Download this article now for $15.00.
Limitations in the Measurement of Urine Ethanol in Clinical Trials to Monitor Ethanol Consumption
Edward M. Sellers, Karen E. Kadlec, Howard L. Kaplan, Claudio A. Naranjo
Self-report of daily alcohol consumption has been used as the dependent variable in clinical trials to assess the effects of two serotonin uptake inhibitory drugs, zimelidine and citalopram. The validity of the diary data was established by correlating concentration of ethanol in daily urine samples with number of reported standard alcoholic drinks ( r = 0.62 for the 934 subject-days in the zimelidine study; r = 0.54 for the 3,103 subject-days in the citalopram study, both p < 0.0001). The effects of factors other than inaccurate reporting, such as the range of values for reported daily drinks and subjects’ drinking patterns, on the correlation coefficients for all subject-days and for individual subjects in the citalopram study are discussed. Sampling 50% or fewer of the 84 days of the citalopram study for each subject is economically advantageous and did not significantly change the values of the correlation coefficients or the rank positions of subjects but did increase the 95% confidence intervals for the correlation coefficients, indicating less certainty about the actual correlation coefficients and, therefore, the accuracy of a subject’s self-report. The ability of urine ethanol concentration to validate objectively diaries of alcohol consumption is limited by factors that must be considered but are likely to be out of the investigator’s control.
