Abstract
Background and aims
Anti-platelet and anti-thrombotic therapy are well-established treatments in acute coronary syndromes. Highly sensitive assays have diminished the positive predictive value of troponin in acute coronary syndromes and increased the importance of the clinical assessment in interpreting positive results. This cohort study sought to investigate over-treatment of non-coronary troponin rises and associated adverse outcomes.
Methods and results
We reviewed 223 consecutive patients presenting to Queen Elizabeth University Hospital, Glasgow, with suspected acute coronary syndromes over a six-week period. Of these, 27 (12%) met our ‘inappropriate therapy’ criteria. This group had a low ischaemic risk (HEART score: 4.2 ± 1.4) (GRACE score: 117 ± 30.8) but an intermediate-high bleeding risk (CRUSADE score: 34 ± 14.5). Approximately half of the patients (14/27, 52%) reported chest pain, with only 4/27 (15%) having ischaemic ECG changes. There were three intracranial haemorrhages, each after the patient had received a single dose of aspirin, ticagrelor and fondaparinux.
Conclusion
The combination of injudicious high-sensitivity troponin testing with potent anti-platelet and anti-thrombotic therapy was associated with possible over-treatment of patients and associated harm. Emphasis on interpretation of troponin in the context of clinical presentation and improved awareness of type 2 myocardial infarction are essential to limit iatrogenic pharmacological harm.
Background and aims
Acute chest pain accounts for 20–30% of all emergency medical admissions in the UK. 1 The early exclusion of acute coronary syndrome (ACS) is a priority, to ensure early treatment and, where appropriate, early discharge of patients.
The use of anti-platelet therapy is well established in the treatment of ACS.2,3 The early VA Cooperative study (1983) and ISIS-2 (1988) trials established the mortality benefits of aspirin.4,5 P2Y12 inhibitors form the basis of dual anti-platelet therapy after the positive findings involving clopidogrel use in the CURE trial. 6 The PLATO and TRITON trials subsequently compared ticagrelor and prasugrel, respectively, to clopidogrel. While demonstrating improved efficacy, these more potent agents were associated with increased bleeding.7,8 The most recent Scottish Intercollegiate Guideline Network (SIGN) and European Society of Cardiology (ESC) guidelines recommend the use of ticagrelor, but with the caveat that clopidogrel may be preferred in patients with higher bleeding risk.9,10 Anticoagulants supplement anti-platelet activity, non-heparin anti-coagulants like fondaparinux having been shown to be the more efficacious agents over low molecular weight heparin (OASIS-5 trial). 11
Types of MI.
The therapeutic approach to type 2 MI is very different to that of type 1. The absence of intraluminal thrombus means that prescribing anti-platelet therapy will expose patients to needless risk of bleeding.
We hypothesised that recent changes in these diagnostic and therapeutic approaches may be associated with over-treatment of patients presenting with possible ACS. Our study aims were to assess the clinical presentation, risk and adverse outcomes of a consecutive cohort of patients assessed through our local ACS pathway.
Methods and results
This was a prospective cohort study of consecutive patients presenting with suspected ACS to the Queen Elizabeth University Hospital, Glasgow. All patients referred for cardiology assessment, either through the emergency department (ED) or acute medical unit (AMU), over a six-week period were included.
Patients who were treated out with established guidelines were identified using ‘appropriateness criteria’. This included patients who were treated with full ACS therapy with a negative troponin and those with clear non-coronary troponin rises.
Each case was adjudicated by a consultant and registrar. Each case designated as having received ‘inappropriate therapy’ was then followed through until discharge with discharge diagnosis and adverse events recorded. Each patient had an objective evaluation of ischaemic (HEART and GRACE score) and bleeding risk (CRUSADE score).14–16
Over our pre-specified six-week period (24/10/2016 until 5/12/2016), 223 patients were referred with suspected ACS. Of these, 27 (12%) were treated with full ACS therapy despite not fulfilling appropriate criteria.
Patient characteristics.
