Visual tubal patency tests for tubal occlusion and hydrosalpinx

Background Around 18% to 33% of couples with infertility present with tubal occlusion (blocked or damaged fallopian tubes) or hydrosalpinx (fluid‐filled tubes). Diagnostic laparoscopy (keyhole surgery) with chromopertubation (methylene blue dye tubal patency testing) is considered the reference standard for ruling out tubal occlusion and hydrosalpinx. However, due to its invasiveness and high costs, alternative, less invasive tests have been carried out using imaging techniques. They include sono‐hysterosalpingography (sono‐HSG), hysterosalpingography (HSG), outpatient transvaginal hydrolaparoscopy (THL), and magnetic resonance hysterosalpingography (MR‐HSG). The choice of test varies in different settings; the choice of contrast, operator skill and test technology are factors that can influence diagnostic quality. Furthermore, the performance of the visual tubal patency tests can vary in different populations, depending on whether the test is carried out in an unselected group or in one classified as high or low risk for having tubal pathology.

To determine and compare the diagnostic accuracy of visual tubal patency tests (sono‐HSG, HSG, THL, and MR‐HSG) for the diagnosis of tubal occlusion. Secondary objectives are to determine and compare the diagnostic accuracy of visual tubal patency tests (sono‐HSG, HSG, THL, and MR‐HSG) for the diagnosis of hydrosalpinx and to evaluate heterogeneity concerning population characteristics (population risk stratification) and index test characteristics (contrast media, technology, operator skills). Search methods We searched CENTRAL, MEDLINE, Embase and CINAHL, and two trials registers. We also contacted experts in the field for any additional studies (last date of search: 6 November 2023). Selection criteria We included studies on the diagnostic accuracy of a single index test and studies on the comparative diagnostic accuracy of two or more index tests. Index tests included were: sono‐HSG, HSG, THL, and MR‐HSG. Laparoscopy with methylene blue dye tubal patency testing was the reference standard. We included participants who had been trying to conceive for at least one year. All participants in the included studies should have undergone this reference standard. Target conditions were bilateral tubal occlusion, at least one‐sided tubal occlusion, tubal occlusion by tube, and hydrosalpinx. Data collection and analysis Two review authors independently extracted data. We performed random‐effects meta‐analysis in a bivariate model. For each index test, we presented pairs of sensitivity and specificity with their 95% confidence intervals (CIs) for each study, as well as the pooled sensitivity and specificity in a forest plot. We used the Quality Assessment of Diagnostic Accuracy Studies (QUADAS)‐2 tool for risk of bias assessments. Main results We identified 11,787 records and included 21 studies (1939 participants). Two studies directly compared both sono‐HSG and HSG to the reference test, while the other 19 reported on the diagnostic accuracy of one of the index tests (10 on sono‐HSG, 10 on HSG, and 3 on THL). We did not include any studies of MR‐HSG. Sono‐HSG: for bilateral tubal occlusion, the pooled sensitivity was 0.98 (95% CI 0.19 to 1.00; 3 studies, 259 women; moderate‐certainty evidence) and specificity was 0.99 (95% CI 0.93 to 1.00; 4 studies, 259 women; high‐certainty evidence). Subgroup analyses showed that colour Doppler was associated with higher specificity than standard ultrasound, but with similar sensitivity. We did not find differences in diagnostic accuracy in the use of 3‐dimensional/2‐dimensional or 2‐dimensional ultrasound, or the use of foam or saline as contrast media. We were unable to perform meta‐analysis on the diagnosis of hydrosalpinx. HSG: for bilateral tubal occlusion, the pooled sensitivity was 0.77 (95% CI 0.58 to 0.89; 7 studies, 670 women; very low‐certainty evidence) and the pooled specificity was 0.94 (95% CI 0.87to 0.97; 7 studies, 670 women; moderate‐certainty evidence). For hydrosalpinx by tube, the pooled sensitivity was 1.00 and specificity was 0.96 (95% CI 0.96 to 0.98; 2 studies, 360 tubes). THL: for bilateral tubal occlusion, the pooled sensitivity and specificity were 0.95 (95% CI 0.30 to 1.00; 3 studies, 172 women; low‐certainty evidence) and 0.99 (95% CI 0.84 to 1.00; 3 studies, 172 women; moderate‐certainty evidence), respectively. We were unable to perform meta‐analysis on the diagnosis of hydrosalpinx. There was insufficient information on operator skills and patient risk stratification for all index tests to carry out further analyses. Authors' conclusions The evidence showed that sono‐HSG, HSG and THL are all reliable tests for the diagnosis of double‐sided tubal occlusion. Evidence for comparative diagnostic accuracy studies is very limited. We did not include any studies on the