Correlation between urine and cervical specimens for the detection of cervical Chlamydia trachomatis and Neisseria gonorrhoeae using ligase chain reaction in a cohort of HIV infected and uninfected adolescents
L. Peralta et al., Correlation between urine and cervical specimens for the detection of cervical Chlamydia trachomatis and Neisseria gonorrhoeae using ligase chain reaction in a cohort of HIV infected and uninfected adolescents, J ADOLES H, 29(3), 2001, pp. 87-92
Purpose: (a) To examine the concordance between ligase chain reaction (LCR)
results from urine and cervical samples for Chlamydia trachomatis and Neis
seria gonorrhoeae in HIV infected and uninfected adolescent women, and (b)
to examine factors that may influence the concordance of LCR results in thi
Methods: Baseline specimens from 269 of 334 female subjects enrolled in a l
ongitudinal study of HIV infection in adolescents were analyzed for C. trac
homatis and N. gonorrhoeae using ligase chain reaction (LCR) assays in a ce
ntral laboratory. Concordance was measured using kappa coefficient with per
mutation analyses to calculate the difference between HIV status groups. Di
scordant LCR results were examined for the co-infection with the other micr
oorganism, bacterial vaginosis, or Trichomonas vaginalis.
Results: The prevalence of C. trachomatis detected by LCR in the HIV infect
ed and uninfected groups was 19.3% and 12.2%, respectively (p =.16); the pr
evalence of N. gonorrhoeae was 7.0% and 2.4%, respectively (p =.16). Urine
LCR assay sensitivity to detect cervical C. trachomatis infection was 86% (
95% CI: 68%-96%) in the HIV infected group and 100% (95% CI: 69%-100%) in t
he HIV uninfected group. Urine LCR assay sensitivity to detect cervical N.
gonorrhoeae infection was 92% (95% CI: 62%100%) in the HIV infected group.
There were only 2 N. gonorrhoeae infections in the HIV uninfected group, an
d both were urine LCR positive. Differences in sensitivity between HIV infe
cted and HIV uninfected subjects were not statistically significant. Coinfe
ction with N. gonorrhoeae, bacterial vaginosis or Trichomonas vaginalis was
not associated with the concordance of urine and cervical LCR results.
Conclusion: The relatively high sensitivity of urine LCR testing overall su
ggests that urine screening may be reasonable for sexually active adolescen
t females with or without HIV infection in situations in which urine screen
ing may be more acceptable than pelvic examinations. Society for Adolescent