashion. Next, the associations between the primary outcome and the four congestion groups were analyzed in a univariable and multivariable fashion. Given the limited sample size and the risk of model overfitting, three nested multivariable Cox proportional hazards models were developed, each adjusted for an additional covariate. Adjustment variables included Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) Profiles < 3, age, and right ventricular failure risk score (RVFRS)7. These variables were chosen because they broadly account for multiple potential confounders. The incidence of the composite secondary outcome was compared between groups using Fisher’s exact test in an unadjusted fashion. LOS data were compared across congestion groups with the Kruskal-Wallis test, and between-group comparisons were performed using the Dunn test with the Bonferroni correction for multiple comparisons. A twosided α threshold of less than 0.05 was used to determine significance. All statistical analyses
were performed in R statistical software version 3.5.3 (R Foundation for Statistical Computing, Vienna, Austria).
Results Among the 144 patients who received LVADs at our institution between October 2014 and February 2018, 134 were included in the analysis (Figure 1). The majority of patients were INTERMACS profile 2 or 3 (Table 1). Eighty-six (64%) received the HeartMate II LVAD (Abbott, Abbott Park, USA) and 45 (33%) received the HeartWare HVAD (Medtronic, Minneapolis, USA). Ninety-two (68%) had at least moderate RV dysfunction defined by semi-quantitative grading with transthoracic echocardiography prior to LVAD implantation.
There were important differences in baseline characteristics across congestion groups (Table 1). Patients in the no congestion group were slightly older than in the other three groups. Patients with persistent congestion had significantly higher creatinine and total bilirubin at the time of LVAD implantation compared with other patients. These patients also had more days inhospital prior to their LVAD implantation compared with other patients. No patients received an RVAD prior to LVAD implantation. Two patients congested at admission received either dialysis or ultrafiltration prior to LVAD surgery, both in the resolved congestion group. Baseline characteristics stratified by congestion status at admission and implant are shown in Table 2.
Admission hemodynamic data were available a median of 1 day (interquartile range [IQR] 0–2 days) after admission and implant hemodynamic data were available a median of 1 day (IQR 1–
2 days) prior to LVAD implantation. Comparing baseline characteristics across congestion groups, patients with RAPimplant ≥ 14 mmHg (new and persistent congestion groups) were more often INTERMACS profile 1 compared to the RAPimplant < 14 mmHg groups. The frequency of moderate or severe RV dysfunction by echocardiography did not appear to differ significantly between groups, though the persistent congestion group had a slightly higher median RVFRS compared to the other groups. Patients congested at admission (resolved congestion and persistent congestion groups) had similar mean RAP values (19 ± 4 mmHg vs 21 ± 4 mmHg, Table 3). Among these, patients with resolved congestion had a significantly lower mean RAP compared to patients with persistent congestion (9 ± 3 mmHg vs 16 ± 2 mmHg, p < 0.01) at the time of LVAD implantation. Pulmonary pressures and PCWP followed similar trends between groups. Notably, cardiac index at the time of LVAD implantation was similar across all congestion groups.
Outcomes Median follow up time after LVAD implantation was 306 days (IQR 154–525 days). The primary outcome occurred in 31 (23.1%) patients in the cohort. There were 10 (7.5%) who required RVADs (2 Impella RP [Abiomed, Danvers, USA], 1 Protek Duo [LivaNova, London, UK], 6 Centrimag [Abbott], and 1 Centrimag with ECMO) and 21 (15.7%) who died during follow up. The incidence rate of death or RVAD implantation was compared between groups stratified by congestion status at admission. Patients congested at hospital admission had similar event rates (13/59 patients) compared with those who were not congested at admission (15/75, HR 1.2, 95% CI 0.6–2.6, Figure 2). When stratified by congestion status at implant, patients
congested at the time of LVAD implantation had a significantly higher rate of death or RVAD implantation (9/23) compared with patients not congested at implant (19/111, HR 2.5, 95% CI 1.1–5.6). In a sensitivity analysis, we analyzed RAPadmit and RAPimplant as continuous variables and found that RAPadmit did not appear to be associated with death or RVAD use (HR 1.01, 95%CI 0.94–1.08, p=0.78). The association between RAPimplant and this outcome did not reach statistical significance (HR 1.07, 95%CI 0.99–1.17, p=0.09).
