Acute myocardial infarction (AMI) is common and is a leading cause of death in the UK, responsible for 7.6% of all deaths in England and Wales in 2004.1 Reperfusion therapy reduces mortality from AMI.2^–4 Prehospital thrombolysis (PHT) is a safe and effective way of administering thrombolysis to patients suffering from AMI.5 PHT leads to a reduction in the time taken between a patient making contact with the health service and thrombolysis being given (the “call-to-needle” time) compared with in-hospital thrombolysis (IHT).6 7 Administering thrombolysis earlier is associated with an improved outcome.8^–10

The term “aborted infarction” (AI) was first introduced in 1993 during the reporting of the MITI trial.11 It was diagnosed in patients who had the evolutionary changes of AMI on their ECG associated with a cardiac enzyme (creatine kinase, CK) level of less than twice the upper limit of normal. Subsequent papers have all defined AI in similar ways:

• The combination of chest pain and transient ECG changes suggesting transmural ischaemia, an increase in CK and its isoenzyme less than twice the upper limit of normal and the cumulative ST segment elevation and depression decreased to <50% within 2 h of treatment.12

• Maximal rise in CK rise of less than or equal to twice the upper limit of normal combined with typical evolutionary ECG changes.8

• Major (ie, >50%) ST segment resolution of the initial ST segment elevation on the presenting ECG indicative of transmural myocardial ischaemia and a lack of a subsequent enzyme rise of more than twice the upper normal limit of CK.13

More recently the Scottish Intercollegiate Group Network (SIGN) has defined aborted (or “threatened”) myocardial infarction as a CK level of less than twice the upper limit of normal or a small troponin release of ⩾0.01 ng/ml and <1 ng/ml for troponin T.14 The British Cardiac Society definition also included an AccuTnI (Beckman Coulter) level of up to 0.5 ng/ml (or other equivalent threshold with other troponin I measurements).15 Essentially, it would seem that AI following reperfusion has occurred when there are evolutionary ECG changes consistent with AMI associated with no (or a small) rise in cardiac markers.

The reported incidence of AI has varied across trials—for example, studies by Lamfers et al12 16 17 reported incidences of 8.6%, 12.6% and 10.4% and a trial by Taher et al8 reported an incidence of 13.3%. The incidence of AI in a population presenting with chest pain and ST segment elevation suggestive of AMI seems to be related to the time that has elapsed between onset of symptoms and reperfusion therapy (“pain-to-needle” time), with the greatest incidence for those thrombolysed within 1 h (23.3%12 and 25%8).

AI has been shown to be associated with a very good prognosis, having a much lower mortality than completed AMI (ie, with a significant cardiac marker rise), with a 12-month mortality of 2.2% in one study16 and a baseline adjusted mortality odds ratio of 0.7 (p = 0.035) in another.8 Given that AI has a reduced mortality associated with it and that it can be clearly defined and measured, it has the potential to be used as a favourable outcome measure in reperfusion trials or as a possible standard for clinical practice. Previous studies of myocardial infarction have appropriately used mortality as a primary outcome measure. Mortality is easily defined but is becoming an increasingly rare event (current short-term mortality in STEMI trials lies at around 5%18), and demonstrating significant mortality benefit requires huge numbers of patients. It would appear from existing literature that AI has a significantly higher frequency as an outcome measure than death.

The relationship between the incidence of AI and PHT has not been specifically investigated within a UK population. Given that, as stated above, AI is associated with earlier thrombolysis and that earlier thrombolysis is delivered in the prehospital setting, we aimed to test the hypothesis that the incidence of AI is directly linked with the administration of PHT in a UK population.

METHODS

A retrospective analysis was performed for all patients administered PHT and IHT in the Avon sector of the Great Western Ambulance Service (GWAS) and in the catchment area of Frimley Park Hospital, Surrey between April 2004 and October 2006. The 12 h troponin level, call-to-needle time (CTN), pain-to-needle (PTN) time, door-to-needle (DTN) time and incidence of AI were analysed. ECGs of patients who did not have a rise in troponin were analysed to ensure that they did indeed have evolutionary changes suggestive of acute infarction (rather than longstanding changes or non-ischaemic ST segment changes, ie, “masquerading” myocardial infarction19 which had been inappropriately thrombolysed).

