B-type natriuretic peptide: powerful predictor of end-stage chronic heart failure in individuals with systolic dysfunction of the systemic right ventricle

Aim To assess whether B-type natriuretic peptide (BNP) can serve as a predictor of end-stage chronic heart failure (CHF) in patients with severe systolic dysfunction of the systemic right ventricle (SRV). Methods We performed a retrospective analysis in 28 patients with severe systolic dysfunction of the SRV (ejection fraction 23 ± 6%) who were evaluated as heart transplant (HTx) candidates between May 2007 and October 2014. The primary endpoints of the study (end-stage CHF) were progressive CHF, urgent HTx, and ventricular assist device (VAD) implantation. Plasma BNP levels were measured using a chemiluminescent immunoassay. Results During median follow-up of 29 months (interquartile range, 9-50), 3 patients died of progressive CHF, 5 patients required an urgent HTx, and 6 patients underwent VAD implantation. BNP was a strong predictor of end-stage CHF (hazard ratio per 100 ng/L: 1.079, 95% confidence interval, 1.042-1.117, P˂0.001). The following variables with corresponding areas under the curve (AUC) were identified as the most significant predictors of end-stage CHF: BNP (AUC 1.00), New York Heart Association functional class class III or IV (AUC 0.98), decompensated CHF in the last year (AUC 0.96), and systolic dysfunction of the subpulmonal ventricle (AUC 0.96). Conclusion BNP is a powerful predictor of end-stage CHF in individuals with systolic dysfunction of the SRV.

Transposition of the great arteries (TGA), the aorta and the pulmonary artery, is a congenital heart defect with a morphologically right ventricle pumping blood into systemic circulation. It occurs either as complete TGA (D-TGA) or congenitally corrected TGA (ccTGA). D-TGA is characterized by dextroposition of the systemic ventricle, sinistroposition of the subpulmonal morphologically left ventricle, and rightsided rotation of the great arteries. On the other hand, the term ccTGA describes sinistroposition of the the systemic ventricle, dextroposition of the subpulmonal morphologically left ventricle, and left-sided rotation of the great arteries (1). Preserved left-sided position of the pulmonary venous return and right-sided localization of the systemic venous return in both types of TGA result in almost complete separation of the systemic and pulmonary circulation in D-TGA, which must be surgically corrected in early childhood, and physiological correction of the defect in ccTGA (1). Unfortunately, both the operative correction of D-TGA by atrial tunneling (atrial switch, the Mustard or Senning procedure) and anatomic situation in ccTGA leave the morphologically right ventricle as the systemic ventricle (1)(2)(3)(4)(5). As compared with the morphologically left ventricle, tricuspid valve, thin walls, different ventricular shape, and myocyte architecture, make this type of the systemic ventricle susceptible to dilatation, development of tricuspid regurgitation, and heart failure (1)(2)(3)(4)(5).
Tricuspid regurgitation, systolic dysfunction of the SRV, and chronic heart failure (CHF) were associated with impaired survival both in individuals with D-TGA (2-4) and ccTGA (5). Prognostic stratification of these patients remains notoriously difficult, especially in the presence of SRV systolic dysfunction. Although B-type natriuretic peptide (BNP) provides strong diagnostic and prognostic information in left ventricular systolic dysfunction (6,7), its role in CHF due to SRV failure remains poorly understood. BNP and/or N-terminal pro BNP correlate just modestly with systolic SRV dysfunction (8-11). In addition, there are scarce prognostic data in individuals with SRV (11). Importantly, improved prognostic stratification could facilitate appropriate use of ventricular assist devices (VAD) and/or heart transplant (HTx) in this population (12,13). Therefore, we decided to evaluate whether BNP has a prognostic role in individuals with SRV who were referred to our heart failure clinic for consideration of HTx.

