Early treatment with intravenous thrombolysis is useful in patients presenting with ST-segment elevation myocardial infarction (STEMI), reducing both short- and long-term coronary mortality in comparison with placebo.1^–3 Critical pathways to expedite the evaluation and management of patients with acute myocardial infarction have been developed. One such approach involves prehospital thrombolysis, which is associated with higher survival rates than in-hospital thrombolysis.4 In France, the emergency medical system (Service d’Aide Médicale Urgente) uses mobile intensive care units (MICUs) with a doctor on board, which has facilitated the use of prehospital thrombolysis.5 This approach shortens time to reperfusion and provides clinical results close to those achieved with primary angioplasty.4^–6 Thrombolysis is most beneficial when administered early after the onset of symptoms, probably because this is the time when the potential for myocardial salvage is greatest and also because thrombolysis is probably more effective at dissolving “fresh” clots and restoring arterial patency early after symptom onset.4 7^–9 A subset analysis from the CAPTIM (Comparison of Angioplasty and Prehospital Thrombolysis in Acute Myocardial Infarction) randomised trial suggests that prehospital thrombolysis may even be more effective than primary percutaneous coronary intervention (PCI) in preventing mortality and cardiogenic shock for patients treated within 2 h of symptom onset.10

Few data exist about the outcome of prehospital thrombolysis as measured by early coronary patency. Most studies deal with in-hospital thrombolysis and antedate the use of contemporary protocols and drugs (such as thienopyridines). Yet, the population of patients treated in the prehospital setting is different from that treated in hospital.11 The objective of the Orientation of Patients Treated for Myocardial Infarction After Lysis (OPTIMAL) study was to identify predictors of early epicardial coronary patency (as judged by TIMI 3 flow) after prehospital thrombolysis in a large, contemporary, cohort representative of routine practice.12 A secondary objective was to develop a nomogram for predicting the likelihood of coronary patency (or lack thereof) according to simple clinical and electrocardiographic characteristics available in the emergency setting.

METHODS

Study design

The methods and rationale for the OPTIMAL study have been described elsewhere.12 This was a multicentre, multidisciplinary, prospective, longitudinal, observational, cohort study conducted in hospitals in France with a prehospital MICU and a coronary care unit with 24 h access to coronary angiography. Each centre had a designated referring emergency specialist and an interventional cardiologist who were free to make decisions about patient management. General recommendations about indications for prehospital thrombolysis and coronary angiography were in accordance with usual clinical practice and current guidelines from the European Society of Cardiology (ESC).13 Data were recorded on standardised case report forms by emergency specialists and cardiologists. Completed forms were mailed to the data-processing company. Procedures were implemented to ensure that the patient database guaranteed anonymity. Patients received an information leaflet that outlined their right to access, correct or withhold their data. The study protocol was reviewed by the French Society of Cardiology and was performed in accordance with French law and the Declaration of Helsinki.

Patient population

To be eligible for inclusion in the study, adult patients (⩾18 years) had to present with STEMI determined from a 12-lead electrocardiogram (ECG) and defined as ⩾0.1 mV in at least two peripheral leads or ⩾0.2 mV in at least two precordial leads; to have received prehospital thrombolysis within 6 h of pain onset; and to be scheduled for coronary angiography within 6 h of receiving thrombolytic treatment. Patients were not eligible for inclusion if they had an acute coronary syndrome without elevation of the ST segment, if they did not have a coronary angiography planned within 6 h of the start of thrombolytic treatment, or if they were enrolled in another clinical trial.

Assessment of coronary patency (TIMI flow) and data collection

Patency of the infarct-related coronary artery was evaluated from coronary angiography, which was systematically done within 6 h of thrombolysis and before any PCI. Thrombolysis In Myocardial Infarction (TIMI) flow was determined by a blinded core laboratory analysis. Patency of the infarct artery was defined as a TIMI flow grade of 3. For 194/997 (19%) patients, the blinded and centralised assessment of TIMI flow could not be done for technical reasons and statistical analyses were therefore conducted using the site-reported TIMI flow grade. Additional variables relating to coronary angiography were delay between thrombolysis and coronary angiography; location of the culprit artery; stenosis status (<50% or ⩾50%) and number of vessels with stenosis ⩾50%.

