OUP user menu

★ fast track ★

A comparison of pharmacologic therapy with/without timely coronary intervention vs. primary percutaneous intervention early after ST-elevation myocardial infarction: the WEST (Which Early ST-elevation myocardial infarction Therapy) study

Paul W. Armstrong
DOI: http://dx.doi.org/10.1093/eurheartj/ehl088 1530-1538 First published online: 6 June 2006


Aims Uncertainty exists as to which reperfusion strategy for ST-elevation myocardial infarction (MI) is optimal. We evaluated whether optimal pharmacologic therapy at the earliest point of care, emphasizing pre-hospital randomization and treatment was non-inferior to expeditious primary percutaneous coronary intervention (PCI).

Methods and results Which Early ST-elevation myocardial infarction Therapy (WEST) was a four-city Canadian, open-label, randomized, feasibility study of 304 STEMI patients (>4 mm ST-elevation/deviation) within 6 h of symptom onset, emphasizing pre-hospital ambulance treatment and participation of community and tertiary care centres. All received aspirin, subcutaneous enoxaparin (1 mg/kg), and were randomized to one of three groups: (A) tenecteplase (TNK) and usual care, (B) TNK and mandatory invasive study ≤24 h, including rescue PCI for reperfusion failure, and (C) primary PCI with 300 mg loading dose of clopidogrel. Time from symptom onset to treatment was rapid (to TNK for A=113 and B=130 min and for PCI in C=176 min). The primary outcome, a composite of 30-day death, re-infarction, refractory ischaemia, congestive heart failure, cardiogenic shock, and major ventricular arrhythmia, was 25% (Group A), 24% (Group B), and 23% (Group C), respectively. However, there was a higher frequency of the combination of death and recurrent MI in Group A vs. Group C (13.0 vs. 4.0%, respectively, P-logrank=0.021), yet no difference between Group B (6.7%, P-logrank=0.378) and C.

Conclusion These data suggest that a contemporary pharmacologic regimen rapidly delivered, coupled with a strategy of regimented rescue and routine coronary intervention within 24 h of initial treatment, may not be different from timely expert PCI.

  • ST-elevation myocardial infarction
  • Percutaneous coronary intervention
  • Fibrinolytic pharmacologic reperfusion
See page 1511 for the editorial comment on this article (doi:10.1093/eurheartj/ehl107)


Few issues in contemporary cardiovascular medicine have been subjected to more sustained and vigorous debate as the optimal approach to life-saving reperfusion therapy in patients with acute ST-segment elevation myocardial infarction (STEMI). Once thrombotic occlusion of a major epicardial coronary artery was established as the cause of STEMI, parallel developments in pharmacological and mechanical reperfusion strategies occurred. Each has been shown to successfully salvage left ventricular (LV) myocardium, improve LV function, and enhance long-term survival.

The proponents of pharmacologic therapy for STEMI highlight its widespread availability to the broad cross-section of patients worldwide, the lack of dependence on operator experience or institutional resources, and the ability to rapidly apply such therapy in the emergency department or even prior to hospital arrival without the sometimes-formidable logistics of transfer to an interventional facility.13 Dissatisfaction with pharmacologic therapy is primarily related to: (i) suboptimal reperfusion rates, (ii) the presence of recurrent ischaemia and re-infarction, and (iii) the risk of intracranial and systemic bleeding complications.4

Enthusiasm for primary percutaneous coronary intervention (PCI) has been fuelled in part by a systematic overview indicating that primary PCI was more effective than fibrinolysis in the therapy of STEMI.5 Although a significant mortality difference was evident favouring PCI, the largest contribution of PCI's advantage over fibrinolysis was a reduction in re-infarction. In addition several methodological concerns have been raised about this report leading some to consider it as hypothesis-generating rather than a definitive basis for a change in practice.1,6

Given the uncertainty as to which strategy is optimal, and the appreciation of the importance of mechanical co-intervention for patients with failed fibrinolysis, we designed a feasibility study to compare outcomes in patients who underwent expeditious primary PCI with those receiving pharmacologic therapy at the earliest point of care, with an emphasis on pre-hospital randomization and treatment. Also, within the cohort receiving pharmacologic therapy, standard care was applied to half of the group whereas the other half underwent mandatory invasive management within 24 h of enrolment including protocol-specified rescue PCI.