Next, the association between congestion group assignment (no congestion, resolved congestion, new congestion, or persistent congestion) and the rate of death or RVAD implantation was evaluated. In an unadjusted analysis, patients with persistent congestion had the least favorable outcomes, with a rate of death or RVAD implantation 3-fold higher than patients who were never congested (HR 3.1, 95% CI 1.3–7.6, Table 4). This association persisted after adjustment for potential baseline confounders in sequentially adjusted models. In contrast, patients with resolved congestion had outcomes similar to patients that were never congested.
Severe early RHF occurred in 60 (44%) of the 134 patients in the cohort. Of these, 47 (78%) required prolonged postoperative inotropes alone. Patients with persistent congestion had the highest incidence of severe early RHF (80% vs 31.3% among patients never congested, p<0.01, Figure 3). When comparing the incidence of individual components of the severe early RHF outcome, patients with persistent congestion had a significantly higher incidence of requiring prolonged postoperative inotropes and RVAD implantation. The incidence of death was highest
among persistently congested patients, though this difference did not reach statistical significance. Among all short-term outcomes, patients with resolved congestion had more favorable outcomes when compared to patients with persistent congestion, though these differences did not reach statistical significance. Median intensive care unit length of stay among patients with persistent congestion was longer than among any other group (9 days [IQR 2.75–14.5 days] vs 5 days [IQR 3.0–8.0 days] among never congested, p=0.004). Total postoperative length of stay was similar among congestion groups (Table 5).
Discussion The primary finding of this analysis is that that right heart congestion status assessed at the time of LVAD implantation is more strongly associated with postoperative outcomes than congestion status assessed at the time of hospital admission. This finding is consistent with prior studies that have identified RAP as a strong prognostic factor for post-LVAD RHF and provides new insights into the optimal timing of RAP assessment for assigning preoperative risk. Furthermore, we found that patients who were unable to be decongested during their admission had the highest risk of death or RVAD implantation, highlighting the importance of reevaluating dynamic changes in RAP over time when estimating risk of post-LVAD RHF.
There are a number of reasons why elevated RAP would be expected to be associated with poor clinical outcomes in LVAD recipients. Elevated RAP is closely associated with worsening renal function, particularly in settings of low renal blood flow like those typically found in patients with advanced heart failure.10,11 Renal dysfunction can lead to postoperative volume
retention, necessitating renal replacement therapy to avoid prolonging intubation due to pulmonary congestion.12,13 Volume overload also results in changes to RV geometry, including leftward septal shift, which can reduce RV contractility,14 and tricuspid annular dilation, leading to worsening tricuspid regurgitation. Furthermore, congestion can result in hepatic dysfunction,15 which is associated with poor postoperative outcomes in patients undergoing cardiac surgery,16 likely by increasing the risk of bleeding and infectious complications.17
Despite these data, elevated RAP has only been identified in approximately half of published risk models for post-LVAD RHF.6,18 One possible explanation may be heterogeneity in the time of RAP assessment among the patients in the different cohorts used to derive these models. Our findings suggest that that RAP measured nearest to LVAD implantation following attempts at preoperative optimization are better used to assist with clinical decision making regarding patient selection as well as for preoperative planning regarding the potential needs for RV pharmacologic and mechanical support. Admission RAP in the setting of acute decompensation appears to be less helpful to guide decision making prior to attempts to improve the hemodynamic status of the patient. Only after a trial of pharmacologic or device based optimization does the RAP seem to best predict the risk of post-operative RHF.
This finding has important implications for hemodynamic data collection in cohort or registry studies. Many large registries collect preoperative hemodynamic data without specifying a time point of interest, and recorded data could be have been obtained days to weeks in advance of LVAD implantation. The variability in timing of data collection and may obscure relationships
between hemodynamic parameters and outcomes. Standardization of the time points for preoperative hemodynamic assessment and data collection should generate more informative results for future studies.