Other variables recorded were time between symptom onset and start of thrombolysis; presence of any chest pain (ie, any pain vs none); haemodynamic status (blood pressure, heart rate and Killip class) before thrombolytic treatment and before coronary angiography; and performance and timing of in-hospital PCI. Blinded core laboratory analysis of the ECGs was performed in 800 patients. ST-segment resolution before angiography was assessed using the standard 12-lead ECG. The lead with the greatest magnitude of ST deviation in the infarct territory (measured 20 ms after the J point) was selected on the first ECG performed just before thrombolysis. The percentage of ST resolution was then measured on the same lead on the ECG performed just before coronary angiography.14 Resolution was considered complete when the percentage reduction was at least 70%. When two leads had an identical degree of ST deviation, the first lead encountered (I, II, III, aVL, aVF, V₁–V₆) was taken into account.15

Statistical analysis

Statistical analyses were performed using SAS software V 9.1. Categorical data are summarised as percentages; continuous variables are presented as medians with interquartile range (IQR). Univariate comparisons were carried out using the χ² test for qualitative variables. A Student t test (or a non-parametric test when the distribution of the variable was not normal) was used to compare quantitative data according to coronary patency. Variables associated with TIMI flow with a p value <0.20 in the univariate comparisons were used for backward multivariable logistic regression analysis to determine odds ratios (ORs) that were independently associated with the probability of presenting with TIMI 3 flow.

Two different analyses were carried out. The first was based on prethrombolysis, prehospital variables to identify independent predictors of coronary patency, before thrombolysis. The second used additional variables available at the time of coronary angiography to assess independent predictors of coronary patency, before coronary angiography. The former scenario corresponds to the ability to identify, in advance, which patients are likely to derive benefit a priori from prehospital thrombolysis. The latter scenario reflects the ability to predict for which patients prehospital thrombolysis is likely to succeed or to fail and who may therefore need emergency angiography and rescue PCI. The β coefficients of the final multivariable model assessing the probability of coronary patency before coronary angiography were used to compute an equation giving the predicted probability of coronary patency for a given subject, according to clinical and electrocardiographic characteristics. This equation was used to develop a nomogram for predicting the likelihood of TIMI 3 flow after prehospital thrombolysis. ORs are reported with 95% confidence intervals (CIs) and the C statistic is given for each model. All multivariable analyses refer to the population of patients with full angiographic and ECG data (n = 800).

RESULTS

Study group

Between November 2004 and November 2005, 1159 patients with a recent STEMI were enrolled in the study in 79 hospitals in France; 162 patients were excluded for the reasons outlined in fig 1. A total of 997 patients had TIMI flow data available, of whom 813 had their ECGs analysed by the core laboratory. Assessment of ST resolution was possible, based on core laboratory data, in 800 patients (fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Patient enrolment schematic. ECG, electrocardiogram; PHT, prehospital thrombolysis; STEMI, ST-segment elevation myocardial infarction.

The median age of the patient population (n = 997) was 59 years and 18% were women (table 1). Overall, 91% were in Killip class I, 16% had a history of symptomatic coronary artery disease, 9% had a previous myocardial infarction, 8% had undergone a previous PCI and 44% were current smokers (table 1). The infarct location was anterior in 42% of patients. The median (IQR) number of leads with ST elevation ⩾0.1 mV was 4 (3–5), with a median maximal amplitude of 3 (2–5) mm.

Before the onset of STEMI, 15% of patients were receiving long-term treatment with aspirin or thienopyridine, 14% took an angiotensin-converting enzyme (ACE) inhibitor, 13% were receiving a β blocker, 10% a calcium channel blocker and 15% a statin (table 1).