PatIeNts aNd MetHods study population and data collection
We performed a retrospective analysis in all patients with SRV who were referred to the Heart Failure Clin-ic of our institution (Institute for Clinical and Experimental Medicine, Prague, Czech Republic) between May 2007 and October 2014. Available demographic, clinical, laboratory, electrocardiographic, and echocardiographic data from the first visit to our clinic were collected from the institutional database. Doses of angiotensin converting enzyme inhibitors, angiotensin receptor blockers, and beta-blockers were expressed as a percentage of the maximum recommended daily dose according to the latest European Society of Cardiology guidelines (14). The patients were followed until July 30, 2015. Vital status and the necessity of urgent HTx or VAD implantation were recorded. The primary endpoint of this study included death, urgent HTx, or VAD implantation. It was considered an equivalent of end-stage CHF.
The study group consisted of 28 consecutive patients with SRV. This group included 21 patients (75%) after atrial switch due to D-TGA (14 patients with the Mustard procedure, 7 patients with the Senning procedure) and 7 patients (25%) with ccTGA. The following associated lesions were corrected during the primary cardiac surgery: ventricular septal defect in 6 patients with atrial switch, ventricular and atrial septal defect in 1 patient with ccTGA, and ventricular septal defect and valvular pulmonary artery stenosis in another patient with ccTGA. 5 patients had a history of reoperation: 3 patients with atrial switch and 1 patient with ccTGA required tricuspid valve replacement, while additional 1 patient after atrial switch underwent aortic valve replacement due to severe regurgitation. All 20 men (71%) and 8 women (29%) with mean age of 34 years (range 15-65) had at the first evaluation severe systolic SRV dysfunction with mean ejection fraction of 23 ± 6% ranging from 14% to 35%. Median BNP plasma level was 618 ng/L (interquartile range [IQR] 86.5-1363.5) ( Table 1). This study conformed to the principles of the Declaration of Helsinki. It was approved by the ethics committee of the Institute for Clinical and Experimental Medicine, Prague, Czech Republic.

Laboratory methods
Blood samples for the measurement of BNP plasma levels were collected into chilled tubes containing EDTA (1 mg/ mL) and aprotinin (500 kallikrein inactivator U/mL). Test tubes were immediately centrifuged. Plasma BNP levels were measured using a chemiluminescent immunoassay (Abbott Laboratories, Diagnostics Division, Abbott Park, IL, USA, Supplier: Abbott, Max-Planck-Ring 2, Wiesbaden, Germany). The lower limit of detection was 10 ng/L, intraassay coefficient of variation (CV)<3.8%, and inter-assay Receiver operating characteristic curve (ROC) analysis and Youden's index (sensitivity + specificity -100) were used to assess and compare the performance of selected variables to predict the primary endpoint. For all tests a probability value of P < 0.050 was considered significant. All analyses were performed using the statistical software SPSS, version 17.0 (SPSS Inc., Chicago, IL, USA).

events during the follow-up
During a median follow-up of 29 months (IQR 9-50), 14 patients experienced the primary endpoint. 3 patients (11%) died of end-stage HF after 5, 14, and 15 months. 5 patients (18%) required an urgent HTx (after 4-14 months) and 6 patients (21%) underwent VAD implantation (HeartMateII, Thoratec Corporation, Pleasanton, CA, USA) after 0-17 months as a bridge to transplant. The remaining 14 individuals were free of the primary endpoint (group A, favorable outcome), whereas the above mentioned 14 patients were included in the group B (unfavorable outcome).

Predictors of the unfavorable outcome
Unfavorable outcome in group B was predicted by the following signs of advanced CHF: a history of decompensated HF in the last 12 months, higher New York Heart Association (NYHA) functional class, higher heart rate, lower systolic blood pressure, markedly elevated plasma levels of BNP (Figure 1), higher creatinine serum levels, worse tolerance of angiotensin converting enzyme inhibitors or sartans, and higher daily doses of furosemide. Advanced stage of cardiac remodeling in group B correlated with a broader QRS complex on ECG, lower ejection fraction of the SRV, higher occurrence of severe regurgitation on the systemic atrioventricular valve, and the presence of systolic dysfunction of the subpulmonal ventricle. Similar results were revealed using univariate Cox regression analysis ( Table 2). Efficacy of clinically relevant predictors of the unfavorable outcome was compared using the ROC curve analysis ( Table 3). The best area under the curve (1.00) and the Youden's index (100%) was found for BNP (optimal cutpoint of 618 ng/L), followed by clinical parameters such as NYHA class, a history of decompensated CHF, and systolic dysfunction of the subpulmonal ventricle. Other variables had a suboptimal predictive capacity (Table 3). Figure 2 illustrates significant separation of Kaplan-Meier curves in individuals with BNP plasma levels above and below the median.

