Left ventricular assist device (LVAD) implantation is the most common mechanical circulatory therapy for advanced heart failure.1,2 Unfortunately, an estimated 10–40% of patients demonstrate some degree of right ventricular (RV) dysfunction after LVAD placement.2 RV failure is characterized by an elevation of central venous pressure (CVP) with manifestations of peripheral edema, ascites, or worsening renal and hepatic function.3,4 This complication is associated with longer intensive care unit (ICU) lengths of stay (LOS), increased mortality, and worse posttransplant outcomes.2,5,6
Although current literature lacks an explicit uniform definition of RV failure, many sources describe this phenomenon by prolonged inotrope infusion requirements, the need for inhaled pulmonary vasodilator therapy, ventilator support for greater than 1 week, or requirement of a RV assist device (RVAD).5 Pharmacologic management of RV failure generally includes the use of diuretics, inotropes, and inhaled pulmonary vasodilator therapy. In addition to these treatment modalities, phosphodiesterase-5 inhibitors (PDE5-i) have gained interest as adjunctive therapy for the treatment of postoperative RV failure after LVAD implantation. PDE5-i prolong the activity of cyclic guanosine monophosphate (cGMP) in smooth muscle cells leading to vasodilation and antiproliferative effects within the pulmonary vasculature.7–9 Furthermore, effects on cGMP may lead to downstream inhibition of phosphodiesterase-3 enzymes and direct enhancement of RV contractility.7,8
The 2013 International Society of Heart and Lung Transplantation Guidelines for Mechanical Circulatory Support provide a weak, class IIb recommendation for PDE5-i use in the setting of persistent pulmonary hypertension and RV dysfunction after LVAD placement.3 This recommendation is based upon limited data, comprised of case reports and case series, demonstrating improved hemodynamic parameters such as pulmonary vascular resistance (PVR) and pulmonary artery pressure (PAP), which may serve as markers for RV function.10,11,9,12,13 Whether these improvements in pulmonary hemodynamics influence longer term clinical outcomes has not been extensively evaluated. Therefore, the purpose of this investigation was to assess the real-world clinical effectiveness of the addition of PDE5-i therapy to routine care for RV dysfunction after LVAD implantation.
Study Design and Population
This was a single-center, retrospective cohort study at a tertiary care academic medical center. All LVAD recipients between January 2011 and May 2015 were assessed for eligibility. Inclusion criteria were: 1) age ≥ 18 years and 2) initial LVAD implantation at the study institution. Patients in the PDE5-i group were required to receive sildenafil or tadalafil for ≥ 48 hours while hospitalized post-LVAD implantation to be included. Exclusion criteria were: 1) LVAD exchange; 2) concomitant receipt of RVAD or extracorporeal membrane oxygenation (ECMO); 3) expiration within 48 hours of procedure; and 4) LVAD follow-up outside of the study institution. The study protocol was approved by the Barnes-Jewish Hospital and Washington University in St. Louis institutional review board.
Data Collection and Study Outcomes
Electronic medical records were reviewed for demographic, clinical, and laboratory data. Perioperative degree of RV dysfunction was estimated qualitatively based on an observed assessment of the tricuspid annulus movement and the RV free wall movement and categorized as normal, mild, moderate, or severe based on transthoracic (TTE) and transesophageal (TEE) echocardiographic evaluations performed by expert echocardiographers during routine clinical care. The primary study outcome was all-cause hospital readmission at 30 days after discharge from the index hospitalization. This end-point was selected because it is of clinical importance and occurs in a sufficient frequency to allow for potential detection of meaningful differences between patient groups, should they exist. Secondary outcomes were duration of intravenous (IV) inotrope or inhaled pulmonary vasodilator therapy, hospital and ICU LOS, duration of mechanical ventilation, overall post-LVAD survival, time to readmission, and improvement in the degree of postoperative RV dysfunction which was determined by the comparison of intraoperative TEE data to postoperative TTE findings.
