OUP user menu

★ Editor's Choice ★

Early invasive compared with a selective invasive strategy in women with non-ST-elevation acute coronary syndromes: a substudy of the OASIS 5 trial and a meta-analysis of previous randomized trials

Eva Swahn, Joakim Alfredsson, Rizwan Afzal, Andrzej Budaj, Susan Chrolavicius, Keith Fox, Sanjit Jolly, Shamir R. Mehta, Robbert de Winter, Salim Yusuf
DOI: http://dx.doi.org/10.1093/eurheartj/ehp009 51-60 First published online: 7 February 2009

Abstract

Aims The aim of this study was to compare benefits and risks of a routine invasive compared with a selective invasive strategy in women with non-ST-elevation acute coronary syndromes.

Methods and results We randomly assigned 184 women, either to a routine or to a selective invasive strategy as a substudy to the OASIS 5 trial, who were followed for 2 years. Meta-analysis of data from previous randomized trials was also done. There were no significant differences between the two treatment strategies in the primary outcome death/myocardial infarction (MI)/stroke [21.0 vs. 15.4%, HR = 1.46, 95% CI (0.73–2.94)], in the secondary outcome death/MI [18.8 vs. 14.3%, HR = 1.39, 95% CI (0.67–2.88)], or separately analysed outcomes MI [12.9 vs. 13.3%, HR = 0.95, 95% CI (0.42–2.19)] or stroke [2.3 vs. 4.4%, HR = 0.67, 95% CI (0.12–3.70)]. However, there were significantly more deaths after 1 year (8.8 vs. 1.1%, HR = 9.01, 95% CI (1.11–72.90) and a higher rate of major bleeding at 30 days [8.8 vs. 1.1%, HR = 11.45, 95% CI (1.43–91.96)] in the routine invasive strategy group. A meta-analysis including 2692 women in previous randomized trials, with a gender perspective, showed no significant difference in the composite outcome death/MI, OR = 1.18, 95% CI (0.92–1.53) but a higher mortality with a routine invasive strategy for women, OR = 1.51, 95% CI (1.00–2.29).

Conclusion The rate of death, MI, or stroke in women was not different in patients treated with a routine invasive strategy compared with a selective invasive strategy, but there was a concerning trend towards higher mortality. When combined with data from previous trials, there does not appear to be a benefit of an early invasive strategy in women with ACS, which differs from the results in men. These data emphasize the lack of clear evidence in favour of an invasive strategy in women and suggest caution in extrapolating the results from men to women.

  • Women
  • Non-ST-elevation acute coronary syndrome
  • Invasive strategy

Introduction

Some, but not all, randomized trials have reported a reduction in myocardial infarction (MI), with a neutral effect on mortality, with a routine early coronary angiography followed by appropriate revascularization, compared with a more selective invasive strategy. There is controversy whether benefits of a routine invasive strategy in women differ from that in men.

The FRISC II trial (Fragmin and Revascularization during Instability in Coronary artery disease)1 and the RITA 3 trial (third Randomized Intervention Trial of Unstable Angina)2 reported a benefit of early intervention for death or MI in men but not in women, whereas the TACTICS TIMI-18 trial (Treat angina with Aggrastat and determine Cost of Therapy with an Invasive or Conservative Strategy—Thrombolysis In Myocardial Infarction 18), which was also analysed to look at marker positivity in relation to benefit of an invasive strategy, indicated a beneficial effect of early intervention for death or MI in both men and women, at least in patients with elevated markers.3 Recently, the ICTUS trial (Invasive vs. Conservative Treatment in Unstable Coronary Syndromes) reported no benefit with an early invasive strategy in reducing death, MI, or rehospitalization for angina symptoms, with similar results in both genders.4

The aim of this study in women was to determine if an early invasive strategy with routine coronary angiography (and if appropriate, coronary revascularization within 7 days) was superior to a selective invasive strategy (with coronary angiography only if symptoms or signs of severe ischaemia occurred) in preventing death, MI, or stroke at 2 years.

Methods

The current study was a randomized, prospectively designed substudy of the OASIS 5 trial (Organization to Assess Strategies in Acute Ischemic Syndromes Investigators). OASIS 5 was an international, multi-centre, randomized, double blind trial in which fondaparinux was compared with enoxaparin in patients with unstable angina pectoris or MI without ST-elevation (NSTEMI). Study design5 and results6 of the OASIS 5 main study are presented in detail elsewhere.

Local ethics committees approved the study protocol. All patients provided written informed consent.