diagnostic accuracy of MR‐HSG. The certainty of evidence ranged from very low to high. This was mainly due to unclear or high risk of bias, heterogeneity and imprecision. Limited reporting on population risk and operator experience may reduce the generalisability of the findings to routine clinical practice. Funding This Cochrane review had no dedicated funding. Registration Protocol (2022) available via: https://doi.org/10.1002/14651858.CD014968 Plain language summary How accurate are imaging tests for diagnosing blocked fallopian tubes and fluid‐filled fallopian tubes? Key messages - The imaging tests sono‐hysterosalpingography (sono‐HSG), hysterosalpingography (HSG), and transvaginal hydrolaparoscopy (THL) can reliably show whether the fallopian tubes are blocked. - HSG seems to suggest blockage of the tubes when there is none more often than sono‐HSG and THL. - We need more research comparing one test against another, and on fluid‐filled tubes and magnetic resonance hysterosalpingography (MR‐HSG). Why is it important to diagnose blocked fallopian tubes? The fallopian tubes connect the ovaries to the womb (uterus). Blocked tubes can prevent pregnancy, as the egg and sperm cannot meet. Diagnosing blocked tubes (tubal occlusion) and fluid‐filled tubes (hydrosalpinx) is essential for choosing the right treatment. Traditionally, keyhole surgery (laparoscopy with dye) is used, but this method is invasive and expensive. Alternative tests have been developed using imaging methods, which are less expensive and less invasive, and can be performed during an outpatient clinic visit. What are the imaging tests? - Sono‐hysterosalpingography (sono‐HSG): salt water or foam is placed in the womb, and an internal ultrasound is used to see if the fluid passes through the fallopian tubes. - Hysterosalpingography (HSG): a contrast fluid is placed in the womb, and X‐ray pictures show whether it flows through the tubes. - Transvaginal hydrolaparoscopy (THL): a tiny camera is placed through the vaginal wall. At the same time, salt water is placed in the womb to see if it comes through the tubes. - Magnetic resonance hysterosalpingography (MR‐HSG): similar to HSG, but a magnetic resonance imaging (MRI) scan is used to make detailed pictures of the tubes. What did we want to find out? We wanted to find out how well these tests can show blocked fallopian tubes and fluid‐filled tubes, compared with the traditional keyhole surgery test. What did we do? We looked for studies that investigated these imaging tests compared to traditional keyhole surgery. We brought together the results from all the studies and checked how well each test could correctly show when the tubes are blocked and when they are fluid‐filled. What did we find? The review included 21 studies with a total of 1939 women: - 8 studies on sono‐HSG; - 8 studies on HSG; - 3 studies for TLH; - no studies for MR‐HSG; - 2 studies for sono‐HSG and HSG Only two studies on sono‐HSG and two studies on HSG reported on fluid‐filled tubes. Three studies were not in English; we translated them online. Nine studies did not clearly explain whether the participants were women with a low or high risk of having blocked tubes. Seven studies did not mention whether the tests were done by a trained person. Blocked tubes Sono‐HSG, HSG and THL are reliable tests for blocked fallopian tubes. HSG may be more likely than sono‐HSG and THL to suggest that tubes are blocked when they are not. If 1000 women were tested for blocked fallopian tubes by the different imaging tests, and 182 of them actually had blockages in both tubes (this number was based on how often the condition was actually found in the studies we included), then: - for sono‐HSG: 186 women would be told they have blocked tubes based on sono‐HSG results. Of these, 8 women would actually have open tubes; 814 women would be told their tubes are open, and of these, 4 women would actually have blocked tubes. - for HSG: 189 women would be told they have blocked tubes based on HSG results. Of these, 49 women would actually have open tubes; 811 women would be told their tubes are open, and of these, 57 women would actually have blocked tubes. - for THL: 181 women would be told they have blocked tubes based on THL results. Of these, 8 women would actually have open tubes; 819 women would be told their tubes are open, and of these, 9 women would actually have blocked tubes. Fluid‐filled tubes - One sono‐HSG study reported results by tube and the other reported results by woman, so we could not combine their results. - Two HSG studies reported results by tube. They showed that all women diagnosed with fluid‐filled tubes had this condition, and four out of 100 women diagnosed with fluid‐filled tubes did not have this condition. What are the limitations of the evidence? The included studies looked at blockages in a single tube, on one side, or on both sides. Only a few