Until recently, testing patients for chlamydia and gonorrhea involved collecting samples from the cervix or urethra and sending them for cultures. The discomfort and embarrassment associated with sampling these sites may deter some high-risk patients from seeking testing for these treatable infections. An alternative to cervical or urethral cultures is nucleic acid amplification tests (NAATs) that can be performed on urine samples. Three types of NAAT are commercially available: polymerase chain reaction (PCR), transcription-mediated amplification, and strand displacement amplification. However, concerns remain about the sensitivity and specificity of the NAAT compared with culture techniques. Cook and colleagues conducted a systematic review of the literature to compare the sensitivity and specificity of the NAAT for detecting chlamydia and gonorrhea in urine samples with that of cervical and urethral samples.

A total of 1,150 potentially eligible articles were identified by a MEDLINE search and supplemental hand searches of three journals focusing on sexually transmitted diseases. Twenty-nine studies published between 1994 and 2004 met prespecified inclusion criteria, which required separate data presentation by sex, comparison between urine and cervical or urethral samples, and an appropriate reference standard. The appropriate reference standard required confirmation of a positive test at a separate anatomic site by culture or a different NAAT than the one being studied. Study sizes ranged from 126 to 3,551 participants; 23 out of 29 studies were conducted at self-identified “sexually transmitted disease” or “family planning” clinics. The prevalence of infections varied widely (3.3 to 21.5 percent for chlamydial infections and 1.2 to 24.0 percent for gonococcal infections), and most studies did not distinguish between asymptomatic and symptomatic patients.

Data were pooled for specificity and sensitivity into four groups: chlamydia in women, chlamydia in men, gonorrhea in women, and gonorrhea in men. All three NAATs had greater than 95 percent specificity for both infections in urine, cervical, and urethral samples. Transcription-mediated amplification and strand displacement amplification had sensitivities between 80 and 93 percent. Urine samples were comparable in sensitivity to cervical or urethral samples, with one notable exception: PCR of urine samples detected only 56 percent of gonococcal infections in women, compared with 94 percent for cervical samples.

The authors conclude that all three commercially available NAATs have high sensitivity and specificity for chlamydia regardless of the site of sample collection. With the exception of PCR for detection of gonorrhea in women, urine samples perform as well as samples collected from other sites. The authors suggest that the widespread use of NAATs on urine samples may improve adherence to screening recommendations for these common sexually transmitted diseases.