Treatments in the MICU and in hospital

Prehospital thrombolysis was performed a median (IQR) of 110 (75–165) min after symptom onset, with 82% of patients receiving treatment within 3 h. The median time from receipt of thrombolysis to coronary angiography was 105 (76–145) min. Figure 2 shows the median cumulative delay between onset of pain and receipt of treatments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Median delays from symptom onset to patient telephoning mobile intensive care unit (MICU), from phone call to arrival of MICU, from MICU arrival to receipt of thrombolysis and from thrombolysis to coronary angiography. IQR, interquartile range.

Tenecteplase was used as the thrombolytic drug in 96% of patients (table 1). Other treatments given in MICUs were heparin for 97% of patients (66% unfractionated heparin, 34% low molecular weight heparin), aspirin in 94% and IV nitrate in 40% (table 1).

Coronary angiography outcomes

In 60% of patients (n = 582), angiography was performed via the femoral approach and in 40% (n = 392) via the radial artery. Overall, 51% (508/997) had a TIMI flow grade of 3 (table 1). PCI was performed in 90% of patients (48% rescue, 52% delayed elective or immediate and performed despite TIMI 3 flow at coronary angiography), with a median (IQR) of 1 (1–1) stent implanted. Among patients who underwent PCI, the proportion of TIMI 3 flow after the procedure was 81% (726/898).

In-hospital clinical outcomes

In the sample, 4.1% of patients (40/973) died after hospital admission, 5.3% had recurrent ischaemia and 6.0% developed cardiogenic shock. The rates of death and death or shock during hospitalisation were lower in patients who achieved TIMI 3 flow (2.8% vs 5.4%, p = 0.039 for death and 4.0% vs 10.2%, p<0.001 for death or shock). Overall, 1.6% of patients developed a cerebrovascular complication (stroke or haemorrhage), and an additional 1.7% had a major haemorrhage during hospitalisation.

Predictors of coronary artery patency (TIMI 3 flow)

In univariate analysis, TIMI 3 flow on initial angiography was associated with Killip class I (assessed either in the MICU or after acute phase treatment), peripheral arterial disease, current or past smoking, no hypertension, five or fewer leads with ST-segment elevation before thrombolysis, relief of chest pain and ST-segment resolution before angiography (table 1). There was a trend towards an association with age (p = 0.054). Delay between symptom onset and thrombolysis was not significantly associated with TIMI flow: median (IQR) delay equal to 111 (75–165) and 110 (75–160) min in patients with TIMI 0–2 and TIMI 3 flow, respectively (p = 0.457). Current and past smokers were considered together in the analyses, as the percentage of TIMI 3 flow was not significantly different in current (55%) and past smokers (53%) (p = 0.642).

Multivariable logistic regression analyses were carried out to assess independent predictors of epicardial coronary patency before thrombolysis (in order to identify, in advance, which patients are likely to derive benefit a priori from prehospital thrombolysis) and before coronary angiography (in order to predict for which patients prehospital thrombolysis is likely to succeed or to fail and who may therefore need emergency angiography and rescue PCI) (table 2). Before prehospital thrombolysis (ie, using the variables available in the prehospital setting), independent predictors of TIMI 3 flow were Killip class I, current or past smoking and five or fewer leads with ST elevation on the ECG. Before coronary angiography, independent predictors of TIMI 3 flow were current or past smoking, five or fewer leads with ST elevation before thrombolysis, Killip class I, chest pain relief and ST-segment resolution ⩾70%. The OPTIMAL nomogram, with the following steps, can be used to estimate the probabilities of achieving or not achieving coronary patency of the infarct-related artery before coronary angiography, using variables available at the time of coronary angiography.

• Find the number of points corresponding to each predictive factor.

• Sum points for all predictive factors (table 3).

• Find the corresponding predicted probability of coronary patency (TIMI 3 flow) (P) (table 4).

• The predicted probability of lack of patency is (1−P).

As an example, for a patient with the following characteristics: complete ST resolution, current smoking, chest pain relief, three leads with ST elevation before thrombolysis and Killip class I, the sum is 10 + 8 + 9 + 8 + 12 = 47 (table 3); the probability of coronary patency is 0.71 (table 4); the probability of lack of coronary patency is 0.29.