studies reported each of these things for each test, so it was difficult to draw conclusions. Also, only a few studies compared one imaging test against another, so we could not say for sure which test is better. How up to date is this evidence? The evidence is up to date to November 2023. Authors' conclusions Summary of findings | Population: women presenting with infertility Prior testing: medical history, physical examination, with or without serological testing (chlamydia antibody testing) Setting: fertility clinics or outpatient Index test: sono‐HSG Reference standard: diagnostic laparoscopy (video‐assisted) with tubal testing with methylene blue Target condition: bilateral tubal occlusion Studies: one direct comparison study, comparing sono‐HSG and HSG (Rezk 2015), and three single‐test studies (Battaglia 1996, Kaur 2015, Soliman 2015) | ||||| | Number of women | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Prevalencea | Certainty of evidence | What do the results mean? | |---|---|---|---|---|---| | 259 women | 0.98 (0.19 to 1.00) | 0.99 (0.93 to 1.00) | 18.2% (182/1000) | Sensitivity: moderateb Specificity: high | With a prevalence of 18.2%, 182 out of 1000 women have bilateral tubal occlusion. - Out of 186 women with bilateral tubal occlusion on sono‐HSG, this is correct in 178 women (96%) and incorrect in 8 women (4%). - Out of 814 women without bilateral tubal occlusion on sono‐HSG, this is correct in 810 women (99.5%), and incorrect in 4 women (0.5%). | | CI: confidence interval; HSG: hysterosalpingography; sono‐HSG: sono‐hysterosalpingography | ||||| | aThis prevalence is based on the actual prevalence observed across the included studies. | | Population: women presenting with infertility Prior testing: medical history, physical examination, with or without serological testing (chlamydia antibody testing) Setting: fertility clinics or outpatient Index test: HSG Reference standards: diagnostic laparoscopy (video‐assisted) with tubal testing with methylene blue Target condition: bilateral tubal occlusion Studies: one study directly compared sono‐HSG and HSG (Rezk 2015), six single‐test studies (Allam 2014; Foroozanfard 2013; Kehila 2014; Sakar 2008; Sharma 2023; Tvarijonaviciene 2008) | ||||| | Number of women | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Prevalencea | Certainty of evidence | What do the results mean? | |---|---|---|---|---|---| | 670 women | 0.77 (0.58 to 0.89) | 0.94 (0.87 to 0.97) | 18.2% (182/1000) | Sensitivity: very lowb,c Specificity: moderateb | With a prevalence of 18.2%, 182 out of 1000 women have bilateral tubal occlusion. - Out of 189 women with bilateral tubal occlusion on HSG, this is correct in 140 women (74%) and incorrect in 49 women (26%). - Out of 811 women without bilateral tubal occlusion on HSG, this is correct in 769 women (95%), and incorrect in 42 women (5%). | | CI: confidence interval; HSG: hysterosalpingography; sono‐HSG: sono‐hysterosalpingography | ||||| | aThis prevalence is based on the actual prevalence observed across the included studies. | | Population: women presenting with infertility Prior testing: medical history, physical examination, with or without serological testing (chlamydia antibody testing) Setting: fertility clinics or outpatient Index test: THL Reference standards: diagnostic laparoscopy (video‐assisted) with tubal testing with methylene blue Target condition: bilateral tubal occlusion Studies: three single‐test studies (Casa 2002; Darai 2000; Watrelot 2003) | ||||| | Number of women | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Prevalencea | Certainty of evidence | What do the results mean? | |---|---|---|---|---|---| | 172 women | 0.95 (0.30 to 1.00) | 0.99 (0.84 to 1.00) | 18.2% (182/1000) | Sensitivity: lowb,c Specificity: moderateb | With a prevalence of 18.2%, 182 out of 1000 women have bilateral tubal occlusion. - Out of 181 women with bilateral tubal occlusion on THL, this is correct in 173 women (96%) and incorrect in 8 women (4%). - Out of 819 women without bilateral tubal occlusion on THL, this is correct in 810 women (99%), and incorrect in 9 women (1%). | | CI: confidence interval; THL: transvaginal hydrolaparoscopy | ||||| | aThis prevalence is based on the actual prevalence observed across the included studies. |

