OBJECTIVES: Chlamydia trachomatis (CT) IgG serology is used in many fertility clinics in order to estimate the risk for tubal factor infertility (TFI) in the fertility work-up. The predictive value for TFI of the currently used mono-target CT serology test should be improved. This study compares the performance of the new multi-target Mikrogen recomWell CT IgG ELISA with the Mikrogen recomLine CT immunoblot and visualizes distribution of individual antibodies in serum with the immunoblot in order to potentially improve the current CT IgG serology test that is clinically used.
METHODS: Study population consisted of 183 Dutch Caucasian infertile women who underwent laparoscopy and/or hysterosalpingography. 48 women had TFI, 135 were controls. Serum was tested with Mikrogen CT IgG ELISA, which detects 3 CT IgG antibodies in one well, and Mikrogen CT immunoblot, which can individually detect 5 CT IgG antibodies. Tests were compared based on the results in general and in the case and control group also taking the individual antibodies into account. Sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), Kappa value and distribution of individual antibodies in positive samples were calculated.
RESULTS: In 183 patients 51% tested positive in the ELISA versus 35% in the immunoblot. 32% versus 65% tested negative. Difference between PPV was not statistically significant (33% and 39% respectively) and NPV in both tests was 81%. Difference in sensitivity and specificity was statistically significant, respectively 65% vs. 52% and 54% vs. 71%. Kappa was only 45%. 64.5% of samples that tested positive with ELISA were positive for at least 4 individual CT antibodies with the immunoblot.
CONCLUSION: The concordance between CT ELISA and CT immunoblot is moderate. Due to separate criteria for positivity of both tests there is a significant difference in sensitivity and specificity. PPV and NPV, the most relevant characteristics for clinicians, of both tests did not differ significantly. The distribution of individual antibodies and the adjustment of the immunoblot algorithm will be further explored in the future in order to develop a potentially better prediction method for TFI with a higher clinical accuracy.