Background: Testing of urine samples is noninvasive and could overcome several barriers to screening for chlamydial and gonococcal infections, but most test samples are obtained directly from the cervix or urethra.
Purpose: To systematically review studies that assessed the sensitivity and specificity of nucleic acid amplification tests for Chlamydia trachomatis and Neisseria gonorrhoeae in urine specimens and to compare test characteristics according to type of assay, site of sample collection, presence of symptoms, disease prevalence, and characteristics of the reference standard.
Data sources: Relevant studies in all languages were identified by searching the MEDLINE database (January 1991 to December 2004) and by hand-searching the references of identified articles and relevant journals.
Study selection: Studies were selected that evaluated 1 of 3 commercially available nucleic acid amplification tests, included data from tests of both a urine sample and a traditional sample (obtained from the cervix or urethra), and used an appropriate reference standard.
Data extraction: From 29 eligible studies, 2 investigators independently abstracted data on sample characteristics, reference standard, sensitivity, and specificity.
Data synthesis: Articles were assessed qualitatively and quantitatively. Summary estimates for men and women were calculated separately for chlamydial and gonococcal infections and were stratified by assay and presence of symptoms. The pooled study specificities of each of the 3 assays exceeded 97% when urine samples were tested, for both chlamydial infection and gonorrhea and in both men and women. The pooled study sensitivities for the polymerase chain reaction, transcription-mediated amplification, and strand displacement amplification assays, respectively, were 83.3%, 92.5%, and 79.9% for chlamydial infections in women; 84.0%, 87.7%, and 93.1% for chlamydial infections in men; and 55.6%, 91.3%, and 84.9% for gonococcal infections in women. The pooled specificity of polymerase chain reaction to gonococcal infections in men was 90.4%. In subgroup analyses, the sensitivity did not vary according to the prevalence of infection or the presence of symptoms but did vary according to the reference standard used.
Limitations: Few published studies present data on the transcription-mediated amplification or strand displacement amplification assays, and few studies report data from asymptomatic patients or low-prevalence groups.
Conclusions: Results of nucleic acid amplification tests for C. trachomatis on urine samples are nearly identical to those obtained on samples collected directly from the cervix or urethra. Although all 3 assays can also be used to test for N. gonorrhoeae, the sensitivity of the polymerase chain reaction assay in women is too low to recommend its routine use to test for gonorrhea in urine specimens.