To Transfer or Not to Transfer? The debate between transfer for PCI versus local thrombolysis.

1
To Transfer or Not to Transfer?The debate
between transfer for PCI versus local
thrombolysis.

• Todd Ring, BSc., MD, CCFP
• March 11, 2004
• University of Calgary
• Emergency Medicine Grand Rounds

2
An Area of Controversy
3
Overview

• Is PCI better than thrombolysis?
• Evidence behind transfer for PCI
• Is transfer safe?
• Is timing everything?
• Issues closer to home

4
Rationale

• Minority of patients with AMI present directly to
  PCI center
• Reality most present to EMS or local hospital
  (non PCI center)
• Results from prior trials comparing PCI to local
  thrombolysis difficult to extrapolate to non PCI
  center
• Treatment bias
• Center and operator experience
• Effect of treatment delay unknown

5
Is PCI Better?
6
Primary Angioplasty Versus Intravenous
Thrombolytic Therapy for AMI A Quantitative
Review of 23 Randomized Trials. The Lancet 36.
2003

• Meta-analysis of 23 RCT
• 7739 TL eligible patients
• 3872 PCI
• 3867 TL (67 TPA)
• Short (4 6 week) and long term (6 18 month)
  outcomes

7
p lt .0001
p .003 (excluding SHOCK)
8
Major bleed only sig. negative result for PCI
Similar results short and long term favouring PCI
9
Problems with Evidence Favoring PCI

• If SHOCK data is excluded and look at subgroup
  receiving aTPA
• Mortality 5.5 PCI vs 6.7 TL p .08
• Definition of re-infarction
• Majority of cases of re-infarction in TL group
  occurs in 1st hour
• At this time patients many patients still in cath
  lab demonstrating low flow, spasm, dissection,
  distal embolization
• Only 2 large trials gt1000 pts 15 trials lt 200
  pts
• No weighting of outcome data

10
Conclusions Regarding PCI

• Evidence favours PCI over all forms of
  thrombolysis
• ? Evidence is not as convincing at it may appear
• Bias from pro-lytic and pro-interventionalists
• TL has higher complications of stroke and
  re-infarction and PCI higher bleeding risks
• Both groups agree that even despite the large
  number of trials confirmation in a large trial
  comparing mortality for PCI vs. modern quick
  infusion TL is needed

11
What is the Evidence Supporting Transfer for PCI?
12
PRAGUE Multicenter RCT comparing PCI vs. TL vs.
combined strategy for patients with AMI
presenting to a community hospital. EHJ 21. 2000

• 1st randomized study to compare transfer for PCI
  vs. thrombolysis June 97 March 99
• 17 community referral centers 4 PCI centers
• Patients randomized into one of three groups
• Group A TL at local hospital remained at local
  hospital
• Group B TL en route angiography?angioplasty if
  necessary
• Group C transfer for PCI

13
PRAGUE

• 1588 pts with STEMI/new LBBB 300 randomized
• Within 6h Sx onset
• Endpoints combined end point (CEP)
  death/re-infarction/stroke
• Transport distance 5 75 km

14
Re-infarction rate only sig. result
15
(No Transcript)
16
(No Transcript)
17
PRAGUE Discussion/Limitations

• Trial only enrolled 300/1588 eligible patients
• lt 6h from Sx onset
• Transport time lt60min distance lt75km
• CEP largely driven by re-infarct
• Support data regarding experienced labs/operators
• No evidence to support facilitated PCI

18
PRAGUE-2 Long distance transport for PCI vs.
immediate thrombolysis for AMI. EHJ 24. 2003

• Based on results of PRAGUE and LIMI (Vermeer)
  larger, nationwide, 30 d mortality as primary
  endpoint
• Sept 99 Jan 02
• 41 community hospitals and 7 PCI centers
• 4853 patients with MI 850 randomized (target
  sample 1200)
• 2 groups
• TL streptokinase (remain in first hospital)
• PCI transport to tertiary center PCI

19
PRAGUE-2

• Based on safety concerns regarding treatment
  delay subgroup analysis
• lt 3 h and 3 12 h after Sx onset
• Transport distance 5 120 km
• Study prematurely stopped
• 2.5 fold excess mortality in TL group treated gt3h

20
p lt .02
Trend p .12
No difference
21
PRAGUE 2Discussion/Limitations

• No difference in lt 3 h group
• Distance lt 120 km
• Streptokinase TL agent
• TL patients remained at local hospital
• Physician at local hospital could elect to send
  patients directly for PCI
• One reason trial stopped early
• ? Source of bias

