PURPOSE: The novel high-sensitivity troponin T assay (hs-cTnT) has been validated for diagnosing AMI in the emergency room. However its utility in high-risk in-patient populations is unknown. METHODS: We retrospectively reviewed admissions to a general cardiology unit that had 2 hs-cTnT measurements in the first 12 h of presentation. We assessed 8 diagnostic algorithms that used hs-cTnT concentration and changes in concentration (including the 99th percentile cut-off of 14 ng/L) for their diagnostic utility in separating AMI patients from cardiac/nonACS and non-cardiac chest-pain patients. UA was excluded. RESULTS: There were 233 patients (mean age 67 years, 153 were males (66%)) admitted over a 2 month period, with AMI diagnosed in 118 of these patients (51%). The recommended 99th percentile cut-off had modest accuracy (65%), good sensitivity (88%), and poor specificity (25%); a higher cut-off of 75 ng/L had a better diagnostic accuracy of 73%, p < 0.05. While some hs-cTnT algorithms were either highly sensitive or specific, none were both. CONCLUSION: In high-risk cardiology in-patients, no hs-cTnT concentration cut-off or change more accurately diagnosed and excluded AMI, although higher cut-offs had better diagnostic utility.
PURPOSE: The novel high-sensitivity troponin T assay (hs-cTnT) has been validated for diagnosing AMI in the emergency room. However its utility in high-risk in-patient populations is unknown. METHODS: We retrospectively reviewed admissions to a general cardiology unit that had 2 hs-cTnT measurements in the first 12 h of presentation. We assessed 8 diagnostic algorithms that used hs-cTnT concentration and changes in concentration (including the 99th percentile cut-off of 14 ng/L) for their diagnostic utility in separating AMI patients from cardiac/nonACS and non-cardiac chest-painpatients. UA was excluded. RESULTS: There were 233 patients (mean age 67 years, 153 were males (66%)) admitted over a 2 month period, with AMI diagnosed in 118 of these patients (51%). The recommended 99th percentile cut-off had modest accuracy (65%), good sensitivity (88%), and poor specificity (25%); a higher cut-off of 75 ng/L had a better diagnostic accuracy of 73%, p < 0.05. While some hs-cTnT algorithms were either highly sensitive or specific, none were both. CONCLUSION: In high-risk cardiology in-patients, no hs-cTnT concentration cut-off or change more accurately diagnosed and excluded AMI, although higher cut-offs had better diagnostic utility.