Study patients

In the OASIS 5 trial, patients were randomly assigned to receive fondaparinux or enoxaparin within 24 h after onset of symptoms. Inclusion and exclusion criteria for the substudy were identical to those of the OASIS 5 main study. Patients were eligible if they met at least two of the following three criteria: an age of at least 60 years, an elevated level of troponin or creatinine kinase MB isoenzyme above upper limit of normal (ULN), or electrocardiographic changes indicative of ischaemia (defined as ST depression at least 1 mm in two contiguous leads or T-wave inversion >3 mm or any dynamic ST shift or transient ST elevation). Patients with contraindications to low-molecular-weight heparin, recent haemorrhagic stroke, indications for anticoagulation other than an acute coronary syndrome, age <21, pregnancy, co-morbid condition with life expectancy <6 months, or severe renal insufficiency (defined as a serum creatinine level of at least 265 µmol per litre) were excluded.

At the same time as randomization to fondaparinux or enoxaparin; female patients participating in this substudy (from select centres) were randomized to a routine coronary angiography (within 4 days of admission and, if appropriate, revascularization within 7 days of admission) or to a selective invasive strategy with coronary angiography only if they experienced symptoms or signs of severe ischaemia. Indications for coronary angiography in the selective invasive arm were: (i) refractory ischaemia, defined as recurrent chest pain/ischaemic symptoms (with documented characteristic ECG changes: horizontal ST depression >1 mm indicative of ischaemia) lasting >5 min, while on ‘optimal’ medical therapy [defined as at least two anti-anginal treatments (nitrate, β-blocker, calcium antagonist)]; (ii) new ST-elevation in two contiguous leads, without Q-waves or T-wave inversion (over 3 mm); (iii) development of haemodynamic instability, or severe heart failure (Killip class IV); (iv) intractable life-threatening arrhythmia; (v) incapacitating angina or severe ischaemia at a stress test before discharge or later during follow-up; (vi) re-infarction during follow-up. In both study groups, the mode of revascularization, percutaneous coronary intervention (PCI) or coronary artery by-pass grafting (CABG), was left to the discretion of the attending physicians and was based on patient characteristics and preferences, extent of disease, co-morbidity, and left ventricular function. If PCI was performed, the use of stents was strongly encouraged. At the start of the PCI, patients in the enoxaparin group received no additional anticoagulant if they had received their subcutaneous enoxaparin injection 6 h or less before PCI; if the interval was >6 h, the dose of weight-adjusted unfractionated heparin (UFH) was 65 IU/kg intravenously if a glycoprotein (GP) IIb/IIIa antagonist was used, and 100 IU/kg intravenously if no GP IIb/IIIa antagonist was used. Doses of heparin were chosen based on recommendations from the sixth ACCP Consensus Conference on Antithrombotic Treatment.7 Patients undergoing PCI within 6 h of the last subcutaneous dose of fondaparinux, in the absence of a GP IIb/IIa antagonist, received an additional 2.5 mg intravenous fondaparinux. If the PCI was done 6 h or more after the last subcutaneous dose of fondaparinux, 2.5 mg of intravenous fondaparinux (with a GP IIb/IIIa antagonist) or 5.0 mg of intravenous fondaparinux (without a GP IIb/IIIa antagonist) was administered.

Prior to PCI, all patients should be pre-treated with aspirin and clopidogrel. Aspirin dose was left to the discretion of the investigator, but doses <100 mg was recommended. It was also strongly recommended that if the patient was not previously on clopidogrel, then a 300 mg loading dose should be given prior to catheterization. If the patient was already on clopidogrel, no loading dose was required prior to catheterization, but if PCI was performed at the same sitting, a loading dose of 300 mg should be given before the procedure. Treatment with intravenous GP IIb/IIIa was encouraged in association with PCI.

Outcomes

The primary outcome was the composite of death, MI, or stroke at 2 years. Secondary outcomes included: (i) each of death, MI, and stroke evaluated separately; (ii) composites of death, MI and death, MI, stroke, or refractory ischaemia (i.e. ischaemia-driven revascularization).

A central committee of clinicians blinded to the allocated management strategy adjudicated death classified by cause, MI, refractory ischaemia, stroke, and major bleedings.

Either one of the following two criteria was required to meet the diagnosis of MI: (i) typical rise and fall of biochemical markers of myocardial necrosis (including troponin, CK-MB, CK) to greater than two times ULN (three times ULN if within 48 h of PCI; 5× ULN if within 48 h of CABG; if markers were already elevated, >50% of the lowest recovery enzyme level from the index infarction) with at least one of the following: (a) ischaemic symptoms, (b) development of pathological Q-waves on the ECG, (c) ECG changes indicative of ischaemia (ST-segment elevation or depression), (d) coronary artery intervention; (ii) findings of an acute MI at autopsy.