22
DANAMI A comparison of coronary angioplasty
with fibrinolytic therapy in AMI. NEJM 349 (8).
2003

• Danish trial Dec97 Oct 01
• 24 referral centers 5 PCI centers
• 62 Danish population
• 2 groups
• TL remained at local hospital
• PCI
• CEP (death, re-infarct, stroke) at 30 days
• Distance 3 150 km (mean 50 km)

23
DANAMI

• 2 concurrent study groups
• Referral hospital
• Invasive
• 1527 pts?1129 from 24 referral hospitals
• ? 443 from 5 invasive centers

24
CEP driven by 75 reduction in re-infarction,
BUT 30 day mortality 24 (re-infarct) vs. 6.5
NNT 18
NNT 17
25
(No Transcript)
26
DANAMI Repeat Revascularization

• 26 of 782 patients (3.3 ) in TL group underwent
  repeat TL within 12 h 15 (1.9) rescue
  angioplasty
• Over 30 days of follow up 148 (18.9) of patients
  in TL vs. 72 (9.1) of PCI underwent mechanical
  revascularization (plt.001)

27
DANAMI Discussion/Limitations

• Primary endpoint CEP
• Excluded high risk patients
• ? Benefit most
• Short transport distance
• Only 2/5 PCI centers performed PCI prior to study
• ? Greater benefit than reported
• Sickest patients not transported

28
CAPTIM Primary angioplasty vs. prehospital
fibrinolysis in AMI a randomised study. The
Lancet 360. 2002

• Randomized, multi-center trial based in France
  June 97 Sept 01
• 840 patients (1200 plannedlack of funding)
• 27 hospitals and associated EMS
• Presented within 6 h
• Two groups
• Pre-hospital fibrinolysis (419) alteplase
• PCI (421)
• Primary endpoint CEP (death, re-infarct, stroke)
  at 30 d

29
Mortality benefit favouring TL group (trend)
CEP favouring PCI group (trend)
30
Trend (p .29)
Trend (p .61)
Mortality rates significantly lower than other
trials
31
CAPTIM Discussion

• Low mortality rate in TL group
• Early TL, transfer to invasive center, liberal
  rescue angioplasty (25 patients rescue
  angioplasties), low risk patients
• Benefit of early TL
• Mortality reduction if treated lt 2 h (57)
• 2.2 TL vs. 5.7 PCI (p.04)
• Well equipped ambulance
• ACLS crew/physician on board

32
Limitations to Generalizations

• 26 of patients need rescue angioplasty
• Only 4 of ambulance calls for CP are STEMI
  eligible for TL
• ½ of patients with STEMI drive themselves to
  hospital
• Physician in ambulance

33
Transfer for Primary Angioplasty Versus Immediate
Thrombolysis in Acute Myocardial Infarction A
Meta-Analysis. Circulation 108. 2003

• 6 RCTs identified from Jan 85 Sept 02
• 3 significantly favour transfer for PCI and 3
  non-significant or no trend
• 2 limited by sample size, 1 a feasibility study,
  1 hampered by recruitment
• Primary endpoint CEP
• Excluded trials or arms of facilitated PCI

34
(No Transcript)
35
NNT 30
NS (with CAPTIM) RR .76 p.03 (Exclude CAPTIM)
36
NNT 33
NNT 86
37
Conclusion Effectiveness of Transport for PCI

• Overall PCI probably the best option
• BUT not always achievable
• TL effective in early MI gt 6h largely
  ineffective ? Very early TL as effective as PCI
• With TL significant number of patients will need
  to go on to further angiography/plasty
• Need to consider other issues surrounding
  transport
• Safety
• Timing/Distance
• Availability

38
Safety and Quality of Transport
39
(No Transcript)
40
Quality of PCI

• Impact of Routine Duty Hours vs. Off Hours JACC
  41(12). 2003
• 1,702 consecutive patients at one center
• Failure rate 3.8 (routine) vs. 6.9 (off) p
  lt.01
• Mortality rate (30d) 1.9 vs. 4.2 p lt.01
• Relationship between volume and mortality JAMA
  284(24). 2000
• NRMI database
• Mortality rate PCI vs. TL high volume 3.4 vs 5.4
  plt.001 intermediate 4.5 vs. 5.9 p lt.001 low
  volume 6.2 vs. 5.9
• More experienced operators shorter door-balloon
  times

