Cassandra J Rowe

Screening for Asymptomatic Carotid Artery Stenosis: A Systematic Review and Meta-analysis for the U.S. Preventive Services Task Force

Stroke is a leading cause of death and disability (1). An estimated 7 million U.S. adults have had a stroke, and roughly 75% were first attacks (2). Ischemic strokes account for nearly 90% of all strokes in the United States (3). Carotid artery stenosis (CAS) causes approximately 10% of ischemic strokes (4). Carotid artery stenosis refers to atherosclerotic narrowing of the extracranial carotid arteriesspecifically, the internal carotid arteries or the common and internal carotid arteries. The best available data for the prevalence of asymptomatic CAS from large U.S.-based studies of the general population were published in the 1990s and enrolled adults aged 65 years or older (5, 6). Data published in 1992 showed a prevalence of just more than 1% for CAS of 75% to 99% (6), and those published in 1998 suggested a prevalence of 0.5% for CAS of 70% to 99% (5). Several studies have attempted to estimate the rate of progression of asymptomatic CAS and predict neurologic events (5, 711). The best available data from large U.S.-based studies of the general population revealed a 5-year risk for ipsilateral stroke of 5% for CAS of 70% or greater (5441 participants) (5). The main purpose of this review is to evaluate the current evidence on whether screening asymptomatic adults for CAS reduces the risk for ipsilateral stroke and on harms associated with screening and interventions for CAS. We also evaluated evidence on the incremental benefit of medical therapy and on risk-stratification tools. Despite a D recommendation from the U.S. Preventive Services Task Force in 2007 (12), many surgeries or interventions for asymptomatic CAS continue to be performed, and free or cash-on-the-barrel screenings are offered in public locations across the country (13). Methods We developed an analytic framework (Supplement 1) and key questions (Table 1 of Supplement 2) that guided the review. Detailed methods and additional results are publicly available in our full evidence report (www.uspreventiveservicestaskforce.org) (14). Supplement 1. Analytic Framework for Screening for Asymptomatic Carotid Artery Stenosis Supplement 2. Tables Data Sources and Searches We searched MEDLINE, the Cochrane Library, and EMBASE for English-language articles published through September 2013 (Tables 2 and 3 of Supplement 2). We conducted a targeted update search of MEDLINE for trials published through 31 March 2014 and searched clinical trial registries for unpublished literature. To supplement electronic searches, we reviewed reference lists of included studies and literature suggested by reviewers. Study Selection Two investigators independently reviewed abstracts and full-text articles against prespecified eligibility criteria (Table 4 of Supplement 2). We included studies that focused on asymptomatic adults with CAS and studies that analyzed the asymptomatic group separately. We included randomized, controlled trials (RCTs) of screening for CAS, RCTs and systematic reviews of treatment effectiveness, multi-institution trials or cohort studies that reported harms, and studies that attempted to externally validate risk-stratification tools. For evaluation of accuracy and reliability of ultrasonography, we focused on systematic reviews but also included primary studies that were published after the literature search cutoff of the most recent good-quality systematic review. Data Extraction and Quality Assessment One investigator extracted pertinent information from each article. Another investigator reviewed extractions for completeness and accuracy. Two independent investigators assigned quality ratings (good, fair, or poor) for each study using predefined criteria (14, 15). Disagreements were resolved with team discussion. Poor-quality studies are described in the full report (14). Data Synthesis and Analysis We qualitatively synthesized findings for each key question by summarizing the characteristics and results of included studies in tabular or narrative format. To determine whether meta-analyses were appropriate, we assessed the clinical and methodological heterogeneity of the studies following established guidance (16). We conducted meta-analysis of RCTs that compared carotid endarterectomy (CEA) with medical therapy for relevant outcomes reported by several studies. We used DerSimonianLaird random-effects models to estimate pooled effects (17) and calculated risk differences between CEA and medical therapy to show the absolute differences between groups. Absolute measures are more easily interpreted, show more directly relevant information, and better allow decision makers to assess tradeoffs between benefits and harms (1820). We calculated chi-square and I 2 statistics to assess statistical heterogeneity in effects among studies (21, 22). To allow the comparison of rates of perioperative harms reported in RCTs with those from sources that may be more representative of real-world clinical practice, we conducted meta-analyses of cohort studies that reported perioperative (30-day) stroke or death rates. We also conducted meta-analyses of such rates reported in trials that involved CEA or carotid angioplasty and stenting (CAAS), regardless of the comparator. We conducted sensitivity analyses using profile likelihood random-effects methods when our meta-analyses included few studies (2326). We did not include poor-quality studies in our analyses. Analyses were conducted using Stata, version 11.1 (StataCorp). Role of the Funding Source The Agency for Healthcare Research and Quality funded the review. Members of the U.S. Preventive Services