Major bleeding was defined as clinically overt bleeding with at least one of the following criteria: fatal, symptomatic intracranial haemorrhage, intraocular haemorrhage leading to significant vision loss or decrease in haemoglobin (Hb) of ≥30 g/L, (with each blood transfusion unit counting for 10 g/L) or requiring transfusion of ≥2 U of blood.

Refractory ischaemia was defined as recurrent chest pain or ischaemic symptoms lasting >5 min with characteristic ECG changes indicative of ischaemia while on optimal medical therapy (at least two anti-anginal treatments) and leading to additional intervention (such as thrombolytic therapy, cardiac catheterization, or insertion of intra-aortic balloon pump) within 48 h of onset of the episode.

Meta-analysis

A computerized literature search was conducted from 1950 through November 2008 of the MEDLINE database using the search terms invasive strategy, conservative strategy, selective invasive strategy, acute coronary syndromes, NSTE MI, non-Q-wave MI, and unstable angina. We restricted the search to controlled clinical trial or meta-analysis. We also contacted experts in the field and performed hand searching through cross-references.

An invasive strategy was defined as the referral of all patients with non-ST-elevation acute coronary syndrome (NSTE ACS) for coronary angiography, followed by revascularization if suitable. A selective invasive strategy was defined as a strategy with initial pharmacological treatment followed by coronary angiography (and subsequent revascularization if suitable) only in patients with symptoms of ischaemia or severe signs of ischaemia on non-invasive stress testing.

Trials were considered for inclusion if they enrolled patients with NSTE ACS (NSTEMI or unstable angina) and randomly allocated patients to receive a routine invasive strategy or a selective invasive strategy. Trials were excluded if the majority of patients enrolled had stable angina or ST-elevation MI. Trials were not considered for inclusion if fibrinolytic therapy was administered to all patients, if coronary angiography was required before randomization, or if randomization process was inadequate.

Finally, we initially excluded all trials without published sex-specific outcome data (death and MI). In addition, we obtained information (by personal communication with the principal investigator) on sex-specific outcome data from the latest trial that fulfilled the criteria to be considered for inclusion.

Sample size estimate and statistical analysis

Based on the overall results of the FRISC II trial, we assumed a reduction in the rate of death or MI from 17.5 to 12.5% at 3 years with a routine invasive strategy. To detect a relative risk reduction of 28.5%, with 80% power, a sample size of 1600 was planned. Because of slow inclusion rate in the substudy, only 184 patients were recruited when the OASIS 5 main trial was stopped. However, a decision was made to follow, in a blinded fashion, all randomized patients in the substudy for 2 years.

At the same time as randomization to enoxaparin or fondaparinux, patients included in the substudy were randomized to routine or selective invasive strategy. The randomization scheme was stratified by site using blocks of size 2 and 4. The curtailment in the sample size as well as follow-up time reduced the effective power substantially to about 12% to detect the hypothesized reduction in outcome (two-sided α = 0.05), but we continued the study as the data could contribute usefully to a meta-analysis of the relevant data from previous trials with a gender perspective.

Means and percentages are used to describe baseline and other characteristics. Outcomes are presented as hazard ratios with 95% confidence intervals. Furthermore, selected outcomes are summarized by Kaplan–Meier curves. No adjustment was made for multiple testing.

A meta-analysis of randomized trials of routine invasive vs. selective invasive strategies, with separate data for the genders, was undertaken using the method described by Yusuf et al.8 and follows the approach utilized and extensively described by Mehta et al.9 This approach uses an assumption-free model and weights the analyses in proportion to the information (number of events) contributed from each trial. For FRISC II, RITA 3 (published results), ICTUS (personal communication by R.W.) and OASIS 5 women substudy, 1-year outcomes were used and for TACTICS TIMI-18, the published 180-day outcomes were used. Data on outcome from each of the published studies were independently extracted by two investigators. Odds ratios (OR) with 95% confidence interval (CI) for the composite outcome death or MI and for mortality were reported. Heterogeneity was assessed with Q-statistic. The level of significance used was 0.05 (two-sided). Statistical software used for the meta-analysis was Comprehensive Meta-analysis V 2.0 (Biostat Inc., Englewood, NJ, USA).

