Association of Physician Variation in Use of Manual Aspiration Thrombectomy With Outcomes Following Primary Percutaneous Coronary Intervention for ST-Elevation Myocardial Infarction: The National Cardiovascular Data Registry CathPCI Registry | Acute Coronary Syndromes | JAMA Cardiology | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Temporal Trends in Use of Manual Aspiration Thrombectomy During Primary Percutaneous Coronary Intervention (PCI)
Temporal Trends in Use of Manual Aspiration Thrombectomy During Primary Percutaneous Coronary Intervention (PCI)

Rates of manual aspiration thrombectomy use from quarter (Q) 3 2009 through Q2 2016 during primary percutaneous coronary intervention for patients presenting with ST-elevation myocardial infarctions. Dots represent the timing of guideline or trial electronic publication. The Thrombus Aspiration in Myocardial Infarction (TAPAS) trial, which demonstrated improvement in myocardial blush with manual aspiration thrombectomy, was published in February 2008 and as such, was not included in the Figure. Manual aspiration thrombectomy use increased from Q3 2009 (10.0%) through Q4 2011 (13.8%) which was followed by a decline. By the end of the study period in Q2 2016 manual aspiration thrombectomy was used in only 4.7% of primary percutaneous coronary interventions. ACCF indicates American College of Cardiology Foundation; AHA, American Heart Association; INFUSE-AMI, Intracoronary Abciximab and Aspiration Thrombectomy in Patients With Large Anterior Myocardial Infarction; SCAI, Society for Cardiovascular Angiography and Interventions; TASTE, Thrombus Aspiration during ST-Segment Elevation Myocardial Infarction; TOTAL, Trial of Routine Aspiration Thrombectomy with PCI versus PCI Alone in Patients With STEMI.

Figure 2.  Operator Use of Manual Aspiration Thrombectomy (AT) During Primary Percutaneous Coronary Intervention (pPCI)
Operator Use of Manual Aspiration Thrombectomy (AT) During Primary Percutaneous Coronary Intervention (pPCI)

Proportion of pPCIs among individual operators in which manual AT was used to treat patients with ST-elevation myocardial infarction. Overall preference for manual AT varied widely across individual operators (0%-83.3%).

Figure 3.  Longitudinal Outcomes Associated With Manual Aspiration Thrombectomy (AT) Use Estimated From the Instrumental Variable Analysis
Longitudinal Outcomes Associated With Manual Aspiration Thrombectomy (AT) Use Estimated From the Instrumental Variable Analysis

Outcomes associated with manual AT use estimated from the instrumental variable analysis at 30 days and 180 days. Manual AT was not associated with any outcome at 30 days or 180 days. Error bars represent 95% confidence intervals.