Baseline characteristics and clinical outcomes were compared between the 2 groups using the χ2 test for categorical data and Mann–Whitney U test for continuous data. To address potential confounding by indication, propensity score (PS) matched analyses were performed.14 Propensity scores were derived by unconditional logistic regression controlling for variables associated with PDE5-i treatment (p < 0.2) and Interagency Registry for Mechanically Assisted Circulatory Support score, which has been shown to potentially influence clinical outcomes post-LVAD implantation.15 Patients receiving PDE5-i treatment were matched 1:1 to patients not receiving PDE5-i, using a greedy nearest neighbor algorithm with a caliper width of 0.2. Balance was assessed in the matched cohort by comparing the standardized differences in baseline covariates between treated and untreated subjects, with an absolute difference > 0.25 signifying meaningful imbalance.16 All outcomes were analyzed in both overall and PS matched cohorts.
To evaluate the relationship between in-hospital PDE5-i treatment and 30 day readmission, multivariable log-binomial regression analysis was performed with PDE5-i treatment status forced into the model as the exposure of interest.17 Other variables were selected for inclusion in this model using a backward stepwise approach. Post-LVAD survival and time to first readmission were evaluated using the Kaplan–Meier method and log-rank test. The association between in-hospital PDE5-i use and post-LVAD survival was further assessed by Cox regression.
A series of secondary analyses were designed a priori. This included a comparison of 30 day readmission between patients who received in-hospital PDE5-i only versus those who continued therapy postdischarge. Additionally, we evaluated the association between PDE5-i treatment and 30 day readmission in subgroups of patients classified as having moderate or severe RV dysfunction on intraoperative postcardiopulmonary bypass TEE and those requiring concomitant inhaled epoprostenol therapy. The moderate-to-severe RV dysfunction subgroup was further evaluated for improvement to no-or-mild RV dysfunction. Statistical analyses were performed using SPSS (IBM Corporation, version 22, Armonk, NY) and SAS software (SAS Institute, version 9.3, Cary, NC) with the level of statistical significance set at < 0.05.
A total of 445 primary LVAD implantations were performed during the 4 year study period. After exclusion of 127 recipients, 318 patients were included in the final analysis (Figure 1). Of these, 208 patients (65.4%) received PDE5-i therapy for RV dysfunction post-LVAD implantation and 110 patients (34.6%) received standard of care only. Baseline characteristics of patients receiving PDE5-i therapy post-LVAD were compared with those not receiving PDE5-i (Table 1). A significantly larger proportion of patients who received PDE5-i were African American, had higher postoperative total bilirubin, chronic kidney disease, moderate-to-severe intraoperative RV dysfunction, and received LVAD implantation as a bridge to transplant (all p < 0.05). These patients were also more likely to receive postoperative dobutamine and inhaled epoprostenol therapy. In the PS-matched cohort (n = 154), these differences were no longer apparent (Table 1) and the 2 groups were well matched (absolute standardized differences <0.25 for all covariates).
The median time to initiation of postoperative PDE5-i therapy was 2 days (interquartile range [IQR], 1–4 days). Sildenafil 20 mg orally 3 times per day was the most commonly prescribed PDE5-i regimen (Table 2). The median in-hospital duration of therapy was 14.0 days (IQR, 8.3–22.8 days) for sildenafil and 13.5 days (IQR, 8.0–18.3 days) for tadalafil (Table 2).
