Characteristics and outcomes of Stanford type A aortic dissection patients with post-operation severe hyperbilirubinemia: a retrospective cohort study

Hyperbilirubinemia is one of the common complications after cardiac surgery and is associated with increased mortality. However, to the best of our knowledge, the report on clinical signicance of postoperative severe hyperbilirubinemia in Stanford type A aortic dissection (AAD) patients is limited. Patients who underwent surgical treatment for AAD in our center between January 2015 and December 2018 were retrospectively screened. In-hospital mortality, long-term mortality, acute kidney injury (AKI), and the requirement of continuous renal replacement therapy (CRRT) were assessed as endpoints. Univariate and multivariate regression models were employed to identify the risk factors of these endpoints.


Introduction
Aortic dissection (AD) is an acute life-threatening condition with a prevalence of about 3/100,000 per year. The International Registry of AD revealed that 67% of AD patients presented with Stanford type A aortic dissection (AAD), which was characteristic as the involvement of the ascending aorta. And, approximately 86% of AAD patients required swift open cardiac surgery to avoid fatal complications such as aortic rupture and cardiac tamponade. [1][2][3] In spite of the improvement in medical management and surgical technique, AAD surgery was associated with early mortality as high as 20%. [4][5][6][7][8] Hyperbilirubinemia is a common severe complication after cardiac operation. [9] The reported incidence of post-operation hyperbilirubinemia varied widely (10-57%) and related to the severity of cardiac diseases and the type of cardiac surgery.
Although hyperbilirubinemia was reported to be associated with in-hospital mortality and mortality after discharge, the effects of hyperbilirubinemia on patient prognosis were heterogeneous. [10][11][12][13][14] For patients undergoing cardiac surgery with cardiopulmonary bypass (CPB), recent studies suggested that severe hyperbilirubinemia (5 times the normal upper limit) instead of mild bilirubin signi cantly increased patient mortality. [15] Additionally, the occurrence of acute kidney injury (AKI) and the acceptance of continuous renal replacement therapy (CRRT) after cardiac surgery were suggested to be associated with the decrease of patient survival proportion. As we know, the pathogenesis, disease severity classi cation, prognosis, and operation method are different between different cardiac diseases. The inclusion of all kinds of cardiac surgery in one study most likely would reduce the speci city and repeatability of the conclusions. Up to now, the reports on the characteristics and outcomes of AAD surgery patients who had post-operation severe hyperbilirubinemia are limited.
Therefore, the purpose of our present study is to describe the clinical characteristics and to investigate outcomes of AAD surgery patients who had post-operation severe hyperbilirubinemia. The results of our present study might be useful for decision makers regarding their treatment strategy and for patients and their families regarding the expected progression.

Study design and patients selection
Our present study was retrospectively designed. Consecutive patients who underwent surgery for AD in our center between January 2015 and December 2018 were screened. AD was proven by enhanced computed tomography and de ned as type A or type B according to the Stanford classi cation. Postoperative severe hyperbilirubinemia was de ned as occurrence of a serum TB concentration of more than 85.5 µmol/l (5 times the normal upper limit) in any measurement during the hospital staying after AAD surgery. Patients were excluded if they had any of the following conditions: (1) serum TB concentration < 85.5 µmol/l, (2) age < 18 years; (3) Stanford type B aortic dissection; (4) the occurrence of severe hyperbilirubinemia before surgery; (5) severe hyperbilirubinemia caused by the reoperation during hospitalization. The local Institutional Review Board of the hospital approved this retrospective study and waived the requirement of informed consent for the use of patients' medical data.

Surgical procedure
The use of surgical techniques has developed during the study period. However, the basic principles for repair of AAD have no change, including (1) prompt establishment of CPB, (2) resection of the primary entry site by open distal anastomosis under deep hypothermic circulatory arrest, (3) preservation of the aortic valve whenever possible, and (4) aortic arch replacement in patients with an entry site located in or extending into the aortic arch. Aortic root replacement with a composite prosthesis and reimplantation of the coronary arteries by the modi ed Bentall technique was performed in patients with conspicuous dilatation of the aortic root.

Data collection
Demographic data, comorbidities, and operation details were retrieved from our hospital's electronic medical record system. All the routine laboratory data was recorded before operation (the nearest to the time of surgery) and in the postoperation period. The severity of illness before and after surgery was assessed by using the acute physiology and chronic health evaluation II (APACHE II), sequential organ failure assessment (SOFA) score, and model for end-stage liver disease (MELD) score. Urine output was recorded every day after the surgery.

