Repeated carotid endarterectomy versus carotid artery ... - Surgery

4 downloads 0 Views 345KB Size Report
Purpose. Carotid restenosis (CRS) after carotid endarterectomy (CEA) is an issue that cannot be ignored. This study was undertaken to compare the outcomes ...
Vascular Repeated carotid endarterectomy versus carotid artery stenting for patients with carotid restenosis after carotid endarterectomy: Systematic review and meta-analysis Jian Tu, MD,a,b Siwen Wang, MD,a Zijun Huo, MD,a,b Ridong Wu, MD,a Chen Yao, MD,a and Shenming Wang, MD, FACS,a Guangzhou City, Guangzhou, China

Purpose. Carotid restenosis (CRS) after carotid endarterectomy (CEA) is an issue that cannot be ignored. This study was undertaken to compare the outcomes of repeated CEA (redo CEA) and carotid artery stenting (CAS) for CRS after CEA. Methods. We performed a systematic analysis using the search terms ‘‘CEA restenosis,’’ ‘‘carotid restenosis,’’ or ‘‘CEA recurrent stenosis’’ in the MEDLINE, EMBASE, PubMed, and Cochrane Library databases. After applying the inclusion criteria, all available data were summarized to evaluate the effects of redo CEA and CAS for patients with CRS after prior CEA. Results. Fifty articles (9 comparative studies and 41 noncomparative studies) involving 4,399 patients were included. No differences were observed in the 30-day perioperative mortality, stroke and transient ischemic attack rates in the comparative studies (P > .05) and the noncomparative studies (P > .05). Patients undergoing redo CEA suffered more cranial nerve injuries (CNIs) than those undergoing CAS (P < .05), but most of these cases recovered within 3 months. Patients treated with redo CEA exhibited similar myocardial infarction (MI) rates to those treated with CAS in the comparative studies (P = .53), but the rate was higher in the noncomparative studies (P < .01). However, a nonsignificant difference was noted in freedom from stroke at 36 months in the comparative studies (P = .47) and at 12 months in the noncomparative studies (P = .89). The risk of restenosis was greater in the CAS patients than in the redo CEA patients (P < .05 for comparative and noncomparative studies). Conclusion. Both redo CEA and CAS are safe and feasible for CRS after CEA. Although the incidences of CNI and MI were increased in the redo CEA group, most of the CNI cases were reversible. Patients treated with CAS were more likely to develop restenosis than those treated with redo CEA over long-term followup. (Surgery 2015;157:1166-73.) From the Department of Vascular Surgery,a The First Affiliated Hospital of Sun Yat-sen University, Guangzhou City; and the 8-year Program,b Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou, China

CAROTID ENDARTERECTOMY (CEA) is the most effective procedure to prevent stroke in patients with carotid stenosis.1 However, 0.6–8% of patients suffer from restenosis in the late follow-up period

after primary CEA.2 Restenosis has become an important issue that can affect the efficacy of CEA. Carotid restenosis (CRS) after primary CEA is the result of neointimal hyperplasia during the

J.T., S.W., and Z.H. contributed equally to this work.

Guangzhou City, Guangdong Province 510080, China. E-mail: [email protected]; and Chen Yao, MD, Department of Vascular Surgery, The First Affiliated Hospital of Sun Yat-sen University, No. 58, Zhongshan 2nd Road, Yuexiu District, Guangzhou City, Guangdong Province 510080, China. E-mail: [email protected].