Patient population

The Which Early ST-elevation MI Therapy (WEST) study involved four metropolitan Canadian communities (Edmonton, Halifax, Montreal, and Vancouver). Patients with STEMI in whom reperfusion therapy (primary PCI, fibrinolysis or transfer for rescue PCI) was feasible within 3 h of randomization were enrolled. The protocol emphasized expedited care with ECG, randomization and therapy undertaken pre-hospital where possible, and direct communication to PCI teams to enhance their state of readiness.

Eligible patients were male or non-pregnant females (≥18 years) with symptoms presumed secondary to STEMI lasting at least 20 min accompanied by ECG evidence of high risk. These included: ≥2 mm of ST-elevation in two or more contiguous precordial leads or limb leads; or ≥1 mm ST-elevation in two or more limb leads coupled with ≥1 mm ST-depression in two or more contiguous precordial leads (total ST-deviation ≥4 mm) or presumed new left bundle branch block. Patients were excluded if primary PCI was deemed to be available within 1 h of diagnosis, or if contraindications to fibrinolysis, prior coronary bypass grafting (CABG), or glycoprotein IIb/IIIa antagonist use within 7 days existed.

Study design, treatments, and endpoints

We intended to enrol 100 patients in each of the three treatment arms in this feasibility study. All patients received aspirin (160–325 mg) and subcutaneous enoxaparin (1 mg/kg) at randomization with subsequent use recommended every 12 h for a minimum of 72 h; additional intravenous enoxaparin (0.3–0.5 mg/kg) was permitted during PCI in Group C and its use post-PCI was discretionary. Patients were randomized in open label fashion to one of three treatment groups. Group A received weight-adjusted tenecteplase (TNK) followed by the usual standard of care. Group B also received weight-adjusted TNK but underwent mandatory invasive management within 24 h of enrolment including protocol-specified rescue PCI, if the admission ST-elevation failed to decrease by ≥50% at 90 min after TNK therapy or if haemodynamic or electrical instability occurred. Group C patients underwent primary PCI with a clopidogrel 300 mg loading dose administered along with ASA and enoxaparin on study entry. Abciximab was recommended for Group C and for use in all PCI procedures in each treatment group unless performed within 3 h of fibrinolytic therapy. Clopidogrel use was employed in Groups A and B according to ACC/AHA PCI guidelines.

The primary efficacy endpoint of this study was a 30-day composite of death, re-infarction, refractory ischaemia, congestive heart failure, cardiogenic shock and major ventricular arrhythmia. The individual components were also examined. Definitions are provided in Appendix 2. Two experienced observers blinded to treatment allocation undertook central adjudication of re-infarction, refractory ischaemia, major ventricular arrhythmia, and the indications for rescue PCI.

Secondary efficacy endpoints included 90- and 180 min ST-resolution (according to the Schroeder method) and the infarct size was assessed using the Selvester QRS score and peak creatinine kinase (CK).7,8 Core laboratories were established where blinded assessments were undertaken for ECG, NT-pro-brain natriuretic peptide (NT-proBNP) and angiographic measurements as noted in Appendix 3.

Safety outcomes included the occurrence of intracranial haemorrhage, disabling stroke, and major systemic bleeding. Finally, we assessed the composite of the primary efficacy and safety endpoints.

The protocol was approved by each institutional Ethics Review Board and a Data and Safety Monitoring Board oversaw the study. WEST was registered at www.ClinicalTrials.gov (identifier NCT00121446) on 13 July 2005.

Statistical analysis

Data were analysed according to intention-to-treat principles and are presented as proportions for categorical variables and as medians and interquartile ranges (IQR) for continuous variables. Data were compared across the treatment groups using χ2 tests, Fisher's exact test, and Kruskal–Wallis tests where appropriate. Repeated measures techniques were used to analyse the variance of NT-proBNP measures across the three collection periods (baseline, 72 h and 30 days).

The WEST study was designed to test the non-inferiority of Groups A and B relative to Group C. In the event of no difference in the primary efficacy composite endpoint between the two fibrinolytic arms, i.e. Groups A and B, we pre-specified a comparison between their combined result and the primary PCI arm, i.e. Group C. Comparisons were undertaken using confidence intervals (CI) with the expectation that a relative minimally important difference (rMID) in the incidence of the primary endpoint of 15% would be considered as evidence for the non-inferiority of pharmacologic therapy relative to primary PCI.