Outcomes and de nition
Post-operation outcomes including the amount of blood transfusion, mechanical ventilation time, the use of extracorporeal membrane oxygenation (ECMO), intra-aortic balloon pump (IABP), AKI, the usage of CRRT, bilirubin adsorption (BA) or plasma exchange (PE), and vasoactive agent, the duration of hospitalization, ICU stay time, and inhospital mortality. For those patients who were alive on discharge, telephone survey was performed to obtain the patients long-term outcome.
Kidney Disease Improving Global Outcomes (KDIGO) criteria [16] based on SCr or urine output was employed to diagnose and grade AKI. The latest SCr concentration before surgery was de ned as preoperative SCr concentration. The decision to start CRRT was made at the discretion of the attending nephrologist. Main indications for starting CRRT were progressive AKI, uid overload, hyperkalemia, and severe metabolic acidosis. [16] 2.5. Statistical analysis Data were analyzed using SPSS version 22.0 software (SPSS, Inc, Chicago, IL, USA). Continuous variables are presented as mean ± standard deviation. Categorical variables are presented as frequencies with percentages. To evaluate the differences between groups, the independent sample t-test was used for continuous variables, whereas the chi-square test or Fisher's exact test was used for categorical variables. Factors signi cantly associated with these endpoints in univariate analysis were included in the multivariate logistic regression analysis or Cox proportional hazard analysis to identify the independent risk factors. Accumulated survival proportion was estimated with the Kaplan-Meier method, and the between-group differences of survival proportion were assessed using the log-rank test. Area under the receiver operating characteristic curve (AUC-ROC) was calculated to assess peak TB concentration on the ability to detect inhospital mortality. Youden index was used for assessment of optimal cut-off values. For all analyses, all statistical tests were 2-sided, and a P-value < 0.05 was considered as statistically signi cant.

Patient characteristics
Of the 2210 screened patients, 1931, 2, 3, and 3 were excluded because of TB concentration < 85.5umol/l, Stanford type B aortic dissection, preoperative TB ≥ 85.5 µmol/l, and post-operation severe hyperbilirubinemia caused by reoperation during this hospitalization, respectively. Ultimately, 271 AAD surgery patients with post-operation severe hyperbilirubinemia were included in our present study (Fig. 1).
The baseline characteristics of the included patients were summarized in Table 1. There were 223 male and 48 female, and the mean age of the patients was 49.1 ± 11.0 years. AAD surgery involved the aortic valve in 83.8% patients, aortic arch in 86% patients, and coronary artery in 17% patients. Preoperative TB concentration was 25.0 ± 15.0 µmol/l. The mean onset time of severe hyperbilirubinemia was 2.8 ± 1.3 days after AAD surgery. The mean peak serum TB concentration was 150.9 ± 93.0 µmol/l and the averaged time to peak TB concentration was 3.8 ± 3.0 days after AAD surgery. The change of TB concentration during the seven days after surgery was showed in Fig. 2. The median follow-up time was 18.63 (0.5-55.9) months. And, 3 patients were lost to follow-up. The time of lost to follow-up was 16-, 95-, and 365-day after the surgery, respectively.

Postoperative AKI
Of the 271 included patients, 222 patients (82.1%) had AKI after AAD surgery, of which 102 (40.2%), 34 (12.5%), and 84 (29.2%) were stage 1, stage 2, and stage 3 AKI (Table 2), respectively. The results of univariate and multivariate logistic regression are presented in Table 3. In the univariate analysis, age, hypertension, preoperative SCr concentration, APACHE score, and MELD score, operation time, and CPB time were associated with the occurrence of postoperative AKI.

Postoperative CRRT
Of the included patients, 50 (18.5%) patients received CRRT after AAD surgery. Univariate analysis indicated that male, preoperative SCr concentration, APECHE score, MELD score, and aortic cross-clamp (ACC) time were associated with the need for CRRT. In the multivariate logistic regression analysis, only preoperative SCr concentration (OR 1.011, 95%CI 1.002-1016, P = 0.003) was identi ed as an independent predictor of the acceptance of CRRT (Table 3). ROC analysis (Fig. 3) identi ed that peak TB concentration was associated with increased mortality, and the optimal cutoff value identi ed by the Youden index was 121.2 µmol/l (sensitivity: 72%, speci city: 60%). And, the multivariate analysis revealed that patients with peak TB concentration ≥ 121.2 µmol/l had a signi cantly higher risk of in-hospital mortality (OR = 2.681, 95%CI 1.119-6.425, P = 0.027, Table 4, model 2), compared with the patients with peak TB concentration < 121.2 µmol/l.

