Target condition being diagnosed

Tubal pathology of the fallopian tubes is a major cause of subfertility in women; prevalence in western countries is around 10% to 30% (Evers 2002). Tubal pathology is often the result of pelvic inflammatory disease (PID), caused by Chlamydia trachomatis (Paavonen 1996), which is the most common sexually‐transmitted disease in western countries. C. trachomatis infection remains asymptomatic in 70% to 80% of women, and so is not treated, thereby increasing the risk of the infection ascending to the internal female genitalia.  C. trachomatis mainly invades the columnar epithelium of the distal tubes, leading to inflammation, fibrosis (formation of fibrous tissue) and occlusion of tubes and/or peritubal/peri ovarian adhesions. Tubal testing is part of the basic fertility check‐up process designed to detect such abnormalities. A large variety of tubal tests are available to the clinician, the characteristics of which have been summarized recently (Papaioannou 2004).

Laparoscopy with dye patency testing is regarded as the reference standard of the available tests for detecting tubal pathology. It allows the clinician to examine tubal patency (function), possible endometriosis, and whether peritubal/peri ovarian adhesions are present under direct view (Swart 1995). Furthermore, if necessary, treatment procedures can be carried out directly. It is a costly and invasive procedure, however, which carries some serious risks, such as potential damage to the bowel, bladder and blood vessels, although these complications are rare. Moreover, as theatre time is required for laparoscopy, there is a limit to the number that can be performed in any healthcare setting, thus restricting widespread use.

Index test(s)

In view of these limitations, Chlamydia antibody titer (CAT) testing is propagated as a primary investigation to identify the women most likely to have, or have had, tubal disease. The rationale behind CAT testing is the detection of immunoglobulin‐G (IgG) antibodies that are formed as a result of infection with C. trachomatis. These antibodies are more likely to be present in patients with tubal damage than in patients without tubal damage (Punnonen 1979), and so the test should enable diagnostic use of laparoscopy to be prioritised

Nowadays, the majority of CAT testing is performed with (micro)‐immunofluorescence (MIF) and enzyme‐linked immunosorbent assays (ELISA). A MIF essay is an indirect IgG antibody test, in which C. trachomatis serovars or elementary bodies are used as antigens. A positive reaction is characterized by specific fluorescence of the elementary bodies. For quantitative determination, serial dilutions are made. In ELISA testing (synthetic) peptides of the cellular components of C. trachomatis, which bind with the patients’ antibodies, are used. Anti‐human IgG antibodies, that form a complex with IgG antibodies, are added. The intensity of this reaction can be quantified by using a spectrophotometer.

Since the description of the association between C. trachomatis IgG antibodies and tubal pathology, several studies have reported on the predictive value of CAT testing, the results of which have been summarized in a meta‐analysis reported over a decade ago (Mol 1997). This study reported on 23 studies, representing 2,279 patients and evaluated the performance of four different techniques, namely microimmunofluorescence (MIF), immunofluorescence (IF), enzyme‐linked immunosorbent assay (ELISA) and immunoperoxidase (IP) assays. Participants included in these studies were subfertile women, and accuracy was expressed as sensitivity, specificity, likelihood ratios and prevalence, calculated from the data in the primary articles. Studies in which it was not possible to (re)construct a 2 x 2 table to cross‐tabulate CAT results and the findings of the reference standard were excluded. Data extraction was performed by two independent reviewers, with the availability of a third reviewer to resolve any disagreements. This meta‐analysis showed that the sensitivity of Chlamydia antibody testing for tubal pathology varied between 0.21 and 0.90, while the specificity varied between 0.29 and 1.00. There was significant heterogeneity among the included studies, with logistic regression analysis indicating that the discriminative capacity of the test varied with the definition of tubal pathology, and type of assay used. There was no difference between studies that verified tubal pathology with laparoscopy only and studies verifying tubal disease with both laparoscopy and/or hysterosalpingography (HSG) (radiography of the uterus and fallopian tubes after the injection of a contrast medium).

As a result of the variable accuracy in the subgroups mentioned above, a subgroup analysis was performed according to assay and definition of tubal pathology. Despite this, a significant proportion of heterogeneity remained and could not be explained.

Alternative test(s)

HSG is another option for a primary investigation to detect tubal pathology. HSG is a more invasive procedure than CAT, which only requires a blood sample. HSG is an imaging technique in which oil, or water–based, contrast medium is injected slowly through the cervical canal into the uterine cavity and fallopian tubes. The uterine cavity, fallopian tubes and abdominal cavity (if the tubes are patent) can be visualized by means of X‐rays. HSG is an outpatient procedure, but may be uncomfortable for the patient (Liberty 2007). In addition to its diagnostic capacity, use of oil‐contrast medium during hysterosalpingography enhances the chance of spontaneous pregnancy (Johnson 2005).

In 1995 a meta‐analysis assessed the value of the hysterosalpingogram in diagnosing tubal patency and peritubal adhesions, compared with laparoscopy with dye patency testing as the reference standard (Swart 1995). It reported on 19 studies (3,964 women) that evaluated tubal patency, and 13 studies (1,894 women) that evaluated peritubal adhesions. Homogeneity could not be rejected for a subgroup of three studies that judged HSG and laparoscopy independently, and pooled summary estimates of the sensitivity and specificity were calculated for these studies. In all studies of tubal patency, the sensitivity ranged from 0.34 to 1.00 while the specificity ranged from 0.55 to 0.99. For the three studies for which homogeneity was not rejected, a summary point estimate of 0.65 for sensitivity was calculated (95% confidence interval (CI) 0.50 to 0.78) versus 0.83 (95% CI 0.77 to 0.88) for specificity. This review did not supply an overview of study characteristics, which might explain some of the heterogeneity found between articles.

Since the publication of these two systematic reviews, there have been several reports on the diagnostic accuracy of CAT and HSG for tubal pathology in subfertile women. No direct comparative meta‐analysis has yet been made of the diagnostic characteristics of CAT and HSG. We therefore propose a comparative systematic review of these two techniques with laparoscopy as the reference standard, using up to date meta‐analytical methods developed specifically for test‐accuracy studies.