Kaplan–Meier estimates were plotted for the primary efficacy endpoint and 30-day death/re-MI according to the treatment groups. Unadjusted Cox regression with 90% CI was also used in the survival analysis of 30-day death and death/re-MI. Rates of the primary efficacy endpoint were also risk-adjusted to explore the possibility of imbalance of baseline characteristics among the study treatment groups. Using logistic regression techniques, two approaches were applied to develop a model for the primary efficacy endpoint: (i) simplified model based on age, heart rate, systolic blood pressure, Killip class, and anterior MI (top five predictors in the GUSTO I 30-day mortality model);9 (ii) full model considering all clinically and statistically significant predictors of the outcome. The primary efficacy endpoints were then adjusted by the observed vs. expected ratio.

Pre-specified subgroup analyses included pre-hospital vs. inhospital randomization and time from symptom onset to randomization (≤2 h vs. >2 h).

The data were housed and analysed at the University of Alberta. All analyses were performed using SPSS 13.0 with the exact tests module (SPSS Inc., Chicago, IL, USA).


In Table 1, the baseline characteristics for the three treatment groups are shown. Key baseline characteristics were well balanced across the three treatment groups, although patients in Group B were slightly younger, had fewer anterior MIs and were more frequently in Killip Class I. As illustrated in Table 2 there was excellent adherence to protocol-mandated pharmacotherapy, including near universal use of clopidogrel and 97% use of abciximab in the 91 patients undergoing primary angioplasty in Group C. In contrast, 48% of the 91 patients undergoing PCI within 24 h of randomization in Group B received abciximab. In Table 3, median times from symptom onset to protocol-mandated therapy and procedures are shown for the overall population, as well as for those patients randomized pre-hospital (n=121) vs. in-hospital (n=183). Once randomization occurred, study drugs were administered expeditiously and the overall median time from symptom onset to PCI in Group C was rapid at 176 min. Group C patients randomized pre-hospital also received their PCI ~1 h earlier than those randomized in hospital.

View this table:
Table 1

Baseline patient characteristics according to study treatment groups (intention to treat)

Intention to treatAll
Group AGroup BGroup C
Age, median (IQR) (years)58 (51–69)57 (50–67)60 (49–71)58 (50–69)
Gender (female)25 (25.0)19 (18.3)18 (18.0)62 (20.4)
Hypertension (yes)37 (37.0)56 (53.8)*33 (33.0)126 (41.4)
Diabetes mellitus, yes (total)18 (18.0)8 (7.7)16 (16.0)42 (13.8)
Family history of early CAD (yes)39 (39.0)44 (42.3)34 (34.0)117 (38.5)
History of angina (yes)25 (25.0)31 (29.8)18 (18.0)74 (24.3)
Previous MI (yes)14 (14.0)12 (11.5)12 (12.0)38 (12.5)
Previous PCI (yes)9 (9.0)8 (7.7)5 (5.0)22 (7.2)
Smoking status (current smoker)45 (45.0)53 (51.0)38 (38.0)136 (44.7)
MI location on Q-ECG (anterior)42 (42.0)37 (35.6)42 (42.0)121 (39.8)
Killip class
 I93 (93.9)99 (97.1)93 (95.9)285 (95.6)
 II6 (6.1)3 (2.9)4 (4.1)13 (4.4)
Systolic BP, median (IQR) (mmHg)140 (12–160)140 (128–160)145 (124–160)141 (124–160)
Diastolic BP, median (IQR) (mmHg)83 (71–92)85 (71–98)80 (70–96)83 (70–96)
Pulse, median (IQR) (bpm)75 (65–85)72 (62–84)75 (62–88)74 (63–86)
Weight, median (IQR) (kg)78 (63–87)84 (71–97)80 (67–90)80 (67–91)
Height, median (IQR) (cm)169 (162–177)173 (166–180)173 (162–178)173 (163–178)
  • *P=0.006.

