General Review The Influence of Study Design on the Evaluation of Ruptured Abdominal Aortic Aneurysm Treatment Rodolfo Pini, Ganluca Faggioli, Matteo Longhi, Raffaella Mauro, Antonio Freyrie, Mauro Gargiulo, Enrico Gallitto, Chiara Mascoli, and Andrea Stella, Bologna, Italy
Background: The best strategy in the treatment for ruptured abdominal aortic aneurysm (RAAA) is an ongoing matter of debate. Differently from several retrospective studies, recent randomized controlled trials (RCTs) failed to demonstrate the superiority of endovascular repair (EVAR) over open repair (OPEN). The aim of the present study was to compare 30-day mortality of EVAR and OPEN in RAAA according to different study designs through a systematic review and meta-analysis. Methods: A systematic literature search of all series comparing the outcome of EVAR and OPEN in RAAA was performed. Studies on symptomatic aneurysms without frank ruptures were excluded. The analyses evaluated the effect of the study design on EVAR versus OPEN 30-day mortality. The pooled mortality risk was expressed as odds ratio (OR) with a 95% confidence interval (CI) by random effect model. Results: Four different study designs were evaluated. 1) Patients allocation in EVAR or OPEN was ‘‘unbiased’’ (3 studies, 2 RCTs): there was no superiority treatment in EVAR versus OPEN (OR, 1.58; 95% CI, 0.82e3.06; P ¼ 0.17). 2) Patients submitted to EVAR were compared with a historical OPEN group (2 studies): no difference between EVAR and OPEN (OR, 3.55; 95% CI, 0.47e26.62; P ¼ 0.22). 3) EVAR was the preferential treatment and OPEN was confined to patients with unsuitable anatomy for endovascular procedures (18 studies): in this type of study OPEN had a higher risk of 30-day mortality (OR, 2.18; 95% CI, 1.61e2.96; P < 0.00001). 4) The 30-day mortality after EVAR introduction in centers using both EVAR and OPEN was compared with the only OPEN treatment (7 studies): the latter had higher mortality compared with the protocol with both EVAR and OPEN options (OR, 2.26; 95% CI, 1.41e3.63; P ¼ 0.0007). Conclusions: Only few studies are available to compare EVAR and OPEN in an ‘‘unbiased’’ cohort, with no significant differences between the 2 treatments. However, after the introduction of EVAR and OPEN protocols, the overall mortality for RAAA was reduced compared with the only OPEN option, suggesting a beneficial effect of EVAR in selected cases.
INTRODUCTION
Department of Vascular Surgery, University of Bologna, Policlinico S. Orsola Malpighi, Bologna, Italy. Correspondence to: Rodolfo Pini, Policlinico S. Orsola Malpighi, via Massarenti 9, Bologna 40138, Italy; E-mail: [email protected] Ann Vasc Surg 2014; -: 1–13 http://dx.doi.org/10.1016/j.avsg.2014.03.017 Ó 2014 Elsevier Inc. All rights reserved. Manuscript received: January 7, 2014; manuscript accepted: March 23, 2014; published online: ---.
The rupture of an abdominal aortic aneurysm (AAA) is a catastrophic event with a preehospital admission mortality reaching 80% and an inhospital mortality ranging from 30% to 60%.1 Conventional open repair (OPEN) is either the procedure of choice or the only treatment available in most centers for ruptured abdominal aortic aneurysms (RAAAs). Because endovascular repair (EVAR) has shown favorable perioperative outcome compared with OPEN in the elective treatment of 1
2 Pini et al.
AAA, some centers have adopted specific protocols for EVAR procedures also in RAAA.2 Evidence of efficacy of the latter approach is not reached yet, and the role of EVAR in RAAA is still a matter of debate. Many retrospective studies and multicenter experiences reported higher survival rates in patients with RAAA treated by EVAR compared with OPEN.3,4 Differently, the few published randomized controlled trials (RCTs) reported similar 30-day mortality in the endovascular and surgical arms.5,6 These contradictory results may be due to a variety of causes, including different clinical and anatomic characteristics and different enrollment protocols. Moreover, many studies included also symptomatic AAA with no evident rupture, further determining heterogeneity into the sample. We aimed therefore to systematically review and meta-analyze OPEN and EVAR results in the treatment of RAAA, considering 30-day mortality, clinical characteristics, and enrollment methods.
