Purpose: Established diagnostic thresholds for high-sensitivity cardiac troponins (hs-cTn) might not apply for elderly patients as they are elevated irrespective of the presence of an acute myocardial infarction (AMI). Aim of the present study was to investigate hs-cTnI in elderly patients with suspected AMI and to calculate optimized diagnostic cutoffs. Material and methods: Data from a prospective multi-centre study and from a second independent prospective single-centre cohort study were analysed. A number of 2903 patients were eligible for further analysis. Patients > 70 years were classified as elderly. hs-cTnI was measured upon admission. Results: Around 34.7% of 2903 patients were classified as elderly. Around 22.5% of elderly patients were finally diagnosed with AMI. Elderly patients had higher hs-cTnI levels at admission irrespective of the final diagnosis (p < 0.001). According to the AUROC, hs-cTnI was a strong marker for detection of AMI in elderly patients. Application of the 99th percentile cutoffs showed a substantially lower specificity in elderly. By using optimized thresholds, specificity was improved to levels as in younger patients in both cohorts but accompanied with a decrease in sensitivity. Conclusions: hs-cTnI levels have a lower specificity for detecting AMI in elderly patients. This lower specificity can be improved by using hs-cTnI thresholds optimized for elderly patients.