Results

A total of 92 patients were randomly assigned to a routine invasive strategy and 92 patients to a selective invasive strategy. Almost all patients were recruited from Sweden (108) and Poland (73). Nine patients were lost to long-term follow-up, but they were equally distributed between routine invasive (five patients) and selective invasive (four patients) strategy. Baseline characteristics are shown in Table 1. Mean age was 68 years in both groups. Almost 80% of the patients had troponin or CKMB above ULN and 86% had any ECG abnormality. During initial hospitalization, 88 patients (96%) in the routine invasive arm and 37 patients (40%) in the selective invasive arm underwent coronary angiography, followed by revascularization in 53 patients (58%) and 28 patients (30%), respectively (Table 2). During the whole study period, 89 patients (97%) in the routine invasive and 60 patients (65%) in the selective invasive group had an angiography performed. In the routine invasive group, 44 patients (48%) had a PCI performed and 15 patients (16%) underwent CABG compared with 36 patients (39%) and 11 patients (12%), respectively, in the selective invasive group. The proportion of patients receiving thienopyridines before or during PCI was similar in the two arms, but a larger proportion of patients in the routine invasive arm received GP IIb/IIIa inhibitors and a larger proportion of patients in the selective invasive arm received UFH. The median time from randomization to angiography, PCI, and CABG differed between the two study groups (51 vs. 156 h, 70 vs. 147 h, and 196 vs. 381 h, respectively) (Table 3). Among patients subjected to PCI, there was a higher success rate (96.2 vs. 87.9%) and use of stents (82.7 vs. 72.7) in the selective invasive group when compared with the routine invasive group (Table 4).

View this table:
Table 1

Baseline characteristics of women entering the study

Routine invasive (n = 92)Selective invasive (n = 92)
Age (years) (mean ± SD)68.2 ± 9.267.8 ± 8.8
Heart rate (b.p.m.) (mean ± SD)72.2 ± 15.473.2 ± 14.5
Systolic blood pressure (mmHg) (mean± SD)139.4 ± 25.7138.3 ± 21.6
Time from onset of pain to randomization [(h, median (interquartile range)]11 (7–19)13 (8–19)
Comorbidity and risk factorsn (%)n (%)
 History of MI22 (24)18 (20)
 Previous PCI7 (8)11 (12)
 Previous CABG5 (5)3 (3)
Stroke4 (4)4 (4)
 Hypertension57 (62)63 (67)
 Diabetes mellitus19 (21)27 (29)
 Current smoker9 (10)24 (26)
Medications at the time of randomization
 Aspirin70 (76)71 (77)
 Clopidogrel or ticlopidine27 (29)28 (30)
 Unfractionated heparin4 (4)2 (2)
 Low-molecular-weight heparin27 (29)26 (28)
 ACE-inhibitor or ARB44 (48)45 (50)
 β-blocker62 (67)61 (66)
 Lipid-lowering drug26 (28)37 (40)
 Calcium channel blocker11 (12)14 (15)
 Digoxin22 (7)10 (4)
 Diuretics22 (24)25 (27)
 Nitrates55 (60)57 (62)
Ischaemic symptoms and signs
 ST-depression (≥1 mm)44 (48)42 (46)
 Any ECG abnormality79 (86)79 (86)
 Troponin or CKMB elevation73 (79)72 (78)
Diagnosis at study entry
 Unstable angina33 (36)27 (29)
 Suspected MI59 (64)65 (71)
  • MI, myocardial infarction; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CKMB, creatinine kinase muscle/brain.

View this table:
Table 2

Medication, procedures, and complications during initial hospitalization

Routine invasive (n = 92)Selective invasive (n = 92)
Medication
 Aspirin91 (100)91 (98.9)
 Clopidogrel or ticlopedin79 (85.9)70 (76.1)
 Dual antiplatelet therapy79 (85.9)69 (75.0)
 Unfractionated heparin iv+sc11 (12)7 (8)
 Low-molecular-weight heparin22 (23.9)20 (21.7)
 Enoxaparin (by allocation)38 (41)54 (59)
 Fondaparinux (by allocation)54 (59)38 (41)
 GP IIb/IIIa11 (12)7 (7.6)
 ACE-inhibitor or ARB65 (70.7)74 (80.4)
 β-blocker87 (94.6)86 (93.5)
 Lipid-lowering drug81 (88)79 (85.9)
Cardiac procedures
 Coronary angiography88 (95.7)37 (40.2)
 PCI42 (45.7)22 (23.9)
 CABG11 (12.0)6 (6.5)
Complications
 Killip class I1 (1.1)5 (5.4)
 Killip class II4 (4.3)12 (13.0)
 Killip classes III and IV2 (2.2)0
  • GP, glycoprotein; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting.

