The performance of six recently introduced highly sensitive enzyme-linked immunosorbent (ELISA) urine pregnancy test reagent kits was evaluated for false positive results using 100 male and 100 postmenopausal female urine specimens and the findings were compared with those of a qualitative radioimmunoassay (RIA) procedure. Based on the findings, the most suitable pregnancy test reagent kit was then selected for doing routine pregnancy testing of urine samples of premenopausal patients. The less sensitive PregnaSTICK ELISA method and Concept RIA procedure did not give any positive results. Positive results for postmenopausal female and male urine samples were obtained as follows: Testpack, 2 and 0 (greater than 50 mIU/mL [IU/L]); Icon, 2 and 0 (greater than 50 mIU/mL [IU/L]); Quest, 4 and 2 (greater than 50 mIU/mL [IU/L]); Nimbus, 17 and 4 (greater than 25 mIU/mL [IU/L]); and Sensi-Chrome, 33 and 19 (greater than 50 mIU/mL [IU/L]), respectively. The medical records of the patients whose urine samples gave positive results were examined for information that would have explained the positive results, but no clear-cut reasons were found. Comparison of the routine urinalysis findings showed that there was a correlation between the mucus content of female (but not of male) urine samples and the incidence of false positive human chorionic gonadotropin results. During 12 months of routine use of the Icon reagents for pregnancy testing of premenopausal urine samples, the University of Texas Medical Branch staff has not reported any suspected false positive findings to the authors.