Long-term mortality
Of the 12 patients who died after discharge, 4 had non-AKI, 3 had stage 1 AKI, 3 had stage 2 AKI, and 2 had stage 3 AKI during their hospital stay, respectively. The accumulated 1-year, 2-year, and 3-year mortality proportions were 32.9%, 33.9%, and 35.3%, respectively (Fig. 4A). The risk factors of long-term mortality were presented in Table 5. Univariate analysis revealed 20 factors, including the post-operation AKI, the acceptance of CRRT and ECMO, and the peak TB concentration ≥ 121.2 µmol/l ( Fig. 4B-E), were signi cantly related to patient long-term mortality. Multivariate COX regression analysis revealed that stage 3 AKI (HR 12.604, 95%CI 5.002-31.762, P < 0.001) signi cantly increased long-term mortality, compared with patients without AKI. The use of ECMO (HR 12.167, 95%CI 4.588-32.264, P < 0.001) was identi ed as an independent predictor of long-term mortality as well. In contrast, postoperative MAP (HR 0.979, 95%CI 0.962-0.995, P = 0.012) was identi ed as an independent protective factor of long-term mortality.

Discussion
It has been reported that severe hyperbilirubinemia was associated with increased mortality in patients underwent cardiac surgery. [11,15,17] Up to now, limited information was available on the risk factors associated with in-hospital and longterm mortality in AAD patients with post-operation severe hyperbilirubinemia. Our presented study had several ndings.
Firstly, the occurrence of AKI, the requirement for CRRT, and the in-hospital mortality were higher than previous studies of patients undergoing AAD surgery without severe hyperbilirubinemia. Secondly, age, preoperative SCr concentration, and CPB time were independent risk factors for postoperative AKI, and preoperative SCr concentration was an independent risk factor for post-operation CRRT as well. Finally, the peak TB concentration, post-operation stage 3 AKI, the total amount of blood transfusion after AAD surgery, the use of ECMO, and low MAP after surgery were signi cantly associated with mortality.

AAD patients with severe hyperbilirubinemia were associated with worse prognosis
The analysis of our present cohort of AAD patients with post-operation severe hyperbilirubinemia showed an overall incidence of post-operation AKI of 81.9%, a requirement for CRRT of 18.5%, and the in-hospital mortality rate of 30.3%. In previous studies of AAD patients underwent surgical treatment without post-operation severe hyperbilirubinemia, the reported AKI incidences were ranged from 40-78%, [5][6][7][18][19][20][21][22][23] the reported incidences of requirement for CRRT were ranged from 3% to 8%, [5,7] and the reported in-hospital mortalities were ranged from 15-26%. [4,5,7,19,24] The discrepancy most likely attributed to the fact that all of the included AAD patients in our study developed severe hyperbilirubinemia. The development of severe hyperbilirubinemia after AAD surgery related to the severity of the AAD and the severity of injury during AAD operation. In animal model, hyperbilirubinemia was proved to have pro-apoptotic effects and aggravate renal ischemia-reperfusion injury. [25] Additionally, high concentration of bilirubin could lead to in ammatory response and cell apoptosis of the brain, [26] which might be one of the potential mechanism of the high mortality of our present cohort.

Risk factors of postoperative AKI and CRRT
In our present cohort, older age, high preoperative SCr concentration, and pronged CPB time were identi ed as independent risk factors for AKI. Meanwhile, preoperative SCr concentration was also associated with the requirement of CRRT after AAD surgery. A recent meta-analysis of patients underwent AAD surgery [18] showed that older age was identi ed as an independent risk factor of AKI as well. Older age-mediated adverse renal structural and functional changes might contribute to the high development of postoperative AKI after AAD surgery in older patients. The aforementioned metaanalysis showed that preoperative SCr concentration did not correlate with postoperative AKI with signi cant heterogeneity (I 2 = 72.8%). In our opinion, pre-operation elevated SCr concentration might indicate structural kidney damage or hemodynamic derangements in AAD patients, which further aggravated the development of post-operation AKI and the requirement of CRRT. Longer CPB time can lead to more hemolysis, and a longer time on the circuit can lead to changes in perfusion to the viscera and to more in ammatory action. A CPB itself might also induce hypoperfusion of abdominal organs, hypoxia or an in ammatory reaction, which causes liver and kidney damage. Therefore, surgeons most likely could reduce the AKI risk of AAD patients by the improvement of their operation strategies and the reduction of CPB time.

