Meta-analysis and Evidence-based Medicine Training in Cardiology
The website is dedicated to meta-analysis and evidence-based medicine training in cardiology. It is sponsored by the Meta-analysis and Evidence-based medicine Training in Cardiology (METCARDIO) Group, currently headquartered in Turin, Italy, and formerly known as the Center for Overview, Meta-analysis, and Evidence-based Medicine Training (COMET). The focus of the website is on clinical research methods and evidence-based cardiovascular medicine with a specific interest in interventional cardiology and peripheral cardiovascular interventions. Nonetheless, there is plenty of training and research opportunities for other evidence-based endeavors, eg in anesthesiology, critical care medicine, and psychiatry.

Another month and another year have almost passed, and we are still recovering from the major breakthroughs reported at the American Heart Association (AHA) and at the European Society of Cardiology (ESC) congresses. Indeed, several important studies were reported at these major venues, including the PLATO trial, focusing on ticagrelor, and the RE-LY trial, focusing on dabigatran, which, together with the SYNTAX, the FAME, and the BARI 2D studies, likely represent the most important pieces of clinical evidence of the whole year.
We thus wish happy holidays to all our readers, colleagues, and friends, urging them to read again these important scientific endeavors, in order to improve their clinical and research skills for the next year.

We apologize for this somewhat late monthly newsletter, but the arrival of 2 very junior members of METCARDIO and the made awaiting worthwhile.
A major novelty of this congress has been the disappointing outcome of the CHAMPION trial, a randomized study appraising the intravenous antiplatelet agent cangrelor, which has not proved superior to the much cheaper clopidogrel in patients with acute coronary syndromes. This appears even more striking as previous data in support of another antiplatelet agent, ticagrelor, have instead been confirmed, with significant benefits of this agent in comparison to clopidogrel in the PLATO study in several patient subsets, including those with ST-elevation myocardial infarction.
The radial and ulnar arteries represent an established means to gain access to the cardiovascular system, and enable diagnostic and interventional procedures minimizing the risk of local vascular complications. Wheareas the radio-ulnar approach is rather common in Europe, in some countries physicians are still reluctant to use this access site. The recent RAPTOR trial has shown that even among inexperienced cardiologist the radio-ulnar access can provide significant benefits without overly detrimental effects on procedure duration, radiation exposure, or clinical success.
Cardiac assist devices are becoming a mainstay in the management of end-stage heart failure, notwithstanding their cost implications. The recently reported HeartMate II trial just confirmed the important role of this devices, as continuous flow assist devices proved superior to pulsative flow devices, as the former were associated with lower mortality and fewer complications.

A few weeks after the Transcatheter Cardiovascular Therapeutics (TCT) congress and a few days before the GISE congress, it is timely to provide an overview of the main news in interventional cardiology.
Coronary interventions are still dominated by drug-eluting stents, and the main novelty in this setting has been the very promising data reported on Abbott Xience/Boston Scientific Promus and Medtronic Endeavor. Specifically, the superiority of Xience/Promus versus Boston Scientific Taxus has been now largely proved in the SPIRIT IV trial and in the COMPARE study. Medtronic Endeavor also was shown similar in terms of repeat revascularization but superior in terms of safety in comparison to Taxus in the ENDEAVOR IV study. Another key piece of evidence was provided by Bhatt et al, which demonstrated in the COGENT trial the safety of concomitant administration of proton pump inhibitors such as omeprazole in patients already taking aspirin and clopidogrel. The PROSPECT registry thoroughly evaluated with coronary imaging subjects with acute coronary syndromes, and found that many of them experience recurrent events in the following years, but, with currently available medical therapy, most of these events are were non-fatal (as cardiac death at 3 years occurred in only 1.9%). The dogma of the safety and efficacy of intra-aortic balloon pumps (IABP) in subjects undergoing high-risk coronary angioplasty was openly challenged by the results of the BCIS-1 trial, which suggested that bail-out use, rather than default use, of IABP should better serve patients. Finally, a thought-provoking study has questioned the clinical role of kissing balloon inflation in patients with coronary bifurcation lesions (NORDIC-BALTIC Bifurcation Study III).

