Recent evolution in use and effectiveness in mainland China of thoracic endovascular aortic repair of type B aortic dissection

A meta-analysis was performed on 175 studies selected among those published in mainland China between 2008 and 2015 on thoracic endovascular aortic repair (TEVAR) for type B aortic dissection (AD). Most TEVAR were performed in Shanghai, Beijing, Hubei and Guangdong in patients with mean age of 53.9 years, and acute (70%) or chronic (30%) type B AD. Procedural success rate was 99.1 ± 0.8%. Major complication rate was 1.7 ± 2.3%, with paraplegia in 0.4 ± 0.0%. Overall in-hospital mortality rate was 1.6 ± 0.9% with AD rupture in 30% (about 40% during first postoperative day); follow-up mortality rate was 2.3 ± 1.1%, with AD rupture in 39.2% (50% within first year). Compared with 2001–2007 data from China, there appeared to be improvement in rates of major complications, paraplegia and in-hospital mortality. Compared with 1999–2004 Western data, rates of procedural success, stroke, and paraplegia appeared similar, while those for major complications, in-hospital mortality, retrograde type A dissection and follow-up mortality appeared lower. Compared with more recent Western data (2006–2013) on acute complicated type B AD, stroke, paraplegia, in-hospital mortality and follow-up mortality appeared lower. Therefore, in mainland China, safety for TEVAR of type B AD appeared better between 2008 and 2015 than in previous periods in China or Western countries.

Data availability. All relevant data are within the paper.

Definitions
Type B AD was classified based on the Stanford classification 4 . Dissection was considered as an acute event if occurring within the first 14 days after symptom onset, and chronic if beyond 14 days. Complications were classified as either major when life-threatening or prompting major therapeutic interventions (e.g., access complications requiring surgical revision), or minor when not requiring further treatment (e.g., transient renal failure not requiring dialysis). Stent-graft implantation was based on indications described by Nienaber and colleagues 5 , although some, such as distance between landing zone not <1.5 cm, maximal aortic diameter >5.5 cm, and no bilateral iliac tortuosity and stenosis, were not followed because of more positive attitude about TEVAR 6,7 , accumulating experience and device improvement 8,9 . For left subclavian artery (LSA) revascularization, use of surgical debranch or chimney technique was at operator's discretion before intentional LSA coverage. Technical details of TEVAR were as described by Nienaber and colleagues 5 . Procedural success was defined as the technically successful deployment of the stent-graft at the intended target location. Any death that occurred suddenly or could not be related to other causes was defined as aortic rupture related death. Reintervention was defined as the need for any surgical conversion or additional endovascular stent-graft procedures. Procedure-related deaths refer to those related to endovascular stent placement (e.g., retrograde dissection, organ or peripheral artery malperfusion, or dissection rupture). Non-procedure-related deaths refer to those (e.g., myocardial infarction, cancer or multiple organ failure) unrelated to endovascular stent-graft placement.

Study Selection.
After excluding 91 studies with non-relevant abstract and 54 studies with <5 cases reported or traumatic dissection reports, 368 studies were selected for further full text review. Among the latter, 193 were excluded because of duplicate publications or overlapping data, and 175 studies were eventually selected for data extraction and analysis (supplemental references). A total of 11879 patients who underwent TEVAR were included in this meta-analysis ( Fig. 1).