This study was supported by Guangdong National Science Foundation of China (Grant No. S2013040012593). Disclosure Summary: The authors have disclosed no relevant financial relationship. Accepted for publication February 13, 2015. Reprint requests: Shenming Wang, MD, FACS, Department of Vascular Surgery, The First Affiliated Hospital of Sun Yat-sen University, No. 58, Zhongshan 2nd Road, Yuexiu District,

1166 SURGERY

0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.02.005

Surgery Volume 157, Number 6

early postoperative period (within 24 months) and of recurrent atherosclerosis after 24 months.3 CRS can be treated with repeated CEA (redo CEA) or CAS. Redo CEA was first reported for revascularization of CRS after CEA in 1976.4 Redo CEA has been regarded as a high-risk procedure because it can cause more perioperative cranial nerve injuries (CNIs) and local complications.5,6 Redo CEA also carries an increased risk of stroke compared with primary CEA. However, various studies have demonstrated acceptable outcomes of redo surgery.7 Additionally, over the last decade, CAS has been applied for patients with CRS as well as for patients with neck irradiation or high cervical lesions by an increasing number of clinicians.7,8 However, precise systematic reviews or welldesigned, randomized, controlled trials to determine which method is better for the treatment of CRS after CEA are lacking. There are limited data about the incidence of tertiary stenosis after primary CEA. It is of great importance to perform such a systematic review to compare the early and late outcomes between redo CEA and CAS for postendarterectomy CRS. METHODS Literature sources. We searched carefully the MEDLINE, EMBASE, PubMed, and Cochrane Library databases for English-language studies using the search terms ‘‘CEA restenosis,’’ ‘‘carotid restenosis,’’ or ‘‘CEA recurrent stenosis.’’ Conference abstracts and reference lists of the studies identified were also included in the search. The last search date was September 30, 2013. Our study included randomized, controlled trials, prospective or retrospective nonrandomized comparative studies, and case reports that fulfilled the following criteria: (1) studies were written in English, (2) studies aimed to assess the effects of interventions for CRS after CEA, (3) studies included >10 patients, and (4) 30-day perioperative data were reported in the studies. The newest and most informative articles were included in cases of duplicate data from the same center. Studies that focused on the effects of redo CEA or CAS between primary CRS and secondary CAS were excluded. Commentaries and reviews were excluded. Definitions. Transient ischemic attack (TIA) was defined as a new neurologic deficit that resolved completely within 24 hours. The primary outcomes were defined as any stroke or death occurring during the procedure or within 30 days after the procedure. Secondary outcomes included TIA, myocardial infarction (MI), CNI occurring during the procedure or within 30 days after the procedure

Tu et al 1167

and TIA, stroke, and disease-related death occurring after 30 days. Restenosis was defined as a 50% diameter (75% area) reduction based on duplex ultrasonography. Severe restenosis was defined as >70% diameter reduction based on duplex ultrasonography. Tertiary stenosis was defined as restenosis after redo CEA or CAS for patients with primary CEA. Reversible CNI indicated CNI symptoms that disappeared within 3 months after CEA. Data extraction. For all of the selected articles, 2 authors independently extracted the following data: clinical characteristics, procedural data, early complications, and late complications. When discrepancies occurred, a decision was reached through further discussion. Statistical analysis. Methodological Index for Non-randomized Studies (MINORS) was used to evaluate the quality of prospective or retrospective nonrandomized comparative studies.9 This index involved 12 items, the first 8 of which were applied for noncomparative studies; the remaining 4 items were designed for comparative studies. Items were scored as 0 (not reported), 1 (reported but inadequate), or 2 (reported and adequate). For studies comparing redo CEA and CAS, we used Rev Man (version 5.0, the Cochrane Collaboration), using 2-sided hypothesis testing. For noncomparative studies, the standardized protocol used for statistical analysis followed the method introduced by Eggebrecht et al.10 The rates of late outcomes were compared by person-year analyses.11 The number of late outcome events was divided by the product of the number of treated patients multiplied by the median or median follow-up duration, as appropriate. We used 2tailed Chi-square statistics for categorical data and Student’s 2-sided t test for continuous variables, which we performed with SPSS software (version 13; SPSS Chicago, IL). RESULTS Overview of studies and patient characteristics. A total of 2,148 articles were reviewed, and 50 publications met the inclusion criteria. Among the excluded studies, 1,634 articles concentrating on primary carotid stenosis were excluded, and 3 articles were excluded because they compared primary CAS with secondary CAS after prior CEA. Of these studies, 9 studies directly compared redo CEA and CAS,3,6,1218 23 studies focused on redo CEA,1,7,19-39 and 18 studies investigated CAS2,40-56 (Figure). A total of 4,399 patients and 4,576 procedures were involved in the included studies. There were 1,137 patients (546 for redo CEA and 591 for CAS) with 1,158 procedures (556 for redo CEA and 602 for CAS) in the

