Are guideline recommended single sample rule-in values for acute coronary syndrome optimum? an analysis of two laboratory high sensitive cardiac troponins in 3543 consecutive presentations

Abstract Background Whilst effective rule-out has improved, single sample rule in (SSRI) by high sensitive troponin remains a challenge for patients with suspected acute coronary syndrome. Purpose To determine the optimal single sample rule in (SSRI) threshold for index type 1 myocardial infarction (T1MI) . Our aim was to maximise the proportion of T1MIs triaged to rule-in whilst retaining as high a positive predictive value (PPV) for SSRI. Methods This analysis was performed using initial sample results from patients recruited to the Mersey Acute Coronary syndrome Rule Out Study. This is a dual-centre, implementation, randomised controlled trial comparing the safety and feasibility of the European Society of Cardiology (ESC) 0/1 hour to an established 0/3 hour pathway. Following enrolment serum samples were analysed in the local laboratory for a hs-cTnT (Roche Elecsys) result. T1MI was adjudicated blinded using the hs-cTnI (Abbott Alinity) results and two consultant cardiologists, with a third acting as a tie breaker. Receiver operator characteristic (ROC) curves were used to determine the optimum SSRI threshold by assay. We studied both the hs-cTnT (Roche Elecsys) and hs-cTnI (Siemens Atellica) assays that have published SSRI values. Results A total of 3543 individual patient samples were used for this analysis. Table 1 shows the demographics of the population. The mean age was 59 years and 53% were male. The prevalence of index T1MI was 7%. Figure 1 plots the proportion of T1MIs triaged to rule in and PPV for an increasing SSRI threshold. At the currently ESC endorsed threshold of 52ng/L 63% of the total T1MIs were triaged to rule in and the PPV was 61%. Our data suggests the optimal SSRI threshold that best balances a high proportion of total T1MIs triaged to rule in and PPV is 60ng/L. Beyond this value the proportion of patients triaged to rule in with T1MIs decreases at a greater rate than the improvement in PPV. We undertook a similar exercise for hs-cTnI. The optimum rule-in value was 30ng/L higher compared to the recommended value. Conclusions Optimum SSRI values can differ from those published or guideline recommended. It is important for laboratories to use local data for SSRI value optimisation.