Table 1.  Operator Characteristics Stratified by Groups of Increasing Operator Manual Aspiration Thrombectomy Use
Operator Characteristics Stratified by Groups of Increasing Operator Manual Aspiration Thrombectomy Use
Table 2.  Patient and Procedural Characteristics of pPCIs Stratified by Groups of Increasing Operator Manual Aspiration Thrombectomy Use
Patient and Procedural Characteristics of pPCIs Stratified by Groups of Increasing Operator Manual Aspiration Thrombectomy Use
1.
Levine  GN, Bates  ER, Blankenship  JC,  et al; American College of Cardiology Foundation; American Heart Association Task Force on Practice Guidelines; Society for Cardiovascular Angiography and Interventions.  2011 ACCF/AHA/SCAI Guideline for Percutaneous Coronary Intervention: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines and the Society for Cardiovascular Angiography and Interventions.  J Am Coll Cardiol. 2011;58(24):e44-e122. doi:10.1016/j.jacc.2011.08.007PubMedGoogle ScholarCrossref
2.
O’Gara  PT, Kushner  FG, Ascheim  DD,  et al; American College of Emergency Physicians; Society for Cardiovascular Angiography and Interventions.  2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.  J Am Coll Cardiol. 2013;61(4):e78-e140. doi:10.1016/j.jacc.2012.11.019PubMedGoogle ScholarCrossref
3.
Resnic  FS, Wainstein  M, Lee  MK,  et al.  No-reflow is an independent predictor of death and myocardial infarction after percutaneous coronary intervention.  Am Heart J. 2003;145(1):42-46. doi:10.1067/mhj.2003.36PubMedGoogle ScholarCrossref
4.
Kotani  J, Nanto  S, Mintz  GS,  et al.  Plaque gruel of atheromatous coronary lesion may contribute to the no-reflow phenomenon in patients with acute coronary syndrome.  Circulation. 2002;106(13):1672-1677. doi:10.1161/01.CIR.0000030189.27175.4EPubMedGoogle ScholarCrossref
5.
Svilaas  T, Vlaar  PJ, van der Horst  IC,  et al.  Thrombus aspiration during primary percutaneous coronary intervention.  N Engl J Med. 2008;358(6):557-567. doi:10.1056/NEJMoa0706416PubMedGoogle ScholarCrossref
6.
Burzotta  F, Trani  C, Romagnoli  E,  et al.  Manual thrombus-aspiration improves myocardial reperfusion: the randomized evaluation of the effect of mechanical reduction of distal embolization by thrombus-aspiration in primary and rescue angioplasty (REMEDIA) trial.  J Am Coll Cardiol. 2005;46(2):371-376. doi:10.1016/j.jacc.2005.04.057PubMedGoogle ScholarCrossref
7.
Silva-Orrego  P, Colombo  P, Bigi  R,  et al.  Thrombus aspiration before primary angioplasty improves myocardial reperfusion in acute myocardial infarction: the DEAR-MI (Dethrombosis to Enhance Acute Reperfusion in Myocardial Infarction) study.  J Am Coll Cardiol. 2006;48(8):1552-1559. doi:10.1016/j.jacc.2006.03.068PubMedGoogle ScholarCrossref
8.
Sardella  G, Mancone  M, Bucciarelli-Ducci  C,  et al.  Thrombus aspiration during primary percutaneous coronary intervention improves myocardial reperfusion and reduces infarct size: the EXPIRA (thrombectomy with export catheter in infarct-related artery during primary percutaneous coronary intervention) prospective, randomized trial.  J Am Coll Cardiol. 2009;53(4):309-315. doi:10.1016/j.jacc.2008.10.017PubMedGoogle ScholarCrossref
9.
Vlaar  PJ, Svilaas  T, van der Horst  IC,  et al.  Cardiac death and reinfarction after 1 year in the thrombus aspiration during percutaneous coronary intervention in acute myocardial infarction study (TAPAS): a 1-year follow-up study.  Lancet. 2008;371(9628):1915-1920. doi:10.1016/S0140-6736(08)60833-8PubMedGoogle ScholarCrossref
10.
Stone  GW, Maehara  A, Witzenbichler  B,  et al; INFUSE-AMI Investigators.  Intracoronary abciximab and aspiration thrombectomy in patients with large anterior myocardial infarction: the INFUSE-AMI randomized trial.  JAMA. 2012;307(17):1817-1826. doi:10.1001/jama.2012.421PubMedGoogle ScholarCrossref
11.
Fröbert  O, Lagerqvist  B, Olivecrona  GK,  et al; TASTE Trial.  Thrombus aspiration during ST-segment elevation myocardial infarction.  N Engl J Med. 2013;369(17):1587-1597. doi:10.1056/NEJMoa1308789PubMedGoogle ScholarCrossref
12.
Jolly  SS, Cairns  JA, Yusuf  S,  et al; TOTAL Investigators.  Randomized trial of primary PCI with or without routine manual thrombectomy.  N Engl J Med. 2015;372(15):1389-1398. doi:10.1056/NEJMoa1415098PubMedGoogle ScholarCrossref
13.
Jolly  SS, Cairns  JA, Yusuf  S,  et al; TOTAL Investigators.  Outcomes after thrombus aspiration for ST elevation myocardial infarction: 1-year follow-up of the prospective randomised TOTAL trial.  Lancet. 2016;387(10014):127-135. doi:10.1016/S0140-6736(15)00448-1PubMedGoogle ScholarCrossref
14.
Levine  GN, Bates  ER, Blankenship  JC,  et al.  2015 ACC/AHA/SCAI focused update on primary percutaneous coronary intervention for patients with ST-elevation myocardial infarction: an update of the 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention and the 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Society for Cardiovascular Angiography and Interventions.  Circulation. 2016;133(11):1135-1147.PubMedGoogle ScholarCrossref
15.
Ge  J, Schäfer  A, Ertl  G, Nordbeck  P.  Thrombus aspiration for ST-segment-elevation myocardial infarction in modern era: still an issue of debate?  Circ Cardiovasc Interv. 2017;10(10):e005739. doi:10.1161/CIRCINTERVENTIONS.117.005739PubMedGoogle ScholarCrossref
16.
Vlaar  PJ, de Smet  BJGL, van den Heuvel  AFM,  et al.  Operator dependence of outcome after primary percutaneous coronary intervention.  EuroIntervention. 2011;6(6):760-767.PubMedGoogle ScholarCrossref
17.
Stukel  TA, Fisher  ES, Wennberg  DE, Alter  DA, Gottlieb  DJ, Vermeulen  MJ.  Analysis of observational studies in the presence of treatment selection bias: effects of invasive cardiac management on AMI survival using propensity score and instrumental variable methods.  JAMA. 2007;297(3):278-285. doi:10.1001/jama.297.3.278PubMedGoogle ScholarCrossref
18.
Secemsky  EA, Kirtane  A, Bangalore  S,  et al.  Practice patterns and in-hospital outcomes associated with bivalirudin use among patients with non-ST-segment-elevation myocardial infarction undergoing percutaneous coronary intervention in the United States.  Circ Cardiovasc Qual Outcomes. 2017;10(9):e003741. doi:10.1161/CIRCOUTCOMES.117.003741PubMedGoogle ScholarCrossref
19.
Secemsky  EA, Kirtane  A, Bangalore  S,  et al.  Use and effectiveness of bivalirudin versus unfractionated heparin for percutaneous coronary intervention among patients with ST-segment elevation myocardial infarction in the United States.  JACC Cardiovasc Interv. 2016;9(23):2376-2386. doi:10.1016/j.jcin.2016.09.020PubMedGoogle ScholarCrossref
20.
Brindis  RG, Fitzgerald  S, Anderson  HV, Shaw  RE, Weintraub  WS, Williams  JF.  The American College of Cardiology-National Cardiovascular Data Registry (ACC-NCDR): building a national clinical data repository.  J Am Coll Cardiol. 2001;37(8):2240-2245. doi:10.1016/S0735-1097(01)01372-9PubMedGoogle ScholarCrossref
21.
Weintraub  WS, McKay  CR, Riner  RN,  et al; American College of Cardiology Database Committee.  The American College of Cardiology National Database: progress and challenges.  J Am Coll Cardiol. 1997;29(2):459-465.PubMedGoogle ScholarCrossref
23.
Austin  PC.  Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research.  Commun Stat Simul Comput. 2009;38(6):1228-1234. doi:10.1080/03610910902859574Google ScholarCrossref
24.
McClellan  M, McNeil  BJ, Newhouse  JP.  Does more intensive treatment of acute myocardial infarction in the elderly reduce mortality? analysis using instrumental variables.  JAMA. 1994;272(11):859-866. doi:10.1001/jama.1994.03520110039026PubMedGoogle ScholarCrossref
25.
Brookhart  MA, Rassen  JA, Wang  PS, Dormuth  C, Mogun  H, Schneeweiss  S.  Evaluating the validity of an instrumental variable study of neuroleptics: can between-physician differences in prescribing patterns be used to estimate treatment effects?  Med Care. 2007;45(10)(suppl 2):S116-S122. doi:10.1097/MLR.0b013e318070c057PubMedGoogle ScholarCrossref
26.
Rassen  JA, Brookhart  MA, Glynn  RJ, Mittleman  MA, Schneeweiss  S.  Instrumental variables II: instrumental variable application-in 25 variations, the physician prescribing preference generally was strong and reduced covariate imbalance.  J Clin Epidemiol. 2009;62(12):1233-1241. doi:10.1016/j.jclinepi.2008.12.006PubMedGoogle ScholarCrossref
27.
Wimmer  NJ, Secemsky  EA, Mauri  L,  et al.  Effectiveness of arterial closure devices for preventing complications with percutaneous coronary intervention: an instrumental variable analysis.  Circ Cardiovasc Interv. 2016;9(4):e003464. doi:10.1161/CIRCINTERVENTIONS.115.003464PubMedGoogle ScholarCrossref
28.
Brookhart  MA, Rassen  JA, Schneeweiss  S.  Instrumental variable methods in comparative safety and effectiveness research.  Pharmacoepidemiol Drug Saf. 2010;19(6):537-554. doi:10.1002/pds.1908PubMedGoogle ScholarCrossref
29.
Imbens  GW, Angrist  JD.  Identification and estimation of local average treatment effects.  Econometrica. 1994;62(2):467-475. doi:10.2307/2951620Google ScholarCrossref
30.
Stock  JH, Yogo  M. Testing for weak instruments in linear iv regression (September 10, 2004): identification and inference for econometric models: essays in honor of Thomas Rothenberg. https://ssrn.com/abstract=1734933. Published 2005. Accessed October 16, 2018.
31.
Leavy  REG, Nancy  AD, Michelle  B.  Registries for Medical Devices. US: Agency for Healthcare Research and Quality; 2014.
32.
Patel  RB, Vaduganathan  M, Samman-Tahhan  A,  et al.  Trends in utilization of surrogate endpoints in contemporary cardiovascular clinical trials.  Am J Cardiol. 2016;117(11):1845-1850. doi:10.1016/j.amjcard.2016.03.021PubMedGoogle ScholarCrossref
33.
Wennberg  DE, Lucas  FL, Birkmeyer  JD, Bredenberg  CE, Fisher  ES.  Variation in carotid endarterectomy mortality in the Medicare population: trial hospitals, volume, and patient characteristics.  JAMA. 1998;279(16):1278-1281. doi:10.1001/jama.279.16.1278PubMedGoogle ScholarCrossref
34.
MacIntyre  K, Capewell  S, Stewart  S,  et al.  Evidence of improving prognosis in heart failure: trends in case fatality in 66 547 patients hospitalized between 1986 and 1995.  Circulation. 2000;102(10):1126-1131. doi:10.1161/01.CIR.102.10.1126PubMedGoogle ScholarCrossref
35.
De Carlo  M, Angelillis  M.  Aspiration thrombectomy in 2015: a TOTAL defeat?  J Thorac Dis. 2016;8(1):10-13.PubMedGoogle Scholar
36.
Diletti  R, van Mieghem  NM, Zijlstra  F.  Primary PCI with or without thrombectomy.  N Engl J Med. 2015;373(7):680. doi:10.1056/NEJMc1507110PubMedGoogle ScholarCrossref
37.
Sirker  A, Mamas  M, Kwok  CS, Kontopantelis  E, Ludman  P, Hildick-Smith  D; British Cardiovascular Intervention Society (BCIS).  Outcomes from selective use of thrombectomy in patients undergoing primary percutaneous coronary intervention for ST-segment elevation myocardial infarction: an analysis of the British Cardiovascular Intervention Society/National Institute for Cardiovascular Outcomes Research (BCIS-NICOR) Registry, 2006-2013.  JACC Cardiovasc Interv. 2016;9(2):126-134. doi:10.1016/j.jcin.2015.10.047PubMedGoogle ScholarCrossref
38.
Hizoh  I, Banhegyi  G, Domokos  D, Gulyas  Z, Major  L, Kiss  RG.  Impact of selective aspiration thrombectomy on mortality in patients with ST-segment elevation myocardial infarction.  Am J Cardiol. 2018;121(7):796-804. doi:10.1016/j.amjcard.2017.12.033PubMedGoogle ScholarCrossref
39.
Hernán  MA, Robins  JM.  Instruments for causal inference: an epidemiologist’s dream?  Epidemiology. 2006;17(4):360-372. doi:10.1097/01.ede.0000222409.00878.37PubMedGoogle ScholarCrossref
Original Investigation
January 9, 2019

