Objectives: To compare echocardiography (ECHO) and radionuclide ventriculography (RVG) in the monitoring of left ventricular systolic function during doxorubicin therapy in adult lymphoma patients.
Design: Prospective study.
Settings: University hospital.
Subjects: A total of 28 adult patients who received doxorubicin to a cumulative dose of 400-500 mg m(-2).
Main outcome measures: ECHO and RVG were performed at baseline and after cumulative doxorubicin doses of 200, 400 and 500 mg m(-2).
Results: At baseline, the mean (+/-SE) left ventricular ejection fractions (LVEF) were 58 +/- 1.3, 71 +/- 1.8 and 58 +/- 1.7% as determined by RVG, M-mode ECHO and two-dimensional (2D) ECHO, respectively. After the cumulative doxorubicin dose of 500 mg m(-2) LVEF decreased to 49.6 +/- 1.7% (RVG) (P < 0.001), 62 +/- 1.6% (M-mode) (P=0.006) and 52.5 +/- 1.3% (2D ECHO) (P=0.036). Although a significant correlation between LVEF determined by RVG and M-mode ECHO (r=0.615, P=0.002) and a trend between RVG and 2D ECHO (r=0.364, P=0.096) were observed, there were substantial differences in the results of individual patients. In the agreement analysis using the method of Bland and Altman there was a mean difference of 12% units with the upper limit of agreement +26% units and the lower limit of agreement -2.1% units for LVEF determinations with M-mode ECHO and RVG, and a mean difference of 3.3% units with upper and lower limits of agreement +19.6 and -13.1% units for LVEF determinations with 2D ECHO and RVG, respectively.
Conclusion: We found only a moderate agreement between left ventricular systolic function determined by ECHO and RVG methods. Thus, in the follow-up of left ventricular function in adult patients during doxorubicin therapy, the guidelines based on LVEF measurement by RVG cannot be applied to ECHO. Consequently, RVG remains the method of choice in this context.