Endocrine Abstracts

Urinary steroid profiling by capillary gas chromatography is useful for identification of disordered steroid metabolism and, when 24h collections are made, for determination of steroid production rates. We have noted a substantial clinical demand for analysis of random collections, most commonly from children with precocious puberty and a general belief that relating steroid levels to creatinine would offset the disadvantage of a short collection. Given the known circadian variations in steroid but not creatinine excretion, we have considered this unlikely. This study set out to examine the variability of individual values obtained in sequential 3h urine collections over one day in adult volunteers (10m, 10f) in comparison with mean values calculated from the total daily steroid excretion. For each individual, highest and lowest values over the day and the value for the period 0600-0900 (approximating to an 'early morning' collection) were expressed as multiples of the mean. Androgen metabolites (AM, androsterone + aetiocholanolone) and total cortisol metabolites (FM) were calculated. Mean values for males for highest, lowest and 0600-0900 periods were AM, 1.63, 0.56 & 1.01, FM, 1.85, 0.48 & 0.85, FM/AM, 1.38, 0.62 & 1.26, AM/creatinine, 1.30, 0.75 & 1.11, FM/creatinine, 1.55, 0.53 & 1.23. Mean values for females were AM, 2.19, 0.51 & 1.12, FM, 1.77, 0.47 & 0.98, FM/AM, 1.54, 0.50 & 1.27, AM/creatinine, 1.68, 0.49 & 1.02, FM/creatinine, 1.63, 0.39 & 1.35. Although some of the mean values for 0600-0900 were close to unity, their CVs ranged from 13 to 55 percent. There is thus very substantial variability in short collections of urine, even when collection times are specified. Expressing steroid levels relative to creatinine provides no benefit. Although these data were obtained in adults, it may be concluded that assessing androgen production in children using cortisol/androgen metabolite ratios is potentially very misleading.