Association of Physician Variation in Use of Manual Aspiration Thrombectomy With Outcomes Following Primary Percutaneous Coronary Intervention for ST-Elevation Myocardial Infarction: The National Cardiovascular Data Registry CathPCI Registry

Author Affiliations
  • 1Smith Center for Outcomes Research in Cardiology, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts
  • 2Harvard Medical School, Boston, Massachusetts
  • 3Division of Cardiology, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts
  • 4Duke Clinical Research Institute, Duke University, Durham, North Carolina
  • 5Center for Interventional Vascular Therapy, Division of Cardiology, Department of Medicine, Columbia University, New York, New York
  • 6Associate Editor, JAMA Cardiology
  • 7Center for Healthcare Delivery Innovation, Minneapolis Heart Institute, Minneapolis Heart Institute Foundation, Minneapolis, Minnesota
  • 8St Luke's Mid America Heart Institute, University of Missouri, Kansas City
JAMA Cardiol. 2019;4(2):110-118. doi:10.1001/jamacardio.2018.4472
Key Points

Question  What are the temporal trends and clinical benefits of manual aspiration thrombectomy use during primary percutaneous coronary intervention for ST-elevation myocardial infarction?

Findings  In this cohort study of 683 584 primary percutaneous coronary interventions, aspiration thrombectomy use increased from 2009 to 2011, followed by a decline through 2016. Aspiration thrombectomy use was associated with no difference in mortality, a small increase in in-hospital stroke, and no difference in cumulative events at 180 days.

Meaning  Aspiration thrombectomy use is declining nationwide, with trends in use corresponding with evolving trial data; there was no clinical benefit of selective aspiration thrombectomy use during primary percutaneous coronary intervention for ST-elevation myocardial infarction.

Abstract

Importance  Following negative randomized clinical trials, US guidelines downgraded support for routine manual aspiration thrombectomy (AT) during primary percutaneous coronary intervention (pPCI) for ST-segment elevation myocardial infarction (STEMI). However, some PCI operators continue to endorse a clinical benefit with AT use despite the lack of supportive data.

Objective  To examine temporal trends and comparative outcomes of AT use during pPCI for STEMI.

Design, Setting, and Participants  Retrospective cohort study of the National Cardiovascular Data Registry (NCDR) CathPCI Registry from July 1, 2009, to June 30, 2016, to assess temporal trends and in-hospital outcomes associated with AT use. To evaluate outcomes through 180 days, a subanalysis was conducted among Centers for Medicare and Medicaid Services–linked patients from July 1, 2009, through December 31, 2014. The comparative effectiveness analysis was performed using instrumental variable analyses to account for treatment selection bias. The instrumental variable was operator’s preference to use AT during pPCI. Data were analyzed between February 1, 2017, and April 1, 2018.

Exposures  Aspiration thrombectomy use during pPCI for STEMI.

Main Outcomes and Measures  Primary outcomes included in-hospital stroke and death. Secondary outcomes included heart failure, stroke, all-cause rehospitalization, and death through 180 days of follow-up.

Results  Among all pPCIs performed (683 584), the mean (SD) age of patients was 61.7 (12.8) years, 489 257 were male (71.6%), and 596 384 were white (87.2%). Among patients undergoing pPCI, AT use increased from 2009 through 2011, with peak use of 13.8%. This was followed by a decline of more than 9%, reaching 4.7% by mid-2016. Overall, AT was used in 10.8% of pPCIs (lowest operator group median, 0%; highest operator group median, 33.8%). After instrumental variable analysis, AT use was associated with no difference in in-hospital death (adjusted absolute risk difference, −0.18%; 95% CI, −0.53% to 0.16%; P = .29) and a small increase in in-hospital stroke (adjusted RD, 0.14%; 95% CI, 0.01%-0.30%; P = .03). Among Centers for Medicare and Medicaid Services–linked patients, AT use was not associated with differences in death, heart failure, stroke, or rehospitalization at 180 days.

Conclusions and Relevance  In this large, nationwide analysis, AT use during STEMI pPCI declined by more than 50% since 2011, with use as of mid-2016 at less than 5%. Selective AT use was associated with a small excess risk of in-hospital stroke and no difference in other outcomes through 180 days of follow-up.

Introduction

Primary percutaneous coronary intervention (pPCI) remains the treatment of choice for patients presenting with ST-elevation myocardial infarction (STEMI).1,2 However, pPCI during STEMI has been associated with embolization of thrombus, which can result in decreased microvascular perfusion, increased postprocedural myocardial infarction, and worse survival.3,4 Manual aspiration thrombectomy (AT), developed to reduce the amount of embolized material during pPCI, has shown improvement in surrogate end points, including more complete resolution of ST-segment elevation and greater tissue perfusion, compared with conventional pPCI.5-8 In addition, in a single study,9 AT use demonstrated better long-term survival. As a result, the 2011 PCI1 and 2013 STEMI2 US guidelines assigned a class IIa recommendation for routine AT use during pPCI for STEMI.