Which cardiology meeting best fulfills your educational needs? Obviously this depends a lot on your knowledge and competence level as a clinician and/or researcher, as well as your specific goals. However, the recent European Society of Cardiology (ESC) Congress is surely posing more and more as the leading cardiology meeting worlwide. Indeed, several pivotal randomized trials were reported. We limit ourselves to the handful studies which are likely to change clinical practice in the next few months or years.
The CURRENT-OASIS 7 trial randomized over 25,000 patients with acute myocardial infarction or unstable angina pectoris to 600 mg (followed by a short course of 150 mg daily) vs. 300 mg (followed as routine by 75 mg daily) clopidogrel, as well as high vs. low dose aspirin. They found significant clinical benefits associated with a higher loading and maintenance dose of clopidogrel, whereas high dose aspirin did not prove more effective or more hazardous than low dose aspirin.
The PLATO trial randomized again over 18,000 patients with acute coronary syndromes to clopidogrel vs. ticagrelor (a novel reversible platelet antagonist), showing that this new drug confers significant benefits in comparison to clopidogrel, including reductions in mortality, despite moderate, but not overcoming, increases in bleeding. Notwithstanding some adverse effects in patients (i.e. dyspnea), this drug seems to be a major step forward in the management of patients with unstable coronary artery disease.
The RE-LY trial also challenges current clinical practice (namely the use of warfarin for the prevention of thrombotic events in patients with atrial fibrillation). In this randomized study more than 18,000 subjects were allocated to dose-adjusted warfarin, low dose dabigatran (an oral direct thrombin inhibitor) or high dose dabigatran. After a median of 24 months, low dose dabigatran proved superior to warfarin in reducing bleeding, while achieving a similar antithrombotic effect, whereas high dose dabigatran proved equivalent to warfarin in bleeding risk, but significantly superior in antithrombotic effect. Thus, if reassurances on adverse effects (e.g. dyspnea) can be provided by further studies, dabigatran could become a major treatment option in patients with atrial fibrillation, as well as, in other conditions (e.g. prosthetic heart valves).

Hopeful that everybody can enjoy his/her vacation time, we would still suggest some interesting studies recently reported in the cardiovascular literature, with a specific focus on risk prediction and prognostication.
Melander et al (Journal of the American Medical Association 2009;302:49-57) have performed an extensive analysis to identify predictors of adverse cardiovascular events among 5,067 apparently healthy Swedish men and women. After a median follow-up of 13 years, most adverse events could be prognosticated effectively by means of standard cardiovascular risk factors (i.e. age, sex, systolic blood pressure, diastolic blood pressure, use of antihypertensive therapy, current smoking, diabetes, levels of low-density and high-density lipoprotein cholesterol, and body mass index). Measurements of C-reactive protein, N-terminal pro-B-type natriuretic peptide, and midregional proadrenomedullin did help in the reclassification of some subjects at intermediate risk, but it is debatable whether their incremental prognostic role justifies their incremental cost in this setting.
Eggers et al (Journal of the American College of Cardiology 2009;54:357-64) performed a somewhat similar research endeavor, but focused on 877 patients with acute non-ST elevation acute coronary syndromes and followed them for 5 years. Both C-reactive protein and N-terminal pro-B-type natriuretic peptide levels were significantly associated with the occurrence of death or myocardial infarction. However, after multivariable adjustment for standard cardiovascular risk features (including age, sex, diabetes, a history of heart failure and a history of acute myocardial infarction), only levels of N-terminal pro-B-type natriuretic peptide measured 6 weeks after the index event could provide independent prognostic information.
Finally, Prasad et al (Journal of the American College of Cardiology 2009;54:477-86) have provided novel insights on the prognostic impact of peri-procedural and spostaneous myocardial infarctions in patients undergoing percutaneous coronary intervention for acute coronary syndromes. Evaluating a total of 7,773 subjects, they found that peri-procudural myocardial infarction occurred in 6.0% and spontaneous myocardial infarction in 2.6%. Despite an apparently evident association between both types of myocardial infarction and prognosis up to 1 year, extensive multivariable analysis showed that peri-procedural myocardial infarction was mostly a proxy of other adverse prognostic features, and thus not an independent and significant outcome predictor by itself. Thus, this finding challenges also the prognostic impact of treatments (by them pharmacologic agents such as prasugrel or based on devices such as thrombectomy devices) whose main effect is reducing peri-procedural myocardial infarctions.

