Percutaneous Coronary Interventions in Facilities without On-Site Cardiac Surgery: A Report from the National Cardiovascular Data Registry (NCDR)

1
Percutaneous Coronary Interventions in
Facilities without On-Site Cardiac Surgery A
Report from the National Cardiovascular Data
Registry (NCDR)

• ACC/SCAI i2 Summit
• Late Breaking Clinical Trials
• March 29, 2008

2
On Behalf of the National Cardiovascular Data
Registry

• Michael A. Kutcher, MD
• Lloyd W. Klein, MD
• Thomas P. Wharton, Jr., MD
• Mandeep Singh, MD, MPH
• Gregory J. Dehmer, MD
• H. Vernon Anderson, MD
• John S. Rumsfeld, MD, PhD
• William S. Weintraub, MD
• Eric D. Peterson, MD, MPH
• Fang-Shu Ou, MS
• Sarah Milford-Beland, MS
• Al Woodward, PhD. MBA
• Ralph G. Brindis, MPH

• Wake Forest University Health Sciences
• Rush University School of Medicine
• Exeter Hospital, Exeter, NH
• Mayo Clinic
• Texas AM School of Medicine
• Univ Texas Health Science, Houston
• Chief Science Officer, NCDR
• Christiana Health Care, Wilmington, DE
• Duke Clinical Research Institute (DCRI)
• DCRI
• DCRI
• NCDR
• Chief Executive Officer, NCDR

3
Special Thanks

• Jessica Morris
• Data Clarification Project Contact Staff
• Kristi Mitchell, MPH
• Data Clarification Project Coordinator
• NCDR and DCRI support staff
• Matthew Sacrinty, MS
• Wake Forest University Health Sciences
• All the hospitals and their staff that have
  committed to participate in the NCDR

4
No DisclosuresRelated to this presentation
5
Background

• There are few published large studies that have
  examined whether the procedural outcomes at PCI
  facilities that do not have surgery on-site are
  as safe and effective compared to those
  facilities that have cardiac surgery on-site.
• Wennberg DE et al. JAMA 20042921961-68.
• Ting HH et al. J Am Coll Cardiol 2006471713-21.
• Carlsson J et al. SCARR. Heart 200793335-8.

6
Background

• The National Cardiovascular Data Registry (NCDR)
  CathPCI Registry is a large ongoing multi-center
  database that offers a unique opportunity to
  provide contemporary insights into this
  controversial issue.
• Standard data sets
• Written definitions
• Uniform data entry
• Secure transmission requirements
• Data quality and auditing checks
• Risk adjustment algorithms

7
Study Population
8
Off-Site Data Clarification Project

• A Data Clarification (DC) Project was undertaken
  to address potentially ambiguous data issues
  unique to Off-Site PCI centers.
• Sites with questionable data were sent a Data
  Clarification Form (DCF) to clarify whether a
  patient transferred for CABG was elective or
  emergency and to verify eventual survival.
• An additional Off-Site Capabilities Survey (OSCS)
  was developed to gather information regarding
  organization, staffing, and logistics.

9
Off-Site Data Clarification Project

• Each Off-Site PCI program was formally contacted
  with follow-up by NCDR staff over a 4 month time
  period.
• 44 sites with 174 patients had data points that
  required verification.
• 38 sites (86) were able to fill out the DCF to
  reconcile transfer and/or mortality data on 153
  patients (88).
• 49 out of 61 sites (80) filled out the Off-Site
  Capabilities Survey (OSCS).

10
Statistical Analysis by DCRI

• Major endpoints
• In-hospital death from all causes following PCI
• Incidence of emergency surgery (version 3.04
  definitions)
• Emergency CABG performed within lt24 hours
  following PCI in which there was evidence of
  active ischemia or mechanical dysfunction.
• Emergent/Salvage patient required
  cardiopulmonary resuscitation en route to the OR
  or before anesthesia.
• Secondary endpoints
• Cerebrovascular accident
• Renal failure
• Hemorrhage
• Myocardial infarction
• Reperfusion time in cases of primary PCI

11
Off-Site Capabilities SurveyTransportation
Logistics
12
Off-Site Capabilities SurveyOrganization and
Staff
13
Institutional Characteristics
Two sites had missing CMS bed data
14
MI Presentation
15
Procedural Success and Complications
16
Observed Outcomes All PCI Patients
(Plt.0001)
(P0.3560)
(P0.8838)
(Plt.0001)
17
Risk Adjusted Outcomes
Odds Ratio (OR) outcomes for patients at On-Site
(vs. Off-Site) facilities Adjusting for site
correlations and potential confounding variables
18
Limitations

• In-hospital outcomes were analyzed long term
  follow-up was not available.
• Definitions did not discriminate whether
  emergency surgery was performed for complications
  of a PCI or whether PCI was a temporizing measure
  prior to staged surgery.
• Our study was based on a voluntary observational
  registry and a selection bias cannot be excluded.

19
Discussion

• Our study involves the largest clinical analysis
  and comparison of diverse PCI centers in the
  United States with and without on-site cardiac
  surgery support.
• The results of the Off-Site Capabilities Survey
  provides detailed information regarding the
  organization and logistics of the Off-Site PCI
  programs participating in the NCDR.

20
Conclusions

• Off-Site PCI centers participating in the NCDR
  are well organized with good logistical plans
• Dedicated staff and facilities.
• Travel time, distances, and modality of transport
  are generally within range for timely transfer to
  the off-site surgery center.
• 92 of sites provide 24/7 coverage.
• All sites are committed to provide primary PCI
  for STEMI.