METHODS The process of the systematic review and metaanalysis was performed according to the Cochrane Collaboration recommendations.7 A rigorous protocol was established before the analyses, concerning all the objectives, inclusion and exclusion criteria, primary and secondary outcomes, and synthesis methods. We reported the results according to the Meta-analysis of Observational Studies in Epidemiology8 and Preferred Reporting Items for Systematic Reviews and Meta-Analyses9 recommendations.
Annals of Vascular Surgery
from the articles obtained were also analyzed. The literature search focused solely on articles published in peer-reviewed journals to enhance the methodologic rigor of studies examined. The inclusion criteria were studies with ruptured infrarenal abdominal aortic or infrarenal aortoiliac aneurysms treated with OPEN or EVAR with aortoiliac endoprosthesis or with standard aortouniiliac endoprosthesis and femorofemoral bypasses, studies providing the diagnostic method for ruptured infrarenal AAA with description of retroperitoneal hematoma or expansion of contrast medium intraperitoneally or extraperitoneally, and studies reporting 30-day mortality. The exclusion criteria were the following: case reports, noneEnglish language publications, studies with no available data on 30-day mortality, and multiple publications on overlapping populations. Also excluded were studies with no explicit and clear methods of patients’ allocation in EVAR or OPEN groups. Unpublished data or data reported only in abstract form were excluded. Also, studies including thoracoabdominal or thoracic and abdominal aneurysms or pararenal aortic aneurysms, or with no diagnostic imaging of RAAA, or reporting painful but unruptured aneurysms were not included in the analysis. Inclusion and exclusion criteria did not include the type of endoprosthesis used or the anatomic indication for EVAR. The final inclusion of the studies was based on agreement between the reviewers. The disagreement was resolved by discussion and consultation with the other coauthors (F.G., R.M., A.F., M.G., A.S.) if necessary.
Research Protocol, Eligibility Criteria, and Study Selection
Data Extraction and Quality Assessment
A comprehensive and systematic literature review was performed through PubMed and Scopus for any English language study comparing the outcome in term of 30-day mortality of patients with RAAA treated with EVAR and OPEN. Two investigators (R.P. and M.L.) independently performed the research. The terminology used for search purposes included ‘ruptured abdominal aortic aneurysm’ [All Fields] AND ‘endovascular aortic repair’ OR ‘EVAR’ [All Fields] AND ‘open repair’ OR ‘surgical repair’ [All Fields]. No limitations or restrictions such as publication year or filters for study design were used in search strategy. The investigators independently reviewed the titles and abstracts of all citations to identify potentially relevant studies and to exclude duplicates. They reviewed the full text of the corresponding publications to assess if the studies met the inclusion criteria. The references
Two investigators (R.P., M.L.) used a standardized preformatted extraction database to extract information on the primary outcome (30-day mortality) and numbers of patients for each group, EVAR and OPEN. The moderators selected to evaluate the influence on the primary outcome in the regression analyses were the publication year, the percentage of male sex, the mean age, the percentage of EVAR feasibility, the percentage of EVAR and OPEN hemodynamic unstable patients, and the mean time elapsed from the admission to hospital and intervention. Characteristics of the studies (period, state of development, design, and enrollment) were also extracted. In the EVAR group, the endoprosthesis type, the type of anesthesia, the usage of balloon aortic clamping, the presence of 30-day postoperative type I endoleak, and the 30-day
Vol.
-,
No.