Summer is approaching and we are still recovering from the wealth of news and data burgeoning at EuroPCR 2009. Nonetheless, a pivotal piece of clinical evidence has been reported in th New England Journal of Medicine by the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trialists (2009;360:2503-15). In this 2368-patient randomized controlled study, subjects with type 2 diabetes mellitus and concomitant coronary artery disease were randomized to an initial medical management versus coronary revascularization, with coronary artery bypass surgery or percutaneous coronary revascularization at the physician's discretion. After 5 years, survival was similar in the revascularization and medical therapy groups, even if significant benefits were conferred in terms of freedom from major cardiovascular events by surgical revascularization.
Thus, despite the lack of a formal randomization to percutaneous versus surgical revascularization, the results of the BARI 2D trial reinforce previous findings on medical versus revascularization therapy for stable coronary artery disease (eg from the COURAGE trial, New England Journal of Medicine 2007;356:1503-16, and the Katritsis et al meta-analysis, Circulation 2005;111:2906-12). In conclusion, previous and current findings from high-quality clinical trials suggest that a trial of maximal medical therapy should be offered to all patients with stable coronary artery disease, with the notable exception of those with diabetes mellitus and eligible to coronary artery bypass surgery. This does not mean that revascularization (either percutaneous or surgical) has no role in such subjects, but it should be reserved to those patients unable to comply with such aggressive medical therapy or those failing to improve with it.

Awaiting the upcoming EuroPCR meeting in Barcelona, it is timely to reconsider the whole saga of drug-eluting stent safety. Since 2006, the safety of drug-eluting stent had indeed been called into question, as also extensively covered in The strongest blows to our confidence in the safety of drug-eluting stents have been specifically the meta-analyses from Nordmann et al and Camenzind et al, and the observational Swedish Coronary Angiography and Angioplasty Registry (SCAAR). Whereas subsequent meta-analyses largely refuted previous works based on incorrect and/or incomplete datasets (eg Biondi-Zoccai et al for a poignant synthesis of pertinent systematic reviews), the SCAAR report had been largely unchallenged, except from their very same authors at public
In-stent optical coherence tomography
In-stent optical coherence tomography
gatherings, but not in peer-reviewed journals.
In particular, in the article published in 2007 in the New England Journal of Medicine, Lagerqvist et al had exploited sophisticated landmark analyses (ie subgroup analysis of time to event datasets using time cut-offs to separate subgroups), and suggested that "at 6 months, there was a trend toward a lower unadjusted event rate in patients with drug-eluting stents than in those with bare-metal stents. However, after 6 months, patients with drug-eluting stents had a significantly higher event rate. At 3 years, mortality was significantly higher in patients with drug-eluting stents and from 6 months to 3 years, the adjusted relative risk for death in this group was 1.32 (95% confidence interval, 1.11 to 1.57)". A novel report from the SCAAR group by James et al, including over 47,000 patients followed for almost 3 years, states that "there was no overall difference between the group that received drug-eluting stents and the group that received bare-metal stents in the combined end point of death or myocardial infarction or the individual end points of death and myocardial infarction, and there was no significant difference in outcome among subgroups stratified according to the indication for stent implantation. Patients who received drug-eluting stents in 2003 had a significantly higher rate of late events than patients who received bare-metal stents in the same year, but we did not observe any difference in outcome among patients treated in later years". These findings clearly put into a different perspective previous data from the SCAAR group, and specifically highlight the hypothesis-generating role of landmark analyses, which should never regarded per se as conclusive.
Further corroborating evidence in support of drug-eluting stents have been recently provided by the Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) trial, which has demonstrated the mid-term safety and efficacy of paclitaxel-eluting stents even in the thrombotic context of myocardial infarction, including favorable 3.5% (vs 3.5%) rates of death, 3.7% (vs 4.5%) rates of myocardial infarction, 8.1% rates of major adverse cardiac events (vs 8.0% with bare-metal stents), and 2.6% (vs 3.0%) rates of definite stent thrombosis.

