PCR and culturing were compared for the routine diagnosis of Chlamydia trac
homatis infections. Two laboratories experienced in both techniques partici
pated in the study, which included 513 specimens. Both techniques were perf
ormed on each specimen; the portion of the specimen used for PCR was divide
d in two, and each half was sent to one of the two laboratories, where the
tests were run in a blinded fashion. The PCR primers used by the two labora
tories matched different parts of the bacterial genome. PCR inhibitors were
looked for in all specimens. Overall, PCR was more sensitive than culturin
g; the difference was marked for sperm and endopelvic specimens and nonsign
ificant for urethral and cervical specimens. False-positive PCR results wer
e few in number, there were no consistent false-positive results when each
specimen was amplified twice. PCR inhibitors were rarely present in urethra
l and cervical specimens but were found in 7% of sperm and endopelvic speci
mens. PCR inhibitors should be looked for routinely during PCR testing of s
perm or endopelvic specimens.