View this table:
Table 2

Summary of protocol-mandated pharmacotherapy

Group AGroup BGroup CAll
Given within 12 h or at index event99 (99.0)104 (100)98 (98.0)301 (99.0)
Clopidogrel given instead of ASA1 (1.0)0 (0.0)2 (2.0)3 (1.0)
Discharged on ASA91 (91.0)101 (97.1)92 (92.0)284 (93.4)
Discontinued prematurely7 (7.0)4 (3.8)4 (4.0)15 (4.9)
Given at index event98 (98.0)102 (98.1)98 (98.0)298 (98.0)
Discontinued prematurely35 (35.7)39 (38.2)13 (13.0)87 (29.2)
 Bleeding [n (%) of those discontinued prematurely]3 (8.6)5 (12.8)2 (13.3)10 (11.2)
 Physician discretion23 (65.7)24 (61.5)7 (53.8)54 (62.1)
 Other9 (25.7)10 (25.6)3 (23.1)22 (25.3)
Given at randomization98 (98.0)103 (99.0)
Given at index event97 (97.0)
View this table:
Table 3

Median (IQR) minutes from symptom (Sx) onset to treatment in all patients and according to pre-hospital vs. in-hospital randomization

Group AGroup BGroup C
Overall (n=304)100104100
Sx onset to randomization105 (63–158)114 (67–172)100 (70–160)
Sx onset to first study druga113 (74–179)130 (75–185)112 (80–164)
Sx onset to PCI395 (294–3711), n=58425 (288–1331), n=81b176 (140–280), n=91
Sx onset to rescue PCI299 (270–325), n=14277 (213–381), n=29
Sx onset to non-rescue PCI1498 (341–5465), n=44926 (398–1454), n=52
First medical contactc to first study druga51 (37–75)54 (38–77)54 (36–69)
First medical contactc to PCI350 (245–3561)324 (218–1216)127 (93–159)
Pre-hospital (n=121)424237
Sx onset to randomization85 (63–147)76 (50–114)72 (49–98)
Sx onset to first study drug91 (70–156)91 (57–142)82 (58–103)
Sx onset to PCI364 (273–2056), n=27371 (215–912), n=31140 (115–171), n=33
Sx onset to rescue PCI274 (254–351), n=8217 (193–342), n=15
Sx onset to non-rescue PCI484 (282–3076), n=19810 (393–1343), n=16
First medical contactc to first study druga46 (35–55)44 (36–57)41 (32–53)
First medical contactc to PCI331 (223–1720)245 (196–856)104 (88–126)
In-hospital (n=183)586263
Sx onset to randomization105 (74–160)135 (91–178)124 (89–185)
Sx onset to first study drug119 (86–180)146 (102–189)137 (103–201)
Sx onset to PCI1441 (303–5494), n=31532 (338–1414), n=50207 (167–292), n=58
Sx onset to rescue PCI302 (293–333), n=6343 (275–419), n=14
Sx onset to non-rescue PCI3216 (376–5547), n=251137 (406–1474), n=36
First medical contactc to first study druga61 (43–83)67 (42–92)61 (40–85)
First medical contactc to PCI1396 (260–5403)606 (259–1308)143 (100–169)
  • aFirst study drug refers to TNK in Groups A and B and to enox in Group C.

  • bProtocol-mandated procedure.

  • cFirst medical contact refers to ambulance arrival or hospital arrival.

Of the Group B patients, 102 underwent cardiac catheterization within 24 h, 89 of whom received in-hospital revascularization with one additional patient revascularized by 30 days—of these eight patients received CABG. In Group C, 98 patients underwent angiography, 91 underwent in-hospital PCI, and 93 patients had revascularization by 30 days—of these three patients had CABG. Revascularization was also performed in-hospital in 60 Group A patients and by 30 days in 65 patients.

Protocol-mandated rescue angioplasty was undertaken in 28% of Group B patients. An additional nine patients in Group B met the ECG criteria for early rescue: three of these had early angiography without PCI, four had angiography and PCI within the 24 h post-randomization window, and two patients did not undergo angiography. 14 patients (14%) in Group A underwent rescue PCI, a median of 197 min after randomization (IQR 172–280 min). Coronary stents were used in over 97% of all patients undergoing PCI. Initial angiographic findings prior to PCI in Group C patients following ASA, enoxaparin, and 300 mg clopidogrel loading revealed that 30% had TIMI 2 or 3 flow. In the 90 patients who underwent PCI (with complete angiographic data), 90% had TIMI 3 and 7% had TIMI 2 flow at the conclusion of their procedure. In the 85 Group B patients undergoing PCI, 81% had TIMI 3 and 12% had TIMI 2 flow at the conclusion of their procedure.

In Table 4, the primary efficacy and safety endpoints are shown. No statistically significant differences were observed in the primary composite endpoint or any of its components. The 30-day mortality rate was generally low, particularly in Groups B and C. Group B tended to have lower rates of heart failure and cardiogenic shock, whereas the lowest rate of recurrent MI was in Group C. There were no intracranial haemorrhages, infrequent major systemic bleeding, and no significant safety differences between treatment groups.

