Temporal (circadian) and functional relationship between prostate-specific antigen and testosterone in healthy men

To study the circadian relationship between serum prostate-specific antigen (PSA) and total testosterone in men without clinically evident prostate disease.

Blood samples were collected every 3 hours for 24 hours (eight per subject) from 11 clinically healthy men, ages 46 to 72 years. PSA was also monitored once a week for 6 weeks in 16 additional healthy men. PSA, testosterone, and age were correlated by linear regression, and 3-hourly PSA and testosterone values normalized to percent of individual mean were analyzed for circadian rhythm by the least squares fit of a 24-hour cosine.

Mean PSA correlated positively (P < 0.001) and testosterone correlated negatively (P = 0.014) with age and inversely with each other (P < 0.001). The mean circadian range of change (ROC) from lowest to highest values for PSA was 0.37 +/- 0.07 ng/mL (28 +/- 9%), and for testosterone it was 202 +/- 23 ng/dL (53 +/- 7%). The mean ROC over 6 weeks was 0.32 +/- 0.04 ng/mL. A significant circadian rhythm was found for PSA (P = 0.011, amplitude = 5.4 +/- 1.8%, acrophase = 5:02 AM; 95% limits, 2:40 to 7:24 PM) and testosterone (P < 0.001, amplitude = 9.4 +/- 1.8%, acrophase = 8:38 AM; 95% limits, 7:12 to 10:04 AM).

The temporal relationship between circadian rhythms in PSA and testosterone suggests different physiologic states over the 24 hours, which may be of chronopharmacologic interest with regard to dosing time of drugs or hormonal treatments intended to affect prostate growth and function. Within-day variation in PSA is of little diagnostic significance and does not prevent accurate clinical classification when a single specimen is used.