Surgical outcomes analysis in patients with uncomplicated acute type A aortic dissection: a 13-year institutional experience

This retrospective study aimed to clarify the short-term and mid-term outcomes of and prognostic factors for patients who underwent surgical repair for uncomplicated acute type A aortic dissection (ATAAD). Between January 2007 and June 2019, 603 consecutive patients underwent ATAAD repair at our institution. According to patients’ preoperative presentations and imaging studies, uncomplicated ATAAD was found in 276 (45.8%) patients by excluding preoperative complicated factors. Patients with uncomplicated ATAAD were classified into the survivor (n = 243) and non-survivor (n = 33) groups. Clinical features, surgical information, and postoperative complications were compared. Three-year survival and freedom from reoperation rates for survivors were analyzed using the Kaplan–Meier actuarial method. The in-hospital surgical mortality rate of uncomplicated ATAAD patients was 11.9%. The non-survivor group had a higher rate of postoperative malperfusion-related complications, and a multivariate analysis revealed that repeat surgery, retrograde cerebral perfusion, and intraoperative extracorporeal membrane oxygenation support were predictors of in-hospital mortality. In the survivor group, 3-year cumulative survival and freedom from aortic reoperation rates were 89.6% (95% confidence interval [CI] 84.8–92.9%) and 83.1% (95% CI 76.8–87.7%), respectively. In conclusion, uncomplicated and complicated ATAAD rates were similar; the short-term and mid-term surgical outcomes in patients with uncomplicated ATAAD were generally acceptable.

www.nature.com/scientificreports/ octogenarian population with a small sample size. In the present study, by excluding most of the preoperative complicated factors to obtain a study group of ATAAD patients with the lowest surgical risk, we generated a retrospective analysis of the experiences of an individual aortic surgery centre and aimed to clarify the short-term and mid-term outcomes of and prognostic factors for patients who underwent surgical repair for uncomplicated ATAAD.

