Potential performance of a 0 h/1 h algorithm and a single cut-off measure of high-sensitivity troponin T in a diverse population: main results of the IN-HOPE study

Abstract

Abstract Aims Chest pain is a major cause of medical evaluation at emergency department (ED) and demands observation to exclude the diagnosis of acute myocardial infarction (AMI). High-sensitivity cardiac troponin assays used as isolated measure and by 0- and 1-h algorithms are accepted as a rule-in/rule-out strategy, but there is a lack of validation in specific populations. Methods and results The IN-HOspital Program to systematizE Chest Pain Protocol (IN-HOPE study) is a multicentre study that prospectively included patients admitted to the ED due to suspected symptoms of AMI at 16 sites in Brazil. Medical decisions of all patients followed the standard approach of 0 h/3 h protocol, but, in addition, blood samples were also collected at 0 and 1 h and sent to a central laboratory (core lab) to measure high-sensitivity cardiac troponin T (hs-cTnT). To assess the theoretical performance of 0 h/1 h algorithm, troponin < 12 ng/L with a delta < 3 was considered rule-out while a value ≥ 52 or a delta ≥ 5 was considered a rule-in criterion (the remaining were considered as observation group). The main objective of the study was to assess, in a population managed by the 0 h/3 h protocol, the accuracy of 0 h/1 h algorithm overall and in groups with a higher probability of AMI. All patients were followed up for 30 days, and potential events were adjudicated. In addition to the prospective cohort, a retrospective analysis was performed assessing all patients with hs-cTnT measured during the year of 2021 but not included in the prospective cohort, regardless of the indication of the test. A total of 5.497 patients were included (583 in the prospective and 4.914 in the retrospective analysis). The prospective cohort had a mean age of 57.3 (± 14.8) and 45.6% of females with a mean HEART score of 4.0 ± 2.2. By the core lab analysis, 74.4% would be eligible for a rule-out approach (45.3% of them with a HEART score > 3) while 7.3% would fit the rule-in criteria. In this rule-out group, the negative predictive value for index AMI was 100% (99.1–100) overall and regardless of clinical scores. At 30 days, no death or AMI occurred in the rule-out group of both 0/1 and 0/3 h algorithms while 52.4% of the patients in the rule-in group (0 h/1 h) were considered as AMI by adjudication. In the observation group (grey zone) of 0 h/1 h algorithm, GRACE discriminated the risk of these patients better than HEART score. In the retrospective analysis, 1.091 patients had a troponin value of <5 ng/L and there were no cardiovascular deaths at 30 days in this group. Among all 4.914 patients, the 30-day risk of AMI or cardiovascular death increased according to the level of troponin: 0% in the group < 5 ng/L, 0.6% between 5 and 14 ng/L, 2.2% between 14 and 42 ng/L, 6.3% between 42 and 90 ng/L, and 7.7% in the level ≥ 90 ng/L. Conclusion In this large multicentre study, a 0 h/1 h algorithm had the potential to classify as rule-in or rule-out in almost 80% of the patients. The rule-out protocol had high negative predictive value regardless of clinical risk scores. Categories of levels of hs-cTn T also showed good accuracy in discriminating risk of the patients with a very favourable prognosis for cardiovascular death in the group with value < 5 ng/L. ClinicalTrials.gov NCT04756362