Our results also suggest that considering trends in RAP in response to decongestive therapies may more comprehensively assess the risk of postoperative right ventricular decompensation. These data demonstrate that patients who were unable to be decongested during the course of their admission prior to LVAD implantation had the highest risk of death or RVAD implantation. Since RAP reflects ventricular compliance and ventricular-vascular coupling in addition to volume status, the inability to optimize RAP prior to surgery identifies a high risk substrate, likely with a severely myopathic RV at high risk of post-operative decompensation. Among those patients who were able to be decongested prior to surgery in this study, we are unable to establish whether the decongestion itself was responsible for the more favorable outcomes we observed. Nonetheless, these data suggest that the ability to acutely reduce RAP during HF decompensation provides reassurance regarding favorable RV reserve and response to therapeutic interventions. Of note, our cohort included a small number of patients that developed congestion during the pre-LVAD period. Because of the small number of these patients, it is difficult to draw conclusions about the relationship between this trajectory of RAP and postoperative outcomes.
It is important to note that these results do not suggest that RAP is the preferred hemodynamic parameter to guide risk stratification. A single predictor variable is unlikely to be identified that
sufficiently stratifies risk. As in nearly all published RHF risk models to date,4–7 a multivariable prediction model will likely yield more accurate predictions than using a single variable; however, accurate prognostication relies on the complex and dynamic interaction of multiple variables that is difficult to capture by traditional modeling techniques.19 Furthermore, risk models are not able to incorporate perioperative events that contribute to acute RV dysfunction, including the impact of blood loss and transfusions as well as vasoplegia induced by prolonged cardiopulmonary bypass time.20,21 The current data, however, suggest that hemodynamic parameters measured more proximally to the time of LVAD implant are more closely associated with outcomes, and that consideration of dynamic hemodynamic trends in response to attempts at preoperative optimization may more accurately classify patient risk. In particular, patients who are unable to be hemodynamically optimized preoperatively appear to be an especially high risk cohort. Defining the true impact of preoperative optimization on postLVAD outcomes, however, requires prospective investigation.
The generalizability of our study is limited due to its small sample size, single center status, and observational nature of the data. Strictly defined protocols to achieve RAP < 14mmHg are not utilized in clinical practice. Furthermore, practice patterns regarding post-operative weaning of inotropic support vary across centers. Despite attempts at multivariable adjustment, the risk of residual confounding remains, particularly given a smaller sample size that limits the number of adjustment variables. In addition, RAPadmit reflects the earliest hemodynamic measurement available at our center, and for patients transferred to us from another facility, is a measurement of congestion status after a period of treatment at that facility for which data are
unavailable. We were also unable to evaluate associations between the intensity of decongestive therapy, including use of device-based therapies, and postoperative outcomes. Whether there are deleterious effects of escalating diuretic or inotrope/vasodilator doses, mechanical circulatory support, or ultrafiltration, that mitigate the benefits of decongestion is unknown. Finally, as the results of the sensitivity analysis suggest, the risk associated with elevated RAPimplant exists on a continuum, and any particular threshold used to define transitions between congested and decongested states should be validated with multicenter data with the understanding that such dichotomization inadequately captures the spectrum of risk.
In conclusion, measuring RAP immediately prior to LVAD implantation appears more informative than admission RAP for predicting postoperative morbidity and mortality. Furthermore, among patients congested at hospital admission, resolution of congestion prior to LVAD implantation is associated with a more favorable postoperative course, and that failure to respond to attempts at decongestion is associated with a high risk of poor postoperative outcome.
Acknowledgements: None
Sources of Funding: Dr. Gulati was supported by NIH grants 1TL1TR002546-01 and 1F32HL14925101.
Disclosures: M S Kiernan: consulting from Medtronic (modest), travel support from Abbott (modest)
The remaining authors report no relevant financial conflicts of interest.
References 1.
Asleh R, Schettle S, Briasoulis A, Killian JM, Stulak JM, Pereira NL, Kushwaha SS, Maltais S, Dunlay SM. Predictors and Outcomes of Renal Replacement Therapy After Left Ventricular Assist Device Implantation. Mayo Clin Proc. 2019;94:1003–1014.
2.
Joly JM, El-Dabh A, Kirklin JK, Marshell R, Smith MG, Acharya D, Rajapreyar IN, Tallaj JA, Tresler M, Pamboukian S V. High Right Atrial Pressure and Low Pulse Pressure Predict Gastrointestinal Bleeding in Patients With Left Ventricular Assist Device. J Card Fail. 2018;24:487–493.
3.
Kormos RL, Teuteberg JJ, Pagani FD, Russell SD, John R, Miller LW, Massey T, Milano CA, Moazami N, Sundareswaran KS, Farrar DJ. 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.