View this table:
Table 4

Efficacy and safety endpoints

Group AGroup BGroup CAll
Primary efficacy endpoint at 30 days25 (25.0)25 (24.0)23 (23.0)73 (24.0)
 Death4 (4.0)1 (1.0)1 (1.0)6 (2.0)
 Re-MI9 (9.0)6 (5.8)3 (3.0)18 (5.9)
 Heart failure15 (15.0)15 (14.4)18 (18.0)48 (15.8)
 Cardiogenic shock6 (6.0)4 (3.8)7 (7.0)17 (5.6)
 Refractory ischaemia0 (0.0)3 (2.9)0 (0.0)3 (0.1)
 Major ventricular arrhythmias1 (1.0)1 (1.0)1 (1.0)3 (1.0)
Primary safety endpoint during index hospitalization
 Intracranial haemorrhage0000
 Non-haemorrhagic stroke01 (1.0)1 (1.0)2 (0.1)
 Major systemic bleeding1 (1.0)2 (1.9)1 (1.0)4 (1.3)
Composite of primary efficacy and safety endpoint26 (26.0)26 (25.0)24 (24.0)76 (25.0)

Figure 1 demonstrates the Kaplan–Meier curves for the 30-day composite primary efficacy endpoint. Given that there was no significant difference in the primary efficacy endpoint between Groups A and B (25 vs. 24%) these groups were combined when assessing the non-inferiority between fibrinolysis and primary PCI. As shown in Figure 2, the relative difference in the primary efficacy endpoint between Groups A, B and Group C fell within the 15% rMID (Figure 3), however, the 90% confidence limits were broad and hence, this finding is not definitive. The primary efficacy endpoint, risk-adjusted for the only significant predictor i.e. age, was similar across the three treatment groups (Group A: 25.0%; Group B: 24.6%; Group C: 22.5%). No significant mortality difference existed between groups {Group B vs. A: unadjusted hazard ratio (HR) 0.23 [90% CI (0.04–1.46)]; Group C vs. A: 0.25 (0.04–1.54); Group B vs. C: 0.95(0.09–9.7)}. In Figure 3, the composite of 30-day death and recurrent MI is shown. Patients in Group A were significantly more likely to experience a combination of death and recurrent MI than in Group C (Group C vs. A: unadjusted HR 0.29, 90% CI(0.11–0.74); P-logrank=0.021) whereas there was no difference between Group B and C (Group B vs. C: 1.73 (0.62–4.8); P-logrank=0.378).

Figure 1

Kaplan–Meier curves of the primary efficacy endpoint according to study treatment groups. Group A (n=100) is represented by the small dashed line; Group B (n=104) by the solid line; Group C (n=100) by the large dashed line.

Figure 2

Relative difference in the primary efficacy endpoint between Groups A and B combined and Group C with 90% confidence limits and pre-specified 15% rMID boundary.

Figure 3

Kaplan–Meier curves of 30-day death/re-MI according to study treatment groups. Group A (n=100) is represented by the small dashed line, Group B (n=104) by the solid line, Group C (n=100) by the large dashed line. Group C vs. A was statistically significant (logrank P=0.021).

In the nearly 40% of patients randomized to treatment in the pre-hospital setting, 22.3% experienced the primary efficacy endpoint compared with 25.1% in those randomized upon arrival to hospital (P=0.587). There was no statistically significant difference across the treatment groups in either setting.

As pre-specified, the primary efficacy endpoint occurred in 38 (20.7%) of the 184 patients who were randomized ≤2 h of symptom onset, compared to 35 (29.4%; P=0.09) of the 119 patients randomized beyond the 2 h mark.This difference was especially evident within Group C patients where there was a two-fold increase in rates in patients randomized beyond 2 h [16.4% (≤2 h) vs. 34.2% (>2 h); P=0.052].

In Table 5, the ECG data are shown. The baseline sum of ST-deviation was similar across the treatment groups. At 180 min after randomization there was a trend towards more patients achieving complete, i.e. 70% resolution in Group B as compared with Groups A and C. Discharge ECG QRS scores indicated a trend (P=0.14) towards a higher score i.e. a greater% of LV infarction in Group C. Median peak (IQR) CK data within the three treatment groups revealed IU values of 1199(548–2351), 1590(771–2624), and 1833(852–3649) for Groups A, B, and C respectively, P=0.045.

