Article Text
Abstract
Objectives: To describe a prehospital thrombolysis (PHT) and expedited inhospital thrombolysis (IHT) programme in south-east Scotland using prehospital 12-lead ECG recordings transmitted by telemetry and autonomous paramedic-administered thrombolysis with decision support being provided by coronary care nurses.
Design: Retrospective observational study.
Setting: Three hospitals in south-east Scotland covering a population of 778 468 served by 54 ambulance vehicles.
Patients: 11 840 patients who telephoned the ambulance service with “chest pain” over 20 months, during which 812 patients were admitted with ST segment elevation myocardial infarction (STEMI).
Main outcome measures: All calls and cardiac/potential cardiac calls to the ambulance service, type/time of patient presentation, symptoms/call/door-to-thrombolysis times.
Results: Of the 11 840 calls to the ambulance service for chest pain over 20 months of the initiative, 60% were cardiac/potentially cardiac-related by Scottish Ambulance Service triage. ST segment elevation was present in 8% of the 5150 12-lead ECGs transmitted by paramedics to the ECG receiving station in the CCU. Over the 20 months, 812 patients were admitted to the three hospitals with STEMI and 71% received thrombolysis. Median symptom-to-thrombolysis times were 91, 148 and 184 min, respectively, in the PHT, telemetry-facilitated IHT and self-presenting IHT groups. Median call-to-needle time for the PHT group was 40 min. In 2/146 cases the cardiologists judged that the patient should not have been administered PHT.
Conclusions: Based on prehospital 12-lead ECG telemetry, it is possible for paramedics and CCU nurses to conduct live reperfusion decision-making in patients with STEMI, with resultant benefits in symptoms-to-thrombolysis time.
Statistics from Altmetric.com
Clinical trials in the treatment of acute ST elevation myocardial infarction (STEMI) have consistently demonstrated the benefits of thrombolysis on mortality. These benefits are time-dependent and, in meta-analyses1 as well as in clinical practice,2 are consistently reported to result in significant reductions in left ventricular systolic impairment and mortality through early administration. It has been argued that very early thrombolysis, whether in the first 60 min,3 70 min4 or 120 min1 following symptom onset, offers the greatest benefit when compared with thrombolysis given more than 2 h after the onset of symptoms.
Various strategies have been designed to reduce delays between diagnosis and administration of treatment with thrombolysis.5 These range from GP diagnosis and administration without 12-lead ECG recording6 to autonomous paramedic administration following a 12-lead ECG recording in the community.7 To reduce inhospital delay (“door-to-needle” time), early warning or pre-alert of the emergency department (ED) of incoming patients allows different models of thrombolysis to be delivered.8 In our own hospital a “fast track” response system was developed to reduce door-to-needle time in the ED.9
In 2002–3 the Scottish Ambulance Service (SAS) began moves towards prehospital thrombolysis (PHT). Paramedics were trained to a nationally accepted standard before being allowed to administer thrombolysis. Legislation has been in place since April 2004 approving both tenecteplase and reteplase under medicines provisions for paramedics. At the same time, the SAS became equipped with ECG machines capable of transmitting the 12-lead ECG by mobile telephone technology. We report our experience of working with the SAS in providing decision support from coronary care unit (CCU) nurses, facilitating PHT as well as expedited inhospital thrombolysis (IHT).
METHODS
Prehospital 12-lead ECG recordings are performed on patients aged over 16 years presenting with non-traumatic chest pain or other symptoms which the paramedic suspects to be cardiac-related and potentially ischaemic (for brevity, the term “paramedic” has been used interchangeably to include either paramedics or ambulance technicians who perform and transmit 12-lead ECGs but are not licensed to administer thrombolysis). The ECG is transmitted to a receiving station in the CCU by mobile telephone technology.
A Lifenet RS 12-lead ECG receiving station (Medtronic Physio-Control, Basingstoke, UK) was installed in The Royal Infirmary of Edinburgh which is the tertiary referral centre for south-east Scotland, with a 10-bed CCU, 2 catheter laboratories, 24 hour/7 day ED and 24 hour/7 day CCU junior grade doctor. The population of 778 468 is served by 3 acute receiving hospitals, 7 ambulance stations, 54 vehicles, 160 paramedics and 200 technicians. In addition to standard equipment, the ambulance thrombolysis pack includes a 10 000 unit vial of tenecteplase and a 5000 unit vial of heparin sodium. Responsibility for dealing with the 12-lead ECG arriving at the receiving station in the CCU is assigned either to the nurse in charge or a junior grade doctor.