Overall, 30 day all-cause readmission was 35.5%. Use of in-hospital PDE5-i therapy did not influence 30 day all-cause readmission in the overall cohort (36.1% vs. 34.5%; risk ratio [RR], 1.02; 95% confidence interval [CI], 0.86–1.21; p = 0.788; Table 3) or the PS-matched cohort (36.4% vs. 32.5%; RR, 1.06; 95% CI, 0.84–1.33; p = 0.611; Table 3). Factors significantly associated with 30 day readmission in univariable analysis were increased median postoperative aspartate aminotransferase (p = 0.006) and comorbid hypertension (odds ratio [OR], 2.82; 95% CI, 1.54–5.16; p = 0.001). After adjusting for confounding variables in multivariable log-binomial regression, the lack of association between PDE5-i therapy and 30 day all-cause readmission persisted (adjusted RR, 1.03; 95% CI, 0.76–1.40; p = 0.856). Only comorbid hypertension was associated with 30 day readmission in this model (adjusted RR, 2.07; 95% CI, 1.30–3.29; p = 0.002). There was no association between PDE5-i therapy and readmission because of bleeding, thrombosis, anemia, arrhythmia, infection, heart failure, syncope/falls, or other causes in overall and PS-matched cohorts (Table 3).
With regard to secondary outcomes, patients receiving PDE5-i therapy required significantly longer durations of IV epinephrine and inhaled epoprostenol (Table 3). Additionally, median hospital LOS, ICU LOS, and duration of mechanical ventilation were significantly prolonged in patients receiving PDE5-i therapy in univariable analysis. However, after PS matching there were no significant differences between groups with regard to any of these secondary outcomes (Table 3).
The median duration of post-LVAD follow-up was 2.33 years (IQR, 1.45–3.26 years). In the overall cohort, no association between in-hospital PDE5-i therapy and post-LVAD survival was observed (hazard ratio, 1.24; 95% CI, 0.84–1.84; p = 0.296; Figure 2A). This lack of association persisted in the PS-matched cohort as well (hazard ratio, 0.98; 95% CI, 0.52–1.83; p = 0.945; Figure 2B). The median time to first readmission was 39.89 days (IQR, 16.07–106.09 days) overall and did not significantly differ between patients who received in-hospital PDE5-i therapy versus those received standard of care only before or after PS matching (Table 3). In Cox regression accounting for the duration of PDE5-i exposure, there was no difference in time to first readmission between treatment groups before (log-rank p = 0.700) or after PS matching (log-rank p = 0.848). Mortality at 90 days and at 6 months did not differ between the groups before or after PS matching (Table 3).
A total of 150/208 (72.1%) patients continued PDE5-i therapy after hospital discharge, 53/208 (25.5%) patients had therapy discontinued before or at discharge, and 5/208 (2.4%) patients expired during the index hospitalization. Hypotension was the documented reason for PDE5-i discontinuation in 11/53 (20.8%) patients. For patients continuing therapy after discharge, the median duration of therapy was 230 days (IQR, 69–478 days). All-cause readmission at 30 days did not significantly differ between patients who continued PDE5-i therapy postdischarge compared with those patients who received in-hospital PDE5-i therapy only (34.0% [51 of 150] vs. 45.3% [24 of 53]; OR, 0.62; 95% CI, 0.33–1.18; p = 0.143).
In the subgroup of patients receiving concomitant inhaled epoprostenol, there was no difference in 30 day readmission between patients who received PDE5-i therapy versus those who did not (38.8% [54 of 139] vs. 33.9% [20 of 59]; OR, 1.08; 95% CI, 0.86–1.35; p = 0.510). There was also no benefit of PDE5-i therapy on 30 day readmission observed in the subgroup of patients with moderate or severe RV dysfunction (36.5% [50 of 137] vs. 34.0% [17 of 50]; OR, 1.04; 95% CI, 0.82–1.32; p = 0.753).
Improvement in the degree of postoperative RV dysfunction was assessed by echocardiogram (median POD 6) in all but 1 patient in the subgroup of patients with moderate or severe RV dysfunction on perioperative echocardiogram. There was no benefit of PDE5-i therapy on improvement in RV function observed in this subgroup of patients pre- or post-PS matching (43.1% [59 of 137] PDE5-i vs. 40.8% [20 of 50] no PDE5-i; p = 0.785 or 51.1% [23 of 45] PDE5-i vs. 36.6% [15 of 42] no PDE5-i; p = 0.175). Tricuspid annular plane systolic excursion (TAPSE) was documented on postoperative echocardiograph for 42 patients included in the subgroup analysis and also showed no difference between groups (0.95 cm PDE5-i vs. 0.80 cm no PDE5-i, p = 0.370).