dIsCUssIoN
To the best of our knowledge, this is the largest clinical study that evaluated the prognostic value of BNP in individuals with severe systolic dysfunction of the SRV. First, BNP was found to be a powerful predictor of end-stage CHF in individuals with systolic dysfunction of the SRV. Second, NYHA functional class III or IV and a history of decompensated CHF, well-known predictors of adverse outcome in patients with SRV, and any dysfunction of the subpulmonal ventricle provided additional important prognostic information.
Systolic SRV dysfunction limits the long-term prognosis both in individuals with atrial correction of D-TGA and ccTGA who frequently die of CHF or malignant arrhythmias. Recent studies demonstrated a cumulative survival after Mustard correction of D-TGA in the range of 80% and 68% at 20 and 39 years, respectively (3,17). Another  study (18) identified CHF in approximately 20% of individuals after atrial switch due to D-TGA and in 30% of patients with ccTGA. The most important risk factors of CHF were NYHA functional class and systolic dysfunction of the SRV. Available studies demonstrated just a modest correlation of BNP and/or N-terminal proBNP with the functional status and systolic function of the SRV (8-11,19). In addition, another two reports found a correlation between worsening of systolic function of the SRV and increasing levels of N-terminal proBNP (20,21). However, just two studies were prognostic in nature (11,22). The study of Larson et al (11) was underpowered, with 21 patients with systolic dysfunction of the SRV (ejection fraction ≤45%) and 1 death during the follow-up. Not surprisingly, this study failed to detect prognostic relevance of natriuretic peptides. On the other hand, in a study of Haberer et al (22), BNP higher than 85 ng/L predicted 1 death, necessity of HTx in 3 patients, and hospitalization due to decompensated CHF in 8 patients. These patients were classified as NYHA class I and II (91%) with missing echocardiographic data. On the contrary, more than 50% of patients in our study group were in NYHA class III or IV and they had severe systolic dysfunction of the SRV (ejection fraction ≤35%).
Although most patients with systolic dysfunction of the SRV receive conventional pharmacotherapy for CHF, it should be noted that in this population beta-blockers, angiotensin converting enzyme inhibitors, and angiotensin receptor blockers showed mixed results (23)(24)(25)(26)(27). Management of adult patients with severe systolic dysfunction of the SRV is difficult, because even stable patients in NYHA functional class II may progress very quickly to end-stage CHF and require advanced non-pharmacological methods. On the other hand, congenital heart disease has been identified as an important risk factor for early mortality after HTx. However, the long-term survival of these individuals is satisfactory due to young age and few comorbidities (12,13). Early mortality may be increased due to complicated anatomic situation or sequalae after previous corrective cardiac surgery or VAD implantation. These complications include tissue adhesions, disappearance of the pericardial space, and allosensitization with resulting risk of antibody-mediated rejection. In addition, patients with congenital heart disease may spend longer time on the waiting list due to specific demands on donors and availability of experienced surgeons (28-32). Some authors suggest early enlistment of patients with complex congenital heart disease for heart transplant to prevent VAD implantation and reduce the number of reoperations (31). BNP measurement might facilitate risk stratification in patients with failing SRV and could improve the timing of non-pharmacological treatment.
There are several limitations of this study. First, the limited number of patients and retrospective study design that reflect the rare occurrence of CHF due to SRV failure may limit general applicability of the study results. A larger study, perhaps a multicenter one, might increase the strength of our findings. However, relatively high mortality and morbidity of the studied population allowed even prognostic assessment. Second, limited number of events precluded the use of multivariate models. Already two variables in one model would result in the risk of "overfitting" (33). Finally, the selected cut-off points predicting the study endpoint have to be considered specific for the studied population and may vary in different populations.
In conclusion, BNP is a powerful predictor of end-stage CHF in individuals with congenital heart disease and systolic dysfunction of the SRV. NYHA class III or IV, a history of decompensated CHF, and any systolic dysfunction of the subpulmonal ventricle provided important addition- al prognostic information. These findings might facilitate prognostic stratification in patients with CHF due to the SRV failure and could improve the timing of heart transplant enlistment or VAD implantation. declaration of authorship MH, JB, and MK conceived the study, collected and analyzed the data and wrote the manuscript. RK performed imaging studies, JF was responsible for laboratory analysis, MK and VL performed the statistical analysis. All authors contributed ideas for discussion, drafted and reviewed the manuscript, and approved the final version.
Competing interests: All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.