4.
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;96:857–864.
5.
Kiernan MS, Grandin EW, Brinkley M, Kapur NK, Pham DT, Ruthazer R, Rame JE, Atluri P, Birati EY, Oliveira GH, Pagani FD, Kirklin JK, Naftel D, Kormos RL, Teuteberg JJ, DeNofrio D. Early Right Ventricular Assist Device Use in Patients Undergoing Continuous-Flow Left Ventricular Assist Device Implantation: Incidence and Risk Factors From the Interagency Registry for Mechanically Assisted Circulatory Support. Circ Heart Fail. 2017;10:e003863.
6.
Soliman OII, Akin S, Muslem R, Boersma E, Manintveld OC, Krabatsch T, Gummert JF, de
By TMMH, 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 Sc. Circulation. 2018;137:891–906. 7.
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.
8.
Fitzpatrick JR, Frederick JR, Hsu VM, Kozin ED, O’Hara M Lou, 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;27:1286–92.
9.
Morine K, Annamalai S, Jorde L, Razavi A, Pedicini R, Esposito M, Gobeil K, HernandezMontfort J, Garan AR, Mahr C, Kapur N. Right Atrial Pressure Predicts Mortality in Cardiogenic Shock: Insights from the Cardiogenic Shock Working Group Registry. J Am Coll Cardiol. 2018;72:B196–B197.
10.
Damman K, Navis G, Smilde TDJ, Voors AA, van der Bij W, van Veldhuisen DJ, Hillege HL. Decreased cardiac output, venous congestion and the association with renal impairment in patients with cardiac dysfunction. Eur J Heart Fail. 2007;9:872–878.
11.
Mullens W, Abrahams Z, Francis GS, Sokos G, Taylor DO, Starling RC, Young JB, Tang WHW. Importance of Venous Congestion for Worsening of Renal Function in Advanced
Decompensated Heart Failure. J Am Coll Cardiol. 2009;53:589–596. 12.
Williams JB, Peterson ED, Wojdyla D, Harskamp R, Southerland KW, Ferguson TB, Smith PK, Milano CA, Lopes RD. Central venous pressure after coronary artery bypass surgery: Does it predict postoperative mortality or renal failure? J Crit Care. 2014;29:1006–1010.
13.
Youssefi P, Timbrell D, Valencia O, Gregory P, Vlachou C, Jahangiri M, Edsell M. Predictors of Failure in Fast-Track Cardiac Surgery. J Cardiothorac Vasc Anesth. 2015;29:1466–1471.
14.
Houston BA, Shah KB, Mehra MR, Tedford RJ. A new “twist” on right heart failure with left ventricular assist systems. J Hear Lung Transplant. 2017;36:701–707.
15.
Nikolaou M, Parissis J, Yilmaz MB, Seronde MF, Kivikko M, Laribi S, Paugam-Burtz C, Cai D, Pohjanjousi P, Laterre PF, Deye N, Poder P, Cohen-Solal A, Mebazaa A. Liver function abnormalities, clinical profile, and outcome in acute decompensated heart failure. Eur Heart J. 2013;34:742–749.
16.
Murata M, Kato TS, Kuwaki K, Yamamoto T, Dohi S, Amano A. Preoperative hepatic dysfunction could predict postoperative mortality and morbidity in patients undergoing cardiac surgery: Utilization of the MELD scoring system. Int J Cardiol. 2016;203:682–689.
17.
Araujo L, Dombrovskiy V, Kamran W, Lemaire A, Chiricolo A, Lee LY, Lemaire A. The effect of preoperative liver dysfunction on cardiac surgery outcomes. J Cardiothorac Surg. 2017;12:10–16.
18.
Kalogeropoulos AP, Kelkar A, Weinberger JF, Morris AA, Georgiopoulou V V., 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 Hear Lung Transplant. 2015;34:1595–1603.
19.
Rich JD. The Inherent Fallacy of Predicting RV Failure Following Left Ventricular Assist Device Implantation. J Card Fail. 2019;25:629–630.
20.
Konstam MA, Kiernan MS, Bernstein D, Bozkurt B, Jacob M, Kapur NK, Kociol RD, Lewis EF, Mehra MR, Pagani FD, Raval AN, Ward C. Evaluation and Management of Right-Sided Heart Failure: A Scientific Statement From the American Heart Association. Circulation. 2018;137:e578–e622.