1168 Tu et al

Surgery June 2015

Figure. PRISMA flow diagram of this meta-analysis.

studies comparing redo CEA and CAS. In addition, 1,746 patients with 1,846 procedures were included in the redo CEA studies, and 1,516 patients with 1,572 procedures were included in the CAS studies (Supplementary Table I, II, III). Only 4 of the included studies were designed prospectively; the others were designed retrospectively. The MINORS scores of the comparative studies was 18.87 ± 0.99 (mean ± SD; Supplementary Table IV). The baseline characteristics of the studies included in this analysis are summarized in Table I. No differences were demonstrated in age, male ratio, symptomatic patient rate, coronary artery disease rate, smoking rate, diabetes mellitus rate, or hypertension rate between redo CEA and CAS in either the comparative group or the noncomparative group. However, patients treated with CAS had significantly reduced follow-up durations than those treated with redo CEA in the noncomparative groups. The indications for interventions of CRS after CEA were >50% restenosis in symptomatic patients and >70 or 80% restenosis in asymptomatic patients in the majority of studies. Outcomes of comparative studies. Primary outcomes. The 30-day mortality could not be estimated by meta-analysis owing to insufficient data. The stroke

rate was also not significantly different (odds ratio [OR], 1.32; 95% CI, 0.62–2.78; P = .47; Supplementary Figure 1). For subgroup analysis, the stroke rates within 30 days were also similar in symptomatic and asymptomatic patients (P > .05; Table II). Secondary outcomes. No differences in TIA rate (OR, 0.87; 95% CI, 0.59–1.29; P = .50; Supplementary Figure 2) or MI rate (OR, 1.71; 95% CI, 0.43–6.69; P = .44; Supplementary Figure 4) were noted between redo CEA and CAS. However, the pooled OR of CNI between these 2 treatments was 7.04 (95% CI, 1.45–34.26; P = .02; Supplementary Figure 3). Moreover, 85.7% of the CNIs caused by redo CEA recovered after 3 months. Significant differences were noted in freedom from restenosis at 36 months (OR, 2.39; 95% CI, 1.13–5.07; P = .02; Supplementary Figure 5). Freedom from stroke at 36 months was similar in the groups directly comparing redo CEA and CAS (OR, 0.87; 95% CI, 0.45–1.70; P = .69; Supplementary Figure 6). Disease-related mortality, stroke, TIA, and tertiary intervention after 30 days were not analyzed owing to insufficient data. Outcomes of noncomparative studies. Early complications. Comparing the redo CEA group and the CAS group patients, neither mortality

Tu et al 1169

Surgery Volume 157, Number 6

Table I. Characteristics of patients in comparative and noncomparative studies Redo CEA

CAS

P value

546 556 69.22 ± 4.54 63.69 ± 13.62 50.34 ± 18.94 39.8 ± 11.14 64.43 ± 26.30 25.08 ± 5.42 72.14 ± 25.02

591 602 70.86 ± 2.83 60.07 ± 12.34 46.44 ± 18.68 42.38 ± 21.43 59.08 ± 26.53 27.70 ± 10.31 79.74 ± 16.85

— — .46 .99 .91 .11 .91 .20 .78

1,746 1,846 66.08 ± 2.36 61.11 ± 9.52 56.98 ± 5.58 50.1 60.01 ± 16.71 66.63 ± 15.22 25.47 ± 15.03 72.79 ± 10.02

1,516 1,572 69.21 ± 2.11 56.98 ± 17.78 36.47 ± 21.85 23.9 54.73 ± 16.88 58.99 ± 22.99 26.02 ± 9.60 76.56 ± 11.22

— — .40 .37 .16