View this table:
Table 5

Electrocardiographic data: ST-deviation, ST-resolution, and QRS score

% of LVGroup AGroup BGroup C
Sum of ST-deviation
 Baseline11.8 (7.5–17)10 (6.5–15.5)11.3 (7–16.5)
 90 min
  ≥50% resolution62 (68.9)59 (60.8)
  ≥70% resolution39 (43.3)44 (45.4)
180 min
 ≥50% resolution58 (73.4)71 (83.5)62 (78.5)
 ≥70% resolution48 (60.8)59 (69.4)44 (55.7)
QRS score on discharge ECG
 0–10–320 (20.0)15 (14.4)10 (10.0)
 2–46–1227 (27.0)36 (34.6)22 (22.0)
 >4>1238 (38.0)41 (39.4)49 (49.0)

Figure 4 illustrates the median NT-proBNP in pg/mL acquired at baseline, 72 h and 30 days. No significant difference in baseline median NT-proBNP measures was observed across the three treatment groups. In the repeated measure analysis, which was based on 184 patients with complete NT-proBNP at all 3 collection times, there was a significant trend in the mean NT-proBNP values from baseline to 30 days (P<0.001) such that an increase was observed from baseline to 72 h but then a decline at 30 days. NT-proBNP values were significantly higher in Group C relative to Group A (P=0.019) and to a lesser extent, Group B (P=0.092).

Figure 4

Median NT-proBNP in pg/mL at baseline, 72 h, and 30 days according to study treatment groups. Group A (n=72–84) is represented by the small dashed line; Group B (n=75–83) by the solid line; Group C (n=70–87) by the large dashed line. NT-proBNP values were significantly higher in Group C both at 72 h (P=0.03) and 30 days (P=0.007) relative to Group A.


Our study of STEMI patients focusing on early reperfusion treatment provides several novel findings. Reperfusion therapy with a contemporary pharmacologic regimen, coupled with a strategy of regimented rescue and routine invasive evaluation within 24 h of treatment, provides an excellent standard of care that is not different from timely expert PCI at experienced centres. We found no difference in the primary composite outcome across the three treatment groups even after adjustment for minor baseline differences. Moreover, there were no intracranial haemorrhages and a small but comparable incidence of non-haemorrhagic stroke and major bleeding.

What might account for our findings that differ from most prior comparisons in larger populations? Our strategy of establishing the diagnosis of STEMI at the first medical contact contributed to expedited randomization with a median time of 105 min from symptom onset. Early diagnosis was followed by expedited treatment, followed by triage and transport in the pre-hospital population. This contributed to the remarkably short times to both fibrinolysis and PCI treatments rarely achieved in prior STEMI clinical trials. The approximate 46 min interval between fibrinolytic treatment in Group B and primary PCI delivery in Group C is well aligned with prior evidence suggesting similar outcomes from both reperfusion methods evaluated in a systematic overview of prior comparisons.10 Hence, the tendency towards more shock and CHF in Group C is of interest and reminiscent of the findings in the Comparison of Angioplasty and Pre-hospital Thrombolysis In Acute Myocardial Infarction trial (CAPTIM) where patients treated with fibrinolysis <2 hrs from symptom onset had lower mortality and cardiogenic shock than those randomized to PCI.11 It is also aligned with the reduced prevalence of cardiogenic shock in the pre-hospital fibrinolysis vs. primary PCI group observed in the French nationwide USIC 2000 Registry.12 These suggestive clinical trends are further supported by the higher peak CK data, discharge ECG QRS scores and elevated NT-proBNP values at days 3 and 30 found in Group C. Mandatory systematic and timely rescue PCI in Group B performed in 28% of patients may have further contributed to their favourable outcomes and is supported by the trend towards superior ST-resolution at 180 min.