This study evaluated clinical outcomes associated with PDE5-i therapy for post-LVAD RV dysfunction, which are not well-described in the existing literature. In the largest and most extensive investigation to date, we found that the addition of in-hospital PDE5-i treatment to standard of care did not reduce 30 day readmission in this propensity-matched population. Ravichandran et al. also reported no difference in 30 day readmissions with the use of post-LVAD sildenafil therapy; however, this evaluation was primarily focused on safety outcomes and occurred in a smaller cohort of patients.18 Moreover, PDE5-i therapy did not appear to facilitate quicker weaning of IV inotropes, inhaled epoprostenol, or mechanical ventilation. Finally, in-hospital PDE5-i therapy did not lead to a difference in overall survival, hospital LOS, ICU LOS, time to first readmission, or improvement in the degree of RV dysfunction in this large cohort of LVAD recipients.
The pharmacologic activity of PDE5-i aims to disrupt the pathogenesis of pulmonary hypertension and RV dysfunction by increasing intracellular concentrations of cGMP and prolonging nitric oxide signaling, thereby increasing pulmonary arterial vasodilation and reducing RV afterload.4,7–9,11 Existing data also suggest that PDE5 enzyme expression is upregulated during RV hypertrophy, so inhibition of this enzyme may be directly associated with enhanced myocardial contractility and improvement of RV output in this setting.19 Yet despite this pathophysiologic basis, the use of post-LVAD PDE5-i therapy has only been supported by weak evidence.3 A previous case report and small case series (n = 10) described the successful use of PDE5-i therapy to acutely facilitate postoperative weaning of inhaled vasodilators and IV inotropes.11,20 In the current study, the addition of PDE5-i therapy failed to reduce durations of IV inotropes and inhaled epoprostenol compared with standard of care only, after adjusting for differences in baseline characteristics via PS matching.
A recent investigation of 14 LVAD recipients showed lower rates of RV failure at 48 hours with the use of PDE5-i, as demonstrated by reductions in tricuspid regurgitation (TR), PAP, transpulmonary gradient (TPG), central venous pressure, and increased TAPSE.12 In the current study, TAPSE was collected when documented in patients with moderate-to-severe RV dysfunction and was not significantly different in the PDE5-i therapy group. Additionally, in-hospital PDE5-i therapy was not associated with improved postoperative RV function, as assessed by formal echocardiography, or 30 day readmission in the subgroup of patients with moderate-to-severe perioperative RV dysfunction.
Reductions in mean PAP and PVR have been noted in patients who received PDE5-i therapy 1–2 weeks after LVAD placement in the setting of persistent pulmonary hypertension defined as PVR ≥ 3 Wood Units despite normalization of pulmonary capillary wedge pressure (PCWP) to < 15 mm Hg.10 Sustained effects throughout a 15 week follow-up period suggest this intervention may be useful as chronic therapy.3,10 In order to assess longitudinal clinical effects of PDE5-i therapy in the current study, we evaluated overall post-LVAD survival and censored patients at the time of PDE5-i discontinuation. Despite a median follow-up period of more than 2 years, PDE5-i therapy did not influence post-LVAD survival. Therefore, PDE5-i therapy also does not appear to affect longer term outcomes in this population. A practical issue associated with prolonged treatment involves the cost of PDE5-i therapy. Patients often report barriers to PDE5-i use for this indication in the outpatient setting, including lack of insurance coverage and high out-of-pocket costs. Thus, pharmacoeconomic implications of PDE5-i therapy in this patient population should also be carefully considered.