Methods
Definitions. After referencing the definitions in previous literature [1][2][3][4] , complicated ATAAD was defined as presenting with preoperative shock status, organ malperfusion, haemopericardium, ventilator support, requirement for CPR, or a combination of these conditions before undergoing ATAAD repair surgery. Shock was defined as systolic blood pressure (SBP) < 90 mmHg or requirement for inotropic medication. Organ malperfusion syndrome was defined according to evidence of a lack of blood flow to the defined organ system on imaging studies in accordance with clinical symptoms regarding the involved organ. Organ malperfusion syndromes included cerebral, spinal, myocardial, limb, renal, and mesenteric malperfusions 8 . Without symptoms of malperfusion, the independent radiographic evidence of dissection flap in the arterial branch vessels was not considered as malperfusion syndrome. In contrast, patients who did not present any features associated with complicated ATAAD were classified as having uncomplicated ATAAD. patient enrollment and preoperative management. The study protocol was approved by the Chang-Gung Medical Foundation Institutional Review Board (No.202000118B0), and all methods were performed in accordance with the relevant guidelines and regulations. With approval from the Chang-Gung Medical Foundation Institutional Review Board, the need for informed consent was waived due to the retrospective nature of the study. Overall, 603 consecutive adult patients underwent emergency ATAAD repair at our institution between January 2007 and June 2019. All patients were diagnosed via helical computed tomography to confirm ATAAD in the emergency department, and the extent of aortic dissection, presence of organ malperfusion, and haemopericardium were analyzed. When the diagnosis of ATAAD was confirmed, patients were emergently transferred to the operating room within 30 min without delay, irrespective of whether the preoperative condition was complicated or uncomplicated. If patients presented with hypertension or tachycardia before surgery, then their hemodynamics were stabilised with intravenous beta-blockers to maintain SBP < 120 mmHg and a heart rate of 60-70 bpm, according to the 2010 American College of Cardiology/American Heart Association guidelines for thoracic aortic disease 9 . When patients presented with shock, medical resuscitation was applied first, including intravenous fluid supplementation and inotropic infusion. Rescue procedures, including pericardiocentesis and subxiphoid pericardiotomy, were performed if unstable hemodynamics persisted and cardiac tamponade was found by on-site echocardiography. For refractory hemodynamic instability or cardiac arrest, CPR was performed. All management strategies were performed on an emergency basis. The overall number of patients per year who underwent ATAAD repair surgery and the number of that with uncomplicated type per year during the study period are illustrated in Fig. 1. As illustrated in Fig. 2, 327 patients www.nature.com/scientificreports/ with complicated ATAAD were excluded, including 150 with shock, 20 with CPR, 33 with preoperative ventilator support, 91 with organ malperfusion, and 196 with haemopericardium. Given that a single patient can have more than one etiologies of complicated ATAAD, the sum of all these categories was not equal to 327. A total of 276 patients who underwent aortic repair surgery for uncomplicated ATAAD were included. The included patients were dichotomised into the following groups according to whether they survived to discharge: survivor (n = 243) group and non-survivor (n = 33) group.
AtAAD repair procedures and postoperative care. The general principles of aortic repair procedures are detailed in previous studies reported by this institute 10,11 . For uncomplicated ATAAD patients who were considered relatively stable, double arterial cannulation with antegrade cerebral perfusion (ACP) strategy was usually implemented. The right axillary and femoral arteries were cannulated with an 8-mm ring-reinforced polytetrafluoroethylene graft and connected with a Y-shape circuit. Following sternotomy, the right atrium was cannulated and cardiopulmonary bypass (CPB) with deep hypothermia was initiated. In general, the dissected aorta was replaced with a Dacron prosthetic graft based on the location of the entry tear and preoperative presentation. The proximal anastomosis was usually performed first, followed by open distal anastomosis under circulatory arrest. During circulatory arrest, the femoral arterial flow was temporarily suspended and selective ACP through the right axillary artery was used. Concomitant aortic root replacement with a composite Valsalva graft and frozen elephant trunk procedure with a covered stent graft were performed if the extent of aortic dissection involved the aortic root and descending thoracic aorta, respectively. After undergoing surgical repair for ATAAD, all patients were transferred to a specialized cardiovascular intensive care unit (ICU) for further treatment and observation. Without unstable haemodynamics, persistent arrhythmia, signs of organ malperfusion, or active bleeding, a ventilator-weaning protocol was initiated at 12-24 h post surgery. Renal replacement therapy was applied according to the Acute Kidney Injury Network criteria if acute renal failure developed after surgery 12 .
Statistical analyses. Statistical analyses were performed using SPSS for Windows (version 22.0; IBM Corp., Armonk, NY, USA) and SAS for Windows (version 9.4; SAS Institute Inc., Cary, North Carolina, USA). Data were presented as means ± standard deviation for continuous variables and as numbers (n) and percentages (%) for categorical variables. For comparing the intergroup disparities between the survivor and non-survivor groups, we used univariate tests, including the independent t test for continuous variables and the chi-square or Fisher's exact test for categorical variables, respectively. The multivariate logistic regression analysis was used to identify the independent predictors of in-hospital mortality. The variables for multivariate logistic regression analysis were first selected from those with a P < 0.05 in univariate tests mentioned above and re-tested by the univariate logistic regression method. Continuous variables were dichotomised based on cut-off values determined by receiver operating characteristics curve analyses before conducting to the logistic regression analyses. Given that the number of events (in-hospital mortality) per covariate was ≤ 10 in some variables, we used penalization through data augmentation to perform multivariate logistic regression to avoid sparse data bias 13,14 . The log-F(m,m) penalized-likelihood method with the shrinkage parameter (m) was implemented by fitting a standard logistic regression to a dataset augmented with pseudo-individuals per covariate. The family of log-F priors will shrink the biased estimate of the regression coefficient in the logistic regression towards the prior odds ratio. In this study, we followed the suggestion by Greenland 15 to set the prior odds ratio being 1 www.nature.com/scientificreports/ (uncertain direction) for brain stroke; 2 (probably positive) for repeat surgery, CPB time, aortic clamping time, circulatory arrest time, hypothermia temperature, renal failure; and being 4 (probably strong) for retrograde cerebral perfusion, extracorporeal membrane oxygenation (ECMO) support. In contrast, penalization was not performed in covariates with a number of events > 10, including axillary arterial cannulation, isolated ascending aorta replacement, aortic arch replacement, total arch replacement, ACP, and malperfusion-related complications. The Kaplan-Meier method was used to estimate the 3-year cumulative survival and freedom from aortic reoperation rates for the survivor group. In addition, the independent t test, Chi-square and Fisher's exact test were also used to compare the differences between patients with complicated and uncomplicated ATAAD. For all analyses, statistical significance was set at P < 0.05. Table 1, the clinical demographics, comorbidities, preoperative conditions, and clinical presentation were generally homogenous between the survivor and non-survivor groups, except for the rate of repeat surgery with previous cardiac surgery. In the non-survivor group, the prevalence of repeat surgery was higher than that of the survivor group (2.9% vs. 18.2%; P < 0.001). Overall, 20.7% of patients were female, and the mean age was 52.9 ± 12.0 years. Hypertension was a prevalent comorbidity that was observed in approximately 70% of patients in both groups. Chest or back pain was the most common clinical presentation, accounting for > 80% of patients in both groups. A total of 16.7% of the patients were diagnosed with intramural haematoma rather than typical aortic dissection. Table 2 provides detailed information regarding surgical variables. Femoral arterial cannulation was a common vascular access used for > 90% of patients in both groups, and the rates of additional axillary arterial cannulation and ACP strategy applied during circulatory arrest were lower in the non-survivor group. A higher rate of aortic arch replacement was found in the non-survivor group than in the survivor group. However, there was no inter-group disparity regarding entry tear exclusion. The time spans of CPB, aortic crossclamping, and circulatory arrest were generally longer in the non-survivor group. A total of 9.8% of the patients required Kerlix packing due to coagulopathy with uncontrolled bleeding. A higher rate of ECMO implementation due to intraoperative myocardial failure was found in the non-survivor group.

