Urogenital specimens of 2,485 patients were examined by an enzymeimmunoassay for the detection of Neisseria gonorrhoeae antigen (Gonozyme, Abbott). The results of the original test (Gonozyme A) and a modified version (Gonozyme B) were compared to bacterial culture. Three different groups were examined by Gonozyme A: 526 men and 464 women from the clinic for sexually transmitted diseases (STD) and 548 registered prostitutes. Sensitivity in men was 98%, in women from the STD clinic 89%, and in prostitutes 81%. Specificity in men was 99.5%, in women from the STD clinic 94%, and in prostitutes 88%. One of the reasons for the decreased specificity in women could be cross reacting bacteria from the genital secretions of the women. The assay was modified by optimizing the antigonococcal antibody and increased incubation times. These modifications improved the specificity of Gonozyme B by reducing the number of cross reactions with other bacterial antigens. Four different groups were examined by Gonozyme B: 261 men and 220 women from the STD clinic, 121 women from an infertility clinic, and 345 registered prostitutes. Sensitivity in men was 100%, in women from the infertility clinic 100%, in women from the STD clinic 94%, and in prostitutes 77%. Specificity in men was 99%, in women from the infertility clinic 100%, in women from the STD clinic 99%, and in prostitutes 96.5%. Analysis of the Gonozyme-positive/culture-negative test results suggested that most patients (85%) did not have gonorrhoea. In 15% of these specimens it is possible that patients had pretreated gonorrhoea which was only detected by Gonozyme.(ABSTRACT TRUNCATED AT 250 WORDS)