-, -
2014
conversion rate were also evaluated (considered as moderators). In the OPEN group, suprarenal clamping and type of prosthesis were also evaluated and considered as moderators. Descriptive variables were expressed as percentage, median, and interquartile range (IQR). To reduce possible clinical selection bias, a subgroup evaluation was performed to identify the patients treated with hemodynamic instability in EVAR and OPEN groups. To have higher sensitivity, different definitions of hemodynamic instability were accepted for the meta-analysis. We assessed interstudy quality by evaluating the inclusion criteria in the studies, the method used for RAAA diagnosis, the indication for EVAR or OPEN, the hemodynamic stability of patients treated, and the follow-up. The Newcastle-Ottawa Scale (NOS)10 was applied to evaluate the methodology quality of the observational studies. This scale was developed to assess the quality of studies using a ‘‘star system’’ (maximum 9 stars), in which a study is judged on 3 broad perspectives: 1) the selection of the study groups, 2) the comparability of the groups, and 3) the ascertainment of outcome of interest. The threshold of 6 stars or greater has been considered as indicative for high quality in other meta-analyses as was chosen as well for this review; studies with NOS
Background: The best strategy in the treatment for ruptured abdominal aortic aneurysm (RAAA) is an ongoing matter of debate. Differently from several retrospective studies, recent randomized controlled trials (RCTs) failed to demonstrate the superiority of endovascular repair (EVAR) over open repair (OPEN). The aim of the present study was to compare 30-day mortality of EVAR and OPEN in RAAA according to different study designs through a systematic review and meta-analysis. Methods: A systematic literature search of all series comparing the outcome of EVAR and OPEN in RAAA was performed. Studies on symptomatic aneurysms without frank ruptures were excluded. The analyses evaluated the effect of the study design on EVAR versus OPEN 30-day mortality. The pooled mortality risk was expressed as odds ratio (OR) with a 95% confidence interval (CI) by random effect model. Results: Four different study designs were evaluated. 1) Patients allocation in EVAR or OPEN was ‘‘unbiased’’ (3 studies, 2 RCTs): there was no superiority treatment in EVAR versus OPEN (OR, 1.58; 95% CI, 0.82e3.06; P ¼ 0.17). 2) Patients submitted to EVAR were compared with a historical OPEN group (2 studies): no difference between EVAR and OPEN (OR, 3.55; 95% CI, 0.47e26.62; P ¼ 0.22). 3) EVAR was the preferential treatment and OPEN was confined to patients with unsuitable anatomy for endovascular procedures (18 studies): in this type of study OPEN had a higher risk of 30-day mortality (OR, 2.18; 95% CI, 1.61e2.96; P < 0.00001). 4) The 30-day mortality after EVAR introduction in centers using both EVAR and OPEN was compared with the only OPEN treatment (7 studies): the latter had higher mortality compared with the protocol with both EVAR and OPEN options (OR, 2.26; 95% CI, 1.41e3.63; P ¼ 0.0007). Conclusions: Only few studies are available to compare EVAR and OPEN in an ‘‘unbiased’’ cohort, with no significant differences between the 2 treatments. However, after the introduction of EVAR and OPEN protocols, the overall mortality for RAAA was reduced compared with the only OPEN option, suggesting a beneficial effect of EVAR in selected cases.
INTRODUCTION
Department of Vascular Surgery, University of Bologna, Policlinico S. Orsola Malpighi, Bologna, Italy. Correspondence to: Rodolfo Pini, Policlinico S. Orsola Malpighi, via Massarenti 9, Bologna 40138, Italy; E-mail: [email protected] Ann Vasc Surg 2014; -: 1–13 http://dx.doi.org/10.1016/j.avsg.2014.03.017 Ó 2014 Elsevier Inc. All rights reserved. Manuscript received: January 7, 2014; manuscript accepted: March 23, 2014; published online: ---.
The rupture of an abdominal aortic aneurysm (AAA) is a catastrophic event with a preehospital admission mortality reaching 80% and an inhospital mortality ranging from 30% to 60%.1 Conventional open repair (OPEN) is either the procedure of choice or the only treatment available in most centers for ruptured abdominal aortic aneurysms (RAAAs). Because endovascular repair (EVAR) has shown favorable perioperative outcome compared with OPEN in the elective treatment of 1
2 Pini et al.
AAA, some centers have adopted specific protocols for EVAR procedures also in RAAA.2 Evidence of efficacy of the latter approach is not reached yet, and the role of EVAR in RAAA is still a matter of debate. Many retrospective studies and multicenter experiences reported higher survival rates in patients with RAAA treated by EVAR compared with OPEN.3,4 Differently, the few published randomized controlled trials (RCTs) reported similar 30-day mortality in the endovascular and surgical arms.5,6 These contradictory results may be due to a variety of causes, including different clinical and anatomic characteristics and different enrollment protocols. Moreover, many studies included also symptomatic AAA with no evident rupture, further determining heterogeneity into the sample. We aimed therefore to systematically review and meta-analyze OPEN and EVAR results in the treatment of RAAA, considering 30-day mortality, clinical characteristics, and enrollment methods.