Surgical information.
postoperative complications. The overall in-hospital mortality rate was 11.9%. The annual in-hospital mortality rates in overall, complicated, and uncomplicated ATAAD patients during the study period are illustrated in Fig. 3. In general, patients with uncomplicated ATAAD showed lower annual in-hospital mortality rates compared to that in the overall and complicated populations. By dividing the study group into the first (2007-2010, n = 62) and second periods (2011-2019, n = 214), the in-hospital mortality rates in two periods were 22.6% (14/62) and 8.9% (19/214), respectively. As illustrated in Table 3, myocardial failure was the most common cause of mortality (39.4%), followed by brain stem failure (24.2%), sepsis (21.2%), and bleeding (15.2%). The blood transfusion volumes were generally similar in both groups. In the non-survivor group, malperfusionrelated complication rates including postoperative renal failure and brain stroke were significantly higher com- Prognostic factors associated with in-hospital mortality. Table 4 shows the logistic regression analysis results for patients who underwent surgical repair for uncomplicated ATAAD. Axillary arterial cannulation and ACP were highly correlated in the aspect of surgical technique (100% of axillary arterial cannulation underwent ACP; 98.3% of ACP was implemented with axillary arterial cannulation). Furthermore, axillary arterial cannulation and ACP were the exact opposite of retrograde cerebral perfusion. To avoid the problem of col-  www.nature.com/scientificreports/ linearity among these three variables, axillary arterial cannulation and ACP were not included in the multivariate logistic regression analysis. The following three significant prognostic factors for in-hospital mortality were identified: repeat surgery (adjusted odds ratio (aOR) 3.01, 95% confidence interval (CI) 1.12-8.09, P = 0.029), retrograde cerebral perfusion (aOR 6.07, 95% CI 2.55-14.47, P < 0.001), and ECMO support post surgery (aOR 6.31, 95% CI 1.97-20.19, P = 0.002).

