Optimisation of the single-sample rule-out pathway for acute myocardial infarction
Z Li, R Wereski, A J F Thurston, Y Y Tew, T Fujisawa, P Fineran, D Harry, G M S Myrmel, A V Ferry, A R Chapman, A Gray, M Tada, K M Aakre, N L MillsAbstract
Background
Single-sample rule-out pathways for myocardial infarction are safe and effective, reducing length of hospital stay and unnecessary admission. However, one in four patients require a second high-sensitivity cardiac troponin (hs-cTn) measurement as they present early or have intermediate values. We evaluated novel approaches to further optimise the performance of the single-sample rule-out pathway.
Methods
In patients with suspected acute coronary syndrome undergoing hs-cTnI measurements in the Emergency Department, we evaluated whether use of the limit of detection (<2 ng/L) in early presenters and earlier serial measurements (1 versus 2 hrs) would increase the proportion ruled out and reduce length of stay without compromising diagnostic performance for spontaneous myocardial infarction or cardiac death at 30 days (Figure 1). Performance was validated in three external cohort studies and the effectiveness was modelled in a real-world cohort of consecutive patients.
Results
In 1,840 patients (59 [51-71] years, 64% males), 495 presented within 2 hours of symptom onset and 301 (16%) had a spontaneous myocardial infarction or cardiac death at 30 days. The proportion ruled out using a threshold of <5 ng/L at presentation was 39% for a negative predictive value (NPV) of 99.6% (95% confidence interval 98.9 to 99.9%) and a sensitivity of 98.8% (97.4-99.7%). Use of a lower 2 ng/L threshold in early presenters increased the proportion ruled-out at presentation from 39% to 47% with similar NPV (99.5 [98.9-99.8]%) and sensitivity (98.7 [96.9-99.5]%). In 2,846 patients from the external validation cohorts, the optimised pathway increased the proportion ruled out at presentation by 2-13% with similar diagnostic performance compared to the original single sample rule-out pathway (Figure 2). In 6,096 consecutive patients evaluated using the original pathway, applying the optimised pathway would reduce the proportion of patients requiring a second measurement from 33% to 29% and the average length of hospital stay by 24 mins (6.8±4.1 versus 6.4±3.9 hrs, P<0.001).
Conclusions
An optimised single-sample rule-out pathway incorporating a lower risk-stratification threshold for early presenters and earlier serial hs-cTn measurements in those not ruled out at presentation could further improve effectiveness and reduce length of hospital stay without compromising diagnostic performance.Figure1.pngFigure2.png