Diagnosis of Genital Tuberculosis in Infertile Women by Using the Composite Reference Standard

Female genital tuberculosis (FGTB) can be asymptomatic or even masquerade as other gynecological conditions. Conventional methods of FGTB diagnosis include various imaging, bacteriological, molecular, and pathological techniques that are only positive in a small percentage of patients, leaving many cases with undiagnosed condition. In the absence of a perfect diagnostic method, composite reference standards (CRSs) have been advocated in this diagnostic study. This study assesses the agreement between traditional diagnostic modalities using CRS and prevalent TB groups among different fallopian tube infertility manifestations. A total of 86 women with primary and secondary infertility were included in the study and subjected to bacteriological, pathological, and radiological examination for the diagnosis of FGTB. Results were evaluated statistically for concordance of the diagnostic tests to the CRS by sensitivity and specificity, while PPV and NPV were calculated for the performance of diagnostic tests of FGTB. We observed that 11.2% of women were found to be true positives by means of CRS. The positive findings by CRS were as follows: ultrasonography (13.9%), laparoscopy (14%), hysteroscopy (12%), GeneXpert (4.8%), culture (4.8%), polymerase chain reaction (4.8%), and histopathology (6.4%). GeneXpert and culture were found to have a perfect agreement with CRS. Hysterosalpingography, laparoscopy, and hysteroscopy have a fair agreement with CRS. Out of 43 women with tubal factor infertility, 6 women were found in the definitive TB group with mixed conditions of tubal manifestations. This study evaluates and demonstrates the reliability of the collective assessment of various diagnostic methods with CRS findings that help in identifying different TB groups of genital tuberculosis patients from all infertile patients by applying the criteria of CRS.


Introduction
FGTB is still a serious concern in low-income nations, causing substantial morbidity, particularly infertility at reproductive age [1]. Due to underreporting of cases, asymptomatic incidences, ambiguous symptomatology, and lack of effective diagnostics with high sensitivity, the exact prevalence of FGTB remains unknown [2,3]. The reported incidence varies by country: 1% in US infertility clinics [4], 1% in Scandinavian countries [5], and 4-8% in Pakistan [6]. South Africa's share is 1% [7]. In different parts of India, the rates range from 16.1% to 19% [8]. The prevalence of female genital TB recently reported in India ranges from 45.1 cases per 100,000 women in the community-based research in the Andaman Islands to 48.5 percent among infertile women in north India [9,10]. The pathophysiology of FGTB in infertile female patients may be considerably influenced by genetics, immunology, environment, and infection [11]. FGTB usually occurs secondary to pulmonary TB [12]. It causes a variety of nonspecific symptoms in women, ranging from infertility to irregular menstruation and pelvic pain [13]. FGTB is regarded as a chronic and asymptomatic or low symptomatic disease, so it may be difficult to diagnose in women with infertility [14]. Due to its paucibacillary nature, the diagnostic dilemma of FGTB continues to be a challenge [15]. A timely diagnosis and effective treatment may prevent it. Bacterial cultures and PCR-based diagnostics are two instances of the more cutting-edge and effective diagnostic methods that are increasingly accessible for the detection of tuberculosis [16]. The sensitivity and specificity are significantly influenced by technical factors as well, including the use of appropriate controls, standard strains, adequate conditions, and the retesting of samples with suspicious positive results [17]. The present study is assessing the combined diagnostic modality for early detection of genital tuberculosis with accuracy. According to the TB guidelines, the diagnosis of FGTB should be made based on any one of the laparoscopic features typical for FGTB, any gynecological specimen positive for acid-fast smear, or positive for mycobacterium tuberculosis (MTB) on culture, any findings consistent with FGTB on histopathology [18]. In the CRS, we included culture, GeneXpert, PCR, histology, radiography, imaging, and history for the diagnosis of FGTB. TB will be confirmed if there are two of the following AFB microscopy/histopathology/consistent feature on USG and HSG/ laparoscopic or hysteroscopy features typical for FGTB or GeneXpert or culture positive in individuals with suspicion of FGTB. PCR or history with imaging features of FGTB will be considered in probable TB category, whereas only imaging features suggestive to FGTB will be grouped into possible TB group and patient tested negative for all tests will be considered as non-TB case. The primary objective of this prospective diagnostic accuracy study was to diagnose FGTB and assess TB in women with tubal factor infertility by means of CRS. We also evaluated the concordance of the diagnostic tests with respect to the CRS by sensitivity and specificity, while positive predictive value (PPV) and negative predictive value (NPV) were calculated for the performance of diagnostic tests of FGTB. women were included in the study with written consent. The unexplained and asymptomatic or low symptomatic infertility and general investigation for infertility helped to make suspicion for secondary tuberculosis. Post hoc sample size for two proportions was calculated. Only 62 women with primary (72%) and secondary (28%) infertility were selected and enrolled according to their clinical presentation. Exclusion criteria are as follows: women over the age of 45, with symptoms suggestive of pulmonary tuberculosis other than infertility, who had taken or were on a regimen of antituberculosis drugs, severe psychiatric dysfunctions, sexual disorders, infertility due to abnormality in ovulation, endocrine problems, pulmonary infections, multiple sclerosis or other autoimmune disorders, human immunodeficiency virus (HIV) and coinfections, diabetes,      3 Disease Markers blockage in the transition zone between the isthmus and the ampulla, calcified lymph nodes, multiple constrictions, and adnexal calcifications that are irregular, linear, or nodular, TB will be highly suspected. Special features such as collarstud abscess, T-shaped uterus, and pseudounicornuate uterus, as well as nonspecific features such as synechiae formation, uterine contour distortion, obliteration of the uterine cavity, and venous and lymphatic intravasations, may be seen as a consequence of tuberculosis. Tubal manifestations were categorized into definitive TB, probable TB, possible TB, and non-TB groups [20].