Cumulative 3-year survival and freedom from reoperation rates.
Follow-up with an average of 4.9 ± 3.5 years (median 4.4, range 0.1-12.7 years) was completed for all patients. As illustrated in Fig. 4

Comparisons between patients with complicated and uncomplicated ATAAD. As illustrated in
Supplementary Table S1, patients with uncomplicated ATAAD were younger with less female compared to the complicated patients. Besides, better preoperative renal function and higher SBP were found in the uncomplicated group than that in the complicated group. In regard to the clinical presentation, the uncomplicated group had a higher rate of typical chest/back pain and a lower incidence of DeBakey type II aortic dissection than the complicated group. Supplementary Table S2 illustrated the surgical information in complicated and uncomplicated ATAAD groups. An higher rate of ACP strategy with a more aggressive aortic arch replacement were found in the uncomplicated group compared with that in the complicated group. As for the postoperative outcomes (Supplementary Table S3), the uncomplicated group showed a lower in-hospital mortality rate (11.9% vs. 20.2%; P = 0.007) and fewer blood transfusions than the complicated group. Finally, the subgroup analyses according to the etiologies of complicated ATAAD were illustrated in Supplementary Fig. S4. Patients that were complicated with preoperative shock, CPR and ventilator support showed higher in-hospital mortality rates compared to the uncomplicated ATAAD patients. In addition, patients who had preoperative organ malperfusion and haemopericardium also showed trends of higher in-hospital mortality rates, although the statistical significances were not reached. www.nature.com/scientificreports/