Infertility, defined as the failure to conceive within 12 months of regular unprotected sexual intercourse, occurs in at least 12% of couples who wish to conceive [1, 2]. Around 18% to 33% of couples with infertility present with tuboperitoneal pathologies such as blocked or damaged fallopian tubes [3, 4, 5]. As the fallopian tubes are essential for transportation of the spermatozoa, the ovum and the embryo [6], bilateral occluded tubes exclude the chance of natural pregnancy. Therefore, bilateral tubal occlusion formed the basis of the development of in vitro fertilisation (IVF) and was the earliest indication for IVF [7]. Most diagnostic protocols for fertility assessment include a test to rule out tubal occlusion [8, 9]. During such tubal patency tests, a contrast agent is flushed into the uterus and through the fallopian tubes, visualising tubal patency. Diagnostic laparoscopy (DLS) with methylene blue dye tubal patency testing, also known as chromopertubation, is generally accepted as the reference standard [9]. However, due to its invasiveness and high costs, alternative, less invasive tests have been carried out. These visual tubal patency tests have evolved alongside the development of radiography, ultrasonography and laparoscopy; they include hysterosalpingography (HSG), sono‐hysterosalpingography (sono‐HSG), magnetic resonance hysterosalpingography (MR‐HSG), and outpatient transvaginal hydrolaparoscopy (THL). The choice of these visual tubal patency tests varies in different settings. Visual tubal patency tests can be used to diagnose tubal, uterine and other pelvic conditions. The most important tubal conditions are bilateral tubal occlusion, unilateral tubal occlusion and hydrosalpinx. The diagnoses of these conditions will directly guide clinical management, so they have been the focus of this Cochrane review. Target condition being diagnosed Target conditions of interest were tubal occlusion and hydrosalpinx. Tubal occlusion Bilateral occlusion Women with untreated bilateral occlusion have no chance of a natural pregnancy, as there is no way for the ovum and spermatozoa to meet. These women benefit from IVF, therefore, women diagnosed with bilateral tubal occlusion are mostly offered IVF directly [7]. However, IVF is not available to all couples worldwide due to differences in healthcare systems and reimbursements. IVF can be preceded by laparoscopic surgery to optimise pelvic anatomy. Unilateral occlusion Management in women with unilateral tubal blockage is more diverse, as in these women, the patent fallopian tube still facilitates transport of the ovum, spermatozoa and embryo. Treatment for a unilateral open tube is largely similar to that for bilateral open tubes, as unilateral tubal occlusion does not reduce pregnancy outcomes significantly [10]. Some studies have reported lower odds of pregnancy when unilateral distal (near the ovary) tubal occlusion is detected in comparison to proximal (near the uterus) tubal occlusion [11]. This observed difference between proximal and distal tubal occlusion may result from inherent diagnostic limitations of HSG or may reflect different underlying pathologies that differentially affect pregnancy outcomes. However, proximal or distal occlusion cannot be identified by all index tests, so we have not differentiated between proximal or distal occlusion in this review. Hydrosalpinx Hydrosalpinx refers to the distension of the fallopian tube due to distal tubal occlusion and fluid accumulation. The most common cause is a previous episode of pelvic inflammatory disease [12]. A hydrosalpinx has a negative impact on fertility outcomes through different mechanisms. Removal or ligation of the hydrosalpinx has a positive effect on clinical pregnancy rates before assisted reproductive technology (ART) [13]. Other conditions Other conditions can be detected during visual tubal patency tests but are not the focus of this review. - Endometriosis can be visualised during THL. It is seen in about 25% to 40% of women with infertility [14, 15]. - Peritoneal and pelvic adhesions can be visualised during sono‐HSG and THL. Pelvic adhesions, caused by previous surgery, pelvic inflammatory disease or endometriosis, may interfere with ovum pickup if they are distorting the anatomy of the ovary and fallopian tube. - Intrauterine pathology can be visualised during HSG and sono‐HSG. Intracavitary conditions such as myomas, polyps or intrauterine adhesions, as well as congenital uterine anomalies, might all have some effect on fertility outcomes [16, 17]. Index test(s) We have considered the following four main groups of index tests. Sono‐hysterosalpingography (sono‐HSG) Sono‐HSG includes both hysterosalpingo‐foam sonography (HyFoSy) and hysterosalpingo‐contrast sonography (HyCoSy). Overall, this test is based on ultrasound, in which an echogenic medium is used to assess the uterine cavity and tubal patency. Many different sono‐HSG techniques are performed, with differences in: - two‐ or three‐dimensional ultrasound modality; - vaginal or abdominal ultrasound; - contrast type (commercially available foam as well as normal saline, saline and air or galactose, or combinations of these); or - the usage of colour Doppler sonography [18]. The advantages of these tests are that they can be performed in an outpatient setting without a radiology department (offering the possibility of a one‐stop fertility evaluation), and are generally well tolerated [19]. Furthermore, when compared to HSG, the procedure does not