We describe an improved procedure using a standard microplate immunoassay reader to measure the concentration of troponin T in human serum. We also describe an immunoassay for troponin I in serum. Only 160 microliters of serum are needed for a single analysis of each troponin. For comparison, creatine kinase MB mass analysis in serum was performed with a commercial luminometric method. From 95 apparently healthy people the following values were obtained: creatine kinase MB mass 2.6 +/- 1.2 micrograms/l, troponin T 0.027 +/- 0.025 microgram/l and troponin I 0.03 +/- 0.031 microgram/l. We compared the results of troponin T and troponin I methods with each other, as well as with those of creatine kinase MB mass measured in 48 patients with verified acute myocardial infarction and in 60 control patients with non-cardiac chest pain. The correlation between troponin T and troponin I values was 0.91 for the total material and 0.94 for 48 patients with acute myocardial infarction. Troponin I showed better earlier sensitivity than troponin T (p = 0.043). In nine patients in the control group, creatine kinase MB mass exceeded the reference limit of 5.0 micrograms/l, while in two patients the cut-off limit of 10.0 micrograms/l was also surpassed, pointing to non-specificity. In the group of infarct patients, the highest serum creatinine value was 193 mumol/l, whereas in the control group it was 406 mumol/l. The sera of patients with impaired renal function without any cardiac failure showed no increase in troponin T and troponin I values. In conclusion, serum creatine kinase MB mass and troponin I seem to confirm an acute myocardial infarction more rapidly than does troponin T; troponin I has the highest cardiac specificity.