Compared with non-smokers, cigarette smokers with ST-segment elevation myocardial infarctions derive greater benefit from fibrinolytic therapy. However, it is not known whether the optimal treatment strategy after fibrinolysis differs on the basis of smoking status. The Trial of Routine Angioplasty and Stenting After Fibrinolysis to Enhance Reperfusion in Acute Myocardial Infarction (TRANSFER-AMI) randomized patients with ST-segment elevation myocardial infarctions to a routine early invasive (pharmacoinvasive) versus a standard (early transfer only for rescue percutaneous coronary intervention or delayed angiography) strategy after fibrinolysis. The efficacy of these strategies was compared in 1,051 patients on the basis of their smoking status. Treatment heterogeneity was assessed between smokers and non-smokers, and multivariable analysis was performed to evaluate for an interaction between smoking status and treatment strategy after adjusting for baseline Global Registry of Acute Coronary Events (GRACE) risk score. Smokers (n = 448) were younger, had fewer cardiovascular risk factors, and had lower GRACE risk scores. They had a lower rate of the primary composite end point of 30-day mortality, reinfarction, recurrent ischemia, heart failure, or cardiogenic shock and fewer deaths or reinfarctions at 6 months and 1 year. Smoking status was not a significant predictor of either primary or secondary end points in multivariable analysis. Pharmacoinvasive management reduced the primary end point compared with standard therapy in smokers (7.7% vs 13.6%, p = 0.04) and non-smokers (13.1% vs 19.7%, p = 0.03). Smoking status did not modify treatment effect on any measured outcomes (p >0.10 for all). In conclusion, compared with non-smokers, current smokers receiving either standard or early invasive management of ST-segment elevation myocardial infarction after fibrinolysis have more favorable outcomes, which is likely attributable to their better baseline risk profile. The beneficial treatment effect of a pharmacoinvasive strategy is consistent in smokers and non-smokers.
When promptly available, primary percutaneous coronary intervention (PCI) is preferred in the contemporary management of ST-segment elevation myocardial infarction (STEMI). Fibrinolysis is still frequently used as initial reperfusion therapy where geography or resources limit timely access to primary PCI. Although cigarette smokers may have higher rates of successful reperfusion after fibrinolysis than non-smokers, smoking is associated with worse outcomes after elective or urgent PCI. When fibrinolysis is readily available but primary PCI is not, the optimal treatment strategy after fibrinolysis may be different for smokers with STEMI. Using data from the Trial of Routine Angioplasty and Stenting After Fibrinolysis to Enhance Reperfusion in Acute Myocardial Infarction (TRANSFER-AMI), we sought to compare the efficacy of routine early invasive management after fibrinolytic therapy in smokers versus non-smokers with STEMI. We also examined whether the “smoker’s paradox,” a phenomenon in which smokers with STEMI have a better short-term prognosis than non-smokers even after adjusting for confounders, persists in the contemporary era of STEMI management.
Methods
TRANSFER-AMI was a randomized trial performed at 52 centers in 3 Canadian provinces (Ontario, Quebec, and Manitoba). Details of the objectives, methodologies, and main results have been described. Participants who received fibrinolysis for acute STEMI at sites without PCI capability were randomized to receive either an early invasive (pharmacoinvasive) strategy of routine early transfer for PCI <6 hours after fibrinolysis with tenecteplase, or to a standard strategy of emergent transfer only for rescue PCI if fibrinolysis failed or delayed angiography <2 weeks after randomization. Eligibility criteria included presentation <12 hours after symptom onset with chest pain lasting ≥30 minutes and ST-segment elevation ≥2 mm in ≥2 contiguous anterior leads or ST-segment elevation ≥1 mm in ≥2 inferior leads with ≥1 of the following high-risk features: systolic blood pressure <100 mm Hg, heart rate >100 beats/min, Killip class II or III, ST-segment depression ≥2 mm in anterior leads, and ST-segment elevation ≥1 mm in right-sided lead V 4 suggestive of right ventricular involvement. Patients with cardiogenic shock before randomization, PCI within the previous month, previous coronary artery bypass grafting, or availability of primary PCI with anticipated door-to-balloon time <60 minutes were excluded. The primary end point was the combined incidence of death, reinfarction, recurrent ischemia, new or worsening heart failure, or cardiogenic shock at 30 days. Secondary end points included death or reinfarction at 30 days, 6 months, and 1 year and the incidence of in-hospital bleeding complications, classified by the Thrombolysis In Myocardial Infarction (TIMI) severity scale. A clinical events committee, blinded to treatment group assignment, centrally adjudicated all end points on the basis of prespecified criteria.
Data on patient demographics, characteristics, hospital management, and outcomes were collected using standardized case report forms, scanned into an electronic database (Teleform version 7.0; Cardiff, San Diego, California) at the Canadian Heart Research Centre (Toronto, Ontario, Canada). At 6 months and 1 year after admission, patients were contacted via telephone interview to ascertain vital status and the occurrence of recurrent cardiovascular events. The protocol was designed by the TRANSFER-AMI steering committee and approved by local ethics review boards at all participating sites. Informed consent was obtained from all patients.