Paramedics are instructed that if they receive no telephone call within 1 min of transmitting the 12-lead ECG, they can assume that the receiving nurse/doctor does not believe the ECG to show changes compatible with STEMI. The ECG receiving station alerts staff via very loud audio and visual alarms and is answered very promptly in every case. As an autonomous practitioner, the paramedic can ultimately choose to administer or not administer tenecteplase regardless of the advice given by the CCU.
Where there is a technical failure to transmit the ECG from the field, paramedics continue with their usual process of care and destination hospital. In the event of ST segment elevation, they are encouraged to telephone the CCU and describe the ECG changes over the telephone, giving PHT where they consider it appropriate. The paramedic is advised to admit the patient directly to the nearest appropriate CCU (with or without PHT) rather than to the ED.
Statistical methods
In order to compare the results between two normally distributed groups, two-sample t tests were used. In cases where the distribution did not appear normal (eg, symptom-to-call time), the Mann-Whitney test was used. In cases where the distribution did not appear normal and there were more than two groups (eg, symptoms-to-thrombolysis time and symptoms-to-hospital time), Kruskal-Wallis tests were used.
RESULTS
The initiative was implemented in October 2004. In the 20 months to May 2006 the SAS attended 11 840 calls within the catchment population where “chest pain” was the main presenting complaint. A total of 7060 calls (60%) were thought to be cardiac or potentially cardiac in nature by SAS triage. A 12-lead ECG was transmitted to the ECG receiving station in 5150 cases, 73% of calls with cardiac/potentially cardiac chest pain.
Table 1 shows the distribution of ECG findings. ST segment elevation of ⩾0.1 mV in leads I–III, aVF, aVL, V4–V6 or ⩾0.2 mV in leads V1–V3 was present in 412 (8%) of the ECGs. Horizontal or downsloping ST segment depression of ⩾0.1 mV in two contiguous leads was present in 1030 (20%). Of the 412 ECGs classified as arrhythmias, 358 (87%) showed atrial fibrillation with ventricular responses between 48 and 178 beats/min.
A total of 6179 patients had a 12-lead ECG recorded, 87% of cardiac/potentially cardiac-related chest pain calls. The 12-lead ECG failed to transmit to the receiving station due to technical problems in 144 cases (2%) and the paramedic did not attempt to transmit it in 885 cases (14%).
During the 20-month study period 812 patients were admitted to the three hospitals with STEMI. Thrombolysis was administered to 575 (71%). Patients presented via the ambulance service in 544 cases (67%), of whom 146 (27%) received PHT, 247 (45%) received IHT and 151 (28%) did not receive thrombolysis (table 2). Twenty patients (2%) with STEMI arrived at hospital within 1 h of the onset of symptoms, 3 were self-presentations and 17 from the SAS. A total of 222 patients presented between 1 and 2 h, 32 (12%) of which were self-presentations and 190 (35%) via the SAS. Fewer patients presented at >6 h via SAS than self-presentations (38 (7%) vs 43 (16%)).
Of the patients transported by the SAS who received thrombolysis either before arriving at hospital or through telemetry-facilitated IHT (thrombolysis given in hospital following a pre-alert by 12-lead ECG transmission and direct admission to the CCU, n = 393), there were no significant differences in symptoms-to-call time (p = 0.067), SAS response time (p = 0.492) or time from symptoms-to-hospital arrival (p = 0.893). Patients in the PHT group had significantly reduced SAS call-to-thrombolysis times (40 vs 86 min, p<0.001), symptoms-to-thrombolysis times (91 vs 148 min, p<0.001) and proportion of patients given thrombolysis within 60 min of calling the SAS (130 (89%) vs 4 (2%)). The SAS spent more time on-scene (26 vs 22 min, p<0.001) and travelling to hospital (47 vs 42 min, p = 0.002) when giving PHT than when transferring for telemetry-facilitated IHT (table 3).
