Background Studies assessing variability of serum testosterone levels associated with seasonal environmental factors have been contradictory. Design We assessed associations between the seasons and changes (δ) in seasonality indices and male serum total testosterone (δTT) variability. Patients and Measurements Data were collected in 144 men with paired serum TT samples (126 non-fasting/18 fasting) analysed at Walsall Manor Hospital, UK (52.3 degrees North). Seasonal factors (ambient temperature within 15 min of sampling, humidity, precipitation, duration of daylight on the day of sampling, monthly average ambient temperature, and precipitation) were obtained from local weather-station archives. Sign-rank test determined inter-sample differences between TT and seasonality indices. Linear regression analyses studied associations between δTT and δ seasonal indices in the total cohort and subgroups (stratified by medians of age, TT and men with paired non-fasting samples). Sign-rank determined whether serum TT differed between the seasons. Results Median inter-sample interval was 63 days. No significant inter-sample differences were evident regarding serum TT levels and seasonality indices. No associations were noted between δTT and δ seasonality indices in the total cohort and subgroups stratified by age and TT. Interestingly, δ ambient temperature (p = 0.012) and daylight duration (p = 0.032) were inversely associated with δTT in the 126 men in the non-fasting group (dependent variable). Only a small degree of the variability in the δTT was accounted by the above-mentioned independent variables. The seasons did not appear to influence serum TT values. Conclusions No relation was shown between seasonality and serum TT in the total cohort, thus possibly eliminating a confounding variable that could affect laboratory and clinical practice. It may be that seasonal variation in length of day is too modest at this latitude to demonstrate significant associations, hence our findings are latitude specific. We suggest that further data analysis to address this question in areas with greater seasonal variation would be appropriate.
Read full abstract