require exposure to radiation or iodine‐containing contrast media [20]. In addition to tubal patency, the uterine cavity and myometrium, as well as both ovaries, can be assessed during the procedure [21]. The choice of contrast, operator skill and test technology have been pointed out as possible confounders, able to influence the diagnostic quality. Hysterosalpingography (HSG) HSG uses serial X‐ray or fluoroscopy images during injection of an iodine‐containing contrast medium through the cervical canal into the uterus and subsequently the fallopian tubes. Different instruments, such as a reusable metal cannula (hysterophore or Jarcho cannula), a 5‐French balloon catheter or a (modified) cervical vacuum cup device, as well as different iodine‐containing contrast media, oil‐based or water‐based, can be used. HSG is contraindicated in women with an allergy to iodine‐containing contrast media. It is a safe and widely accepted procedure in the outpatient setting, but it needs to be performed in a radiology department. HSG is well‐tolerated, although more painful than sono‐HSG [19] or THL [22]. In addition to its advantage of evaluating the uterine cavity and tubal patency, it has a potential therapeutic effect when an oil‐soluble contrast medium is used, with a higher chance of clinical pregnancy and live birth rates [23, 24]. Choice of contrast medium, operator skill and the observer interpreting the HSG have been investigated as potential sources of heterogeneity [25]. Magnetic resonance hysterosalpingography (MR‐HSG) MR‐HSG is similar to HSG. It uses magnetic resonance (MR) imaging instead of X‐ray or fluoroscopy, and the contrast medium is a gadolinium‐based solution, available from different manufacturers and prepared in different ways [26]. Similar to HSG, the procedure can be performed in an outpatient setting, and it is well tolerated, but needs to be performed in a radiology department [27, 28]. Unlike HSG, it avoids exposure to radiation and iodine‐containing contrast media, and can also be used to diagnose (deeply infiltrating) endometriosis, and uterine and ovarian anomalies. In comparison to sono‐HSG, the advantage of MR‐HSG is that it is not operator‐dependent, with better reproducibility [26, 28]. Transvaginal hydrolaparoscopy (THL) Also known as transvaginal endoscopy or fertiloscopy, THL uses hydroflotation for exploration of the pelvic cavity. The pelvis is filled with warm normal saline for pelvic cavity distention, then a small‐diameter optic is inserted transvaginally through an incision in the vaginal posterior fornix. Tubal patency can be tested by using a dye, mostly methylene blue [29]. THL can be carried out with disposable or reusable instruments [30]. It is a known, safe and well‐tolerated procedure, which can be performed in an outpatient setting under local anaesthesia [22, 30, 31]. An advantage is the direct visualisation of the female genital tract, thus allowing the evaluation of hydrosalpinx, endometriosis, and pelvic adhesions next to the tubal blockage. It is possible that the experience of the operator influences the success rate of THL. Clinical pathway There is a wide range of variation in visual tubal patency tests during fertility workup, at both national and international levels [8, 9]. Prior test(s) In general, a comprehensive medical history is obtained as the first step to explore the possible causes of female‐factor infertility. Next, a physical examination and transvaginal ultrasound assessment are performed. In some settings, tubal patency is always tested [8, 9], while in other settings, tubal testing is considered based on findings from medical history, physical examination and serological testing (chlamydia antibody testing (CAT)), and only women with a high risk for tubal pathology will undergo tubal testing [32]. Women are usually considered as being at high risk for tubal pathology when they have had a history of chlamydia infection or a positive CAT, pelvic inflammatory disease or peritonitis, or when they have been diagnosed with endometriosis or have had pelvic surgery in the past [33, 34]. Role of index test(s) A visual tubal patency test can be used as a triage or as a replacement test. When used as a triage test, women will undergo laparoscopy only when occlusion is suspected or the visual tubal patency test shows indeterminate findings. However, the aim of laparoscopy in current practice is more often to select women who may benefit from therapeutic laparoscopy, rather than to select women for diagnostic laparoscopy. For example, in the Federation of Medical Specialists (FMS) guideline [32], visual tubal patency tests are performed in high‐risk women as a triage test to select women who require laparoscopy. Laparoscopy without prior visual tubal patency testing is reserved only for those women with severe endometriosis or hydrosalpinges, where the diagnostic procedure and therapeutic