When comparing those receiving IHT, the median door-to-thrombolysis time in patients receiving telemetry-facilitated IHT was 18 min (range 4–48) compared with 30 min (range 6–94) in patients receiving IHT who self-presented to hospital (p<0.001). Median symptoms-to-thrombolysis times were significantly different between the three groups: 91 min (range 28–527) in the PHT group, 148 min (range 62–442) in the group given telemetry-facilitated IHT and 184 min (range 32–517) in the self-presenting group given IHT (p<0.001).
Emergency percutaneous coronary intervention (PCI), either for failure to achieve ⩾50% ST segment resolution on the 90 min post-thrombolysis ECG or for acute reinfarction within 6 h following initially successful ECG evidence of reperfusion, was performed in 172/812 patients (30%). There were no significant differences between the three groups (32% for PHT, 30% for telemetry-facilitated IHT and 28% for self-presenting IHT).
DISCUSSION
We can find no previous reports of autonomous paramedic-administered thrombolysis with telemedicine decision support provided predominantly by nurses in a CCU. ST segment elevation was present on the ECGs of 3.5% of all calls for chest pain, which corresponds to previous observations.10 A study of advanced medical priority dispatch (AMPDS) codes used by ambulance services has shown that 5% of all calls with “chest pain” have a hospital diagnosis of acute coronary syndrome.11
The observed rate of failure to transmit a prehospital 12-lead ECG due to technical problems of 2% was not unreasonable and is similar to the findings of previous Scottish investigators.12
Patients in this study were much more likely to use ambulance services than North American and Swedish patients,10 13 14 and those using the SAS were more likely to arrive at hospital within 0–2 h than self-presenting patients with STEMI. This may be particularly relevant to discussions surrounding optimal reperfusion strategies in patients presenting early after the onset of symptoms, with particular regard to the benefits of primary PCI or thrombolysis within 3 h of the onset of symptoms.15
Improved time to diagnosis,16 time to treatment17 and outcomes18 have been previously demonstrated by a strategy of prehospital diagnosis. This study shows a twofold increase in median symptoms-to-thrombolysis time from prehospital administration, through telemetry-facilitated inhospital administration to inhospital administration in self-presenting patients. MINAP data reported a PHT rate of 10% in England and Wales in 2005.19 In all cases of STEMI receiving thrombolysis, a prehospital administration rate of 25% in our study appears reasonable given that up to 20% of patients with STEMI may have a contraindication to thrombolysis20 and 10% presented >6 h after the onset of symptoms (and were thus precluded from PHT under the Joint Royal Colleges Ambulance Liaison Committee (JRCALC) guidelines21 in place at the time).
Two previous studies22 23 have emphasised the constraints placed on PHT by the JRCALC guidance. Hanson and Williamson23 recommended revision of the guidelines in order that they become more inclusive. It is encouraging that the JRCALC has made moves to extend the time window from 6 h to 12 h from symptom onset.
SAS response times were excellent and on average were better than the national performance indicator of 8 min. Although there was a statistically significant difference, the numerical difference of 4 min spent on-scene between PHT and non-PHT groups may be an encouraging finding for networks planning primary PCI as either part or all of a reperfusion service in STEMI. The extra 4 min spent delivering the intravenous heparin and thrombolytic agent suggests that, even with PHT, 85% of the ambulance on-scene time is spent reaching the patient, primary and secondary assessment, therapeutic intervention, 12-lead ECG, communication with CCU and decision-making regarding the most appropriate destination for the patient. This may influence those who believe that primary PCI will be a paradigm shift for ambulance services, given that the assessment, diagnosis and treatment processes remain the same, with only the 10 s bolus administration of tenecteplase being omitted.
While the paramedic can ultimately choose to administer or not administer tenecteplase regardless of the advice given by the CCU, this did not happen during the 20-month feedback. Feedback from both paramedics and CCU nurses was that the telephone discussion was mutually beneficial. Paramedics felt supported in the knowledge that they had a live link to experts in the care of STEMI, particularly in the context of cases which were outwith the rigid JRCALC guidance.
