We developed and followed a standard protocol for all steps of this review, which underwent review by a panel of experts in the field as well as the public. The peer-reviewed final report details the methodology including the analytic framework, literature search strategy, and analysis plan as well as evidence tables and is available at http://www.effectivehealthcare.ahrq.gov. The authors have no conflicts of interest.
We refined key questions in collaboration with a panel of technical experts which included cardiologists, internists, and representatives from managed-care organizations. The following key questions in patients with STEMI who are undergoing PCI of native vessels were defined:
1. What are the comparative effects of adjunctive devices from different classes on intermediate [e.g. ST-segment resolution (STSR), myocardial blush grade (MBG), thrombolysis in myocardial infarction (TIMI) 3 flow, ejection fraction, no reflow and distal embolization)] and final health outcomes [mortality, MACE, and health-related quality-of-life (HRQoL)]?
2. How do the rate and type of harms (e.g. coronary dissection, coronary perforation, and prolonged procedure time) differ between device types when compared to PCI alone?
The final report of this comparative effectiveness review, available at http://www.effectivehealthcare.ahrq.gov, includes information and results of analyses specific to patients with non-ST-segment myocardial infarction (NSTEMI), unstable angina (UA), or mixed ACS (STEMI, NSTEMI, and/or UA), results from observational trials with over 500 patients, and information on results in various subpopulations.
We conducted a computerized literature search of Medline, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, and Web of Science databases from January 1996- March 2010, without language restrictions. We restricted the search to 1996 and later to reflect contemporary practice. The complete search strategy is included in Additional File 1. Additionally, in an attempt to locate unpublished studies and increase the sensitivity of our search, we reviewed references from identified studies and systematic reviews. We searched for and reviewed abstracts from major cardiology meetings (American Heart Association, American College of Cardiology, European Society of Cardiology), Transcatheter Cardiovascular Therapeutics (TCT) Conference of the Cardiovascular Research Foundation and the TCTMD http://www.tctmd.com, CardioSource Plus http://www.cardiosource.com, and http://ClinicalTrials.gov
http://www.clinicaltrials.gov web sites. We updated the literature search in March 2011 during the peer review period using the same search strategy.
Two independent reviewers assessed studies for inclusion in a parallel manner using a priori criteria. RCTs of any size were eligible for inclusion if they: 1) compared the use of adjunctive devices plus PCI to PCI alone, 2) included patients with STEMI, 3) had ≤ 5% of the study population receiving PCI of saphenous veins, and (4) reported data on at least one pre-specified intermediate outcome, final health outcome, or harm. Given the known benefit of distal embolic protection devices in patients undergoing PCI of a saphenous vein graft, this review was restricted to a population with lesions primarily in native vessels.
Two reviewers used a standardized data extraction tool to independently extract study data. Data extracted from each study included interventions, study design, inclusion and exclusion criteria, methodological quality criteria, study population, baseline patient characteristics, use of concurrent standard medical therapies, and pre-specified benefits and harms.
Assessment of Study Quality and Strength of Evidence
Two reviewers independently assessed the validity and strength of evidence using recommendations in the Methods Guide for Effectiveness and Comparative Effectiveness Reviews . We assessed each study for the following individual criteria: randomization technique, comparable study groups at baseline, detailed description of study outcomes, blinding of outcome assessors, intent-to-treat analysis, description of participant withdrawals (percent follow-up), and potential conflict of interest. Studies were then given an overall quality score of good, fair, or poor.
We used a modified version of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system to assess the strength of evidence for each outcome of interest. Four required domains were evaluated- risk of bias, consistency, directness, and precision. We classified the strength of evidence for each outcome as high, moderate, low, or insufficient.
Data Synthesis and Analysis
We qualitatively examined data from all identified studies. Six device classes were considered and for each outcome, we conducted separate analyses of studies that compare each device class with control or two device classes to each other. Device classes included catheter aspiration, mechanical thrombectomy, distal filter embolic protection, distal balloon embolic protection, proximal balloon embolic protection, and embolic protection devices combined (distal or proximal, balloon or filter). We conducted meta-analyses when two or more RCTs that were adequate for data pooling were available for any outcome. For dichotomous outcomes, weighted averages are reported as relative risks (RR) and risk differences (RD) with associated 95 percent confidence intervals using a DerSimonian and Laird random-effects model . We used automatic 'zero cell' correction for studies with no events for a particular outcome occurring in one group. We excluded studies with no events occurring in the treatment and control groups. We addressed statistical heterogeneity using the I2 statistic (with a value > 50% deemed to be representative of significant heterogeneity) and we used Egger's weighted regression statistic to assess for the presence of publication bias . We used StatsDirect statistical software, version 2.7.8 (StatsDirect Ltd., Cheshire, England) with a p-value of < 0.05 considered statistically significant.
We defined attainment of optimal myocardial reperfusion as a MBG-3 or TIMI-3 blood flow (or a MBG or TIMI blood flow of at least two in studies not reporting the other endpoint) and ST-segment resolution as 70 percent resolution in peak ST-segments (or at least 50 percent resolution in studies not reporting the other endpoint). We used results for ST-segment resolution at or close to 60 minutes post-PCI and never exceeding 90 minutes post-PCI. For final health outcomes, we defined the base-case analysis using the maximum duration of follow-up for which a final health outcome was reported.
To assess the effect of heterogeneity (both clinical and methodological) on the conclusions of our meta-analysis, we conducted subgroup and sensitivity analyses based on study quality (limited to good quality trials) and duration of follow-up on the efficacy of adjunctive devices. Data at different follow-up times (in-hospital, ≥ 30 days but < 180 days, ≥ 180 days but < 365 days, and ≥ 365 days) were pooled in separate subgroup analyses.
Role of the Funding Source
The University of Connecticut/Hartford Hospital Evidence-based Practice Center prepared this systematic review, with funding from the Agency of Healthcare Research and Quality. The funding source formulated draft research questions and provided the copyright release for this manuscript, but did not participate in the literature search, data collection or analysis, or interpretation of the results.