Endoscopy.
Out of 62 women, 50 women were undergone laparoscopy and hysteroscopy. For the diagnosis of FGTB, microcaseations, and micropolyps, fibrosed ostia, synechia bands, narrow cavity/T-shape cavity, and Asherman's syndrome are considered as diagnostic classification by hysteroscopy. During hysteroscopy, the color of the endometrium, the opening of the endometrial glands, and any TB features such as tubercles, shaggy regions, and intrauterine adhesions were all thoroughly examined. The entire pelvic and abdominal cavity, including the fallopian tubes, uterus, ovaries, Douglas pouch, uterovesical pouch, liver, peritoneum, intestines, and gall bladder, were thoroughly examined for any tuberculous lesions such as tubercles, shaggy areas, pyosalpinx, hydrosalpinx, beading of tubes, pelvic, abdominal or perihepatic adhesions, ovarian tubercu-losis, tube patency, and all other abnormalities carefully studied by laparoscopy. Sacculated tubes, convoluted, fluidfilled vesicles, yellow discoloration of mesosalpinx, hydrosalpinx, lead pipe appearance, encysted fluid collection, tuboovarian mass, pyosalpinx, various grades of pelvic adhesions, and miliary tubercles appearances are considered as diagnostic classification by laparoscopy ( Figure 1) [21].

Processing of Endometrial Tissue Biopsy.
Endometrial biopsies (EMBs) from all 62 women were aspirated using Karman cannula no. 4, between the 20th and 25th day of menstruation in the mini operation theatre of the hospital. In the BSL-3 laboratory, each EMB sample was centrifuged for 20 minutes at 6,000 rpm in a tube containing sterile normal saline, and the pellet (about 1 mL) was transferred to a 1.5-mL Eppendorf tube containing fine glass beads up to one-third of the Eppendorf tube's capacity. For 1 minute, the material was homogenized in a tissue lyser (Bertin Technologies Pvt. Ltd) [22]. Each homogenized tissue sample was divided into three parts: GeneXpert, PCR, and culture. The colony grew on Lowenstein Jensen (L-J) media were again subjected to AFB staining and PCR.
2.5. GeneXpert MTB/RIF Assay. One ml of homogenized EMB was mixed with 2.0 ml of GeneXpert sample reagent. For 30 seconds, the mixture was vortexed. After allowing the sample to stand for 15 minutes at room temperature,  2.9. Histopathological Examination. The biopsy specimens were cut into paraffin-embedded tissue slices and fixed in 10% formalin, hematoxylin, and eosin stains and were used to stain the sections. Caseating granuloma is indicated in samples for the diagnosis of genital tract TB, along with epithelioid cells, giant cells, fibrosis, and lymphocyte proliferation coupled with caseous necrosis (Figure 2)