21.
Ali H-JR, Kiernan MS, Choudhary G, Levine DJ, Sodha NR, Ehsan A, Yousefzai R. Right Ventricular Failure Post-Implantation of Left Ventricular Assist Device. ASAIO J. 2019;Epub ahead of print.
Age (y)
60.2 (9.2)
Resolved congestion N=44 53.9 (12.3)
Female sex
10 (14.9)
10 (22.7)
No congestion N=67
55.9 (9.4)
Persistent congestion N=15 56.4 (11.2)
0.022
2 (25.0)
4 (26.7)
0.61
New congestion N=8
INTERMACS profile
p
0.31
1
6 (9.0)
4 (9.1)
2 (25.0)
3 (20.0)
2
22 (32.8)
26 (59.1)
2 (25.0)
7 (46.7)
3
30 (44.8)
11 (25.0)
2 (25.0)
5 (33.3)
≥4
9 (13.4)
3 ( 6.8)
2 (25.0)
0 (0.0)
BMI (kg/m )
26.2 (5.3)
27.7 (5.5)
27.0 (7.2)
29.7 (6.1)
0.12
BTT strategy
41 (61.2)
28 (63.6)
3 (37.5)
11 (73.3)
0.4
Creatinine (mg/dL)
1.14 [0.89, 1.61]
1.27 [1.06, 1.67]
1.14 [0.92, 1.35]
1.60 [1.25, 2.17]
0.035
AST (U/L)
26.0 [19.0, 37.5]
24.5 [20.8, 31.5]
30.0 [13.0, 39.0]
29.0 [24.0, 37.5]
0.67
ALT (U/L) Total bilirubin (mg/dL) INR
27.0 [17.5, 35.0]
23.0 [16.5, 44.2]
19.0 [12.8, 46.8]
21.0 [12.0, 36.0]
0.80
0.8 [0.5, 1.25]
1.0 [0.7, 1.6]
1.0 [0.8, 1.8]
1.3 [1.0, 2.4]
0.010
1.1 [1.1, 1.2]
1.2 [1.1, 1.3]
1.2 [1.1, 1.3]
1.3 [1.2, 1.4]
0.008
11.3 (6.1)
15.1 (6.4)
11.4 (5.1)
19.1 (6.5)
<0.001
2
MELD score TR severity
0.66 None
1 (1.5)
0 (0.0)
0 (0.0)
0 (0.0)
Mild
33 (49.3)
18 (40.9)
3 (37.5)
5 (33.3)
Moderate
28 (41.8)
17 (38.6)
3 (37.5)
8 (53.3)
5 (7.5)
9 (20.5)
2 (25.0)
2 (13.3)
Severe RV dysfunction
0.25
None
5 (7.5)
0 (0.0)
1 (12.5)
1 (6.7)
Mild
22 (32.8)
7 (15.9)
1 (12.5)
4 ( 26.7)
Moderate
25 (37.3)
21 (47.7)
3 (37.5)
8 ( 53.3)
Severe
13 (19.4)
16 (36.4)
3 (37.5)
2 ( 13.3)
0 [0, 0]
0 [0, 2.5]
0 [0, 1.13]
2.5 [0, 3]
0.02
11 [6.0, 16.0]
15 [9.75, 21.25]
10 [3.5, 15.25]
15 [8.5, 31]
0.027
RV failure risk score Pre-LVAD days in hospital Device type
0.55
HeartMate II
41 (61.2)
31 (70.4)
4 (50)
10 (66.7)
HVAD
24 (35.8)
13 (29.5)
4 (50)
4 (26.7)
2 (3.0)
0 (0.0)
0 (0.0)
1 (6.7)
117.49 (64.47)
128.60 (61.37)
141.50 (69.24)
130.14 (84.75)
0.68
1 (1.5)
0 (0.0)
0 (0.0)
0 (0..0)
0.80
6 (9.0)
7 (15.9)
3 (37.5)
4 (26.7)
0.080
HeartMate 3 Surgical details CPB time (min) Concomitant MV surgery Concomitant TV
surgery Concomitant AV surgery
5 (7.5)
4 (9.1)
1 (12.5)
1 (6.7)
0.95
Table 1: Baseline patient characteristics stratified by congestion group. Values presented as mean (sd), median [IQR], or N (%). ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BTT, bridge to transplant; CPB, cardiopulmonary bypass; INR, international normalized ratio; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; RV, right ventricular; TR, tricuspid regurgitation.
