(424) Variations in Testosterone Levels with Time of Blood Sampling in Real World Ambulatory Care

Introduction It is well understood that the time of blood sampling is quite important in measuring serum total testosterone (TT) levels for proper diagnosis. It is well understood that the time of blood sampling is quite important in measuring serum total testosterone (TT) levels for proper diagnosis. As such, we recognize the diurnal variation in TT levels throughout the day and are concerned about obtaining TT levels through random blood sampling. This variability could potentially have a significant impact on research findings using such data. Objective However, in actual outpatient care, it is often difficult to ensure the blood sampling time to be ideally constant due to various restrictions on both the hospital and patient sides. Therefore, we investigated the degree of diurnal variation in TT values at different sampling times based on a real-world clinical database of prostate cancer cases in our hospital. Methods To make the background of patients as uniform as possible, we referred a clinical database of 615 consecutive cases of localized prostate cancer diagnosed between 2007 and 2014 and treated with definitive IMRT alone in our single hospital. Excluding 33 patients whose pretreatment TT values were deficient, 582 patients were evaluated. Our laboratory system automatically records the time of each blood draw. We retrospectively surveyed the blood sampling times for TT measurement at the first visit of the patients from the database and compared TT values by each time zone. Results A study cohort of 582 patients was stratified into 4 groups according to time periods of blood sampling: 8-10AM (A1, n=47, 8.1%), 10AM-0 PM (A2, n=141, 24.2%), 0-2PM (P1, n=106, 18.2%) and 2PM- (P2, n=294, 50.5%). First, the 4 groups were compared for age, body mass index (BMI), serum PSA, and TT levels. A statistically significantly lower TT value was observed only in the P2 group (A1, A2, P1, P2; 4.22, 4.20, 4.11, 3.73 ng/mL, median, p=0.007; Kruskal-Wallis test). The P2 group was also significantly younger in age (A1, A2, P1, P2; 73, 71, 71, 70 years; p=0.001). Though an initial appointment for consultation of IMRT arrangement are open all day, it is speculated that senior patients tended to come earlier time of day, while relatively younger patients who still might have been in active social life had to visit hospital later time of day. There were no statistical differences in terms of BMI and PSA between the groups. Next, propensity score matching was used to adjust for age and BMI between the two groups; (A1+A2+P1) and P2. In each parameter-matched group consisting with 184 patients, TT was significantly lower in the P2 group (Median P2; 3.90 ng/mL versus A1+A2+P1; 4.38 ng/mL; p=0.004). Conclusions In the outpatient clinic, there was a significant variation in TT values depending on the time of blood sampling, with a median decrease of 11%, especially after 2 PM. This reiterates the need for keeping consistent time for blood sampling when measuring TT even in the ambulatory care. Disclosure No