ObjectivePoint-of-care (POC) cardiac troponin (cTn) assays have a rapid turnaround time but are generally less sensitive than laboratory-based assays. Previous research found that the Abbott i-Stat cardiac troponin I (cTnI) assay has good diagnostic accuracy when used with the Troponin-only Manchester Acute Coronary Syndromes (T-MACS) decision aid and serial sampling over 3 hours. Accuracy of other assays may differ. We therefore evaluated the diagnostic accuracy of a different POC cTnI assay with serial sampling over 3 hours, both with T-MACS and when used alone.MethodsIn a prospective diagnostic accuracy study at eight EDs in England (July 2015–October 2017), we collected clinical data from consenting adults with suspected ACS at the time of assessment in the ED. Blood samples were drawn on arrival and 3 hours later for POC cTnI (Cardio 3 Triage, Alere). The target condition was an adjudicated diagnosis of acute myocardial infarction (AMI), based on reference standard serial laboratory-based cTn testing. We calculated test characteristics for POC cTnI using the limit of detection (LoD, 0.01 µg/L) and the T-MACS decision aid.ResultsOf 347 participants, 59 (14.9%) had AMI. With serial POC cTnI testing over 3 hours, POC cTnI at the LoD cut-off ruled out AMI in 193 (55.6%) patients with 98.1% sensitivity (95% CI 89.9% to 100.0%) and 99.5% negative predictive value (NPV, 95% CI 96.5% to 99.9%). T-MACS ruled out AMI in 117 (33.7%) patients with 98.1% sensitivity (95% CI 89.9% to 100%) and 99.2% NPV (95% CI 94.3% to 99.9%). T-MACS ruled in AMI with 97.9% specificity (95% CI 95.8% to 99.5%) and 83.7% positive predictive value (95% CI 70.6% to 91.7%).ConclusionsWith serial sampling over 3 hours, the Alere Cardio 3 Triage cTnI assay has relatively high NPV for AMI using either the LoD cut-off alone or the T-MACS decision aid. However, wide CIs around the measures of diagnostic accuracy mean that further prospective testing of this strategy is required before clinical implementation.Trial registration numberUKCRN 18000.