View this table:
Table 3

Cardiac procedures, degree of coronary artery disease, medication, and complications during percutaneous coronary intervention during long-term follow-up

Routine invasive (n = 92), n (%)Selective invasive (n = 92), n (%)
Coronary angiography89 (97)60 (65)
Time to angiography [h, median (interquartile range)]51 (26–84)156 (7–1838)
PCI44 (48)34 (39)
Time to PCI [h, median (interquartile range)]70 (23–93)147 (49–1345)
CABG15 (16)11 (12)
Time to CABG [h, median (interquartile range)]196 (147–931)381 (225–1844)
Diseased vesselsa
 Total numbers of lesions6652
 none22 (24.7)10 (16.7)
 142 (47.2)27 (45.0)
 27 (7.9)12 (20.0)
 3 or more18 (20.2)11 (18.3)
 Left main5 (5.6)8 (13.3)
Medication prior to or during PCIb
 Thienopyridine36 (81.8)29 (80.6)
 UFH17 (38.6)17 (47.2)
 GP IIb/IIIa15 (34.1)8 (22.2)
Complications during PCIb
 Vascular site complications3 (6.8)0
 Otherc4 (9.1)0
  • h, hours; SD, standard deviation; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; UFH, unfractionated heparin; GP, glycoprotein.

  • aIn patients that were subjected to coronary angiography.

  • bIn patients that had a PCI performed.

  • cOne new thrombus during angiography, one catheter thrombus, and two dissections.

View this table:
Table 4

Percutaneous coronary intervention analysis by lesion

Routine invasive (n = 66), n (%)Selective invasive (n = 52), n (%)
Successful PCI
 Complete success58 (87.9)50 (96.2)
 Partial success2 (3.0)0
Stent use
 BMS29 (43.9)31 (59.6)
 DES19 (28.8)12 (23.1)
 No stent18 (27.3)9 (17.3)
  • PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; BMS, bare metal stent; DES, drug-eluting stent.

There was no statistically significant difference in the primary outcome of death/MI/stroke [21.0% in the routine invasive group vs. 15.4% in the selective invasive group, HR = 1.46, 95% CI (0.73–2.94)] (Figure 1) or in any of the secondary composite outcomes of death /MI or death/MI/refractory ischaemia (Table 5). We found no significant differences when MI or stroke was evaluated separately (Table 5). However, at 30 days, there was a trend towards a higher rate of death with a routine invasive strategy compared with a selective invasive strategy [4.3 vs. 1.1%, HR = 4.47, 95% CI (0.49–40.70)]. At 1 year, there was a statistically significant difference in mortality, with eight deaths in the routine invasive arm compared with one death in the selective invasive arm [8.8 vs. 1.1%, HR = 9.01, 95% CI (1.11–72.90)]. The observed difference in mortality at 1 year persisted until the end of follow-up at 2 years, although the statistical significance was borderline [8.8 vs. 2.2%, HR = 4.65, 95%CI (0.97–22.20)] (Table 5, Figure 2).

Figure 1

Cumulative risk of death, myocardial infarction (MI), or stroke through long-term follow-up. Hazard ratio for death, myocardial infarction, or stroke at 2 years, 1.46 (95% confidence interval 0.73–2.94).

Figure 2

Cumulative risk of death through long-term follow-up. Hazard ratio for death at 2 years, 4.65 (95% confidence interval 0.97–22.20).

View this table:
Table 5

Outcomes

Routine invasive (n = 92), n (%)Selective invasive (n = 92), n (%)Hazard ratio (95% CI)
2 years
 Death/MI/stroke19 (21.0)14 (15.4)1.46 (0.73–2.94)
 Death/MI/RI19 (21.0)19 (20.8)0.99 (0.52–1.90)
 Death/MI17 (18.8)13 (14.3)1.39 (0.67–2.88)
 Death8 (8.8)2 (2.2)4.65 (0.97–22.20)
 MI11 (12.9)12 (13.3)0.95 (0.42–2.19)
 Stroke2 (2.3)4 (4.4)0.67 (0.12–3.70)
 Major bleeding9 (10.0)2 (2.2)6.90 (1.48–32.13)
30 days
 Death/MI/stroke9 (9.8)4 (4.3)2.18 (0.66–7.17)
 Death/MI/RI11 (12.0)10 (10.9)1.02 (0.43–2.43)
 Death/MI8 (8.7)4 (4.3)1.97 (0.58–6.63)
 Death4 (4.3)1 (1.1)4.47 (0.49–40.70)
 MI4 (4.4)3 (3.3)1.19 (0.26–5.40)
 Stroke1 (1.1)1 (1.1)1.02 (0.06–16.93
 Major bleeding8 (8.8)1 (1.1)11.45 (1.43–91.96)
180 days
 Death/MI/stroke14 (15.2)8 (8.7)1.75 (0.73–4.20)
 Death/MI/RI15 (16.3)13 (14.1)1.08 (0.51–2.30)
 Death/MI13 (14.1)7 (7.6)1.84 (0.73–4.65)
 Death7 (7.6)1 (1.1)7.47 (0.91–61.52)
 MI7 (7.8)6 (6.6)1.08 (0.36–3.25)
 Stroke1 (1.1)3 (3.3)0.41 (0.04–3.99)
 Major bleeding8 (8.8)1 (1.1)11.45 (1.43–91.96)
1 year
 Death/MI/stroke15 (16)11 (12)1.45 (0.66–3.20)
 Death/MI/RI15 (16.3)16 (17.4)0.92 (0.45–1.88)
 Death/MI13 (14.1)10 (10.9)1.36 (0.59–3.13)
 Death8 (8.8)1 (1.1)9.01 (1.11–72.90)
 MI7 (7.8)9 (9.9)0.78 (0.29–2.14)
 Stroke2 (2.3)3 (3.3)0.86 (0.14–5.26)
 Major bleeding9 (10.0)1 (1.1)13.35 (1.68–105.9)
  • MI, myocardial infarction; RI, refractory ischaemia.