There are a number of potential areas for further study. The one-third of patients who continue to receive no thrombolysis regardless of how they present to healthcare services are arguably the most pressing of these. The finding of 29% in this study is similar to findings of the GRACE investigators. Their registry of 16 814 patients with STEMI across 113 hospitals in 14 countries reports that 29% of patients did not receive reperfusion therapy.24
The 25% of patients who either had an ECG performed but not transmitted by the paramedic or who did not have an ECG performed require further investigation. This may simply reflect confidence among paramedics who feel secure in their interpretation of the ECG. With increased pressures on unscheduled/emergency medical services, the 20% of patients who had a prehospital 12-lead ECG showing ST segment depression or T wave inversion (60% of whom had a diagnostic rise in troponin I at 12 h) could perhaps be admitted directly to a cardiology unit for further assessment, thus reducing pressure on the emergency/medical assessment departments.
Although not prospectively measured, the CCU nurse in charge independently received and reported 85–95% of incoming 12-lead ECGs, with 5–15% involving the CCU junior grade doctor. In a study by Quinn et al,25 95% of CCU nurse respondents were found to reach a standard of 12-lead ECG interpretation in patients with ST segment elevation deemed as “gold standard” by a panel of cardiologists. There were six cases during the 20-month study period where there was not a subsequent rise in cardiac enzymes following thrombolysis administration by a paramedic. Four of these were thought by their cardiologist to have clinical myocardial infarction and ECG evidence deserving of PHT and two were thought not to have merited PHT.
The technique of prehospital 12-lead ECG transmission by cellular telephone was first described in 1987.26 Before this, only single-lead rhythm strips could be transmitted to the ED. Previous investigators12 found that 2% of attempts at prehospital 12-lead ECG telemetry were unsuccessful, identical to the failure-to-transmit rate in our study.
We found that 67% of the patients with STEMI presented to hospital via the SAS. In the 329 122 patients sub-studied in the large US NRMI-2 registry over 4 years in 1674 hospitals, 53% of patients were transported to hospital by ambulance.10 The non-users of emergency medical services (self-presentation or self-transport) were younger, had less previous history of cardiovascular disease, presented later after the onset of symptoms, were less likely to have anterior wall infarction or sustained ventricular tachycardia/fibrillation and had a lower killip class, TIMI risk score and hospital mortality (5.5% vs 14.3%). The odds of ambulance use were reported to increase by 21% over each decade of life. Previous North American studies of patients with suspected or confirmed STEMI have reported varying ambulance use of between 33%13 and 59%.27 Similar findings were reported in Sweden where the 34% of patients who used emergency medical services were older, had a greater likelihood of previous cardiovascular disease and were more likely to have a final diagnosis of myocardial infarction (69% vs 38% of patients not transported by ambulance).
Herlitz et al14 concluded that, after correcting for confounding factors, patients admitted to Swedish EDs with chest pain who were transported by ambulance had a much higher mortality during the subsequent 5 years than those not transported by ambulance.
Limitations of the study
Long term outcome data are not reported which may have been useful in determining the benefits (or otherwise) of prehospital 12-lead ECGs, particularly where 92% of patients did not have ST segment elevation. Although data are available to suggest the number of patients receiving PHT who perhaps should not have, there are no data reported on patients who could have received PHT but did not. In governance as well as research terms it is unfortunate that we made no prospective documentation of the decision-making discussion between paramedic and hospital nurse/doctor. This is something we will remedy as a result of this study. The statistically significant differences in SAS on-scene time and travel time may have been a type I error caused by the relatively small sample size.
CONCLUSIONS
Using cellular telephone technology, it is possible for paramedics and CCU nurses to conduct live decision-making based on a transmitted prehospital 12-lead ECG in patients with STEMI. This is an important finding and may create opportunities for reconfiguration of STEMI services. Prehospital administration of thrombolysis was associated with significantly reduced symptoms-to-thrombolysis times. Patients received IHT following direct CCU admission as a result of prehospital diagnosis more expeditiously than patients who self-presented to hospital.
A prehospital diagnosis of STEMI and decision-making process involving paramedics and CCU nurses may aid reductions, not only in symptoms-to-thrombolysis time, but also in the development of primary PCI services by reducing diagnosis to balloon time and creating a hybrid system for optimal reperfusion in STEMI.
Acknowledgments
The authors thank all the paramedics, nurses and junior grade doctors involved in this ongoing initiative. Cat Graham provided statistical support. Ian Archibald of the Scottish Ambulance Service (SAS) was integral in planning, implementing and troubleshooting the SAS aspects of the initiative.
REFERENCES
Footnotes
Competing interests: None declared.
Linked Articles
- Primary survey