OBJECTIVE To determine if an algorithm implementing a serial high-sensitive cardiac troponin T (hs-cTnT) measurement at presentation (0 h) and at 3 hours after presentation (3h) is helpful for early diagnosis of non-ST-elevation myocardial infarction (NSTEMI) in older patients. METHODS Prospective observational cohort study. METHODS An emergency department (ED) of a city hospital covering a population of approximately 1 million in Germany. METHODS A total of 332 consecutive unselected patients were recruited, of whom 25 had one or more of the prespecified exclusion criteria and 1 had a missing hs-cTnT at 3h, resulting in a final population of 306 patients. METHODS In addition to clinical examination, hs-cTnT was measured at 0 h and 3 h. The final diagnosis of NSTEMI was adjudicated by two independent consultants and an algorithm for rule-in and rule-out of NSTEMI was developed using classification and regression tree analysis. All patients were followed-up for cardiovascular outcome within 12 months. RESULTS Among 306 patients (mean age 81 ± 6 years), 38 (12%) patients had NSTEMI. Accuracy to diagnose NSTEMI was significantly higher for hs-cTnT measurements at 3 h versus 0 h (area under the receiver operating characteristic curve [AUC] 0.88 vs. 0.82, P = .0038) and for absolute versus relative hs-cTnT delta changes (AUC 0.89 versus 0.69, P < .001). A diagnostic algorithm using hs-cTnT values at presentation and absolute delta changes values ruled-in NSTEMI in 23% and ruled-out NSTEMI in 35% of patients. For patients neither fulfilling the rule-in nor the rule-out criteria, an observational zone was established. Cumulative 1-year survival was 79.4%, 88.5%, and 99.1% in patients classified as rule-in, observational zone, and rule-out, respectively. CONCLUSIONS In older patients, serial hs-cTnT measurements and absolute delta-changes at 3h were valuable for early diagnosis of NSTEMI. An algorithm ruled-in NSTEMI in one quarter of patients with high risk and ruled-out NSTEMI in one-third with low risk.