DISCUSSION

The OPTIMAL study is the largest prospective observational cohort study to analyse early angiographic patency of the infarct-related artery after prehospital thrombolysis. The results of this study may help clinicians to identify optimal responders to prehospital thrombolysis and decide for which patients the procedure has a high probability of failure and who therefore require emergency coronary angiography and rescue PCI.

There is a strong and direct association between early implementation of reperfusion therapy in acute myocardial infarction and improved clinical outcomes.4 7^–9 Use of prehospital thrombolysis is associated with better survival than in-hospital thrombolysis and, in some settings, provides clinical outcomes similar to those achieved with primary angioplasty.4^–6 Prehospital thrombolysis may also be more effective than primary PCI in preventing death and cardiogenic shock in patients treated within 2 h of symptom onset.10 When appropriate facilities exist and with trained medical and paramedical staff, the ESC guidelines recommend prehospital thrombolysis in patients with STEMI or new or presumed new left bundle branch block.13

Predictors of patency

Before thrombolysis, only Killip class I, history of smoking and a lower number of ECG leads with ST elevation were independent predictors of TIMI 3 flow in the infarct artery. These three variables can be used to identify which patients are likely to derive most benefit from prehospital thrombolysis. When patients were assessed after thrombolysis, before angiography, relief of chest pain and ST resolution were additional predictors of infarct-artery patency, leading to a more precise assessment of the probability of presenting with TIMI 3 flow.

The “beneficial” effect of smoking on TIMI flow may appear to be somewhat paradoxical but has been consistently described in published reports. Because they are usually younger, smokers admitted for acute myocardial infarction experience better outcomes than non-smokers.16 Furthermore, the types of lesions that precipitate myocardial infarction are probably different in smokers and non-smokers, with a greater contribution of thrombus to coronary occlusion, which is the likely explanation of why thrombolysis seems to be consistently more effective in smokers.16 In accordance with the association between smoking and TIMI flow, we observed a significantly higher rate of TIMI 3 flow among patients with a history of peripheral arterial disease compared with other patients, in unadjusted univariate comparison. This statistical association was actually explained by a higher proportion of smokers and past smokers among patients with peripheral arterial disease and it disappeared in multivariable analyses.

A divergence from many of the previously published reports is that prolonged delay from symptom onset to receipt of thrombolysis was not predictive of TIMI 3 flow.1 7 9 17^–19 This finding may be in part related to the fact that in the OPTIMAL study there was a skewed distribution of delays towards short intervals (82% of patients received thrombolysis within 3 h of symptom onset), with few patients presenting after long delays. In the latter, time to thrombolysis might play a more significant role. Time from symptom onset to treatment was also not predictive of TIMI 3 flow in the GUSTO-I (Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries) Angiographic Study.20 In that study, Lundergan et al found use of an accelerated tissue-type plasminogen activator (t-PA) regimen, culprit artery location (right or left circumflex), body weight and history of smoking to be predictive of TIMI 3 flow in the infarct artery at 90 min. t-PA and its derivatives may be less sensitive to time elapsed since symptom onset than streptokinase, for which efficacy decreases with increasing time to treatment.21

Clinical implications

Primary percutaneous coronary intervention is the preferred reperfusion strategy for patients presenting with STEMI and admitted to a hospital with an experienced team of interventional cardiologists routinely performing PCI.22 In areas without PCI facilities, thrombolysis performed within the first 3 h after symptom onset is recognised as a viable alternative, but should be followed by coronary angiography within 24 h.22 These guidelines are based on the results of three randomised controlled trials assessing the superiority of coronary angiography followed by PCI when applicable, as compared with a more conservative strategy based on standard medical treatment unless clinical deterioration occurs.23^–25