METHODS The process of the systematic review and metaanalysis was performed according to the Cochrane Collaboration recommendations.7 A rigorous protocol was established before the analyses, concerning all the objectives, inclusion and exclusion criteria, primary and secondary outcomes, and synthesis methods. We reported the results according to the Meta-analysis of Observational Studies in Epidemiology8 and Preferred Reporting Items for Systematic Reviews and Meta-Analyses9 recommendations.
Annals of Vascular Surgery
from the articles obtained were also analyzed. The literature search focused solely on articles published in peer-reviewed journals to enhance the methodologic rigor of studies examined. The inclusion criteria were studies with ruptured infrarenal abdominal aortic or infrarenal aortoiliac aneurysms treated with OPEN or EVAR with aortoiliac endoprosthesis or with standard aortouniiliac endoprosthesis and femorofemoral bypasses, studies providing the diagnostic method for ruptured infrarenal AAA with description of retroperitoneal hematoma or expansion of contrast medium intraperitoneally or extraperitoneally, and studies reporting 30-day mortality. The exclusion criteria were the following: case reports, noneEnglish language publications, studies with no available data on 30-day mortality, and multiple publications on overlapping populations. Also excluded were studies with no explicit and clear methods of patients’ allocation in EVAR or OPEN groups. Unpublished data or data reported only in abstract form were excluded. Also, studies including thoracoabdominal or thoracic and abdominal aneurysms or pararenal aortic aneurysms, or with no diagnostic imaging of RAAA, or reporting painful but unruptured aneurysms were not included in the analysis. Inclusion and exclusion criteria did not include the type of endoprosthesis used or the anatomic indication for EVAR. The final inclusion of the studies was based on agreement between the reviewers. The disagreement was resolved by discussion and consultation with the other coauthors (F.G., R.M., A.F., M.G., A.S.) if necessary.
Research Protocol, Eligibility Criteria, and Study Selection
Data Extraction and Quality Assessment
A comprehensive and systematic literature review was performed through PubMed and Scopus for any English language study comparing the outcome in term of 30-day mortality of patients with RAAA treated with EVAR and OPEN. Two investigators (R.P. and M.L.) independently performed the research. The terminology used for search purposes included ‘ruptured abdominal aortic aneurysm’ [All Fields] AND ‘endovascular aortic repair’ OR ‘EVAR’ [All Fields] AND ‘open repair’ OR ‘surgical repair’ [All Fields]. No limitations or restrictions such as publication year or filters for study design were used in search strategy. The investigators independently reviewed the titles and abstracts of all citations to identify potentially relevant studies and to exclude duplicates. They reviewed the full text of the corresponding publications to assess if the studies met the inclusion criteria. The references
Two investigators (R.P., M.L.) used a standardized preformatted extraction database to extract information on the primary outcome (30-day mortality) and numbers of patients for each group, EVAR and OPEN. The moderators selected to evaluate the influence on the primary outcome in the regression analyses were the publication year, the percentage of male sex, the mean age, the percentage of EVAR feasibility, the percentage of EVAR and OPEN hemodynamic unstable patients, and the mean time elapsed from the admission to hospital and intervention. Characteristics of the studies (period, state of development, design, and enrollment) were also extracted. In the EVAR group, the endoprosthesis type, the type of anesthesia, the usage of balloon aortic clamping, the presence of 30-day postoperative type I endoleak, and the 30-day
Vol.
-,
No.
-, -
2014
conversion rate were also evaluated (considered as moderators). In the OPEN group, suprarenal clamping and type of prosthesis were also evaluated and considered as moderators. Descriptive variables were expressed as percentage, median, and interquartile range (IQR). To reduce possible clinical selection bias, a subgroup evaluation was performed to identify the patients treated with hemodynamic instability in EVAR and OPEN groups. To have higher sensitivity, different definitions of hemodynamic instability were accepted for the meta-analysis. We assessed interstudy quality by evaluating the inclusion criteria in the studies, the method used for RAAA diagnosis, the indication for EVAR or OPEN, the hemodynamic stability of patients treated, and the follow-up. The Newcastle-Ottawa Scale (NOS)10 was applied to evaluate the methodology quality of the observational studies. This scale was developed to assess the quality of studies using a ‘‘star system’’ (maximum 9 stars), in which a study is judged on 3 broad perspectives: 1) the selection of the study groups, 2) the comparability of the groups, and 3) the ascertainment of outcome of interest. The threshold of 6 stars or greater has been considered as indicative for high quality in other meta-analyses as was chosen as well for this review; studies with NOS