Task Force and Agency for Healthcare Research and Quality assisted in developing the reviews scope and reviewed draft manuscripts, but the authors are solely responsible for the content. Results We included 78 published articles that reported on 56 studies (Figure 1) . Figure 1. Summary of evidence search and selection. WHO ICTRP = World Health Organization International Clinical Trials Registry Platform. Direct Evidence that Screening Reduces Ipsilateral Stroke We found no eligible studies that provided evidence on whether screening reduced ipsilateral stroke. Accuracy and Reliability of Duplex Ultrasonography We included 3 meta-analyses (2729) and 1 fair-quality primary study (30) (Table 5 of Supplement 2). The most recent good-quality meta-analysis (28) included 47 studies published through 2003 that used digital subtraction angiography as the reference standard. It reported sensitivity and specificity for detecting stenosis of 50% or greater (1716 participants) of 98% (95% CI, 97% to 100%) and 88% (CI, 76% to 100%), respectively. Sensitivity and specificity for detecting stenosis of 70% or greater (2140 participants) were 90% (CI, 84% to 94%) and 94% (CI, 88% to 97%). Using data from this meta-analysis, the last evidence report for the U.S. Preventive Services Task Force estimated the sensitivity and specificity for detecting stenosis of 60% or greater as 94% and 92%, respectively (31). The meta-analysis reported wide, clinically important variation in measurement properties among laboratories (28). The findings of the other meta-analyses were generally consistent with these results, but specificity in the primary study was lower (66% for detecting CAS of 70% to 99% [CI, 63% to 71%]; 503 participants) (30). Additional results are provided in our full report (14). Benefits of CEA or CAAS Beyond Medical Therapy We included 3 RCTs (Table 1) described in 12 publications (3243) that compared CEA with medical therapy and 3 systematic reviews described in 5 publications (31, 4447). We found no eligible studies that compared CAAS with medical therapy and no studies that compared CEA with current standard medical therapy. Table 1. Characteristics and Main Results of Included Fair- or Good-Quality Randomized, Controlled Trials of CEA Compared With MM for Asymptomatic CAS* The ACAS (Asymptomatic Carotid Atherosclerosis Study) and the VACS (Veterans Affairs Cooperative Study) were conducted in North America; the ACST (Asymptomatic Carotid Surgery Trial) involved 30 countries, primarily in Europe. Medical therapy varied across trials and was often not clearly defined or standardized. Surgeons with a history of low complication rates were selected. They submitted records of their last 50 cases or previous 24 months of experience with CEA and were selected on the basis of review by a committee or morbidity and mortality rates less than 3%. Our meta-analyses found that fewer persons treated with CEA had perioperative stroke or death or subsequent ipsilateral stroke, perioperative stroke or death or any subsequent stroke, any stroke or death, nonperioperative ipsilateral stroke, and any nonperioperative stroke than those in medical therapy groups (Table 2 and Figure 2). For all-cause mortality, we found no significant difference. Results for sensitivity analyses using profile likelihood methods were very similar to those of our main analyses, with only minor variation in width of CIs (Table 2). Table 2. Summary of Main Results of Meta-analyses Figure 2. Meta-analyses of randomized, controlled trials comparing CEA with medical therapy, by outcome. ACAS = Asymptomatic Carotid Atherosclerosis Study; ACST = Asymptomatic Carotid Surgery Trial; CEA = carotid endarterectomy; MM = medical management; RD = risk difference; VACS = Veterans Affairs Cooperative Study. In the ACST, more than one half (57.8% [166 of 287]) of nonperioperative strokes were disabling or fatal, and the proportional reduction in disabling or fatal stroke (relative risk, 0.61 [CI, 0.41 to 0.92]) was similar to that for any stroke (relative risk, 0.54 [CI, 0.43 to 0.68]) (37). Subgroup analyses of the ACAS showed a statistically significant reduction for men (relative risk reduction, 66% [CI, 36% to 82%]) but not for women (relative risk reduction, 17% [CI, 96% to 65%]) for estimated 5-year rate of perioperative stroke or death and subsequent ipsilateral stroke.

Genetic polymorphisms and response to medications for alcohol use disorders: a systematic review and meta-analysis

AIM To assess whether response to medications for alcohol use disorders varies by genotype. METHODS Systematic review and meta-analysis. RESULTS We found no studies that assessed the clinical utility of genotype-guided dosing strategies or genotype-guided medication selection, and none randomized by genotype. All included studies assessed the association between genotype and response to medication. Of 15 included studies, eight (n = 1365 participants) assessed variation in naltrexone response and polymorphisms of OPRM1. Our meta-analyses for return to heavy drinking found no significant difference between A allele homozygotes and those with at least one G allele, both without (risk difference: 0.26; 95% CI: -0.01-0.53; n = 174) and with inclusion of studies rated as high or unclear risk of bias (risk difference: 0.14; 95% CI: -0.03-0.3; n = 382). For all other polymorphism-medication pairs, we found just one eligible study. CONCLUSION Estimates of effect for return to heavy drinking suggest it is possible that patients with at least one G allele of A118G polymorphism of OPRM1 might be more likely to respond to naltrexone, but confidence intervals were wide; additional studies are needed to improve confidence in the estimates.