Performance of Bacteriology Methods.
Out of 62 samples, 3 (4.8%) samples were found to be positive by GeneXpert and culture method. Only one sample was found to be rifampicin resistant by GeneXpert. Further, 5 isolates were also found positive by PCR (Figure 3). However, when these results were compared with CRS, similar 3 samples are found in the confirmed TB group, whereas 2 samples were found in probable TB group and 57 were in non-TB group, respectively, as shown in Table 3.

Performance of Histopathology Methods.
Of 62 samples, 4 (6.4%) samples were found to be positive by HPE. However, when compared with CRS, 4 samples were found in confirmed TB group, and 58 were in non-TB group (Table 4).

3.4.
Agreement of Imaging, Bacteriology, and Histopathological Results with CRS. When the kappa (k) value was calculated, GeneXpert and culture showed perfect   (Tables 1 and 2).

Tuberculosis in Women with Tubal Factors.
A total of 43 were found to have tubal factor infertility. Out of these, only 6 (13.9%) women were characterized in the definitive TB category, whereas 10 (23.2%) samples were found in the probable TB group and 21 (48.8%) samples were categorized in the possible TB group. Among all women, tubal occlusion  7 Disease Markers and tubal dilation are the most prevalent conditions for TB infection, but tubal conditions fall into probable TB and possible TB cannot be ruled out as negative TB. Three samples from the probable TB group had dual conditions (calcifica-tions+ tubal outline irregular, tubal occlusion + peritubal adhesion, and tubal occlusion+ tubal dilation), whereas only one sample with triple conditions (calcifications + tubal outline irregular + tubal occlusion) falls into the same group of TB. One woman found definitive TB with dual tubal abnormal conditions (tubal occlusion + tubal dilation). Two women were found with normal spillage conditions. Only 6 women were categorized in non-TB group (Table 5).