There are several limitations of this study worth noting. This was a retrospective cohort investigation and suffers from the potential biases inherent to this design. Initiation of PDE5-i therapy, including dose and duration, was left to the discretion of treating physicians, as no standard criteria for use exist for this population at our institution. In general, a combination of echocardiographic findings and an observed report of RV function in the operating room are used to drive this decision. It is also important to recognize that classification of RV dysfunction may have been subject to variability among different reviewing physicians, although echocardiograph readers were not aware of study group and reports were generated as part of routine clinical care. Pulmonary hemodynamic markers like PAP and PVR were not collected because of inconsistent reporting of these results, as formal pre- and postoperative right heart catheterizations were not routinely performed for this patient population at our institution during the study time period. Numerical measurements of RV function, such as TAPSE, were collected for patients classified as having moderate–severe RV dysfunction; however, data were missing for many patients. Despite this limited availability of information, the primary objective of our evaluation was to capture clinical outcomes rather than surrogate markers of RV dysfunction. It seems likely that response to PDE5-i therapy is patient-dependent and those patients exhibiting improved pulmonary hemodynamic markers following PDE5-i introduction may be those more likely to experience benefits in clinical outcomes. Future studies should aim to characterize factors associated with pulmonary hemodynamic PDE5-i response to ascertain a PDE5-i “responder” phenotype which can be applied clinically. Longitudinal clinical outcomes in this subgroup of PDE5-i responders would then be able to be evaluated. Because of the observational nature of this study, we were unable to assess adherence to PDE5-i therapy after hospital discharge. It should also be noted that multiple dosing regimens were employed in our study and it is possible that optimization of dose and duration of PDE5-i therapy may lead to better outcomes. The PDE5-i dosing strategies employed in previous studies are variable and range from a single, 25 mg enteral dose of sildenafil up to 75 mg 3 times daily for a prolonged duration after LVAD implantation.10,12,13,20 Unfortunately, the optimal PDE5-i regimen has yet to be described in this population.
This investigation did not include a detailed safety analysis and future study is warranted to better evaluate potential adverse effects of PDE5-i therapy in the post-LVAD population. Such an investigation would be best assessed in a prospective analysis with uniform evaluation and documentation procedures. In the recently published evaluation by Ravichandran et al.18, the use of sildenafil in 122 LVAD recipients was not associated with an increased risk of stroke or gastrointestinal bleeding; however, other clinically important adverse effects, such as hypotension, were not evaluated. It is plausible that patients with more severe RV dysfunction would be more likely to be treated with adjunctive PDE5-i therapy. We attempted to account for this confounding by indication through PS matching and secondary subgroup analyses; however, these methods are unable to account for unmeasured confounders that may exist. This potential bias can only be overcome with a prospective randomized trial design, which given the low absolute differences in clinical outcomes observed in the current study, would require enrollment of a very large number of patients and come at a significant financial cost. In the absence of such prospective data, the current study provides the most robust clinical evidence to date evaluating the clinical utility of PDE5-i therapy after LVAD implantation.
In summary, the addition of in-hospital PDE5-i therapy to standard of care for RV dysfunction did not result in improved clinical outcomes nor facilitate weaning of mechanical ventilation, IV inotropes, or pulmonary vasodilatory agents in a large cohort of LVAD recipients. Although a subgroup of patients who are able to benefit clinically from PDE5-i therapy post-LVAD may exist, we were unable to uncover such a group in the current study. Considering the lack of improved clinical outcomes observed in the current study and the substantial cost of these medications, we believe additional research is needed to support widespread use of PDE5-i therapy for RV dysfunction post-LVAD implantation.
The authors acknowledge Nicholas Hampton, PharmD, for his contribution to data acquisition and technical support.
1. Kirklin JK, Naftel DC, Pagani FD, et al. Seventh INTERMACS annual report: 15,000 Patients and counting. J Heart Lung Transplant 2015.34: 1495–1504.