The information was collected from the Myocardial Infarction National Audit Project (MINAP) databases at the relevant hospitals, from the patient record forms at GWAS for those given PHT, and from hospital case records. Joint Royal Colleges Ambulance Liaison Committee (JRCALC) guidelines were followed for the administration of PHT in both regions.

AI was defined as having occurred in a patient treated with either IHT or PHT who had evolutionary ECG changes and a 12 h troponin I level of <0.5 ng/ml. This definition was chosen to be consistent with the current British Cardiac Society definition.15 All the patients treated with IHT had their ECGs reviewed by a cardiologist and had been ascribed a discharge diagnosis of AMI. The ECGs for those given PHT that fitted the biochemical definition for AI also had their ECGs reviewed to ensure appropriate evolutionary changes.

Analysis of data

Pearson χ² analysis was used to compare proportions and a two-sample t test with unequal variance was used to analyse the comparisons of the means and their confidence intervals. Significance was defined at an alpha value of 0.05.

RESULTS

A total of 472 patients were identified during the study period. Insufficient data were recorded on 18 patients which precluded their inclusion in further analysis (ie, no troponin data or PTN times available), leaving 454 datasets. A further 47 had no 12 h troponin levels recorded, leaving 407 patients with sufficient data to analyse for AI (fig 1). Of the 454 patients in the dataset, 23 had the troponin level reported but no PTN time leaving 431 for analysis of PTN times, and 91 had no CTN times leaving 363 for further analysis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Analysis of aborted infarction (AI). IHT, in-hospital thrombolysis; PHT, prehospital thrombolysis; MI, myocardial infarction; PTN, pain-to-needle time; TnI, troponin I.

Three hundred and thirty-five patients received IHT during the study period and 119 received PHT. All the patients in the PHT group were thrombolysed based on their first prehospital ECG and all those in the IHT group were thrombolysed based on their first in-hospital ECG. This is a requirement of the MINAP registry.

Patients receiving PHT were significantly younger than those receiving IHT (mean age 62.0 vs 68.1 years, p<0.001, mean difference 6.2 years (95% confidence interval (CI) for difference between means 3.8 to 8.5). There was a significantly higher proportion of men (74% vs 67%, p = 0.003).

PTN times were significantly different between the PHT and IHT groups (median 87 vs 144 min and mean of 126 vs 205 min for PHT and IHT respectively, p<0.001, mean difference 79.6 min (95% CI for difference between means 46.2 to 113.0). 47.6% of all thrombolysis was administered within 2 h of symptom onset: 69.1% of patients in the PHT group received thrombolysis within 2 h compared with 40.4% of patients in the IHT group (p<0.001).

CTN times were also significantly different between the PHT and IHT groups (median 39 vs 59 min and mean 40 vs 65 min, p<0.001, mean difference 24.9 min (95% CI for difference between means 22.1 to 27.7). Of those analysed, 66.4% had a CTN time of ⩽1 h. 93.5% of those administered PHT had a CTN time of <1 h compared with 54.9% administered IHT (p<0.001). No patient in the PHT group had a CTN time of >2 h whereas 10 (3.9%) of those administered IHT had a delay of >2 h. The mean DTN time for patients receiving IHT was 25 min.

Of all those with AI, 54% had a PTN time of <2 h. There was a relationship between PTN and CTN times and the incidence of AI, with higher proportions of AI occurring in the earlier time brackets (fig 2). There was a non-significant trend towards a shorter PTN time being associated with a higher incidence of AI (p = 0.06). When the mean PTN time for patients with AI was compared with the mean PTN time for those with an established MI (145.9 vs 185.6 min, mean difference 39.8 min (95% CI for difference between means 5.3 to 74.2)), the difference was found to be significant (p = 0.009).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Comparison of aborted infarction and established myocardial infarction (MI) with pain-to-needle times.