Risk factors of in-hospital and long-term mortality
In previous studies, older age had been identi ed as an independent risk factor of mortality for patients with hyperbilirubinemia after cardiac surgery. [27] In our present study, older age was established a risk factor for in-hospital mortality for patients with hyperbilirubinemia after AAD surgery as well. Nevertheless, aging indicated diminished functional capacity of the liver and added to the cumulative burden in the case of developed severe hyperbilirubinemia. Additionally, different from the a recent study outlining the increased risk of mortality for patients with hyperbilirubinemia after cardiac surgery associated with time to peak bilirubin, [15] our study indicated that peak TB concentration was an independent driver of in-hospital mortality for patients with severe hyperbilirubinemia after AAD surgery. And, we found out that the optimal cut-off value of peak TB on predicting in-hospital mortality was 121.2 µmol/l. This difference might be attributed to the existence of severe hyperbilirubinemia after AAD surgery and the limitation of small sample size. The high concentration of bilirubin leads to in ammatory response or cell apoptosis in the brain. [26] Cui et al. [28] showed that bilirubin induced lung edema and injury by inducing the apoptosis of alveolar epithelial cells. From the above, severe hyperbilirubinemia induces oxidative stress and apoptosis in many organs, which is associated with poor outcomes in patients with AAD surgery.
In a study for patients with jaundice after open heart surgery, Chu et al. [29] reported patients with more severe hyperbilirubinemia or delayed serum peak TB concentration might be signi cantly associated with more transfused blood and hypotension. Increased amount of blood transfusions contributed to hemolysis, and hypotension would reduce hepatic perfusion [30], which both caused an increased bilirubin load. This might be one of the potential mechanisms that the number of transfused blood units and lower MAP increased mortality in our present study. Previous studies suggested that mechanisms underlying the development of post-operation late peak TB concentration differed from those of the early peak bilirubin concentration. The immediate development of post-operation peak TB concentration and rapid decline thereafter re ected the transient damaging effects by CPB surgery, whereas late development of post-operation peak TB concentration was a consequence of hepatic dysfunction caused by persistent cardiac failure or sepsis. [9,12] Therefore, attention should be paid to the monitoring of heart failure and optimizing hemodynamics after AAD surgery to prevent further deterioration. Furthermore, identifying and implementing effective risk reduction strategies is needed. Molecular adsorbent recirculation system, prometheus therapy, fractionated plasma separation and adsorption have shown promise in reducing bilirubin concentration, [27,31] but further study in this area is needed.
AKI had been incorporated into a risk tool to predict early mortality for patients underwent AAD surgery. [18] Our results added to these ndings and demonstrated stage 3 AKI markedly increased both in-hospital and long-term mortality.
Additionally, our study also showed in-hospital and long-term mortality was independently predicted by the use of ECMO.
Hemolysis is a common complication in ECMO support. [32] Pump head or oxygenator thrombosis and excessive pump speed (typically greater than 3000 rpm) which may destroy blood red cells, [33] which in turn is important factor in the development of hyperbilirubinemia. Therefore, more careful postoperative management is needed to improve prognosis.

Study limitation
Our present study was a retrospective clinical research from a single institution and had some limitations. First, SCr concentration on admission was regard as preoperative renal function. However, some patients might already have AKI on admission. As a result, the number of patients with AKI may be underestimated. Second, the number of major adverse events including the use of ECMO, plasma exchange and bilirubin adsorption was small, which will be likely to reduce the statistical power for risk factor analysis. Finally, the renal prognosis was not regular followed up after the hospital discharge, which is important for the evaluation of the renal outcome. Therefore, further prospective multicenter studies with larger samples are needed to obtain stronger evidences.

Availability of data and materials
The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.

Contributions
Chen, Bai, Zhao, and Sun analyzed and interpreted the data. Li, Yu, Zhang, Ma and Chen performed the statistical analyses. All authors contributed to the study design and interpretation of the study results. All authors read and approved the nal manuscript.

Ethics declarations
Ethics approval and consent to participate The local Institutional Review Board of the hospital approved this retrospective study and waived the requirement of informed consent for the use of patients' medical data, and all experimental methods were performed in accordance with the relevant guidelines and regulations.

Consent for publication
Not applicable.
Competing interests  The averaged serum total bilirubin concentration during the postoperative seven days Long-term survival results of (A) all patients, (B) patients without AKI and those with stage 1, 2, or 3 AKI, (C) patients without the use of CRRT and those with CRRT, (D) patients without the use of ECMO and those with ECMO, (E) patients with post-operation peak TB ≥ 121.2 μmol/l and those with post-operation peak TB < 121.2 μmol/l.