Study and Patient
Characteristics. An overview of study and patient characteristics is presented in Table 1. A total of 173 hospitals and 176 departments were included in the meta-analysis. Among patients with pertinent information: average age (among 8,580 patients) was 53.9 years; 70.6% (among 7,677 patients) were male; 70.2% (among 4,890 patients) had acute dissection; 83.5% (among 6,693 patients) had hypertension; and 5.4% (among 3,402 patients) had diabetes.
Numbers of included studies, patients, hospitals and departments by geographic area are shown in Table 2. Two municipalities and 4 provinces (Beijing, Shanghai, Zhejiang, Yunnan, Liaoning and Hunan) had over 100 patients per department. Among departments with published cases, the department of vascular surgery from Zhongshan Hospital in Shanghai had the largest number (443) of type B AD cases. Figure 2A shows TEVAR case prevalence by geographic region. Number of cases was categorized into 5 levels (green: <50, brown: 50-<100, blue: 100-<500, yellow: 500-<1000, and red: 1000-1500). Except for Tibet, all 22 provinces, 4 autonomous regions and 4 municipalities reported on TEVAR for type B AD. Two municipalities and 2 provinces (Shanghai, Beijing, Hubei and Guangdong) had 1000-1500 cases; three provinces had 500-<1000 cases; eleven provinces, 4 autonomous regions and 2 municipalities had 100-<500 cases; 2 provinces had 50-< 100 cases; and 2 provinces had <50 cases. Figure S1 presents the geographic distribution of cases in the 2001-2007 Chinese data meta-analysis: only 10 provinces, 4 autonomous regions and 2 municipalities reported TEVAR among type B AD patients, and the reported number of cases was only about one-tenth of that between 2008-2015; 2 municipalities and 1 province had 100-<500 cases; 2 provinces and 1 autonomous region had 50-<100 cases; 7 provinces and 3 autonomous regions had <50 cases; and the remaining 12 provinces, 1 autonomous region and 2 municipalities had no studies reporting TEVAR. The proportion of cases treated by each specialist type was calculated (   Procedural Data and In-Hospital Course. Table 3 shows procedural data and in-hospital course. The majority of patients (71.8% among 5314) received general anesthesia. TEVAR was successful in 99.1% of 8885 patients. Regarding adjunctive techniques in TEVAR, covered left subclavian artery was used in 9.5%; debranch technique in 2.8% (among 8885 patients); and chimney technique in 6.1% among 783 patients. In-hospital major complications were reported in 1.7% of 6956 cases. The most critical in-hospital complications were related to retrograde dissection extension into the ascending aorta (0.4% of 6532 cases) and stent-graft related distal AD (0.1% of 6532 cases). In terms of neurologic complications, in-hospital paraplegia occurred in 0.4% of 1106 cases, and stroke occurred in 1.1% of 3695. Total in-hospital surgical conversion rate was 0.3% of 8885 cases. In-hospital adjunctive endovascular procedure was required in 0.3% of 8885 cases. One hundred and thirty-seven of 8316 cases died during the in-hospital period, yielding an overall in-hospital mortality rate of 1.6%. Type and proportion of in-hospital death causes are summarized in Fig. 3. Dissection rupture was the most common cause of death (32.1%). Among the 22 dissection rupture patients with known time of death, 36.4% died during the first postoperative day and 95.5% within postoperative week one. Figure 4 shows the Kaplan-Meier curve for in-hospital death among patients with available data. Forty percent of deaths occurred during the first postoperative day, 50% within the first 3 postoperative days, and ~75% within the first postoperative week. Figure S3 depicts rates of procedure-and non-procedure-related deaths during hospitalization, which were statistically similar (0.8% vs. 0.6%, log-rank p = 0.765).

Data available (n) No. of events [n(% ± SE)]
Anesthesia 5314  Table 3. Procedural data and in-hospital course. AD: aortic dissection.     Figure 6 shows the Kaplan-Meier survival cure for long-term death among patients with available information. Death rate at one-and two-year follow up was about 50% and 75%, respectively; and 90% within 4 years of discharge. Figure S4 depicts procedure-and nonprocedure-related deaths during follow up, which were statistically similar (1.0% vs. 1.3%, P = 0.737).

