Myocardial infarction: rapid ruling out in the emergency room

Patients with symptoms of possible acute coronary syndrome make up a large proportion of people who present to emergency departments, where they undergo lengthy, intensive, and costly assessments. Yet few are finally diagnosed with an acute coronary syndrome. Improvements in methods to exclude acute coronary syndrome are needed to reliably reassure and safely discharge low-risk patients who can then proceed to further investigations as outpatients. High-sensitivity cardiac troponin assays are reliable and have low thresholds of detection. But how to take full advantage of this improved precision in clinical care is unclear...

Patients with symptoms of possible acute coronary syndrome make up a large proportion of people who present to emergency departments, where they undergo lengthy, intensive, and costly assessments. 1,2 Yet few are fi nally diagnosed with an acute coronary syndrome. Improvements in methods to exclude acute coronary syndrome are needed to reliably reassure and safely discharge low-risk patients who can then proceed to further investigations as outpatients. High-sensitivity cardiac troponin assays are reliable and have low thresholds of detection. But how to take full advantage of this improved precision in clinical care is unclear. [3][4][5] In The Lancet, Anoop Shah and colleagues 6 report results of a prospective observational cohort study of 6304 patients presenting at emergency departments with suspected acute coronary syndrome. They identifi ed a threshold for a high-sensitivity cardiac troponin assay below which patients were at low risk of type 1 myocardial infarction and potentially suitable for early discharge from hospital. The primary outcome was index myocardial infarction, or subsequent myocardial infarction or cardiac death at 30 days. The investigators derived 5 ng/L as the optimum cutoff for cardiac troponin I concentration, using a predefi ned negative predictive value of 99·5% as an acceptable threshold of safety. This threshold enabled a large proportion of the total patients presenting to emergency departments (2311 [47%] of 4870) assessed for possible myocardial infarction to be identifi ed on presentation as at low risk for events at presentation and at 30 days. For those patients who were not identifi ed as having a myocardial infarction on presentation (2311 [61%] of 3799), this threshold had a negative predictive value of 99·6% (95% CI 99·4-99·9). The results were validated in two cohorts of 1434 patients presenting to emergency departments with similar positive fi ndings (overall negative predictive value 99·4%, 95% CI 98·8-99·9). The negative predictive value, although an accepted method of evaluating screening tests of exclusion, is aff ected by disease prevalence, and this should be considered in diff erent emergency department populations.
These fi ndings are highly promising. The investigators identifi ed some aspects of clinical practice that might have aff ected the fi ndings. Fewer than 42% (1608 of 3799) patients had serial troponin testing despite guideline recommendations. 7,8 Patients with a delayed increase in troponin after the initial test might therefore not have been identifi ed as having an acute myocardial infarction, and missed events might be more common than reported. Furthermore, the median time for the single troponin test was 54 min (IQR 33-85) after presentation to the emergency department. In systems that support very early blood sampling in the emergency department, the threshold of 5 ng/L might not have such a high negative predictive value. In addition, although early presenters represent only a small proportion of all patients (5%), the use of the single troponin test value failed to meet the predefi ned negative predictive value of 99·5% in these patients, and serial testing should continue in such patients.
Additionally, there are important considerations relating to troponin assays when interpreting (and considering implementation of) the fi ndings of this study. First, all troponin assays are diff erent, and the cutoff and fi ndings described by Shah and colleagues 6 are specifi c to the troponin assay that they used. These do not apply to any other assays, even other highsensitivity assays. 9 Second, high-sensitivity cardiac troponin assays provide improved analytical precision at low

Comment
With the emergence of eff ective treatments for infl ammatory arthritis, the new concept of treat to target has evolved. Treat to target is defi ned as "a treatment strategy in which the clinician treats the patient aggressively enough to reach and maintain explicitly specifi ed and sequentially measured goals, such as remission or low disease activity". 1 A proactive clear endpoint, which is the aim of the tre atment, should be used as a specifi c target algorithm. This endpoint should be supported by fi ndings from randomised controlled trials which suggest that early aggressive treatment approaches are advantageous. In rheumatoid arthritis, this treatment approach has been proven to be eff ective in the Tight Control of Rheumatoid Arthritis (TICORA) trial. 2 Indeed, an international task force has now published treat-to-target recommendations for rheumatoid arthritis. 3 In psoriatic arthritis-an infl ammatory musculoskeletal disease associated with psoriasisascertainment of the treatment target has been more

Is it time for treat to target in psoriatic arthritis?
concentrations of troponin compared with previous generations. Clinicians should be aware that results are usually less reliable at troponin concentrations lower than at the 99th percentile. 9 To be classifi ed as high sensitivity, an assay must have a coeffi cient of variation of at least 10% at the 99th percentile for the assay. When considering local implementation of this diagnostic strategy, the assay's precision should be discussed by the laboratory and clinicians (eg, specialists in emergency medicine, cardiology, and internal medicine) with an assessment made about the ability to maintain acceptable precision. Reassuringly, in this study, 6 the interlaboratory precision across 33 instruments at 3·5 ng/L was good, suggesting that the accuracy of this assay in clinical practice might be reliable. The ultimate validation for the safety and effi cacy of discharging patients with cardiac troponin concentrations less than 5 ng/L will be the report of clinical outcomes after this threshold is implemented in routine clinical practice.
Finally, what further assessment, if any, is needed for those patients identifi ed as low risk and suitable for early discharge? Trials are needed to assess the safety and eff ectiveness of clinical pathways that involve no further testing for such patients.
Shah and colleagues' study 6 is a huge advance in the assessment of patients with possible acute coronary syndrome in emergency departments. We strongly urge close collaboration between front-line clinicians and their laboratory colleagues to identify optimum assessment strategies, including consideration of troponin assay availability and reliability, before local implementation.