First-catch urines from 276 asymptomatic male military recruits were screened by polymerase chain reaction for the detection of Chlamydia trachomatis. Eight initially positive specimens were retested by polymerase chain reaction, ligase chain reaction and transcription-mediated amplification. Urine specimens from six (2.2%) subjects were considered to contain C. trachomatis. However, retesting of serially collected urines from five of these six subjects using different nucleic acid amplification methods showed some discrepancy. This may have a major impact on the efficacy of screening programs for C. trachomatis in low prevalence populations.