Contemporary randomized clinical trials have challenged the safety and effectiveness of routine thrombus aspiration during pPCI.10-12 The largest of these trials, the Trial of Routine Aspiration Thrombectomy With PCI vs PCI Alone in Patients With STEMI (TOTAL),12 randomized 10 732 patients with STEMI to either AT before pPCI or pPCI alone and found no differences in the primary composite clinical end point or its individual components. Furthermore, there was an increased risk of post-PCI stroke, which persisted through 1 year of follow-up.13 Consequently, the routine use of AT was downgraded to a class III recommendation in the 2015 focused update on PCI for STEMI, with selective and bailout AT use designated a class IIb recommendation.14

Despite these negative trial results and guideline changes, some operators continue to endorse more selective AT as a useful strategy when treating patients with STEMI, particularly among those with large thrombus burdens.15 It has been proposed that the negative findings from these trials were influenced by heterogeneity in the experience of the PCI operators, which may be associated with the safety of thrombus aspiration.16 In addition, mandating AT use for all patients with STEMI in randomized clinical trials, irrespective of thrombus burden, potentially diluted the clinical effectiveness of the procedure and increased the likelihood of adverse events. Although there have been calls for additional randomized clinical trials to examine selective rather than routine AT use,15 to our knowledge, no such trials are currently under way.

The use of registry data can fill voids in the clinical trial landscape, particularly when coupled with sound statistical methods. These data are also the best source of information for evaluating how clinical practice responds to changes in guidelines and trial data. In this study, we leveraged data from the National Cardiovascular Data Registry (NCDR) CathPCI registry to accomplish 2 aims. First, we sought to examine temporal trends in nationwide operator use of AT during pPCI for STEMI from 2009 through 2016 and how operator use responded to changes in guidelines and trial data that occurred during the study period. Second, we sought to evaluate the comparative safety and effectiveness of selective AT use in a broad, unselected patient population using instrumental variable analyses (IVA), a statistical method designed to control for hidden bias in observational data.17-19

Methods
Data Source and Study Population

The study population was derived from the NCDR CathPCI Registry, a national quality improvement program that collects in-hospital data on patients undergoing cardiac catheterizations and PCIs.20,21 All pPCIs performed from July 1, 2009, through June 30, 2016, for STEMI with complete data available were analyzed (eFigure, A in the Supplement). Primary PCIs were excluded if nonmanual thrombectomy devices were used or if pPCIs were performed by operators who completed less than 10 pPCIs for STEMI during the study period. To examine longitudinal outcomes, linked data between the CathPCI Registry and Centers for Medicare and Medicaid Services (CMS) data were analyzed. Centers for Medicare and Medicaid Services–linked data were available from June 1, 2009, through December 31, 2014. Similar inclusion and exclusion criteria were applied, with the additional exclusion of pPCIs associated with patients who were not fee-for-service beneficiaries and by operators who could not be linked with CMS data (eFigure, B in the Supplement). Waiver of written informed consent and authorization for this study were granted by Chesapeake Research Review Incorporated.

Patient, Procedural, Operator, and Institutional Characteristics

A full description of the data elements of the NCDR CathPCI registry can be found at the NCDR website.22 Patient and procedural characteristics were analyzed for all pPCIs. Race/ethnicity were reported as collected and classified by the NCDR CathPCI Registry. Operator characteristics included annual volumes of PCI and pPCI for STEMI. Institutional characteristics included hospital community, region, and teaching status.

Study Outcomes

The primary outcomes for this analysis were in-hospital stroke and in-hospital mortality. Stroke was defined as a documented loss of neurologic function caused by an ischemic or hemorrhagic event with residual symptoms lasting at least 24 hours after onset or leading to death. Secondary outcomes were analyzed at 30 days and 180 days and included cumulative stroke (in-hospital and out-of-hospital events); cumulative mortality (in-hospital and out-of-hospital events); postdischarge heart failure hospitalizations; and all-cause readmission. Claims codes used to identify out-of-hospital stroke and heart failure events can be found in the eMethods in the Supplement.

Statistical Analysis

Categorical variables were reported as counts and percentages and continuous variables as means and standard deviations. Kaplan-Meier methods were used to estimate longitudinal event rates. Given the large sample size, standardized differences were reported, with a threshold of at least 10% to define statistical significance.23

For the first aim, quarterly time trends of the proportion of pPCIs for STEMI that used AT were evaluated and plotted from quarter 3 of 2009 through quarter 2 of 2016. Randomized clinical trials of AT and society guidelines published during the study period were identified based on the date of electronic publication. To evaluate for heterogeneity in practice between operators, each operator’s proportional use of AT during pPCI for the total study period was assessed and depicted in a histogram.

For the second aim, an IVA was prespecified as the primary statistical method to compare the safety and effectiveness of selective AT use. Instrumental variable analyses, adopted from econometrics, have been used specifically to remove the effects of treatment selection bias in observational studies.17,24 In our analysis, this bias may be prevalent because the decision to use AT relies on a number of factors that are uncommonly collected in registry data; for example, the burden and characteristics of thrombus seen on initial angiography.15

Operator preference for AT was used as the instrumental variable.18,19,25-27 When meeting the proper assumptions, the instrumental variable assigns patients into treatment groups independent of patient or procedural characteristics. The IVA then compares patients based on the likelihood of receiving the treatment, in this case AT, instead of the actual receipt of AT, similar to an intention-to-treat analysis of a randomized clinical trial. The IVA estimates the treatment effect on the marginal population, namely, patients who would be treated with AT if they presented to a high-AT operator and would not be treated with AT if they presented to a low-AT operator. Because no operator used routine AT, ie, AT for all pPCIs (median operator AT use was 33.8%), this analysis generated a treatment effect for selective AT use. Out of concern for temporal changes in AT use, each operator’s AT use was plotted for the total study period. Only nominal variations in AT use were observed and were seen among both high-AT and low-AT users. As such, each operator’s cumulative AT use was used as the instrumental variable.

To evaluate differences in outcomes with selective AT use, an IVA using the 2-stage least square method28,29 and the PROC REG command in SAS (SAS Institute Inc) was performed. The first-stage linear model tests the strength of the proportion of AT use as an instrumental variable. The outcome of the model is actual AT use (yes/no), and the primary predictor is the proportion of AT use by the operator (continuous). From this model, predicted probabilities of the likelihood of receiving AT were generated. The instrumental variable was evaluated by the Cragg–Donald Wald F statistic of the model. A value less than 10 suggests a weak instrumental variable.30 The effectiveness of the instrument for balancing clinical characteristics was assessed by comparing characteristics of PCIs across groups of increasing operator AT use.25,28 Owing to the nonnormal distribution of AT use, groups for comparison were established based on the overall rate of AT use in the study, with cutpoints set at 0%, half the rate, the rate, and twice the rate. For the second-stage linear model, each end point served as the outcome (yes/no). The primary predictor was the instrumental variable (ie, the predicted probabilities of receiving AT as calculated by the first-stage model). For both stages of the model, patient, procedural, and operator characteristics, as listed in Tables 1 and 2, were used for adjustment.