We reviewed the initial presenting complaint – from either the emergency department notes or GP letter as appropriate. The five most common primary presenting complaints are illustrated in Figure 1. While 14/27 patients (52%) also reported chest pain, only 4/27 (15%) had potentially ischaemic ECG changes. Notably, in 10/27 cases, the presenting complaint was a collapse or fall.
Most common presenting complaints.
Ischaemic risk was assessed with the HEART and GRACE scores. The HEART score is trial validated in all patients presenting with chest pain, predicting risk of major cardiac events at six weeks. The mean HEART Score for the ‘inappropriately treated’ cohort was 4.2 (± 1.4), suggesting a relatively low risk of major cardiac events (low risk score < 4, intermediate risk ≥ 4–7, high risk ≥ 7). 14
GRACE score risk stratifies patients with diagnosed ACS to estimate their in-hospital and six-month to three-year mortality. The GRACE score at six months is generally considered the most clinically relevant and is used locally to risk stratify patients prior to angiography following myocardial infarctions (low risk score ≤ 108, medium risk 109–140 and high risk > 140). By definition our study population does not meet the diagnosis criteria for ACS; however, the GRACE score serves as a useful tool to estimate the ongoing ischaemic risk of the population. The mean baseline GRACE score for the ‘inappropriately treated’ cohort was 117 (±30.8). 15
The CRUSADE bleeding risk score is calculated using baseline patient characteristics on admission to hospital. The study defined very low risk (3.1% risk of major bleeding) as a score of ≤ 20, low risk (5.5%) 21–30, moderate risk (8.5%) 31–40, high risk (11.9%) 41–50, and very high risk (19.5%) > 50. Our group had a mean CRUSADE score of 34 ± 14.5. 16
Three intracranial bleeds were identified over the course of the study. Each patient had been given just one dose of aspirin, ticagrelor and fondaparinux prior to their neurological deterioration. Two of these patients presented as falls with documented head injury, and one required neurosurgical intervention. The third patient presented with acute decompensated heart failure with a history of known left ventricular systolic dysfunction. In each of these cases, CT imaging was obtained after neurological deterioration and after prescription of ACS therapy (Figure 2). Whilst there was a possibility that haemorrhage may be been incipient prior to the clinical presentation (e.g. resulting from the fall), the sequence of events (ACS therapy followed by neurological deterioration and radiological evidence of bleeding) suggests that any intra-cranial bleed could have been worsened by the drug therapy. All three patients survived to discharge.
CT brain (non-contrast) showing large subdural haematoma with evidence of midline shift.
There was a wide range of final discharge diagnoses, illustrated in Figure 3.
Final discharge diagnosis.
Unlicensed combination therapies.
Of the 27 patients only 1 patient was discussed with the cardiology team prior to administration of therapy. The occurred despite 12 patients (44%) being admitted between 08:00 and 17:00 and therefore having readily available on-site senior specialist support.
Discussion
Our study confirms concerns that a large cohort of patients presenting with non-cardiac symptoms and non-coronary troponin rises are being exposed to full dose ACS therapy with significant potential adverse consequences. The observations from our cohort serve as a reminder of the bleeding risk associated with these treatments, even after a single dose of therapy.
In 2012, for the first time on modern record, cardiovascular disease was no longer the leading cause of death in the United Kingdom. Between 1961 and 2012, the overall mortality from cardiac causes fell from 52% to 28%.17,18 The adoption of more sensitive biomarkers and more effective therapies has been accompanied by a move towards protocol-driven decision-making in the management of patients with chest pain. This approach has been effective in reducing hospital admissions and length of stay, but is critically dependent on the clinically assiduous selection of patients appropriate to process through these ‘MI screen’ pathways. Otherwise non-coronary patients, often older with multiple comorbidities and higher bleeding risk, may be exposed to inappropriate anti-platelet and anti-thrombotic therapy based on troponin release attributable to a type 2 MI (i.e. secondary to myocardial demand-supply mismatch).
