Validation of a reference ELISA for estrone glucuronide using urine samples normalized by dilution to a constant rate of urine production

A direct enzyme linked immunosorbent assay (ELISA) system has been optimized as a reference method for the measurement of first statistically significant rises in estrone glucuronide excretion rates in human urine by analysing samples pre-diluted at the time of the collection by the women subjects to a constant urine production rate of 150 mL/h. Validation was achieved by correlation of the individual menstrual cycle profiles with the corresponding estrone glucuronide excretion rates determined by radioimmunoassay (RIA) on the same urine samples for a total of 221 samples from nine cycles. The pre-dilution procedure removed random variations due to fluctuations in the daily rate of urine excretion and minimized between sample matrix effects. When the ELISA data were correlated with the RIA data, Deming regression gave a slope of 1.20+/-0.03 and an intercept of 4.6+/-1.8 nmol/24h (r=0.944) and a random experimental error of 14.2 nmol/24h. The major difference in the measurements was a proportional error of 20%, which was present in either the ELISA or RIA methods or in both. Comparison of the standard normal variate transformation of the ELISA and RIA data gave hormonal profiles of the individual menstrual cycles (N=9) that overlapped almost perfectly. Statistically significant rises or falls in the magnitude of the excretion rate in one profile were mirrored faithfully in the other.