Congestion status at admission Not congested Congested p N=75 N=59 59.7 (9.2) 54.5 (12.0) 0.005
Age (y) Female sex
12 (16.0)
14 (23.7)
INTERMACS profile
0.37
Congestion status at implant Not congested Congested N=111 N=23 57.7 (10.9) 56.2 (10.4) 20 (18.0)
6 (26.1)
0.074
p 0.57 0.55 0.63
1
8 (10.7)
7 (11.9)
10 (9.0)
5 (21.7)
2
24 (32.0)
33 (55.9)
48 (43.2)
9 (39.1)
3
32 (42.7)
16 (27.1)
41 (36.9)
7 (30.4)
≥4
11 (14.6)
3 ( 5.1)
12 (11.8)
2 (8.7)
BMI (kg/m )
26.2 (5.5)
28.2 (5.6)
0.043
26.8 (5.4)
28.8 (6.5)
0.12
BTT strategy
44 (58.7)
39 (66.1)
0.48
69 (62.2)
14 (60.9)
1
Creatinine (mg/dL)
1.14 [0.89, 1.58]
1.37 [1.08, 1.81]
0.016
1.21 [0.92, 1.65]
1.43 [1.06, 1.84]
0.16
AST (U/L)
26.0 [19.0, 37.5]
26.0 [21.0, 33.5]
0.88
26.0 [19.5, 34.0]
29.0 [22.0, 37.5]
0.38
ALT (U/L) Total bilirubin (mg/dL) INR
26.0 [17.0, 36.0]
23.0 [15.0, 42.0]
0.71
25.0 [17.0, 37.0]
20.0 [12.0, 41.5]
0.33
0.8 [0.5, 1.4]
1.1 [0.7, 1.8]
0.010
0.9 [0.5, 1.35]
1.2 [1.0, 2.4]
0.009
1.1 [1.1, 1.2]
1.2 [1.1, 1.3]
0.004
1.1 [1.1, 1.2]
1.2 [1.1, 1.4]
0.021
11.3 (5.9)
16.1 (6.6)
<0.001
12.8 (6.5)
16.4 (7.0)
0.018
2
MELD score TR severity
0.32
0.72
None
1 ( 1.3)
0 (0.0)
1 (0.9)
0 (0.0)
Mild
36 (48.0)
23 (39.0)
51 (45.9)
8 (34.8)
Moderate
31 (41.3)
25 (42.4)
45 (40.5)
11 (47.8)
7 (9.3)
11 (18.6)
14 (12.6)
4 (17.4)
Severe RV dysfunction
0.086
0.82
None
6 (8.0)
1 (1.7)
5 (4.5)
2 (8.7)
Mild
23 (30.7)
11 (18.6)
29 (26.1)
5 (21.7)
Moderate
28 (37.3)
29 (49.2)
46 (41.4)
11 (47.8)
Severe
16 (21.3)
18 (30.5)
29 (26.1)
5 (21.7)
0 [0, 0]
0 [0, 2.5]
0.03
0 [0, 0]
0 [0, 3]
0.01
6.0 [3.0, 9.0]
10.0 [7.5, 16.5]
<0.001
8.0 [5.0, 12.0]
8.0 [3.0, 12.5]
0.80
RV failure risk score Pre-LVAD days in hospital Device type
0.52
0.74
HeartMate II
45 (60.0)
41 (69.5)
72 (64.9)
14 (60.9)
HVAD
28 (37.3)
17 (28.8)
37 (33.3)
8 (34.8)
2 (2.7)
1 (1.7)
2 (1.8)
1 (4.3)
120.1 (64.9)
129.0 (67.1)
0.45
121.8 (63.2)
134.3 (78.0)
0.42
1 (1.3)
0 (0.0)
1
1 (0.9)
0 (0.0)
1
9 (12.0)
11 (18.6)
0.41
13 (11.7)
7 (30.4)
0.049
HeartMate 3 Surgical details CPB time (min) Concomitant MV surgery Concomitant TV
surgery Concomitant AV surgery
6 (8.0)
5 (8.5)
1
9 (8.1)
2 (8.7)
Table 2: Baseline patient characteristics stratified by congestion status at admission and implant. Values presented as mean (sd), median [IQR], or N (%). ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BTT, bridge to transplant; CPB, cardiopulmonary bypass; INR, international normalized ratio; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; RV, right ventricular; TR, tricuspid regurgitation.
