Use of cardiac troponin T in diagnosis and prognosis of cardiac events in patients on chronic haemodialysis.

Background: Patients undergoing chronic haemodialysis frequently have elevated serum cardiac troponin T (cTnT) levels resulting in difficulty in diagnosing acute coronary syndromes (ACS) in these patients. We sought to determine whether: (i) cTnT concentrations were consistent over time; (ii) intradialytic changes in cTnT levels were due to haemoconcentration; (iii) baseline cTnT levels predicted subsequent mortality or ACS.

Methods: We measured serial pre- and post-dialysis cTnT concentrations in 75 asymptomatic patients undergoing chronic haemodialysis at baseline, and at 48 h, 8 months and 15 months. At 15 months, we also measured pre- and post-dialysis haematocrit levels in order to adjust the post-dialysis cTnT concentration for the effect of ultrafiltration. Kaplan-Meier survival curves, log-rank tests and Cox models were employed to determine whether baseline cTnT levels predicted death or ACS within 18 months.

Results: Thirty-five (47%) patients had a baseline pre-dialysis cTnT concentration in the diagnostic range for an ACS (cTnT > or = 0.03 microg/l). There was a strong correlation between serial cTnT concentrations in individual patients (P<0.0001 for each time point). The median cTnT concentration was significantly greater post- than pre-dialysis (P<0.01 for each serial analysis); however, there was no significant difference following correction of post-dialysis cTnT levels for the effect of haemoconcentration (P = 0.48). Elevated baseline cTnT levels were associated with an increased risk of mortality or ACS at 18 months (P = 0.0015).

Conclusion: In asymptomatic patients on haemodialysis, serum cTnT concentrations are frequently elevated, and they rise during dialysis due to haemoconcentration. cTnT levels fluctuate minimally in individual patients in the medium term, therefore annual measurements may be useful reference points in the diagnosis of chest pain and in the prediction of ACS and mortality.