By the end of 2 years, there were eight deaths in the routine invasive arm compared with two deaths in the selective invasive strategy arm. Three cases of fatal asystole, two cases of sudden death, one of cardiogenic shock, and two cases of death due to other vascular causes (one acute intestinal ischaemia and one brain damage because of earlier ventricular fibrillation) were reported in the routine invasive strategy arm. In the selective invasive arm, one death was reported to be caused by haemorrhage and in one case the cause of death was judged to be unknown.

To evaluate whether any of the invasive treatment modalities was associated with worse outcome, we assessed outcomes according to treatment with PCI or CABG, regardless of initial allocation. At 2-year follow-up, there were two deaths (7.7%) among the 26 patients who had a CABG performed, six deaths (7.5%) among 80 patients subjected to PCI, and only two deaths in the 78 patients not revascularized (2.6%).

The rate of major bleedings was substantially higher in patients randomized to the routine invasive strategy at 30 days [8.8 vs. 1.1%, HR = 11.45, 95% CI (1.43–91.96)], and the difference persisted during long-term follow-up at 730 days [10.0 vs. 2.2%, HR = 6.90, 95% CI (1.48–32.13)] (Table 5). Of the 11 patients who had experienced a major bleeding by 2-year follow-up, eight occurred among patients who were subjected to PCI and two among patients who had a CABG performed. However, there were only two deaths among these 11 patients.

Although a larger proportion of patients randomized to a routine invasive as compared with the selective invasive strategy was initially allocated to treatment with fondaparinux (59 vs. 41%) (Table 2), there was no major difference in the primary composite outcome death/MI/stroke (17 vs. 18%) or secondary outcomes MI (14 vs. 13%) or death (4 vs. 7%) between patients randomized to treatment with fondaparinux vs. treatment with enoxaparin. There was, however, less major bleedings among patients allocated to fondaparinux compared with enoxaparin (1 vs.11%).

Three earlier randomized trials have reported outcome in women and men separately: FRISC II, RITA 3, and TACTICS TIMI-18. Data from the ICTUS trial were obtained from the study investigators. We conducted a meta-analysis based on previously published data from the first three trials, data provided to us from ICTUS and regarding women, the present trial. Data from 1-year outcomes were used from FRISC II, RITA 3, ICTUS, and OASIS 5 women substudy. Since 1-year data were not available from TACTICS TIMI-18, outcomes from 6-month follow-up were used. The meta-analysis indicates benefit with a routine invasive strategy in men but not in women. Odds ratio for death/MI, for the routine invasive compared with a selective invasive strategy, showed no significant benefit in women [OR=1.18, 95% CI (0.92–1.53)] compared with men [OR = 0.78, 95% CI (0.66–0.93)] (Figure 3). For men, the routine invasive strategy was associated with lower mortality [OR 0.70, 95% CI (0.51–0.96)], whereas the opposite was found in women [OR 1.51, 95% CI (1.00–2.29)] (Figure 4). Heterogeneity testing indicated heterogeneity between men and women for a routine invasive compared with a selective invasive strategy regarding both death/MI (Q = 6.63, P = 0.01) and death (Q = 6.3, P = 0.012).

Figure 3

Meta-analysis including 1-year outcome in FRISC II, RITA 3, ICTUS, and OASIS 5 women substudy and 180 days outcome in TACTICS TIMI-18. Odds ratio for the composite outcome death/myocardial infarction was calculated in routine invasive compared with selective invasive strategy. Heterogeneity in men vs. women P = 0.01.