The demographics of our ‘inappropriately treated’ cohort show an elderly group with a mean age of 74. This is significantly older than the study populations on which our management strategies have been based, with the mean age in the PLATO (ticagrelor vs. clopidogrel) and OASIS-5 (low molecular weight heparin vs. fondaparinux) trials being 62 and 66, respectively.7,11 The absence of ECG changes (in 85% of our group) taken together with the HEART and GRACE scores suggest our cohort were of low-to-intermediate risk for ischaemia, while their age and CRUSADE scores suggest they were of intermediate-high risk for bleeding. This highlights the importance of careful pharmacological consideration in elderly patients.
Ticagrelor was reported in the PLATO trial to be associated with a higher rate of intra-cranial bleeding than clopidogrel (0.1 vs. 0.02%, p = 0.02). 7 Subsequent registry data from the CHANGE-DAPT study suggests that the true observed bleeding rate with ticagrelor (and clopidogrel) may be higher than initially anticipated from the selective population in the randomised controlled trial (2.7% vs. 1.2%, p = 0.02). 19 All these features suggest that indiscriminate ticagrelor usage may be a potential source of major harm.
Of greatest concern was the frequent use of ACS therapy in patients who presented with unexplained syncope. The widespread heterogeneity of clinical conditions which can lead to a collapse is well recognised (ranging from aortic dissection to pulmonary embolism) and requires careful deliberation as the approaches to anti-coagulation will often be diametrically opposed. Furthermore, elderly patients are more vulnerable to secondary intracranial bleeding after a fall, which can be potentiated by ACS therapy.
Use of unlicensed combination therapies (such as two P2Y12 inhibitors or combining ticagrelor with full dose NOAC or warfarin) suggests further education is required regarding the specific management of patients already on existing anti-platelet therapy and anti-coagulation.
It is notable that ‘inappropriate therapy’ decisions occurred evenly across normal and out-of-hours working, suggesting that general education and support, not only access to specialist cardiology advice (whilst this is essential), is required to reduce the associated risks. It is probable that guidelines need to further emphasise the distinction between type 1 and type 2 MI, support a ‘first do no harm’ strategy in elderly patients, and provide special caveats for therapeutic decision-making in patient groups already on anti-ischaemic therapy.
Our study is limited by the relatively small cohort. Clinical judgement was used to adjudicate ‘inappropriate therapy’, independent assessment confirmed that no patients in the cohort had a type 1 MI. Our study did not assess patients directly discharged from the ED or AMU, but we assume these patients were considered to be non-ACS and thus lower risk. Whilst this is a single centre snapshot, it does provide a signal of activity which is likely to be experienced elsewhere.
It is notable that these therapeutic decisions occurred both during ‘normal hours’ in addition to ‘out-of-hours’. At the Queen Elizabeth University Hospital Glasgow we have well-established acute cardiology service that primarily operates from 08:00 to 17:00. It is led by easily accessible and visible advanced nurse practitioner (ANPs) based out of the acute medical unit and emergency department. The ANPs have immediate access to both cardiology consultant and registrar input in addition to their own vast experience. Outside of this there is 24 hour availability of on-call advice provided by a cardiology specialist registrar. This suggests that general education and lack of awareness of available support could be a key causative factor.
A re-emphasis of the distinction between type 1 and type 2 myocardial infarctions is required, as is the need to provide special caveats for selecting therapies amongst patient groups already on anti-ischaemic therapy. A new updated chest pain guideline which incorporated all of these key recommendations was subsequently produced (Appendix 1). A focused undertaking to update guidelines to emphasise the need for patient selection, heightened caution for non-cardiac presentation and due attention to bleeding risk is now warranted.
Conclusion
Recent advances in the diagnosis and treatment of ACS have improved patient outcomes and experience, but there is a risk of iatrogenic harm as a consequence of inappropriate use of therapy. Better educational awareness of the causes of non-coronary troponin rises, careful deliberation over ambiguous clinical presentations, and consideration of the balance between ischaemic and bleeding risks, are essential to reduce blanket anti-platelet and anti-coagulation prescribing.
Footnotes
Acknowledgements
Many thanks to Ewan Currie (Senior Cardiology Nurse), Tommy Gardiner, Claire Darroch, Donna Campbell (Cardiology Advanced Nurse Practitioners) and Stephanie Lip (Core Medical Trainee) for their contribution to this project. They have each made key contributions to our project with respect to data collection.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.