21
Conclusions

• Compared to On-Site PCI centers, Off-Site PCI
  programs
• Have smaller bed capacities.
• Are predominantly located in rural and suburban
  areas.
• Have lower annual PCI volume.
• Treat a higher percentage of patients who present
  with subsets of MI (STEMI and NSTEMI).
• Have better reperfusion times in primary PCI.

22
Conclusions

• Compared to On-Site PCI centers, Off-Site PCI
  programs have similar observed
• Procedure success
• Morbidity
• Emergency CABG surgery rates
• Mortality in cases that require emergency CABG
• The risk-adjusted mortality rate in Off-Site
  facilities was comparable to those PCI centers
  that have cardiac surgery on-site.

23
Implications

• Off-Site PCI centers can provide excellent care
  to patients if the organization of the program
  is thoughtfully developed.
• The Off-Site programs in our study have
  demonstrated a strong commitment to key
  structure, process, and outcomes measurements.
  Without such a commitment, similar results may
  not be achievable.

24
Implications

• The findings of our study should not be
  extrapolated to encourage the wide-spread
  proliferation of Off-Site PCI programs.
• Our study does confirm the safety of an Off-Site
  strategy at existing PCI centers where rigorous
  clinical, operator, and institutional criteria
  are in place and are monitored to assure high
  quality outcomes.

25
Thank You
26
Backup Slides
27
Sensitivity Analysis

• Since there was some missing data for follow-up
  mortality that was not clarified, a sensitivity
  analysis was performed to assess the stability of
  the risk adjusted results.
• The analysis was comprised of 4 different models
  which imputed missing mortality to various
  potential scenarios.

28
Sensitivity Analysis



Worst case scenario Patients with missing
mortality were considered as all died Best case
scenario Patients with missing mortality were
considered as all alive
29
Sensitivity Analysis

• Although the Odds Ratio could change from 1.1 to
  0.8, the sensitivity analysis of risk adjusted
  mortality for any of the 4 models was not
  statistically significant between Off-Site versus
  On-Site facilities.
• Based on these results, the missing data would
  not have significantly affected the stability or
  the conclusions of the risk adjusted model.

30
Statistical Analysis

• Data Analysis was performed by DCRI
• To test for independence of a patients baseline
  characteristics, in-hospital care patterns and
  outcomes with respect to Off-Site vs. On-Site
  centers were analyzed.
• Mann-Whitney-Wilcoxon nonparametric tests were
  used for continuous variables.
• Pearson chi-square tests were used for
  categorical variables.

31
Statistical Analysis

• A multivariable logistic regression was utilized
  to estimate the association surgical status
  (On-Site versus Off-Site) and outcomes.
• The Generalized Estimate Equation (GEE) method
  was applied to account for within-hospital
  clustering, assuming patients at the same
  hospital are more likely to have similar
  responses relative to patients in other hospitals.

32
Variables in Risk Adjusted Mortality Model

• Age
• Gender
• Insulin treated diabetes
• Hypercholesterolemia
• Hypertension
• GFR/dialysis
• Cerebrovascular disease
• COPD
• PVD
• CHF
• Prior CABG
• Prior PCI
• Prior MI

• Cardiogenic shock
• MI presentation (STEMI, NSTEMI, no MI)
• Preoperative IABP
• PCI status (salvage, emergent, urgent, elective)
• Subacute thrombosis
• Treated left main lesion
• Treated total occlusion
• Treated lesion TIMI flow 0
• Treated lesion High/C
• Total number of lesions treated

33
Variables in Risk Adjusted Emergency Surgery Model

• Cardiogenic shock
• MI Presentation
• STEMI
• NSTEMI
• No MI
• Pre-operative IABP
• PCI status
• Salvage
• Emergent
• Urgent
• Elective
• Any treated left main lesion

34
Results
35
Patient Characteristics
36
Lesion Characteristics
37
Observed Outcomes Primary PCI Patients
(P0.9833)
(P0.1213)
(P0.9439)
(P0.9195)
38
Background

• Since the introduction of PCI in 1977 by Andreas
  Gruntzig, a preferred practice has been to have
  cardiac surgery capabilities on-site to provide
  emergency CABG in the event of life threatening
  acute procedural failures.

39
Background

• Over the last 10 years, as a result of
  improvements in technology and pharmacology
• The incidence of emergency CABG surgery for
  failed PCI is now very infrequent (0.3-0.6)
• Seshadri N et al. Circulation 20021062346-50.
• Yang EH et al. J Am Coll Cardiol 20052004-20.
• Primary PCI has been shown to be superior to
  fibrinolytic therapy for the treatment of STEMI
• Keely et al. Lancet 200336113-20.

40
Background

• These developments have formed the justification
  for some hospitals without on-site cardiac
  surgery to develop PCI programs based on a
  strategy to
• Provide more rapid and superior care for STEMI in
  the form of primary PCI
• Increase the availability of primary and elective
  PCI to patients residing in geographically
  underserved areas.

41
Background

• The safety and efficacy of performing primary PCI
  in facilities without on site surgical back-up
  has been documented in several trials.
• Wharton TP Jr. et al. J Am Coll Cardiol
  1999331257-65.
• Aversano T el. C-PORT trial. JAMA
  20022871943-51.
• Wharton TP Jr. et al. PAMI-NoSOS Study. J Am Coll
  Cardiol 2004431943-50.
• There have been numerous observational reports
  that extend the Off-Site concept to both primary
  and elective PCI.

42
Background

• The ACC/AHA/SCAI 2005 PCI Guidelines designated
  the following indications for PCI at centers that
  do not have surgery on-site
• Primary PCI Class IIb may be considered
• Elective PCI Class III not recommended
• Smith SC Jr. et al. J Am Coll Cardiol
  200647216-35.
• The 2007 Focused PCI Guideline Update did not
  address or change these designations.
• King SB III et al. J Am Coll Cardiol
  200851172-209.