Of the 407 patients for whom 12 h troponin levels were available, 67 had an AI (16.5%); 13 of these had received PHT (11.8% of the PHT group) and 54 had received IHT (18.2% of the IHT group) (fig 1).The proportion of patients fulfilling criteria for AI was therefore higher in the IHT group than in the PHT group, although this difference was not significant (p = 0.124).

DISCUSSION

Consistent with previous studies evaluating early thrombolysis,6 7 we found a significant reduction in PTN and CTN times associated with PHT compared with IHT. Also consistent with other studies,8 16 20 21 we have demonstrated a trend (although not significant) towards an increase in AI with shorter PTN and CTN times.

The overall incidence of AI in our study (16.5%) is comparable to most previous studies (13.3%, 8.5%, 12.5% and 10.4%).8 12 16 17 The GREAT substudy22 reported an AI rate of 46%, much higher than all other trials, but their inclusion criteria were based on ECG changes only and will have included those with electrocardiographic reperfusion who still had a significant rise in cardiac markers. Given its association with good outcome and its reasonably high and measurable incidence, AI could become a useful outcome measure in trials evaluating reperfusion for AMI or a useful audit standard.

We did not confirm our hypothesis of an increased incidence of AI associated with PHT in our two UK settings; indeed, there was a non-significant trend to an increased incidence of AI associated with IHT. This is counterintuitive and is in contrast to the studies performed by Lamfers et al7 16 17 which all showed an increased incidence of AI with PHT. These studies, however, all used the same historical in-hospital comparator group which would not have allowed for changes over time in the in-hospital process that may have taken place to improve DTN, CTN and PTN times.

Clearly there appears to be a relationship between earlier thrombolysis and an increased rate of AI, as supported by our study and data from other trials. It is also clear that PHT reduces PTN and CTN times when compared with IHT and successfully recruits a high proportion of patients into the optimal timeframe for thrombolytic therapy.23^–25 Given these facts, it is unclear why we could not demonstrate a link between PHT and AI; there may be other factors in addition to time to reperfusion which determine which patients are more likely to have an AI.

Not all crews were trained in the administration of thrombolysis (this is a paramedic-only skill in these two ambulance services) and some patients will have been attended by double technician crews; a proportion of these patients may therefore have been eligible for PHT and not received it. We believe that this is unlikely to have affected the AI group disproportionately but do not have the data to confirm this. There were also a significant number of patients bypassing the ambulance service and presenting to the emergency department directly (15% of IHT patients at Frenchay and 20% at Frimley Park during our study), which is an area not studied before.

It is possible that those receiving PHT are more likely to be “barn door” MIs and therefore arguably less likely to have an AI; we do not have the data to support or refute this. It is likely, however, that the patients included in this study who received IHT were also “barn door” given MINAP inclusion criteria: it is the presenting ECG that is acted upon which would exclude “uncertain diagnoses” and “stuttering infarction” (indeed, the mean DTN times for IHT were less than 25 min). We believe that the reasons for receiving IHT rather than PHT are more likely to be related to process issues (eg, double technician crews) and differences between JRCALC guidelines and in-hospital protocols for thrombolysis. Left bundle branch block (LBBB), for example, was not included in the JRCALC guidelines at the time of this study but would have been an indication for IHT. We do not have data on the proportion of patients in the IHT group with LBBB on their presenting ECG.

It should be noted that the PHT and IHT groups were not balanced for age and gender (as reported), and these or other unrecognised factors may be playing a part.

CONCLUSION

PHT is associated with significantly shorter CTN and PTN times than IHT. Our study has confirmed this and is consistent with previous data reporting the overall incidence of AI in patients presenting with AMI receiving thrombolysis. However, despite the known relationship between earlier thrombolysis and AI, we could not demonstrate an association between PHT and AI in our UK population. Indeed, there was a higher proportion of patients with an AI in the IHT cohort. This finding is difficult to explain and may be due to other as yet unrecognised factors which are of importance in understanding the relationship between AI and early thrombolysis.