Our study was not powered to show a definitive impact on mortality or the combination of death and MI. Notwithstanding this, our findings are consistent with the notion that a strategy of early fibrinolysis coupled with routine early invasive management (or timely rescue PCI, if warranted) results in rates of death and re-infarction comparable with those achieved with direct PCI.11 Whereas prior comparative studies frequently indicate that re-MI constitutes the principal efficacy advantage for PCI this difference is known to be mitigated by the use of enoxaparin, clopidogrel, and timely mechanical co-intervention.1316 Interestingly, patients in Group A with later and lower rescue rates and less frequent overall revascularization appeared to fare less well than direct PCI patients, especially relating to the composite of death and re-MI. This finding is consistent with the GRACIA-1 study where patients randomized six hours after fibrinolysis to angiography and intervention within 24 h vs. an ischaemia-guided conservative approach tended to have a reduced rate of death or re-infarction.17 It is noteworthy that coronary interventions in GRACIA-1 occurred a mean of 17 h after fibrinolysis whereas in ASSENT-4PCI a strategy of immediate PCI (1–3 h) after TNK was associated with higher in-hospital mortality, cardiac ischaemic complications, and stroke as compared with those who received direct PCI alone.18 Hence, the optimal frequency and timing of co-intervention in patients receiving fibrinolysis remains uncertain but systematic adherence to STEMI rescue guidelines is likely a key factor. Our data are also well aligned with the recently reported REACT study.16 In that trial, when systematic rescue PCI was employed early after failed fibrinolysis in STEMI patients, their composite outcome of death, re-MI, cerebrovascular accident, and severe heart failure was significantly superior to those receiving repeat fibrinolysis or conservative therapy.

Overall times from symptom onset to randomization were well balanced and short in our study, and the expected advantage of pre-hospital randomization was also evident. Hence, our strategy of pre-hospital randomization and treatment not only substantially shortened the time to pharmacologic therapy but also abbreviated the time to PCI by over 1 h by ensuring enhanced readiness at the receiving PCI centres.

Our choice and dosing of enoxaparin deserves discussion given the findings of ASSENT 3 and 3 PLUS.19,20 We chose to omit the IV enoxaparin bolus and allow for supplemental IV dosing in the setting of PCI. Given the short time lapse from initial dosing of enoxaparin, the time to adequate anticoagulation was likely a more important issue in Group C.21 However, systematic use of abciximab, known to favourably affect PCI outcomes in such patients was also employed.

Our study has some limitations that should be noted: because a screening log was not maintained and it was unblinded we cannot rule out bias. However the blind adjudication of clinical endpoints as well as the core assessments of ECG, angiographic, NT-proBNP and CK data conducted without the knowledge of treatment assignment, make this unlikely. Because of the modest sample size, we cannot exclude the play of chance. For us to have reached a definitive non-inferiority conclusion (based on a 15% rMID and 90% CI with a composite endpoint of 24.0% in Group B and 23% in Group C), a trial would need to enrol 1625 patients in each arm or 2578 patients in each arm (based on a similar rMID and a 24.5% composite endpoint in a combined Group A and B vs. 23% in Group C).

In summary, our investigation provides novel data supporting the efficacy, safety and feasibility of a strategic pharmacologic approach that combines fibrinolysis with timely catheter co-intervention for patients who can be assessed early after symptom onset. This strategy is especially relevant to both pre-hospital and community hospital settings and its use in the current study should provide encouragement to regions where an integrated systems approach to the management of STEMI has not yet been undertaken. Although such a strategy requires 24 h/7 day access to interventional facilities, it may provide a more reasonable balance as it relates to the timing and frequency of their use. In the light of the failure of routine immediate intervention after fibrinolytic therapy recently reported in ASSENT-4 PCI coupled with persisting delays in accessing primary PCI, an approach focusing on pre-hospital care and other methods to ensure early treatment and timely and effective post-fibrinolytic rescue seems well positioned to provide a useful therapeutic option deserving of more definitive large scale investigation.18,22

Our findings may particularly assist informing the design of such an initiative.


We would like to thank the many nurses, paramedical personnel, technical and administrative staff that contributed to WEST and the patients for their volunteer spirit. The study was supported by unrestricted research grants from Hoffman-La Roche and sanofi-aventis Canada and also Eli Lilly Canada.

Supported by an unrestricted grant from Hoffmann-LaRoche Limited, and Aventis Pharma, a member of the sanofi-aventis Group, as well as Eli-Lilly.

Conflict of interest: P.W.A. has received speaker honoraria from Hoffman LaRoche and sanofi aventis; I.B. has received speaker honoraria from Hoffman LaRoche; C.B. has received speaker honoraria from Hoffman LaRoche; R.G. has received speaker honoraria from sanofi aventis; B.O. has received speaker honoraria from Hoffman LaRoche; P.T. has received speaker honoraria from Hoffman LaRoche and sanofi aventis; R.W. has received speaker honoraria from Hoffman LaRoche and sanofi aventis.