2. Lampert BC, Teuteberg JJ. Right ventricular failure after left ventricular assist devices. J Heart Lung Transplant 2015.34: 1123–1130.
3. Feldman D, Pamboukian SV, Teuteberg JJ, et al; International Society for Heart and Lung Transplantation: The 2013 International Society for Heart and Lung Transplantation guidelines for mechanical circulatory support: executive summary. J Heart Lung Transplant 2013.32: 157–187.
4. Alaeddini J, Uber PA, Park MH, Scott RL, Ventura HO, Mehra MR. Efficacy and safety of sildenafil in the evaluation of pulmonary hypertension in severe heart failure. Am J Cardiol 2004.94: 1475–1477.
5. Cypel M, Levvey B, Van Raemdonck D, et al; International Society for Heart and Lung Transplantation: International Society for Heart and Lung Transplantation donation after circulatory death registry report. J Heart Lung Transplant 2015.34: 1278–1282.
6. Romano MA, Cowger J, Aaronson KD, Pagani FD. Diagnosis and management of right-sided heart failure in subjects supported with left ventricular assist devices. Curr Treat Options Cardiovasc Med 2010.12: 420–430.
7. Archer SL, Michelakis ED. Phosphodiesterase type 5 inhibitors for pulmonary arterial hypertension. N Engl J Med 2009.361: 1864–1871.
8. Barnett CF, De Marco T. Pulmonary hypertension associated with left-sided heart disease. Heart Fail Clin 2012.8: 447–459.
9. Baker WL, Radojevic J, Gluck JA. Systematic review of phosphodiesterase-5 inhibitor use in right ventricular failure following left ventricular assist device implantation. Artif Organs 2016.40: 123–128.
10. Tedford RJ, Hemnes AR, Russell SD, et al. PDE5A inhibitor treatment of persistent pulmonary hypertension after mechanical circulatory support. Circ Heart Fail 2008.1: 213–219.
11. Matamis D, Pampori S, Papathanasiou A, et al. Inhaled NO and sildenafil combination in cardiac surgery patients with out-of-proportion pulmonary hypertension: acute effects on postoperative gas exchange and hemodynamics. Circ Heart Fail 2012.5: 47–53.
12. Hamdan R, Mansour H, Nassar P, Saab M. Prevention of right heart failure
after left ventricular assist device implantation by phosphodiesterase 5 inhibitor. Artif Organs 2014.38: 963–967.
13. Klodell CT Jr., Morey TE, Lobato EB, et al. Effect of sildenafil on pulmonary artery pressure, systemic pressure, and nitric oxide utilization in patients with left ventricular assist devices. Ann Thorac Surg 2007.83: 68–71; discussion 71.
14. Glynn RJ, Schneeweiss S, Stürmer T. Indications for propensity scores and review of their use in pharmacoepidemiology. Basic Clin Pharmacol Toxicol 2006.98: 253–259.
15. Boyle AJ, Ascheim DD, Russo MJ, et al. Clinical outcomes for continuous-flow left ventricular assist device patients stratified by pre-operative INTERMACS classification. J Heart Lung Transplant 2011.30: 402–407.
16. Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009.28: 3083–3107.
17. McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003.157: 940–943.
18. Ravichandran AK, LaRue SJ, Novak E, et al. Sildenafil in left ventricular assist device is safe and well-tolerated. ASAIO J 2017.64: 280–281.
19. Nagendran J, Archer SL, Soliman D, et al. Phosphodiesterase type 5 is highly expressed in the hypertrophied human right ventricle, and acute inhibition of phosphodiesterase type 5 improves contractility. Circulation 2007.116: 238–248.
20. Mychaskiw G, Sachdev V, Heath BJ. Sildenafil (viagra) facilitates weaning of inhaled nitric oxide following placement of a biventricular-assist device. J Clin Anesth 2001.13: 218–220.