1
No congestion N=67
Resolved congestion N=44
New congestion N=8
Persistent congestion N=15
p
Admission RAP (mmHg) 9 (3) 19 (4) 9 (2) 21 (4) <0.001 PASP (mmHg) 52 (14) 60 (11) 52 (15) 61 (15) 0.002 PADP (mmHg) 25 (7) 34 (6) 26 (6) 33 (7) <0.001 PCWP (mmHg) 22 (7) 31 (5) 25 (7) 30 (8) <0.001 2 CI (L/min/m ) 2.1 (0.5) 1.8 (0.5) 1.9 (0.3) 1.8 (0.7) 0.001 PAPi 3.4 (2.8) 1.5 (0.6) 2.9 (1.3) 1.4 (0.5) <0.001 Implant RAP (mmHg) 8 (3) 9 (3) 17 (3) 16 (2) <0.001 PASP (mmHg) 50 (13) 50 (10) 63 (17) 58 (13) 0.01 PADP (mmHg) 24 (7) 25 (6) 32 (8) 30 (6) 0.001 PCWP (mmHg) 19 (6) 21 (6) 30 (2) 26 (5) <0.001 2 CI (L/min/m ) 2.2 (0.5) 2.3 (0.4) 2.2 (0.5) 2.2 (0.5) 0.87 PAPi 4.4 (4.2) 3.1 (1.8) 1.8 (0.5) 1.7 (0.5) 0.008 Table 3: Hemodynamic parameters at admission and time of implant stratified by congestion group. Results presented as mean (sd). CI, cardiac index; PADP, pulmonary artery diastolic pressure; PAPi, pulmonary artery pulsatility index; PASP, pulmonary artery systolic pressure; PCWP, pulmonary capillary wedge pressure; RAP, right atrial pressure.
Adjusted Models Unadjusted model
Model A*
Model B**
Model C***
HR (95% CI)
HR (95% CI)
HR (95% CI)
HR (95% CI)
No congestion
ref
ref
ref
ref
Resolved congestion
0.6 (0.2-1.6)
0.6 (0.2-1.7)
0.8 (0.3-2.3)
0.8 (0.3-2.4)
New congestion
0.6 (0.1-4.4)
0.6 (0.1-4.4)
0.6 (0.1-4.9)
0.6 (0.1-5.1)
Persistent congestion
3.1 (1.3-7.6)
3.3 (1.3-8.5)
4.6 (1.7-12.5)
5.6 (1.9-16.9)
Table 4: Univariable and multivariable Cox proportional hazards models of the association between congestion group and time to death or RVAD implantation. Three sequentially adjusted multivariable models are presented. INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; RVFRS, right ventricular failure risk score. * adjusted for INTERMACS profile < 3 ** adjusted for INTERMACS profile < 3 and age ** adjusted for INTERMACS profile < 3, age, and RVFRS
Resolved New Persistent congestion congestion congestion p N=44 N=8 N=15 19.5 [14.75, 21.5 [14.75, Total LOS 16 [14, 22.5] 22 [17.5, 27] 0.17 29] 28.5] ICU 5 [3, 8] 6 [4, 11.25] 6 [2.75, 14.5] 9 [6.5, 19.5] 0.009 Table 5: Postoperative length of stay stratified by congestion group. ICU, intensive care unit No congestion N=67
Figures
Figure 1: Derivation of the cohort and division into four congestion groups. ECMO, extracorporeal membrane oxygenation; LVAD, left ventricular assist device.
Figure 2: Survival free of death or RVAD stratified by congestion status at admission or implant. Freedom from death or RVAD implantation for patients stratified by congestion status at admission (A) or implant (B). Congestion was defined using a cutoff right atrial pressure of 14 mmHg. RVAD, right ventricular assist device.
Figure 3: Incidence of severe early right heart failure stratified by congestion group. Severe early right heart failure was the composite of death within 30 days, need for RVAD within 30 days, or need for postoperative inotropes beyond 14 days. RHF, right heart failure; RVAD, right ventricular assist device.