Statistical analyses were performed using SAS software, version 9.3 (SAS Institute Inc). A 2-sided P value of less than .05 was considered significant.

Results
Temporal Trends and Operator Preference for Manual Aspiration Thrombectomy

During the study period, overall AT use during pPCI was 10.8%. The temporal trend in AT use is displayed in Figure 1. At the beginning of quarter 3 of 2009, AT use was 10.0% and peaked in quarter 4 of 2011 (13.8%), which corresponded with publication of the 2011 American College of Cardiology/American Heart Association/Society for Cardiovascular Angiography and Interventions PCI guidelines.1 Following this, there was a decline in use, which paralleled publication of the negative Thrombus Aspiration During ST-Segment Elevation Myocardial Infarction (TASTE) trial.11 Decreasing AT use continued through the end of follow-up, reaching a nadir of 4.7% by quarter 2 of 2016.

Individual operator’s proportional AT use ranged from 0% to 83.3%, with a median AT use of 3.5%, first quartile of 0% and third quartile of 13.2% (Figure 2). Among predefined groups of operator AT use, 34.1% did not perform AT during the study period (n = 2738), 23.4% used AT in more than 0% to 5.4% of pPCIs (n = 1883), 13.1% used AT in 5.5% to 10.8% of pPCIs (n = 1054), 13.5% used AT in 10.9% to 21.5% of pPCIs (n = 1087), and 15.8% used AT in 21.6% to 83.3% of pPCIs (n = 1268). The median AT use among the highest group of operators was 33.8%. Operator characteristics according to frequency of AT use are displayed in Table 1 and eTable 1 in the Supplement. There was no clear trend in annual or STEMI PCI volume with increasing operator AT use. However, compared with low-AT users, a greater proportion of high-AT users practiced in the Northeast and Midwest, at teaching hospitals, and in urban hospital centers.

Patient and Procedural Characteristics of pPCIs by Manual AT Use

eTables 2 and 3 in the Supplement display patient and procedural characteristics of pPCIs, stratified by AT treatment. Patients treated with AT were younger compared with those not treated with AT. Otherwise, there were similar demographics, comorbidities, and cardiovascular risk factors between groups. In addition, there were similar frequencies of symptomatic angina, heart failure, cardiogenic shock, and cardiac arrest at presentation. When comparing procedural characteristics, patients treated with AT more often received glycoprotein IIb/IIIa inhibitors and more commonly had single-vessel disease, intervention to a single vessel, and shorter total stent lengths (eTable 3 in the Supplement). Similar findings were seen for the CMS-linked population (eTables 4 and 5 in the Supplement).

Assessment of the Instrumental Variable

For the in-hospital IVA, the stage 1 Cragg–Donald Wald F statistic was 176 215.2 (P <.001), consistent with a strong instrumental variable. As displayed in Table 2, the instrumental variable also demonstrated effectiveness at balancing characteristics of pPCIs across operators with different frequencies of AT use. Notably, after application of the instrumental variable, there was improved balance in age, procedure medications, number of diseased vessels, and total length of stents. Among the CMS-linked population, the stage 1 Cragg–Donald Wald F statistic was 20 917.8 (P <.001). Similar balance in clinical characteristics was observed across operator AT groups (eTables 6 and 7 in the Supplement).

In-Hospital Outcomes

Between patients treated with vs without AT, crude rates of in-hospital death and stroke did not statistically differ (death, 4.51% vs 5.76%; absolute risk difference [RD], −1.25%; standardized difference, −5.70%; stroke, 0.53% vs 0.56%; RD, −0.03%; standardized difference, −0.40%). After IVA, the association between AT use and in-hospital death remained nonsignificant (adjusted RD, −0.18%; 95% CI, −0.53% to 0.16%; P = .29), but AT use was associated with an increase in the risk of stroke (adjusted RD, 0.14%; 95% CI, 0.01%-0.30%; P = .03).

Longitudinal Outcomes

Before IVA, the cumulative incidence of heart failure was similar in patients treated with vs without AT (30 days, 3.14% vs 2.88%; 180 days, 6.02% vs 5.92%; log-rank P = .72) (eTable 8 in the Supplement). In addition, the cumulative incidence of stroke did not differ between those treated with vs without AT (30 days, 1.31% vs 1.26%; 180 days, 2.20% vs 2.10%; log-rank P = .56). However, death was less frequent among those who did vs did not receive AT (30 days, 9.63% vs 11.80%; 180 days, 12.90% vs 16.10%; log-rank P < .001), as was all-cause readmissions (30 days, 14.80% vs 15.40%; 180 days, 30.40% vs 32.20%; log-rank P = .007).

After IVA, there were no statistically significant associations observed between AT treatment and any of the outcomes, either at 30 days or 180 days (Figure 3). In particular, the adjusted RD for death at 180 days was 0.22% (95% CI, −0.47% to 0.90%; P = .54), for stroke was −0.08% (95% CI, −1.65% to 1.48%; P = .92), for heart failure was 1.05% (95% CI, −1.15% to 3.25%; P = .35), and for all-cause readmission was 0.55% (95% CI, −0.53% to 1.64%; P = .32).

Discussion

In this large, nationwide analysis of patients undergoing pPCI for STEMI, AT use declined by more than 50% since 2011, with use as of mid-2016 less than 5%. This change in clinical practice followed closely the publication of negative randomized clinical trial data. Notwithstanding these population-level trends, operator preferences for AT varied substantially, with some operators choosing never to use AT during STEMI pPCI and other operators using AT in more than 20% of cases. After adjustment for potential confounding using IVA, there was no evidence of benefit with selective AT use. Conversely, we detected a small increase in the rate of in-hospital stroke, which was attenuated with longitudinal follow-up.

Clinical registries play an important role in the study of device outcomes and physician behavior in a real-world setting. Currently, registries are used for various needs, including prospective device surveillance, fulfillment of postmarket observational study commitments, and evaluation of device safety and effectiveness.31 These databases are critical for understanding how device performance in randomized clinical trials compares with use and outcomes in routine clinical practice.

The temporal change in AT use observed in our registry study has important implications. We found that clinical practice was meaningfully influenced by the publication of randomized clinical trial data. For instance, we noted a near 50% increase in the use of AT in the 2 years following the publication of the Thrombus Aspiration in Myocardial Infarction (TAPAS) trial5 (not included in Figure 1 because it was published in February 2008). The TAPAS trial demonstrated a significant improvement in its primary end point of myocardial blush grade with routine AT use during STEMI. Additionally, we observed a substantial decline in AT use following the publication of the randomized TASTE trial,11 which failed to demonstrate a mortality benefit with AT use. This trend was perpetuated by the larger randomized TOTAL trial,12 which also found no difference in its primary effectiveness end point.

