One hundred fifty-eight patients receiving Adriamycin underwent 226 transjugular biopsy procedures. The specimens were evaluated by electron microscopy for evidence of drug-related cardiotoxicity. Ejection fraction determinations using echocardiographic or nuclear techniques at rest were available for 69% and 81% of the patients, respectively. Analysis of the data revealed a correlation between cumulative Adriamycin dose and biopsy grade (p less than 0.02). No similar relationship existed between cumulative Adriamycin dose and ejection fractions obtained at rest or between biopsy grades and ejection fractions. In patients who underwent serial endomyocardial biopsies and serial ejection fraction determinations, the correlation between changes in biopsy grade and ejection fraction was poor. A change in resting ejection fraction detected by either method did not reliably predict a change in biopsy grade. The poor correlation between ejection fractions and biopsy grades could be due in part to the sensitivity and specificity of the Adriamycin-related structural changes in contrast to the wider range of disease processes that can affect myocardial function, and to the fact that structural changes often precede the ejection fraction abnormalities. The greater sensitivity and specificity of the biopsy grade should prove useful in reducing the risks associated with evaluating new anthracyclines and potential myocardial protectors of Adriamycin toxicity.