Appendix 1

WEST Steering Committee

Paul W. Armstrong (Chair)

Iqbal Bata

Christopher E. Buller

Edward Cain

James W. Christenson

Richard Gallo

G.B. John Mancini (Director of Core Angiographic Laboratory)

Blair O'Neill

David Petrie

Pierre Theroux (Director of Biomarker Core Laboratory)

Andrew Travers

John Webb

Robert C. Welsh

Biostatistical Analysis: Cynthia M. Westerhout and Wei-Ching Chang

Project Management: Monica Adam and Wanda Sutherland

ECG Core Laboratory Yuling Fu

Data Management: EPICORE Centre, University of Alberta

Editorial Assistance Durenda Tremblay


Appendix 2

Endpoint definitions

  1. Refractory ischaemia: Symptoms of ischaemia with ST-deviation or definite T-wave inversion persisting for at least 10 min despite medical management while in hospital.

  2. Recurrent MI (myocardial re-infarction):

    1. In the first 18 h after randomization:

      1. Recurrent signs and symptoms of ischaemia at rest accompanied by new or recurrent ST-segment elevations of ≥0.1 mV in at least two contiguous leads lasting≥30 min.

    2. After 18 h:

      1. New Q-waves (by Minnesota Code Criteria) in two or more leads and/or enzyme evidence of re-infarction: re-evaluation of CK-MB or troponin to above the upper limit of normal and increased by >50% over the previous value.

      2. The total CK must either be re-elevated to two times or more the upper limit of normal and increased by >25% or be re-elevated to >200 U/mL over the previous value.

        1. If re-evaluated to less than two times the upper limit of normal, the total CK must exceed the upper limit of normal by >50% and exceed the previous value by two-fold or be re-elevated to >200 U/mL.

    3. Re-infarction after PTCA (+/−stenting):

      1. CK greater than three times the upper limit of normal and 50% greater than the previous value and/or new Q-waves (Minnesota Code) in two or more contiguous leads.

    4. Re-infarction after CABG surgery:

      1. CK greater than five times the upper limit of normal and ≥50% greater than the previous value and/or new Q-waves (Minnesota Code) in two or more contiguous leads.

  3. Congestive heart failure:

    1. Physician's decision to treat CHF with a diuretic, intravenous inotropic agent or intravenous vasodilator and either

      1. the presence of pulmonary edema or pulmonary vascular congestion on chest X-ray believed to be of cardiac cause or

      2. at least two of the following:

        1. rales greater than one-third up the lung fields believed to be due to CHF.

        2. PCWP >18 mmHg

        3. Dyspnoea, with documented pO2 less than 80 mmHg on room air or O2 saturation <90% on room air, without significant lung disease.

  4. Cardiogenic shock: The manifestation of vascular collapse and shock (SBP<90 mmHg for at least 30 min or SBP>90 mmHg after inotropic or intra-aortic balloon support with a cardiac index <2.2 L/min/m2 or <2.5 L/min/m2 inotropic or intra-aortic balloon support, peripheral signs of hypoperfusion, and chest X-ray with pulmonary edema.

  5. Major ventricular arrhythmias: Ventricular arrhythmias >6 h after randomization requiring electrical cardioversion/defibrillation.

  6. Major bleeding: Bleeding that causes haemodynamic compromise requiring blood or fluid replacement, inotropic support, ventricular assist devices, surgical intervention, or CPR to maintain a sufficient cardiac output.


Appendix 3


NT-pro-brain natriuretic samples were allowed to clot for 35–45 min and the blood centrifuged within 1 h of collection for 15 min at 1700 g and shipped in dry ice to a central laboratory (Montreal Heart Institute) where it was stored at −70°C. Samples were acquired at baseline, 72 h after admission, and at 30 days. Samples were batch-analysed at the end of the study by an electrochemiluminescence immunoassay with Roche Elecsys Instrument and Elecsys NT-proBNP Reagent kit (Roche Diagnostics, Indianapolis, USA) with an analytical range of 5–35000 pg/mL and respective intra-assay and inter-assay variability of 8 and 4%.


ST-segment measurements were acquired manually with magnification and hand-held callipers to the closest 0.05 mV at the J point. Per cent resolution was acquired according to the method of Schroeder. QRS scoring was analysed using the method of Selvester, where each point in the score represents ∼3% of the left ventricle.6,7


All patients in Groups B and C had angiographic analysis undertaken in a core facility to assess culprit coronary artery TIMI flow. These assessments were performed by observers blinded to treatment assignment.


View Abstract