Although the incorporation of contemporary trial data into clinical practice is important, there may be potential negative consequences, in particular when the available trials use surrogate end points. For instance, the TAPAS trial met its primary end point of improving myocardial blush grades, but did not meet its secondary clinical end points of target vessel revascularization, death, or major cardiovascular events at 30 days.5 Only in a 1-year follow-up study were differences in these end points detected.9 Nonetheless, AT use increased until publication of the TASTE trial,11 which showed no associated mortality benefit. The heterogeneity in physician AT use observed in this study may in part reflect the differential influence of these contemporary trial data on physician behavior. Surrogate end points have been increasingly used in randomized clinical trials as a method to decrease costs and length of follow-up, as well as to expedite device approval. However, there has been variable association between surrogate end points and clinical outcomes.32

Another insight from our study is that updates in US guidelines lagged changes in physician behavior. For instance, the American College of Cardiology/American Heart Association/Society for Cardiovascular Angiography and Interventions 2011 PCI guidelines that assigned routine AT use a Class IIa indication1 were released after AT achieved peak use. Similarly, the American College of Cardiology/American Heart Association/Society for Cardiovascular Angiography and Interventions 2015 STEMI PCI update14 that downgraded routine AT use during STEMI to a IIb recommendation was published after a more than 9% fall in AT use. A potential downside of this delayed association is difficulty in relying on registry data to assess physician compliance with guideline recommendations or to measure health care quality. Finding ways to better streamline the writing and publication of future guidelines, as well as releasing more frequent updates to incorporate evolving data, would likely help overcome these trends.

In addition to postapproval monitoring, the use of registry data to perform comparative effectiveness research provides a critical addition to the armamentarium of data needed to evaluate devices and therapies. This is particularly relevant because multiple studies have demonstrated differences between the results of clinical trials and results in actual clinical practice.33,34 In addition, data from registries have been used to fill in gaps for decision makers31 and to study patient populations excluded from clinical trials. Statistical methods used to analyze nonrandomized data have become more sophisticated, and as a result, there has been increasing interest and investment in registries for comparative effectiveness research.31

For this study, we performed a comparative analysis of AT use to fulfill 2 unmet needs. First, we evaluated AT use in a broad, unselected patient and operator population in order to assess the generalizability of trial results. Second, we investigated the association between selective AT use and outcomes because the obligation to use AT in all pPCIs when randomized in a trial has been implicated as a major cause of negative results.15 Furthermore, clinicians have attempted to justify the continued use of AT based on the possibility of clinical benefits that were unexamined in trials conducted within the past few years, such as reductions in hospitalizations owing to heart failure,35 or that may become apparent with greater longitudinal follow-up.36

Using IVA to account for unmeasured confounders, we found no clinical benefit of selective AT use in patients undergoing pPCI for STEMI, including no differences in in-hospital or long-term death, all-cause readmissions, or postdischarge heart failure admissions. Moreover, we found a small increase in the risk of in-hospital stroke, yet no evidence of a cumulative stroke risk at 180 days. These findings are generally consistent with previously published observational analyses of selective AT use,37,38 as well as randomized clinical trials of routine AT use adequately powered to examine clinical end points.11,12

Limitations

The results of this study must be considered in the context of its design. First, this analysis evaluated selective AT use under the assumption that AT was used in clinical scenarios in which the operator felt the patient might benefit from this device, such as high thrombus burden. However, thrombus burden and other patient characteristics that may lead to AT use are not collected in the CathPCI registry. Notably, in our analytic approach, these omissions would only bias the study results if their prevalence differed substantially between operators. Second, we were unable to determine whether AT was used upfront or as a bailout strategy during pPCI. As such, there may remain clinical scenarios in which superselective AT use may provide clinical benefit in which this analysis was not designed to detect. Third, our analysis in part focused on unambiguous clinical end points, including death and readmission, as events in the CathPCI Registry are not adjudicated. However, to be consistent with trial data, we also reported stroke and heart failure hospitalizations, and these outcomes may be susceptible to misclassification bias.

The IVA may also have its own limitations, particularly if operators who used AT more or less often differed in other unmeasured ways. For instance, operator characteristics were not completely balanced across increasing AT groups. In attempt to overcome this, operator characteristics were adjusted for in both stages of the IVA, yet it remains possible that there may be residual influence between operator variables, such as care setting, and the outcomes. Furthermore, the causal interpretation of IVAs rests on certain assumptions that cannot always be verifiable. One issue with preference-based IVAs is that of monotonicity, which assumes that at each operator preference threshold, no operator who prefers AT would use AT unless all operators who prefer AT would do so.39 At the extreme tails of operator preference, this is likely to hold true. However, one must assume that the concept of monotonicity is maintained when differences between operator preferences are more modest. In addition, it may be difficult to identify the marginal population in clinical practice, but this represents the efficacy of use among patients for whom high-use operators would choose AT, while low-use operators would not.

Conclusions

In this nationwide registry analysis, AT use during pPCI for STEMI has declined since mid-2011. This change in physician behavior closely followed trial data demonstrating a lack of benefit and possible risk with AT. Furthermore, in this broad, unselected patient population, selective AT use during pPCI for STEMI was not found to be associated with clinical benefit. Our findings have 2 important implications. First, these data demonstrate an association of physician practice patterns with evolving trial data and support the use of registry data for evaluating postapproval physician behavior. Second, our comparative analysis helps fill a gap in the data regarding the effectiveness of selective AT use during pPCI for STEMI.

Back to top
Article Information

Corresponding Author: Eric A. Secemsky, MD, Smith Center for Outcomes Research in Cardiology, Beth Israel Deaconess Medical Center, 375 Longwood Ave, 4th Floor, Boston, MA 02215 (esecemsk@bidmc.harvard.edu).

Accepted for Publication: November 10, 2018.

Published Online: January 9, 2019. doi:10.1001/jamacardio.2018.4472

Author Contributions: Drs Secemsky and Yeh had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. The statistical analysis was conducted at the Duke Clinical Research Institute by coauthors Ms Zakroysky and Mr Wojdyla.

Study concept and design: Secemsky, Yeh.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Secemsky, Ferro, Zakroysky.

Critical revision of the manuscript for important intellectual content: Secemsky, Ferro, Rao, Kirtane, Tamez, Wojdyla, Bradley, Cohen, Yeh.

Statistical analysis: Secemsky, Tamez, Zakroysky, Wojdyla.

Obtained funding: Secemsky.

Administrative, technical, or material support: Secemsky, Ferro.

Study supervision: Secemsky, Rao, Yeh.