Discussion
The main findings in the present meta-analysis of 2008-2015 publications from mainland China on TEVAR for type B AD are a change in the geographic and specialty practice of TEVAR and apparently improved safety outcomes relative to prior time periods in mainland China and Western countries. In the past eight years, the shift in specialist type treating type B AD in mainland China from interventional radiologists to cardiologists and cardiothoracic surgeons while stable for vascular surgeons may reflect that more cardiologists and cardiothoracic surgeons are now capable of performing TEVAR; moreover, because  interventional radiology is usually not the first specialty to which type B AD patients are referred, cardiologists, and cardiothoracic and vascular surgeons might become the main specialists treating type B AD in China.
The mean age of 53.9 years for type B AD patients in China is 7-9 years younger than those reported by Eggebrecht's 3 , Hiratzka's 10 and Patel's 11 studies, suggesting an earlier onset of type B AD among Chinese compared with Caucasians in Europe and North America. Although published data suggest a lower prevalence of atherosclerosis among Chinese AD patients as compared to Caucasians, rates of postoperative neurological complications, namely stroke (p = 0.899) and paraplegia (p = 0.319) were statistically similar in this and the study by Eggbrecht et al. 3 (Table 5). When compared with the study by Moulakakis et al. on a similar data collection time period (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013), rates of stroke (p = 0.722) and paraplegia (p = 0.049) were statistically similar to those in our study for acute uncomplicated type B AD cases in the study by Moulakakis et al. 12 . However, stroke (p = 0.001) and paraplegia (p = 0.007) rates were lower in our study as compared to the data from acute complicated type B AD in the study by Moulakakis et al. 12 (Table 6).
While the results of the present study indicate that TEVAR procedural success rate in China has remained stable over the past 15 years and comparable to that in Western countries, in-hospital mortality appeared to improve in mainland China from 2001-2007 2 to 2008-2015, becoming significantly lower to that in the study by Eggebrecht et al. 3 . A similar trend was found when data from the present study were compared to those from the study by Moulakakis et (Table 6), and to the high in-hospital mortality (30% in complicated patients aged older than 70 years, and 10.1% in those younger than 70 years) from the International Registry of Acute Aortic Dissection (IRAD) study 11 . The latter differences might be attributed to: continuous indication expansion and technical improvement for TEVAR during the past 8 years 6,13 ; and younger age of patients undergoing TEVAR in China than in Western countries, with reduced risk for postoperative cardiovascular and cerebrovascular adverse events or death.
Incidence of retrograde type A AD after TEVAR was significantly lower (approximately one fifth) both in 2008-2015 and 2001-2007 Chinese studies than in the study by Eggbrecht et al. 3 , despite: (1) respective proportion of acute cases of 70%, 36% and 58.1%, which have been associated with a higher proportion of stent-graft induced new intimomedial tear due to very fragile aortic wall in acute stage of AD [14][15][16][17] ; and (2) less conservative preference by Chinese surgeons during the past 8 years for 15% oversizing of stent-grafts. The lower incidence of post TEVAR retrograde type A AD tear might be explained by younger mean age in the Chinese patient population with less atherosclerotic burden in the landing zone allowing stronger resistance to stent-graft radial pressure, and increased use of domestic stent-grafts (over half of stent-grafts used in 2001-2007) with  Table 6. Results of comparison among the study by Moulakakis et al. (2006Moulakakis et al. ( -2013 and Chinese data (2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015).
more recently optimized curve shaped design to better adapt to the anatomy of arch-descending aorta anatomy 8,18 , and stent-graft struts designed to be in compliance with the upward convexity and forward convexity of arch-descending aorta to reduce intima trauma [19][20][21] . Improvement in TEVAR diagnosis and timely intervention during the last 8 years might have been an important determinant in reducing nonprocedure-related mortality during hospitalization. Because of in-hospital death predominantly (75%) occurring during the first week after TEVAR and secondary to AD rupture, more attention should be paid during this period to signs and symptoms of impending AD rupture, such as persistent and aggressive chest pain, refractory hypertension, pleural effusion and proximal type I endoleak 22 , in which case aorta CTA should be performed as soon as possible to ensure timely re-intervention.
Lower follow-up mortality in 2001-2007 and 2008-2015 Chinese data relative to that in the study by Eggebrecht et al. 3 , and lower follow-up mortality in 2008-2015 Chinese data relative to that in the study by Moulakakis et al. 12 and the rather high 5-year mortality (16%) in the IRAD study 11 , also might be attributed to the younger Chinese population with less comorbidity. Importance of surveillance extends to follow-up mortality with 33% and 60% of deaths occurring within the first 6 and 12 months of follow-up, respectively and 90% up to 4 years follow-up; therefore, continuous monitoring and disease management are most important and great attention should be paid to changes in AD morphology and signs of rapid growth of dissection and impending rupture (intermittent dull chest pain 23 or hoarseness of the voice 24 ) to ensure timely reintervention.
Although the present study provides a contemporary overview of use and effectiveness of TEVAR in treatment of type B AD in mainland China, including data by province, autonomous region and municipality, it has several limitations. Overall, the generalizability of the results of the present meta-analysis on prevalence of type-B dissection and major complication rates after TEVAR is limited by the high proportion of studies included that did not report pertinent data because of their research focus and design; information on several variables therefore was available for about 50% and in some cases less of total patients in the meta-analysis. More robust meta-analyses are warranted including future studies providing more comprehensive datasets. First, many studies did not clarify if type B AD was complicated or uncomplicated, which might influence post-TEVAR complication and mortality rate estimation thereby leading to misclassification bias when comparing to other meta-analyses. Second, most studies did not report stent-graft type, rendering unfeasible to accurately determine the effect of various types on TEVAR outcomes, especially endoleak and stent-graft-induced new entry. Third, because doctors from medical school affiliated, teaching and tertiary hospitals usually are more motivated to publish than those from non-tertiary hospitals; generalizability of study results might be limited. Fourth, beyond traditional hard endpoints, i.e., death and major complications, false lumen remodeling has been proposed as an important long-term outcome post TEVAR; however, few studies have reported on it. Fifth, most studies did not report patient comorbidities, such as cardiovascular and cerebrovascular diseases, which would bias postoperative morbidity and mortality estimation. Sixth, single center series included in this study might include highly selected patients that may introduce publication bias. Seventh, because follow up time in studies in the 2008-2015 Chinese dataset was expressed as a range of time and not a particular timepoint, overall median follow-up duration could not be calculated for comparisons with other published meta-analyses. Finally, the use of adjunctive techniques (such as debranch and chimney) in TEVAR has increased during the past 8 years, which may also increase the risk of procedure-related and postoperative morbidity; however, scarce information on their use did not allow to control for the related confounding effect.
In conclusion, the present study suggests that use and effectiveness of TEVAR for type B AD treatment in mainland China has improved over the years, achieving short-and long-term outcomes that are comparable and some possibly more favorable than those in Western countries. Further effort is required to establish more reliable and informative reporting protocols and systems at a national level to ensure generation of more comprehensive and generalizable evidence to guide future medical development in China.