RESUMO
OBJECTIVE: To compare three traditional measures of compliance with antibiotic therapy (parent report diary, preregimen and postregimen bottle-weight difference, and urine bioassay for antibiotic activity), with a deuterium oxide tracer measure of compliance. METHODS: Clinical trial in which all four compliance measures were used for subjects participating in a comparison of the efficacy of azithromycin and penicillin in treating group A beta-hemolytic streptococcal infection. Subjects were 41 children, aged 3 to 15 years (average age, 7.9 years), in a suburban pediatric private practice, who had positive rapid streptococcal antigen test results. RESULTS: Of the 41 subjects, 20 children were randomly assigned to receive azithromycin and 21 to receive penicillin. Compliance was uniformly high by all four measures. Parent diaries indicated that all doses were administered. Urine bioassays were obtained for 40 subjects, and all showed antibiotic activity. Differences in bottle weights were obtained for 27 subjects and showed that 142% of the prescribed medication was missing from the bottles at the end of the regimen. The deuterium oxide measure was obtained for 40 subjects and showed that 107% of the prescribed azithromycin and 92% of the prescribed penicillin were ingested. The correlation coefficient between measured and expected deuterium enrichment was 0.89. There was no significant correlation between the bottle-weight measure and the deuterium oxide tracer. CONCLUSIONS: The bottle-weight measure overestimates compliance; the deuterium oxide tracer is feasible for use in an office setting and produces a high correlation between the expected urinary enrichment and the measured enrichment. Increased use of this quantitative and direct measure would improve the accuracy of compliance measurement in trials of pediatric liquid medications.