Discussion
In the previous literature, ATAAD with preoperative complications was associated with disastrous outcomes 1-4 . However, the characteristics and surgical results of uncomplicated ATAAD remain to be clarified. Furthermore, the definition and prevalence of uncomplicated ATAAD varied in deferent studies 3,7,16 . In this single-centre study, we excluded most of the preoperative complicated factors that resulted in an inferior outcome, including shock status, organ malperfusion, haemopericardium, ventilator support, and CPR, to obtain a study group of ATAAD patients with the lowest surgical risk. For 276 patients who underwent aortic repair surgery for uncomplicated ATAAD, we found a reasonable in-hospital mortality rate and acceptable mid-term outcomes during the 3-year follow-up. However, it should be possible to optimize the results by improving the surgical strategy and perioperative management. An expected lower in-hospital mortality rate was observed for the uncomplicated ATAAD patients of this study compared to that reported by previous studies performed at this institute 10,11 , which ranged from 15.5 to 16.0% for the general ATAAD population. Furthermore, as reported by Czemy et al., a large retrospective study in regard to this topic conducted by a national data registry (German Registry for Acute Aortic Dissection Type A database) found that the in-hospital mortality rate in uncomplicated ATAAD patients was 12.6% 3 . A similar outcome was observed in this study. However, the postoperative malperfusion-related complication rates, such as acute renal failure and brain stroke, remained considerable in the present study. We suspected several reasons for this finding. First, malperfusion observed in aortic dissection is correlated to its complex anatomic interactions  www.nature.com/scientificreports/ between the true lumen and the false lumen along the entire dissected aorta. This anatomic interaction can be dynamic during aortic repair surgery. A dissected branch vessel without preoperative malperfusion symptoms may not certainly maintain adequate blood flow to the perfused organ during CPB and circulatory arrest, which render a different blood circulation compared to the normal physiologic pattern. Second, the overall entry tear exclusion rate was only 74.6% in the present study. For patients with residual aortic entry tear, the possibility of true lumen compromise and blood flow limitation may be increased. As reported by Inoue et al., aggressive primary entry tear resection could contribute to the satisfactory short-term and long-term outcomes 17 . We believed that this principle should be applied more aggressively among uncomplicated ATAAD patients because their preoperative condition is relatively stable and they may have better ability to endure a complex surgery. Furthermore, patients with branch vessel dissection but without resection of primary entry tear should be managed with a more aggressive strategy for detecting and treating postoperative malperfusion-related complications early. Finally, approximately 30% of patients underwent aortic arch replacement in this study, and only a total of 3.6% (10/276) of patients underwent bilateral ACP during circulatory arrest. Furthermore, the average circulatory arrest times were > 50 and > 70 min for overall patients and those in the non-survivor group, respectively. The clinical benefits of using bilateral ACP for ATAAD surgery remain controversial [18][19][20] . However, as reported by Preventza and Angleitner et al., bilateral ACP was associated with superior outcomes for patients requiring prolonged circulatory arrest time > 30 or 50 min 19,20 . Therefore, this modality may also be applied more aggressively to reduce cerebrovascular complications among uncomplicated ATAAD patients included in this study.
In the present study, repeat surgery after previous cardiac surgery was an independent risk factor for in-hospital mortality. Furthermore, the prevalence of repeat surgery in the non-survivor group was six times greater than that in the survivor group. In the previous literatures, repeat cardiac surgery for valvular disease and coronary artery bypass grafting (CABG) could have results comparable to those of first-time surgery performed at experienced centres 21,22 . However, as reported by Rylski and Estrera et al., open surgical repair of ATAAD after previous cardiac surgery was associated with higher in-hospital mortality and brain stroke rates and a more frequent need for postoperative cardiac support compared to first-time surgery 23,24 . Therefore, in the context of ATAAD, it is widely accepted that previous cardiac surgery represents a significant risk factor for poorer outcomes, similar to the results of the present study. We suspect several reasons for this finding. First, for ATAAD patients, all management strategies were performed on an emergent basis, and preoperative evaluation of coronary artery disease, which is essential for patients with previous CABG surgery or scheduled to undergo elective cardiac surgery, was usually inadequate. Therefore, it may increase the risk of insufficient myocardial protection during the redo ATAAD surgery, which usually comprises prolonged CPB and cardiac arrest times. Furthermore, with a less invasive incision compared to resternotomy, left and right thoracotomies could be safe approaches and provide substantial benefits by reducing CPB time and postoperative bleeding for patients undergoing redo CABG and valvular surgeries 25,26 . However, these approaches were usually not considered as a preferable choice for treating ATAAD due to the unsatisfactory exposure of the aortic arch and aortic root. In the present study, all patients underwent aortic repair surgery via sternotomy. Although endovascular stent graft technology is a developing modality that offers an alternative treatment option for selected high-risk patients with ascending aortic disease, including ATAAD 27,28 , it is not feasible for every patient because of the individual differences in the vascular anatomy affected by aortic dissection. Furthermore, the evidence of its long-term durability is lacking. Conventional open repair remains the standard treatment for ATAAD. Finally, to prevent major bleeding from the ruptured aorta when performing resternotomy and dividing the mediastinal adhesions, prior CPB was usually established by peripheral cannulation. However, it could cause significantly prolonged CPB time and increase relative complications.

Limitations of this study
This study has several limitations. First, as a retrospective and non-randomized control study, bias that could influence the homogeneity of the survivor and non-survivor groups might have existed. Second, because this crossed cohort spans a period of > 12 years, the technology of CPB and myocardial protection, as well as cerebral protection strategies and ICU care protocols for treating ATAAD may have changed over time. Finally, despite the convincing short-term and mid-term results of the present study, an extended follow-up study should be conducted in the future to evaluate long-term outcome in uncomplicated ATAAD population.

conclusions
The prevalence of uncomplicated ATAAD was similar to complicated ATAAD, and its short-term and midterm surgical outcomes were generally acceptable. Even though patients with uncomplicated ATAAD comprise a relatively stable population, accurate surgical strategies for preventing postoperative malperfusion remain crucial to optimize the outcomes.

Data availability
As containing identifying information, including individual patient's age, gender, the specific date and details of hospital admission/surgical procedures, the data that we collected cannot be made publicly available for ethical and legal reasons. It is the requirement of the Chang-Gung medical foundation institutional review board to review any request to share data publicly in order to protect patients' privacy. Requests for data can be sent to the corresponding author B9002078@cgmh.org.tw and the Chang-Gung medical foundation institutional review board at irb1@cgmh.org.tw. www.nature.com/scientificreports/