Conflict of Interest Disclosures: Dr Kirtane reports research grants from Abbott Vascular, Abiomed, Boston Scientific, CathWorks, CSI, Medtronic, Philips, Recor, Siemens, and Spectranetics. Dr Cohen reports consultant fees from Edwards Lifesciences and Medtronic and research grants from Abbott Vascular, AstraZeneca, Biomet, Boston Scientific, CSI, Corvia, Daiichi-Sankyo, Edwards Lifesciences, Eli Lilly, Medtronic, Merck, and V-Wave Medical. Dr Yeh reports consultant fees from Abbott Vascular, Asahi Intecc, Boston Scientific, and Medtronic; research grants from Abiomed and Boston Scientific; and salary from Baim Institute for Clinical Research. Dr Rao reports consultant fees from Amgen, Boehringer Ingelheim, Corindus, CSI, and Medtronic. No other disclosures were reported.

Funding/Support: This analysis was funded by the National Cardiovascular Data Registry.

Role of the Funder/Sponsor: The funding organization had a role in the collection, management, analysis, and interpretation of the data and the approval of the manuscript but not the preparation or review of the manuscript and the decision to submit the manuscript for publication.

Disclaimer: Dr Kirtane is Associate Editor of JAMA Cardiology, but he was not involved in any of the decisions regarding review of the manuscript or its acceptance.