Figure 4

Meta-analysis including 1-year outcome in FRISC II, RITA 3, ICTUS, and OASIS 5 women substudy and 180 days outcome in TACTICS TIMI-18. Odds ratio for death was calculated in routine invasive compared with selective invasive strategy. Heterogeneity in men vs. women P = 0.012.

Discussion

In this small substudy of women randomized to routine invasive strategy compared with selective invasive strategy, there was no benefit with routine invasive strategy on any of the outcomes examined. There was, however, a trend towards a higher rate of the composite outcome death/MI with the routine invasive strategy, which is in accordance with the findings in women enrolled in the FRISC II (12.4% with a routine invasive vs. 10.5% with a selective invasive strategy) and RITA 3 (8.6% with a routine invasive vs. 5.1% with a selective invasive strategy) trials.

Surprisingly, we found a trend towards a higher mortality rate and also a significant excess in major bleeds associated with a routine invasive strategy in women with NSTE ACSs.

Consistent with the overall results of several previous randomized trials comparing a routine invasive with a selective invasive strategy1012 and a meta-analysis of previous trials,9 we observed a trend to an increased risk in early mortality, but unlike the previous trials, we did not observe a compensatory late benefit with a routine invasive strategy.

When these results are put into the perspective of earlier trials addressing the effect of a routine invasive strategy in women, there is a trend towards an adverse effect. The 1-year mortality rates in the routine invasive and the selective invasive arms, respectively, were 4.0 vs. 3.2% in FRISC II and 5.1 vs. 2.4% in RITA 3, and in TACTICS TIMI-18, the corresponding 180-day mortality rates were 3.8 vs. 3.6%.13 These trials are consistent with the present study and collectively indicate little apparent benefit on major clinical outcomes with a routine invasive strategy in women. In the present study, the difference in 1-year death rates between routine invasive and selective invasive strategy was larger; 8.8 and 1.1%, respectively, a concerning trend that strongly merits further studies.

We were surprised that despite recruiting a relatively small number of patients, we observed a significant increase in mortality. The large apparent excess in mortality observed in our trial is likely to be an exaggeration, perhaps due to chance, but our data are directionally consistent with the findings of several previous studies. However, no study had enough power to show a statistically significant difference separately in outcomes in women. Gender-specific data from all presently available studies, taken together in our meta-analysis, indicate that in women there is little benefit with a routine invasive strategy compared with a more selective approach. The meta-analysis includes all patients in the studies and does not account for the possibility of differences in benefit from a routine invasive strategy within certain subgroups, e.g. patients with elevated markers of necrosis. Although our meta-analysis consists of five trials, it is based on only 2692 patients with few events (97 deaths and 263 death/MI), but the lower CI limits indicate that, at best, the benefit with a routine invasive strategy is likely to be small. However, this question is best resolved by a large prospective randomized trial involving women. In men, in contrast to what we found in women, our meta-analysis confirmed a significant benefit for both death and death/MI with an invasive strategy. Hence, our meta-analysis indicates heterogeneity between the genders in the effect of a routine invasive strategy. In a recent meta-analysis by ÓDonoghue et al.,13 where sex-specific data from eight randomized trials had been extracted, the overall analysis showed a non-significant 11% higher odds for death with an invasive strategy in women [OR 1.11, 95% CI (0.72–1.70)] contrasting an also non-significant 11% lower odds for death with an invasive strategy in men [OR 0.89, 95% CI (0.58–1.35)].

The two study groups in the present study were well matched regarding age, myocardial damage marker elevation, and ST-segment depression. There were more current smokers and diabetics in the selective invasive group; otherwise there were no major differences between the groups regarding risk factors and medical history that could explain the large difference in outcome. In our study, the effect of intervention is identical with or without elevated troponin when compared with TACTICS TIMI-18.3 The interaction P-value is 0.508.