In the present study, we examined data from all patients randomized in TRANSFER-AMI with documented smoking status (n = 1,051), after excluding 8 patients who had missing smoking status (4 in each treatment group). We stratified the study population into 2 groups on the basis of current smoking status. For this analysis, former smokers (n = 199) were considered non-smokers.
Continuous variables are presented as medians with interquartile ranges and discrete variables as percentages. We compared group differences in continuous and discrete variables with nonparametric Mann-Whitney U and chi-square tests, respectively. Because randomization in the trial was not stratified by smoking status, we assessed potential differences in baseline characteristics between treatment arms within each smoking group that arose by chance. We examined outcomes between smokers and non-smokers. We used the Breslow-Day test for homogeneity of odds ratios to evaluate whether smoking status modified the effect of treatment assignment (pharmacoinvasive vs standard) on the primary and secondary end points. We performed a sensitivity analysis by grouping former and current smokers together to compare outcomes and treatment heterogeneity. Because of a previously identified interaction between Global Registry of Acute Coronary Events (GRACE) risk score and treatment effect and possible differences in baseline risk profiles between smokers and non-smokers, we calculated the GRACE risk score for each patient. The GRACE risk score for in-hospital mortality consists of age, systolic blood pressure, heart rate, Killip class, cardiac arrest, ST-segment deviation, abnormal levels of biomarkers of myocardial necrosis, and serum creatinine. We performed multivariable analysis to test for any significant interaction between smoking status and treatment assignment, adjusting for GRACE risk score and its interaction. Finally, we used the multivariable models to evaluate the independent prognostic significance of smoking status. We performed statistical analyses using SPSS version 21.0 (IBM, Armonk, New York) and considered 2-sided p values <0.05 as significant.
Results
Of the 1,051 patients in the study cohort, 448 (42.6%) were current smokers. Table 1 lists their baseline characteristics. Current smokers were younger and less likely to have hypertension, diabetes mellitus, or previous heart failure. On admission, smokers had higher weight, a higher prevalence of peripheral vascular disease, lower serum creatinine levels, and lower GRACE and TIMI risk scores. Smokers also had shorter time intervals from hospital arrival to administration of fibrinolytic therapy, although not a significantly different interval between symptom onset and fibrinolytic administration.
| Variable | Current Smoker | p Value | |
|---|---|---|---|
| No (n = 603) | Yes (n = 448) | ||
| Age (years) ∗ | 60 (52–71) | 54 (48–59) | <0.001 |
| Men | 77.4% | 82.4% | 0.051 |
| Systemic hypertension | 38.8% | 25.2% | <0.001 |
| Diabetes mellitus | 18.2% | 10.0% | <0.001 |
| Dyslipidemia | 29.0% | 26.1% | 0.30 |
| Prior angina pectoris | 12.9% | 9.8% | 0.12 |
| Prior myocardial infarction | 11.6% | 8.7% | 0.13 |
| Prior heart failure | 57.4% | 42.6% | 0.010 |
| Prior stroke/transient ischemic attack | 2.3% | 1.3% | 0.25 |
| Prior peripheral vascular disease | 0.5% | 2.2% | 0.012 |
| Family history of premature CAD | 32.0% | 34.8% | 0.36 |
| Prior percutaneous coronary intervention | 5.3% | 5.1% | 0.90 |
| Systolic blood pressure (mmHg) ∗ | 144 (129–162) | 146 (132–161) | 0.34 |
| Diastolic blood pressure (mmHg) ∗ | 83 (73–94) | 86 (75–97) | 0.033 |
| Heart rate (bpm) ∗ | 76 (64–88) | 76 (64–90) | 0.80 |
| Weight (kg) ∗ | 79 (68–90) | 82 (72–91) | 0.002 |
| Killip Class | 0.49 | ||
| I | 91.7% | 91.7% | |
| II | 6.5% | 7.1% | |
| III-IV | 1.8% | 1.1% | |
| Admission creatinine (μmol/L) ∗ | 93 (79–107) | 90 (78–101) | 0.013 |
| Abnormal initial cardiac biomarker | 50.9% | 48.9% | 0.52 |
| Anterior lead ST-segment elevation | 57.4% | 42.6% | 0.069 |
| GRACE risk score ∗ | 130 (112–152) | 116 (103–133) | <0.001 |
| TIMI risk score ∗ | 2 (1–4) | 1 (1–3) | <0.001 |
| Time from symptom onset to fibrinolytic administration (minutes) ∗ | 117 (77–197) | 112 (70–166) | 0.064 |
| Time from hospital presentation to fibrinolytic administration (minutes) ∗ | 28 (18–46) | 24 (15–39) | <0.001 |
Several differences in in-hospital medications and interventions were noted between randomized treatment groups stratified by smoking status ( Table 2 ). Smokers had higher use of clopidogrel in the pharmacoinvasive group. Left main coronary artery stenosis was significantly more frequent in non-smokers who received the pharmacoinvasive treatment strategy compared with the standard treatment strategy. In both smokers and non-smokers, the use of a glycoprotein IIb/IIIa inhibitor and cardiac catheterization were higher in the groups randomized to pharmacoinvasive management. There were no significant differences in TIMI flow of the infarct-related artery before or after PCI between the treatment groups.