The American College of Cardiology and Innovation in Intervention Scientific Sessions (ACC/I2) have brought significant novelties for patients with cardiovascular disease. Among the most notable studies presented are the ABOARD, ARMYDA RECAPTURE, NAPLES II, PROTECT-AF, and TIPS trials.
The TIPS trial examined a group treated with a combination of three blood-pressure-lowering drugs at low doses, with a statin, aspirin, and folic acid in a single tablet (the polypill) in comparison to or to eight other groups treated with either aspirin alone, simvastatin alone, hydrochlorthiazide alone, three combinations of the two blood-pressure-lowering drugs, three blood-pressure-lowering drugs alone, or three blood-pressure-lowering drugs plus aspirin. The study enrolled apparently healthy middle-aged individuals with at least one cardiovascular risk factors, and focused on blood pressure, serum lipid levels, heart rate, and urinary thromboxane B2, also assessing its tolerability, and found that the polypill could be conveniently used to reduce multiple risk factors and cardiovascular risk, without safety or tolerability issues. The TIPS study and the story of the polypill tested in this promising pilot trial is indeed a proof of the great contribute of meta-analyses to clinical practice and research. In 2003, Wald and Law proposed, only basing their arguments on multiple meta-analyses, that a single pill containing moderate doses of different cardiovascular agents could boost compliance, safety and risk factor control, possibly leading to reductions in the risk of cardiovascular events reaching 80% (Wald and Law, BMJ 2003).
Results of the PROTECT-AF trial also warrant more attention, as this is another example of sequential design and analysis based on Bayesian methods. In this study, 800 patients with atrial fibrillation were enrolled and randomized to percutaneous left atrial appendage closure followed by 45 days of warfarin versus long-term warfarin. The study found non-inferiority of the intervential treatment for the primary efficacy end-point (including stroke), and superiority in comparison to medical therapy for the primary safety end-point (including bleeding). Whereas the impact of these promising findings on the routine clinical management of patients with atrial fibrillation remain debatable, they surely warrant further larger studies.
For those interested in Bayesian methods, important resources are the American Food and Drug Administration Draft Guidance for the Use of Bayesian Statistics, the International Society for Bayesian Analysis (ISBA), the Baysian analysis Using Gibbs Sampling (BUGS) project, and the English Bayesian methods for combining multiple Individual and Aggregate data Sources in observational studies (BIAS) project.

The SYNergy Between PCI With TAXUS and Cardiac Surgery (SYNTAX) Trial has been finally been published in the New England Journal of Medicine (2009;360:961-72) further fueling the debate on the appropriate role of percutaneous coronary revascularization versus coronary artery bypass surgery in patients with severe and/or multivessel coronary atherosclerotic disease. Despite the evident inferiority of percutaneous coronary intervention on the risk of repeat revascularization, both
Intracoronary virtual histology
Intracoronary virtual histology
revascularization techniques were associated with similar rates of death, myocardial infarction or stroke, similar rates of death, similar rates of myocardial infarcation, and similar rates of stent thrombosis or symptomatic graft occlusion. Conversely, surgery was associated with a significant increase in the occurrence of stroke.
Indeed, these findings are very important and novel, as they show that even patients with three-vessel disease or unprotected left main stenosis can be treated successfully by means of percutaneous coronary intervention. The excess of repeat revascularizations, despite disappointingly showing that first-generation paclitaxel-eluting stents are not yet the definitive coronary devices, is in line with previous evidence (eg Biondi-Zoccai et al, Ital Heart J 2003;4:271-80) for balloon-only angioplasty and bare-metal stenting.
Other important data have been reported, including the quality of life substudy of the Occluded Artery Trial (OAT, Mark et al, New Engl J Med 2009;360:774-83). In this work, the investigators suggest that late revascularization of the infarct-related occluded artery translates into improvement in quality of life at 4, 12 and 24 months (which are however statistical significant only at 4 months), that it reduces angina at 4 and 12 months, and that it reduces dyspnea at 4, 12, and 24 months. Unfortunately (but obviously) percutaneous coronary intervention appears also expensive, and, according to the authors' analyses and simulations, and, according to the authors' interpretation, not cost-effective or justified according to current conventional cost-effectiveness thresholds. Whether this applies to all individual patients remain highly debatable and hitherto unproven. This appears even more critical given the discrepant results on long-term survival and cardiac remodeling of the OAT study in comparison to other similar trials (Abbate et al, J Am Coll Cardiol 2008;51:956-64; Appleton et al, Catheter Cardiovasc Interv 2008;71:772-81).