The use of evidence-based medications in this study was high, and there were no major differences between the two treatment arms. In the routine invasive strategy and selective invasive arm, respectively, almost all patients received aspirin (100 vs. 99%) and a large proportion was given clopidogrel/ticlopidine (86 vs. 76%), ACE-inhibitor (ACE-I)/angiotensin receptor antagonist (ARB) (71 vs. 80%), β-blocker (95 vs. 94%), and lipid-lowering drug (88 vs. 86%). For comparison, in TACTICS TIMI-18 (where discharge medication is reported), use of aspirin was similarly high (98%), but the use of ACE-I/ARB (18%), β-blocker (61%), and statins (45%) was substantially lower. In TACTICS TIMI-18, GpIIb/IIIa inhibitors were routinely used prior to the angioplasty as compared with 23.8% in our study. However, a meta-analysis of all trials of GP IIb/IIIa inhibitors in ACS has failed to demonstrate any benefit in women.14 The proportion of patients receiving thienopyridines was not reported in earlier trials, but it is likely to have been lower than in the current trial. Could the higher use of effective medications in the present study, compared with previous trials, explain the very low mortality rate in the selective invasive strategy arm? With an optimal medical therapy, there may be little to gain from a routine invasive strategy. In contrast, restricting the use of invasive procedures to patients who are not stabilized on optimal medical therapy may select those who benefit most from, while at the same time minimizing hazards associated with, such procedures. This may be more important in women who have less severe coronary artery disease. As expected, among patients that had an angiography performed, patients in the routine invasive arm had less severe coronary artery disease. Also as expected, by design, there was an important difference between the study groups in time from randomization to invasive procedures, with longer time delays in the selective invasive group. Whether early invasive procedures are associated with lower procedural success rates and higher complication rates as suggested by our data can only be resolved by a large trial that randomizes patients to early vs. more delayed invasive strategy. As mentioned, in an earlier meta-analysis, routine invasive strategy was associated with higher early mortality and MI (early hazard), which appeared to be compensated by a later benefit.9 The early hazard as well as lack of later benefit may be more pronounced in women who tend to be older and experience higher risk of procedural complications with an invasive strategy. Combined with less significant coronary artery disease, women may have less to gain and more risks with invasive procedures.

Patients in the routine invasive group were more often allocated to fondaparinux (59 vs. 41%), but fondaparinux was not associated with worse outcome as there were two deaths in patients treated with fondaparinux (4.3%) and six deaths in patients allocated to enoxaparin treatment (6.5%). Adjustment for this imbalance did not alter the results.

In the FRISC II trial, the mortality rate among women randomized to early invasive strategy and treated with CABG was 9.9% compared with 1.2% among men, whereas the corresponding numbers for patients treated with PCI were 1.5 and 1.0%. The higher risk associated with CABG was suggested as an explanation for lack of benefit with an invasive strategy among women in the FRISC II trial. We assessed mortality according to actual mode of revascularization. There were two deaths among CABG-treated patients (7.7%) and six deaths among PCI-treated patients (7.5%). Hence, CABG-related mortality in our study does not appear to be a major reason for the higher mortality with an invasive strategy in this study. Indeed the lowest mortality rate (2.6%) was observed in those who did not undergo any revascularization procedure.

Another important finding in this study was the high rate of major bleedings associated with the routine invasive strategy. A marked difference in bleeding rates was evident early during the study (at 30 days: 8.8 vs. 1.1%, P = 0.004), with few further additional events later on in either group. Accordingly, increased risk of major bleeding was consistent and statistically significant at all time points during the study. There were 11 major bleedings in the current study, and 8 of these occurred in patients subjected to PCI, 2 in patients who had a CABG performed, and only 1 in patients not revascularized. These higher rates of bleeding with an invasive strategy have been reported in earlier studies comparing invasive and conservative strategies in NSTE ACS.4,1012 Women also are at higher risk for bleeding complications. For example, in the TACTICS TIMI-18 trial, the rate of major bleeding in women undergoing PCI was 8.3% compared with 2.9% in men.3 Increased bleeding rates have been reported in ACS treated with GP IIb/IIIa inhibitors, especially in combination with thienopyridines15 and among women.16 Although the numbers are too small to draw any firm conclusions, we found a three times higher mortality among patients experiencing a major bleeding; there were two deaths among the 11 patients (18%) that experienced a major bleeding and 8 deaths among the 173 patients (6%) that did not experience a major bleeding. However, several recent studies have highlighted that bleeding in patients with ACS increases both short-term and long-term risk of death.1719 The high bleeding rate among women treated invasively is worrying and needs further studies.

Limitations

The small sample size and the shortened follow-up are important limitations. The results should therefore be interpreted with great caution and in the context of similar data from other trials. Hence, our meta-analysis is more reliable than an emphasis on the results of any specific trial in isolation.

Conclusions

Women with NSTE ACS did not seem to benefit from a routine invasive strategy, but instead experienced higher mortality and bleeding. The results taken together with previous trials indicate a lack of benefit with a routine invasive strategy in women. This suggests that the results from men, with regards to invasive procedures, may not necessarily apply to women and that large-scale randomized trials in women are needed to determine the optimal treatment strategy.

Funding

The study was supported by Sanofi-Aventis, Organon, and GlaxoSmithKline. The sponsor had no role in the study design; the collection, analysis, or interpretation of the data; the preparation, review, or approval of the manuscript. E.S. had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Conflict of interest: none declared.

Acknowledgement

We are indebted to patients and investigators who participated in the study.

References

View Abstract