Thus, the need for coronary angiography and PCI after thrombolysis is generally accepted, but the optimal delay between thrombolytic administration and angiography is less well established. In particular, it is not known whether immediate transfer to a PCI facility, irrespective of the patient’s clinical and electrocardiographic presentation, is a better choice than transfer a few hours later. Recently, the CARESS-in-AMI (Combined Abciximab REteplase Stent Study in Acute Myocardial Infarction) study examined this subject, providing evidence that immediate transfer of all patients located far from PCI facilities is a better strategy than early transfer restricted to those needing rescue PCI.26 However, only high-risk patients were enrolled in this study. Furthermore, the thrombolytic regimen used (half-dose reteplase associated with a glycoprotein IIb/IIIa inhibitor, aspirin and unfractionated heparin) was not the standard treatment recommended in guidelines, thus placing the usual-care group at a possible disadvantage. However, a standard dose of tenecteplase was used in the TRANSFER-AMI (Trial of Routine Angioplasty and Stenting After Fibrinolysis to Enhance Reperfusion in Acute Myocardial Infarction) study and the superiority of immediate transfer of all patients compared with elective transfer to PCI was highlighted.27 28 These two trials compared two strategies for transferring patients after thrombolysis, but no comparison was done with a strategy of primary PCI. Primary PCI was better than systematic immediate PCI after thrombolysis in the ASSENT-4 PCI (Assessment of the Safety and Efficacy of a New Thrombolytic Regimen) study, at least when patients did not receive glycoprotein IIb/IIIa inhibitor.29

Our study pertains to a broad population of patients, enrolling subjects irrespectively of infarct size, Killip class or left ventricular ejection fraction. Our results identify patients for whom the transfer to a hospital with PCI available cannot be delayed as they are very unlikely to achieve coronary patency and therefore need rescue PCI. Furthermore, our study explores coronary patency after prehospital thrombolysis, whereas most previous studies have considered in-hospital thrombolysis. Since the early use of fibrinolytic therapy enhances the probability of coronary patency, optimum strategies for triaging patients may differ according to whether prehospital or in-hospital thrombolysis is performed.4

The nomogram should not be considered part of a strategy of facilitated PCI, which is not recommended.22 29 30 A strategy of facilitated PCI is an initially planned PCI after the administration of a clot-dissolving drug, whereas the strategy developed in our paper is to determine candidates for rescue PCI among those who have received prehospital thrombolysis. The clinical usefulness of the nomogram lies in its ability to assist medical decisions and guide doctors in the best strategy for managing patients. Patients with a high probability of lysis failure should be transferred urgently to a catheterisation laboratory using the fastest possible means of transportation (eg, helicopter, if applicable), whereas angiography may be more delayed in patients with a high probability of successful lysis.

Study limitations

Patients who presented with delays greater than 3 h between symptom onset and thrombolysis were under-represented. The use of the nomogram for prediction of TIMI 3 flow may be less accurate in patients treated later after symptom onset; in particular, time from symptom onset to thrombolysis may then play a more significant role. Thus, the accuracy of the nomogram will need to be tested prospectively in a more general population—for instance, in patients treated later because prehospital thrombolysis was not available. Our population may also be biased against sicker or dying patients, including those who were being ventilated or defibrillated and could not be enrolled in the study. The potential benefit of the nomogram for triaging patients should also be tested in a randomised trial performed in areas located far from PCI facilities (or more generally, when delay until PCI may be long), comparing the nomogram with strategies used previously and analysing “hard” clinical-outcome data.

Core laboratory identification of the infarct-related artery was not possible in 194 patients. ECG monitoring was not continuous; instead, measurements to detect ST-segment changes were taken at two time points. ST deviation is, however, a dynamic process with spontaneous intermittent recanalisation and reocclusion occurring in the early phase, hence measurements taken at different times might have yielded conflicting results.31 Chest pain was analysed in a binary fashion, as opposed to quantitative analysis using an analogue visual scale. Importantly, the use of morphine and other painkillers had no impact on the association between chest pain and patency.

CONCLUSIONS

The OPTIMAL study is the largest prospective observational study of angiographic patency of the infarct-related artery after prehospital thrombolysis. The population comprised patients who were “less selected” than those enrolled in clinical trials and reflects “real-life” contemporary clinical practice.32 Our study provides a simple nomogram for predicting the success of prehospital thrombolysis and may assist doctors in triaging patients to emergency angiography and rescue PCI, when thrombolysis is unlikely to achieve effective recanalisation.