Cardiac troponin T (cTnT) levels are widely used to assess for evidence of myocardial infarction. We studied the effect of freezing and long-term storage on the stability of cTnT in blood samples from 178 patients with end-stage renal failure. The serum was separated and divided into multiple aliquots. Baseline cTnT levels were measured in the unfrozen aliquots. The remaining aliquots were frozen using standard techniques. The aliquots were thawed after 3, 6, 12, or 24 months and cTnT levels measured. There were no significant changes in the mean +/- SEM cTnT levels up to 12 months (0.111 +/- 0.098 microg/L) compared with baseline (0.114 +/- 0.098 microg/L); after 24 months, cTnT levels were significantly lower (0.107 +/- 0.095 microg/L) than baseline ( P = .004). The cTnT assay is a reliable method of measuring the cTnT level in human serum up to 12 months of frozen storage. However, after 24 months, the cTnT level was 0.007 microg/L lower than baseline, potentially causing erroneous interpretations. The clinical significance of the change in the cTnT level after long-term frozen storage is unclear. Further studies, including prospective analysis of patient outcomes, should be helpful.