Even if one of the coldest winter of the last years is showing no sign of easing, things are hotting up in the cardiology arena thanks to many interesting recently published studies. Adherence to statin medication and reduced all-cause mortality in primary and secondary prevention is a topic for thought in the study performed by Shalev et al (Arch Intern Med 2009;169:260-268). This study, a retrospective observational analysis of 229,918 individuals enrolled in a health-maintenance organisation, evaluated the effect of statin therapy in patients with and without pre-existing coronary heart disease. In both primary and secondary prevention cohorts, continuous compliance with statins (defined by the authors as compliance 90% of the time during the follow-up period), was associated with 45% and 51% reductions, respectively, in the risk of death, compared with less adherent patients. In our opinion these powerful results, discrepant from the evidence provided by previous clinical trials, must be interpreted cautiously. Indeed, the study design and the evaluation of treatment regimen based on dispensing information might have led to residual confounding and to a nondifferential information bias.
The results from the study of Juurlink et al (CMAJ 2009;180[7]) support suspicions on the potential harmful interaction between proton pump inhibitors and clopidogrel. In a population-based nested-case control study conducted among patients 66 years or older prescribed clopidogrel after acute myocardial infarction, the Canadian researchers showed that concomitant use of a proton pump inhibitor was associated with a significantly increased short-term risk of reinfarction. Behind these observations there is the hypothesis that some proton pump inhibitors significantly reduce or even abolish the cardioprotective effects of clopidogrel by inhibiting its bioactivation mediated by cytochrome P450 2C19 (similar to patients with loss-of-function polymorphisms).
Ypenburg and associates (J Am Coll Cardiol 2009;53:483-490), with an intriguing and well designed study, shed new light on the relationship between echocardiographically evaluated left ventricular reverse remodeling and long-term outcome. They divided 286 patients receiving cardiac resynchronization therapy (CRT) into four subgroups according to their degree of change in left ventricle end-systolic volume (LVESV) at six months and followed them up for a median of 22 months. The findings of this study highlighted the relationship between extensive left ventricular reverse remodeling and better clinical and functional improvement. Remarkably, patients who were super-responders to CRT had the most severe dyssynchrony at baseline, whether defined electrocardiographically (QRS interval prolongation and left bundle branch block [LBBB]) or echocardiographically (tissue Doppler imaging criteria).
Other evidence regarding CRT, pertinent to patients with heart failure, are derived from an insightful analysis of the COMPANION Trial published by Anand et al (Circulation 2009;119:969-977). The authors handle the tricky problem of differential follow-up times due to recurrent hospitalizations and competing risk of mortality with a sophisticated statistical method developed by Ghosh and Lin (Biometrics 2000;56:554-562). Such analysis confirmed that the use of CRT was associated with a true reduction in all-cause, cardiac and heart failure hospitalization rates. It also demonstrated that either resynchronization therapy alone (CRT-P) or in combination with a defibrillator (CRT-D) brought about a similar reduction in hospitalizations.

The new year has just begun, yet few major novel studies have been reported in the last few weeks on interventional cardiology topics.
We wish nonetheless to focus the attention of visitors of this website on 3 recent studies on genetic polymorphisms associated with responsiveness to clopidogrel, a commonly used oral antiplatelet agent (Collet et al, Lancet 2008; Mega et al, New Engl J Med, 2008; and Simon et al, New Engl J Med, 2008).
Porcine coronary artery treated with a stent
Porcine coronary artery post-stent
Collet et al reported on 259 young patients who survived a first myocardial infarction and were exposed to clopidogrel. Among them, variant 681 G>A (*2) of cytochrome P450 2C19 (a protein involved in the enzymatic activation of clopidogrel) showed to be associated death, myocardial infarction or urgent coronary revascularization, as well as stent thrombosis.
Mega et al showed among 162 healthy subjects and 1477 patients with acute coronary syndromes that carriers of at least one cytochrome P450 C19 reduced-function allele (occurring in approximately 30% of the study population) had a relative reduction of 32.4% in plasma exposure to the active metabolite of clopidogrel, and that these patients with this feature had a significant increase in the risk of death from cardiovascular causes, myocardial infarction, or stroke, as well as stent thrombosis.
Simon et al provided data on 2208 subjects with acute myocardial infarction. They found that patients carrying any two CYP2C19 loss-of-function alleles had an increased rate of death, stroke or recurrent infarction than those without such alleles. All these studies provide important evidence on the detrimental prognostic role of specific genetic determinants in patients with coronary artery disease. Whether these features remain only predictive factors or may be targets for specific therapy is however still unclear. Further insights will undoubtedly be provided by dedicated randomized clinical trials focusing on patients with adverse genetic features.
Another very intriguing work on more general cardiology topics has just been reported (Sabatine et al, Eur Heart J 2009). Specifically, Sabatine et al have performed an observational study in patients undergoing stress testing exploiting a dedicated highly sensitive assay for cardiac troponin I. They found that patients experiencing significant, yet reversible ischemia, had troponin I release, thus contrasting with a typical dogma of cardiovascular medicine, ie that only irreversible ischemia and cardiomyocyte death can cause troponin release. The impact of these findings remains to be appraised further, but the possible implications for coronary artery disease and heart failure are far reaching.