Noninvasive Cardiac Allograft Monitoring D. Kniepeiss, F. Iberer, B. Grasser, S. Schaffellner, G. Schreier, and K.H. Tscheliessnigg H EART TRANSPLANTATION (HTX) is a generally accepted therapeutic option in patients with end- stage cardiac disease. Immunosuppressive therapy has been improved, but acute cardiac allograft rejection and infec- tion remain as major complications. Obligatory graft sur- veillance is still achieved with the invasive and expensive endomyocardial biopsy (EMB). The analysis of intramyo- cardial electrograms for noninvasive rejection monitoring has been studied by our group. It has been found that the repolarisation phase of the intramyocardial electrogram contains the most accurate information to indicate the absence of rejection. 1 In a recent study, we assessed the superiority of the ventricular evoked response (VER) with respect to beat-to-beat variability, long-term reproducibil- ity, and the independence of the signal from variations of the electrode position. 2 Constant and comparable in- tramyocardial electrograms obtained by a quickly and fre- quently applicable procedure are required for noninvasive heart allograft monitoring. 3 METHODS During HTX, two unipolar, epimyocardial, fractally coated, screw-in electrodes were implanted. The first was screwed in to the right ventricular outflow tract, about 4 centimeters from the pulmonary valve, as near as possible to the ventricular septum. This is the site of first epicardial breakthrough of the ventricular depolarisation and achieves more comparable spontaneous record- ings and an improved ventricular contraction pattern after stimu- lation. 4,5 The other preferably is implanted to the lateral wall of the left ventricle, which presumably enables detection of left and right ventricular dilatation immediately after HTX. Both electrodes are connected to a telemetric pacemaker (Physios CTM 01, Biotronik, Berlin/Germany), which is implanted subcutaneously in the left upper abdominal wall. The optimum display of the VER is achieved by unipolar pacing and signal detection at the same epicardial fractal electrode (ELC 54-UP, Biotronik, Berlin, Ger- many) and by the capability of the telemetric pacemaker to record low frequencies (0.3 to 200 Hz). As a consequence of different studies, 6 routine measurements are scheduled at the same time of the day and patients have to rest for 5 minutes in supine position before we record 1-minute sequences of the VER at pacing frequencies between 100 and 130 bpm. 7 Intramyocardial electro- grams were transferred via the Internet to the central data pro- cessing site in Graz, Austria. After automatic signal morphology checking to exclude spontaneous, ectopic, or fusion beats, each intramyocardial electrogram sequence was averaged and the pa- rameters were extracted. The T-slew-rate parameter is the maxi- mum of the descending part of the T-wave, and the duration parameter is defined as the time between the pacemaker spike and the crossover of the 0-line with the maximum slope of the t-wave of the VER. Within a few minutes, comprehensive patient reports were returned via the Internet. We have defined those parameters to monitor rejection and infection after HTX. The aim of the present study was to investi- gate the value of these parameters for noninvasive cardiac allograft monitoring. The Physios CTM 01 pacemaker was implanted in 42 patients; mean age at heart transplantation was 55.8 years. All patients received an induction therapy with polyclonal antibodies and were maintained on triple-drug immunosuppressive therapy. A total of 1613 follow-ups with intramyocardial electrogram record- ings and 514 EMBs were performed. Repeated surveillance biop- sies guided the indication for rejection therapy. The biopsy results were classified according to the guidelines of ISHLT. The infection status of the patient was scored before data acquisition. RESULTS The prevalence of rejection grade 2 or higher in our biopsy results was 4%. The differences in the VER T-slew rate parameter and the VER duration parameter between cases with versus without rejection grade 2 or higher are shown in the left of Fig 1. Differences between cases with versus without clinically apparent infection are displayed in the right. The normalised VER T-slew rate parameter values that were associated with rejection grade 2 or higher were lower (84  17% vs 103  18%, P  .0002). During the clinically apparent infection, the values decreased as well (90  23% vs 104  17%, P  .0001). The median of the duration parameter values significantly increased under the influence of infection (359 ms to 367 ms), but did not significantly change under the influence of rejection. The application of the single threshold diagnosis model to the VER T-slew rate parameter values with a diagnostic threshold of 90% allowed detection of rejection grade 2 or higher with a negative predictive value of 99% (P  .000001), a sensitivity of 71%, and a specifity of 76%. From the Department of Surgery, Division of Transplantation, Karl-Franzens University of Graz (D.K., F.I., B.G., S.S.), and Institute of Biomedical Engineering, Technical University of Graz (G.S., K.H.T.), Graz, Austria. Address reprint request to Dr D. Kniepeiss, Karl-Franzens- University, Department of Surgery, Division of Transplantation, Auenbruggerpl 29, 8010 Graz, Austria. 0041-1345/01/$–see front matter © 2001 by Elsevier Science Inc. PII S0041-1345(01)02046-2 655 Avenue of the Americas, New York, NY 10010 2456 Transplantation Proceedings, 33, 2456–2457 (2001)