Background: The unique demands of cardiac transplantation in infancy have led to non-invasive rejection-surveillance strategies. ECHO-A is a multiparametric, two-dimensionally guided, M-mode analysis algorithm that assigns an empirically derived score for deviations of recipient parameters to age-adjusted, population-based normal values. A cumulative ECHO-A score > or =4 is highly predictive of endomyocardial biopsy Grade > or =3 and of cellular rejection.
Methods: This study determined whether modifying ECHO-A to score for deviations of recipient parameters from the recipient's baseline would improve the predictive power of ECHO-A. We reanalyzed 701 consecutive echocardiograms of 18 pediatric cardiac transplant recipients (median age at transplantation, 142 days) and based scoring on significant (Z score > or =1) deviation from the patients' baseline means (ECHO-B).
Results: Eight episodes of treated rejection occurred during the first year after transplantation (median, 1.4 years). Approximately 10% (72) of the analyses had ECHO-A scores > or =4 that were not associated with treated rejection and were considered false positives. We identified parameters that contributed to the false-positive evaluations and calculated patient-specific baseline mean +/- standard deviation. The ECHO-B, in comparison with ECHO-A, decreased the number of false positives from 72 to 10, increased specificity from 90% to 99%, and increased the positive predictive value about 4-fold (10% to 44%). With treated rejection episodes, ECHO-B increased ECHO-A scores in 7 of 8 recipients and increased the mean score from 6 to 8.
Conclusions: analysis algorithm based on change from baseline improved the positive predictive power without reducing the negative predictive value of multiparametric quantitative analyses of echocardiograms following pediatric heart transplantation.