| Variable | Non-smoker | p Value | Current Smoker | p Value | ||
|---|---|---|---|---|---|---|
| Pharmacoinvasive (n = 299) | Standard (n = 304) | Pharmacoinvasive (n = 234) | Standard (n = 214) | |||
| Medication use | ||||||
| Aspirin | 99.3% | 97.7% | 0.098 | 98.3% | 98.6% | 0.79 |
| Clopidogrel | 95.3% | 91.8% | 0.077 | 97.4% | 90.2% | 0.001 |
| β-blocker | 94.3% | 94.7% | 0.82 | 95.7% | 96.3% | 0.77 |
| Glycoprotein IIb/IIIa inhibitor | 70.9% | 56.3% | <0.001 | 76.9% | 58.9% | <0.001 |
| ACE inhibitor/ARB | 87.0% | 87.5% | 0.84 | 89.3% | 90.7% | 0.64 |
| Statin | 93.0% | 94.7% | 0.37 | 93.6% | 94.4% | 0.72 |
| Interventions | ||||||
| Cardiac catheterization | 98.0% | 89.1% | <0.001 | 99.1% | 87.9% | <0.001 |
| Randomization to coronary angiography, (hours) ∗ | 2.6 (1.9–3.5) | 42.6 (4.4–72.0) | <0.001 | 2.5 (1.8–3.5) | 14.5 (3.4–67.3) | <0.001 |
| Infarct-related artery TIMI flow † | 0.094 § | 0.58 § | ||||
| 0 | 17.1% | 18.6% | 17.6% | 22.4% | ||
| 1 | 15.3% | 9.1% | 10.6% | 9.4% | ||
| 2 | 17.1% | 13.9% | 17.6% | 14.1% | ||
| 3 | 50.5% | 58.4% | 54.2% | 54.1% | ||
| Percutaneous coronary intervention | 82.3% | 66.1% | <0.001 | 88.5% | 68.7% | <0.001 |
| Rescue percutaneous coronary intervention ‡ | 29.0% | 40.6% | ||||
| Thrombectomy | 3.3% | 1.6% | 0.18 | 3.8% | 2.8% | 0.54 |
| Coronary artery bypass surgery | 9.7% | 8.9% | 0.73 | 2.6% | 6.5% | 0.042 |
| Infarct-related artery TIMI flow, post-percutaneous coronary intervention † | 0.78 § | 0.33 § | ||||
| 0 | 0.8% | 1.1% | 0.5% | 0.0% | ||
| 1 | 1.2% | 0.5% | 0.5% | 1.4% | ||
| 2 | 6.2% | 4.8% | 5.0% | 2.1% | ||
| 3 | 91.7% | 93.7% | 94.0% | 96.5% | ||
| Coronary arteries involved † | 0.001 § | 0.78 § | ||||
| 0 | 3.4% | 11.0% | 8.7% | 10.2% | ||
| 1 | 48.7% | 53.5% | 56.5% | 59.3% | ||
| 2 | 37.7% | 25.3% | 28.5% | 24.0% | ||
| 3 | 10.2% | 10.2% | 6.3% | 6.6% | ||
| Left anterior descending stenosis ≥70% † | 67.5% | 63.6% | 0.34 | 59.0% | 53.4% | 0.26 |
| Right coronary artery stenosis ≥70% † | 54.3% | 43.8% | 0.015 | 49.8% | 50.8% | 0.83 |
| Left circumflex stenosis ≥70% † | 36.5% | 30.7% | 0.16 | 28.6% | 26.2% | 0.59 |
| Left main stenosis ≥50% † | 9.0% | 3.3% | 0.008 | 1.5% | 6.4% | 0.011 |
∗ Median (25 th –75 th percentiles).
† Angiographic data represent only the percentages among those who underwent coronary angiography.
‡ Among patients who were randomized to standard treatment strategy (n = 515).
Compared with current smokers, non-smokers had higher rates of significant left anterior descending coronary artery stenosis (65.6% vs 56.5%, p = 0.004) and left circumflex coronary artery stenosis (33.7% vs 27.5%, p = 0.047) and underwent coronary artery bypass surgery more frequently (9.3% vs 4.5%, p = 0.003). In patients randomized to the standard treatment strategy, smokers had higher rates of rescue PCI (40.6% vs 29.0%, p = 0.007). Current smokers had lower unadjusted rates of mortality, mortality or reinfarction, and the primary composite end point of mortality, reinfarction, recurrent ischemia, new or worsening heart failure, or cardiogenic shock at 30 days ( Table 3 ). They also had lower mortality or reinfarction rates at 6 months and 1 year.