Discussion
According to the guidelines for extra-pulmonary TB for India, criteria should be used to make a diagnosis of FGTB [18]. But problems arise when the presentation is variable, and a high level of clinical suspicion is necessary to make the diagnosis. Around 11% of individuals report having no symptoms other than infertility, and these patients require a diagnostic workup to rule out all prevalent causes of infertility [27].
The presence of AFB on microscopy, culture, or histopathological evidence of TB granuloma provides a definitive diagnosis, but it is only positive in a small number of cases and forces the use of additional modalities such as PCR, hysteroscopy, or laparoscopy findings to make a timely diagnosis for early treatment. However, there is no gold standard approach to detecting FGTB, so in this study, we shared the experience and reliability of our proposed criteria of CRS in the diagnosis of FGTB (Figure 4). We compared a mix of microbiological, histological, molecular, and radiological methods and history of TB. Kappa is a statistical coefficient that evaluates the degree of agreement between two raters (judges) who classify things into mutually exclusive groups [28].
In our study, the agreement between conventional investigation and CRS corresponds to the collective diagnostic model. In GeneXpert, whereas culture has perfect agreement, PCR and HPE have substantial agreement with CRS. Most of the time, endometrial biopsy does not contain sufficient numbers of bacilli for AFB or culture investigation [29]. Several investigations have been conducted in the past using endometrial tissue as well as tissues from other organs, but due to low pick-up rates (1-18%) of bacilli or most probably as a result of the monthly shedding of the endometrium's superficial layers, identification of TB was very poor [22,30,31].
In the imaging, USG shows moderate agreement with CRS, whereas HSG, laparoscopy, and hysteroscopy each have fair agreement, respectively. The sensitivity of the imaging and pathology tests ranged from 75.0% to 85.71%, whereas specificity was found to range from 70.27% to 98.28%. On the other hand, GeneXpert and culture have 100% sensitivity and specificity, but PCR has 100% sensitivity and 96.61% specificity. The three women who tested positive for GeneXpert also tested positive for culture and PCR which indicates active TB. Of the 5 PCR-positive samples, 3 samples were categorized as definitive TB, and 2 were in the probable TB category (Tables 5 & 6). One patient (OBG19RS23) with tubal blockage went under all diagnostic tests, and the findings in the fallopian tubes were beaded tubes and intrauterine adhesion. She was found to be Gen-eXpert positive with rifampicin resistance, and culture positive, PCR positive, and caseous nodules were found in HPE (Table 6). This trend indicates the presence of active bacilli in the patients. Our study is comparable to a recent study conducted by Sethi et al. [3], where PCR was found 22.39% positive and HPE 2.99%. However, no cases were found culture positive. Our findings suggest high reliability on our criteria of CRS, as clinicians can be quick to identify FGTB patients in order to avoid irreversible damage.
In high prevalence countries, mycobacterium culture and histopathology facilities are inadequate [32]. In that situation, the infection is often detected during HSG for the first time in any of the infertility investigations [33]. Furthermore, HSG is still the gold standard for tubal lumen assessment [20] and is a useful method for the diagnosis of female genital TB [34]. Genital TB causes a range of HSG appearances, from nonspecific to specific findings ( Figure 5). According to Chavhan [34] and Afzali [14], the appearance of calcified lymph nodes in the pelvic or along the length of the fallopian tubes may confirm a diagnosis of tuberculosis. In our study, no women with calcified fallopian tubes were found to have definitive TB. But other samples which lie in the probable and possible groups cannot be ruled out as non-TB group. In tubal outline, caseous ulceration of the tubal mucosa results in an irregular, ragged, or diverticular appearance of the tubal lumen contour. Tubal occlusion (18%) is the most common finding by HSG in TB conditions in our study. Scarring can cause several constrictions along the course of the fallopian tube, giving it a "beaded" pattern. Each condition of tubal occlusion from all tubal blockage cases was also found to be prevalent (46.51%) in our study. We found HSG findings with tubal occlusion and dilation were the most prevalent conditions of FGTB in women with tubal factor infertility. The previous history is not much helpful factor to include in the model as only three women with a history of TB out of 11 women were found to be FGTB positive by CRS. The remaining 8 women with history and suffering from tubal blockage with imaging consistency were categorized in probable TB. Overall, out of 62 patients, 8 (12.9%) patients were found TB positive, in which 6 (9.6%) women with tubal blockage and 2 asymptomatic women were found to be TB positive by using CRS (Table 6). By using HSG, the typical radiographic features of genital TB are reliable indicators. All 62 samples (control) from the patients without mycobacterial infections were found to be negative for both PCR and culture as expected. The PCR assay cannot differentiate live and dead bacteria; hence, it is recommended only for new and active cases [35]. We could only include EMB samples in our study as EMB curettage is less invasive. The number of samples is less as it is a one-year study.

Conclusion
Although composite reference standard can minimize the degree of bias, they cannot entirely eliminate it since a 8 Disease Markers combination of imperfect tests is unlikely to generate a composite standard with perfect sensitivity and specificity. In our study, bacteriology had a perfect and substantial agreement with CRS, so we should collect a sufficient endometrial sample using the right approach, preferably laparoscopic or hysteroscopy guided, from a highly suspicious site. The combination of imaging tests with CRS has a better scope as their agreement is fair with CRS. Due to differential causes or earlier damage of fallopian tubes by TB in infertile women, HSG manifestations have high numbers of false positivity, so high suspicion for each patient is required.
Our study demonstrates that, while still using microbiological, histological, and radiological techniques, the composite reference standard incorporates many least invasive diagnostic modalities and is conducive to conclude FGTB.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.