41
Conclusions Regarding Safety/Quality

• Transport appears to be safe
• Quick and dirty vs. slow and clean
• Quick and dirty fast but only basic equipment
• Slow and clean fully equiped slow deployment
  times
• All studies to date some form of slow and clean
• No studies look at safety of long transports
• PRAGUE2 longest transports but highest death and
  adverse events
• Impact of off hours and cath lab volumes likely
  to affect quality

42
Is Timing Everything?
43
Relationship of Symptom Onset to Balloon Time and
Door to Balloon Time with Mortality in Patients
Undergoing Angioplasty for AMI. JAMA 83(22). 2000

• Prospective observational study of data collected
  in the Second National Registry of MI
• 27,080 consecutive patients with STEMI/ new LBBB
• Only 2230 (8 of patients) underwent PCI within
  60 min of presentation
• In-hospital mortality rate 4.2
• gt 3 h mortality rate 8.5 mortality

44
Relationship of Symptom Onset to Balloon Time and
Door to Balloon Time with Mortality in Patients
Undergoing Angioplasty for AMI. JAMA 83(22). 2000

• Performed logistic regression to adjust for
  baseline differences
• Door to balloon time greater than 2h 41 62
  increased risk of death
• Confounding was serious concern in door to
  balloon times in this study
• Shorter time men, younger, non-DM
• Propensity analysis door to balloon time longer
  than 2h still increased risk of death (28 vs.
  41 62 )
• Did not find an increased mortality associated
  with prolonged Sx onset to balloon time

45
Percutaneous Coronary InterventionVersus
Fibrinolytic Therapy in AMI Is Timing (Almost)
Everything? AJC 92. 2003

• Meta-regression analysis of the Grines
  meta-analysis comparing PCI and TL
• Assess the impact of time delay
• Endpoints were 4-6 week incidence of death and
  CEP of death, re-infarction and stroke
• As PCI related time delay increased, mortality
  reduction favouring PCI decreased
• .94 reduction for every 10 minute delay
• 2 strategies equal after PCI delay of 62 min
• CEP equivalence occurred at 93 min

46
Prague Prague 2 DANAMI
Randomization to Treatment (time)
PCI 95 82 90
TL 22 12 20
Difference 73 70 70
Symptom Onset to Treatment (time)
PCI 215 277 224
TL 132 185 169
Difference 83 92 55
47
Clinical Characteristics and Outcome of Patients
with Early, Intermediate and Late Presentation
Treated by PCI and TL for AMI. EHJ 23. 2002

• 2635 patients in 10 RCTs
• Presentation delay associated with older age,
  female, DM, increased HR
• CEP (death, re-infarction, stroke) at 30d for PCI
  vs. TL
• Early (lt2h) group 5.8 vs. 12.5
• Int. (2-4h) group 8.6 vs. 14.2
• Late (gt4h) group 7.7 vs. 19.4
• With increase in time to presentation adverse
  events increase in TL group (p lt.04) but not in
  the PCI group (p gt.4)

48
Conclusions Regarding Timing

• PCI superior at all time points
• AHA goal TL door-needle time lt 30 min gt 6h
  ineffective
• AHA goal PCI door-balloon time 90 min /- 30 min
• Evidence from transport trials supports
  feasibility with respect to timing
• With increasing delay (60 90 min) for transport
  for PCI mortality benefit may be lost
• ? Most beneficial group late presenters (gt3 6h)

49
Issues Closer to Home

• Limited availability of tertiary care centers in
  Canada (lt 10 of all hospitals)
• Large geographic area
• Substantial disparities in the quality of
  ambulance and pre-hospital services
• Tertiary care center variability
• Cost Effectiveness
• 10,711 PCI vs. 13,664 TL

50
Questions to Ask?

1. What is the time from Sx onset to medical
   contact?
2. What is the risk associated with this MI?
3. What are the risks of TL?
4. What are the risks of transport?
5. What is the time to PCI?

51
Conclusions

• Regardless of strategy early reperfusion
  paramount
• In early presenters TL should not be delayed for
  PCI especially in those patients at low risk for
  TL complications
• Each center must decide which strategy is best
• Transport appears safe but adverse events during
  transport can occur and need to be considered
• Time delays need to be minimized
• Door to needle lt 30min door to balloon lt 90min