References
1.
Levine  GN, Bates  ER, Blankenship  JC,  et al; American College of Cardiology Foundation; American Heart Association Task Force on Practice Guidelines; Society for Cardiovascular Angiography and Interventions.  2011 ACCF/AHA/SCAI Guideline for Percutaneous Coronary Intervention: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines and the Society for Cardiovascular Angiography and Interventions.  J Am Coll Cardiol. 2011;58(24):e44-e122. doi:10.1016/j.jacc.2011.08.007PubMedGoogle ScholarCrossref
2.
O’Gara  PT, Kushner  FG, Ascheim  DD,  et al; American College of Emergency Physicians; Society for Cardiovascular Angiography and Interventions.  2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.  J Am Coll Cardiol. 2013;61(4):e78-e140. doi:10.1016/j.jacc.2012.11.019PubMedGoogle ScholarCrossref
3.
Resnic  FS, Wainstein  M, Lee  MK,  et al.  No-reflow is an independent predictor of death and myocardial infarction after percutaneous coronary intervention.  Am Heart J. 2003;145(1):42-46. doi:10.1067/mhj.2003.36PubMedGoogle ScholarCrossref
4.
Kotani  J, Nanto  S, Mintz  GS,  et al.  Plaque gruel of atheromatous coronary lesion may contribute to the no-reflow phenomenon in patients with acute coronary syndrome.  Circulation. 2002;106(13):1672-1677. doi:10.1161/01.CIR.0000030189.27175.4EPubMedGoogle ScholarCrossref
5.
Svilaas  T, Vlaar  PJ, van der Horst  IC,  et al.  Thrombus aspiration during primary percutaneous coronary intervention.  N Engl J Med. 2008;358(6):557-567. doi:10.1056/NEJMoa0706416PubMedGoogle ScholarCrossref
6.
Burzotta  F, Trani  C, Romagnoli  E,  et al.  Manual thrombus-aspiration improves myocardial reperfusion: the randomized evaluation of the effect of mechanical reduction of distal embolization by thrombus-aspiration in primary and rescue angioplasty (REMEDIA) trial.  J Am Coll Cardiol. 2005;46(2):371-376. doi:10.1016/j.jacc.2005.04.057PubMedGoogle ScholarCrossref
7.
Silva-Orrego  P, Colombo  P, Bigi  R,  et al.  Thrombus aspiration before primary angioplasty improves myocardial reperfusion in acute myocardial infarction: the DEAR-MI (Dethrombosis to Enhance Acute Reperfusion in Myocardial Infarction) study.  J Am Coll Cardiol. 2006;48(8):1552-1559. doi:10.1016/j.jacc.2006.03.068PubMedGoogle ScholarCrossref
8.
Sardella  G, Mancone  M, Bucciarelli-Ducci  C,  et al.  Thrombus aspiration during primary percutaneous coronary intervention improves myocardial reperfusion and reduces infarct size: the EXPIRA (thrombectomy with export catheter in infarct-related artery during primary percutaneous coronary intervention) prospective, randomized trial.  J Am Coll Cardiol. 2009;53(4):309-315. doi:10.1016/j.jacc.2008.10.017PubMedGoogle ScholarCrossref
9.
Vlaar  PJ, Svilaas  T, van der Horst  IC,  et al.  Cardiac death and reinfarction after 1 year in the thrombus aspiration during percutaneous coronary intervention in acute myocardial infarction study (TAPAS): a 1-year follow-up study.  Lancet. 2008;371(9628):1915-1920. doi:10.1016/S0140-6736(08)60833-8PubMedGoogle ScholarCrossref
10.
Stone  GW, Maehara  A, Witzenbichler  B,  et al; INFUSE-AMI Investigators.  Intracoronary abciximab and aspiration thrombectomy in patients with large anterior myocardial infarction: the INFUSE-AMI randomized trial.  JAMA. 2012;307(17):1817-1826. doi:10.1001/jama.2012.421PubMedGoogle ScholarCrossref
11.
Fröbert  O, Lagerqvist  B, Olivecrona  GK,  et al; TASTE Trial.  Thrombus aspiration during ST-segment elevation myocardial infarction.  N Engl J Med. 2013;369(17):1587-1597. doi:10.1056/NEJMoa1308789PubMedGoogle ScholarCrossref
12.
Jolly  SS, Cairns  JA, Yusuf  S,  et al; TOTAL Investigators.  Randomized trial of primary PCI with or without routine manual thrombectomy.  N Engl J Med. 2015;372(15):1389-1398. doi:10.1056/NEJMoa1415098PubMedGoogle ScholarCrossref
13.
Jolly  SS, Cairns  JA, Yusuf  S,  et al; TOTAL Investigators.  Outcomes after thrombus aspiration for ST elevation myocardial infarction: 1-year follow-up of the prospective randomised TOTAL trial.  Lancet. 2016;387(10014):127-135. doi:10.1016/S0140-6736(15)00448-1PubMedGoogle ScholarCrossref
14.
Levine  GN, Bates  ER, Blankenship  JC,  et al.  2015 ACC/AHA/SCAI focused update on primary percutaneous coronary intervention for patients with ST-elevation myocardial infarction: an update of the 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention and the 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Society for Cardiovascular Angiography and Interventions.  Circulation. 2016;133(11):1135-1147.PubMedGoogle ScholarCrossref
15.
Ge  J, Schäfer  A, Ertl  G, Nordbeck  P.  Thrombus aspiration for ST-segment-elevation myocardial infarction in modern era: still an issue of debate?  Circ Cardiovasc Interv. 2017;10(10):e005739. doi:10.1161/CIRCINTERVENTIONS.117.005739PubMedGoogle ScholarCrossref
16.
Vlaar  PJ, de Smet  BJGL, van den Heuvel  AFM,  et al.  Operator dependence of outcome after primary percutaneous coronary intervention.  EuroIntervention. 2011;6(6):760-767.PubMedGoogle ScholarCrossref
17.
Stukel  TA, Fisher  ES, Wennberg  DE, Alter  DA, Gottlieb  DJ, Vermeulen  MJ.  Analysis of observational studies in the presence of treatment selection bias: effects of invasive cardiac management on AMI survival using propensity score and instrumental variable methods.  JAMA. 2007;297(3):278-285. doi:10.1001/jama.297.3.278PubMedGoogle ScholarCrossref
18.
Secemsky  EA, Kirtane  A, Bangalore  S,  et al.  Practice patterns and in-hospital outcomes associated with bivalirudin use among patients with non-ST-segment-elevation myocardial infarction undergoing percutaneous coronary intervention in the United States.  Circ Cardiovasc Qual Outcomes. 2017;10(9):e003741. doi:10.1161/CIRCOUTCOMES.117.003741PubMedGoogle ScholarCrossref
19.
Secemsky  EA, Kirtane  A, Bangalore  S,  et al.  Use and effectiveness of bivalirudin versus unfractionated heparin for percutaneous coronary intervention among patients with ST-segment elevation myocardial infarction in the United States.  JACC Cardiovasc Interv. 2016;9(23):2376-2386. doi:10.1016/j.jcin.2016.09.020PubMedGoogle ScholarCrossref
20.
Brindis  RG, Fitzgerald  S, Anderson  HV, Shaw  RE, Weintraub  WS, Williams  JF.  The American College of Cardiology-National Cardiovascular Data Registry (ACC-NCDR): building a national clinical data repository.  J Am Coll Cardiol. 2001;37(8):2240-2245. doi:10.1016/S0735-1097(01)01372-9PubMedGoogle ScholarCrossref
21.
Weintraub  WS, McKay  CR, Riner  RN,  et al; American College of Cardiology Database Committee.  The American College of Cardiology National Database: progress and challenges.  J Am Coll Cardiol. 1997;29(2):459-465.PubMedGoogle ScholarCrossref
23.
Austin  PC.  Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research.  Commun Stat Simul Comput. 2009;38(6):1228-1234. doi:10.1080/03610910902859574Google ScholarCrossref
24.
McClellan  M, McNeil  BJ, Newhouse  JP.  Does more intensive treatment of acute myocardial infarction in the elderly reduce mortality? analysis using instrumental variables.  JAMA. 1994;272(11):859-866. doi:10.1001/jama.1994.03520110039026PubMedGoogle ScholarCrossref
25.
Brookhart  MA, Rassen  JA, Wang  PS, Dormuth  C, Mogun  H, Schneeweiss  S.  Evaluating the validity of an instrumental variable study of neuroleptics: can between-physician differences in prescribing patterns be used to estimate treatment effects?  Med Care. 2007;45(10)(suppl 2):S116-S122. doi:10.1097/MLR.0b013e318070c057PubMedGoogle ScholarCrossref
26.
Rassen  JA, Brookhart  MA, Glynn  RJ, Mittleman  MA, Schneeweiss  S.  Instrumental variables II: instrumental variable application-in 25 variations, the physician prescribing preference generally was strong and reduced covariate imbalance.  J Clin Epidemiol. 2009;62(12):1233-1241. doi:10.1016/j.jclinepi.2008.12.006PubMedGoogle ScholarCrossref
27.
Wimmer  NJ, Secemsky  EA, Mauri  L,  et al.  Effectiveness of arterial closure devices for preventing complications with percutaneous coronary intervention: an instrumental variable analysis.  Circ Cardiovasc Interv. 2016;9(4):e003464. doi:10.1161/CIRCINTERVENTIONS.115.003464PubMedGoogle ScholarCrossref
28.
Brookhart  MA, Rassen  JA, Schneeweiss  S.  Instrumental variable methods in comparative safety and effectiveness research.  Pharmacoepidemiol Drug Saf. 2010;19(6):537-554. doi:10.1002/pds.1908PubMedGoogle ScholarCrossref
29.
Imbens  GW, Angrist  JD.  Identification and estimation of local average treatment effects.  Econometrica. 1994;62(2):467-475. doi:10.2307/2951620Google ScholarCrossref
30.
Stock  JH, Yogo  M. Testing for weak instruments in linear iv regression (September 10, 2004): identification and inference for econometric models: essays in honor of Thomas Rothenberg. https://ssrn.com/abstract=1734933. Published 2005. Accessed October 16, 2018.
31.
Leavy  REG, Nancy  AD, Michelle  B.  Registries for Medical Devices. US: Agency for Healthcare Research and Quality; 2014.
32.
Patel  RB, Vaduganathan  M, Samman-Tahhan  A,  et al.  Trends in utilization of surrogate endpoints in contemporary cardiovascular clinical trials.  Am J Cardiol. 2016;117(11):1845-1850. doi:10.1016/j.amjcard.2016.03.021PubMedGoogle ScholarCrossref
33.
Wennberg  DE, Lucas  FL, Birkmeyer  JD, Bredenberg  CE, Fisher  ES.  Variation in carotid endarterectomy mortality in the Medicare population: trial hospitals, volume, and patient characteristics.  JAMA. 1998;279(16):1278-1281. doi:10.1001/jama.279.16.1278PubMedGoogle ScholarCrossref
34.
MacIntyre  K, Capewell  S, Stewart  S,  et al.  Evidence of improving prognosis in heart failure: trends in case fatality in 66 547 patients hospitalized between 1986 and 1995.  Circulation. 2000;102(10):1126-1131. doi:10.1161/01.CIR.102.10.1126PubMedGoogle ScholarCrossref
35.
De Carlo  M, Angelillis  M.  Aspiration thrombectomy in 2015: a TOTAL defeat?  J Thorac Dis. 2016;8(1):10-13.PubMedGoogle Scholar
36.
Diletti  R, van Mieghem  NM, Zijlstra  F.  Primary PCI with or without thrombectomy.  N Engl J Med. 2015;373(7):680. doi:10.1056/NEJMc1507110PubMedGoogle ScholarCrossref
37.
Sirker  A, Mamas  M, Kwok  CS, Kontopantelis  E, Ludman  P, Hildick-Smith  D; British Cardiovascular Intervention Society (BCIS).  Outcomes from selective use of thrombectomy in patients undergoing primary percutaneous coronary intervention for ST-segment elevation myocardial infarction: an analysis of the British Cardiovascular Intervention Society/National Institute for Cardiovascular Outcomes Research (BCIS-NICOR) Registry, 2006-2013.  JACC Cardiovasc Interv. 2016;9(2):126-134. doi:10.1016/j.jcin.2015.10.047PubMedGoogle ScholarCrossref
38.
Hizoh  I, Banhegyi  G, Domokos  D, Gulyas  Z, Major  L, Kiss  RG.  Impact of selective aspiration thrombectomy on mortality in patients with ST-segment elevation myocardial infarction.  Am J Cardiol. 2018;121(7):796-804. doi:10.1016/j.amjcard.2017.12.033PubMedGoogle ScholarCrossref
39.
Hernán  MA, Robins  JM.  Instruments for causal inference: an epidemiologist’s dream?  Epidemiology. 2006;17(4):360-372. doi:10.1097/01.ede.0000222409.00878.37PubMedGoogle ScholarCrossref
×