laparoscopy are combined at the same time. In most other settings, a visual tubal patency test is used as a replacement for the reference standard. The outcome of this test will then be used to determine if fertility treatment is necessary. Fertility treatment can be therapeutic laparoscopy or assisted reproduction, depending on the availability and preferences of the doctor and woman undergoing treatment. An example is the National Institute for Health and Care Excellence (NICE) guideline in which women with low risk for tubal pathology are offered a visual tubal patency test and those with high risk are offered a laparoscopy [9]. Depending on the results, women with tubal obstruction can be offered tubal surgery, when appropriate surgical expertise is available, or alternatively, assisted reproduction. The choice of visual tubal patency tests also varies depending on the preference and skills of the clinician, the preference of the couples with infertility, and the availability of tubal testing methods in the clinic. In different geographical and economic contexts, costs, availability and the accessibility of these testing methods will differ. Alternative test(s) Alternative tests were not applicable, as we have included all visual tubal patency tests in this review. Rationale Over the last two decades, new tubal patency tests (e.g. MR‐HSG) have emerged, as well as new contrast media and test technology for existing tubal patency tests. Therefore, it is important to summarise all the evidence on the accuracy of individual tests, and to compare the accuracy of the tests against each other. As visual tubal patency tests are all less invasive than diagnostic laparoscopy and are well‐tolerated in an outpatient setting, it seems reasonable to offer such a test instead of the reference standard diagnostic laparoscopy. Currently, there is no consensus in terms of how different types of visual tubal patency tests compare to each other. Before replacing the reference standard, it is important to understand the diagnostic accuracy of each individual visual patency test and to compare their accuracy when possible. Couples with infertility will benefit from this research as it will guide clinicians to select the most accurate visual tubal patency test.

To determine and compare the diagnostic accuracy of visual tubal patency tests (sono‐HSG, HSG, THL, and MR‐HSG) for the diagnosis of tubal occlusion. Secondary objectives To determine and compare the diagnostic accuracy of visual tubal patency tests (sono‐HSG, HSG, THL, and MR‐HSG) for the diagnosis of hydrosalpinx. To evaluate heterogeneity concerning population characteristics (population risk stratification) and index test characteristics (contrast media, technology, operator skills).

We followed PRISMA‐Diagnostic Test Accuracy (DTA) for reporting [35]. Differences between protocol and review - We did not use MetaDTA to generate the figures [36]; instead, we used the metadta command in Stata, as it became available after protocol publication [37]. In addition to producing figures for DTA meta‐analyses, it also allows for the addition of covariates to the model and the analysis of comparative diagnostic test accuracy. It is the preferred tool for conducting methodologically rigorous DTA meta‐analyses in Stata [38, 39]. - We were unable to perform all analyses prespecified in the protocol due to insufficient data [37]. Criteria for considering studies for this review Types of studies We included studies on the diagnostic accuracy of a single index test and studies on the comparative diagnostic accuracy of two or more index tests. For the diagnostic test accuracy of a single index test, we included single‐gate studies, in which one of the index tests (defined below) was compared with the reference standard within a time frame of three months, in which the tubal status was unlikely to have changed. For the comparative diagnostic accuracy of two or more index tests, we included the following types of studies. - Studies with fully‐paired direct comparisons in a single‐gate design. In these studies, the participants received two or more index tests and the reference standard. - Randomised controlled trials that directly compared two or more index tests. In all arms, the index test should be followed by diagnostic laparoscopy as the reference standard. We excluded two‐gate studies, as these study designs are likely to overestimate sensitivity and specificity. Furthermore, we excluded studies with the primary endpoint of prognostic capacity for fertility outcomes, as well as diagnostic accuracy studies for sterilisation purposes. We excluded studies with a sample size of fewer than 50 participants, given the relatively low prevalence of bilateral tubal occlusion across all risk groups. Although this threshold may be considered arbitrary, both sensitivity and specificity could be unreliable or biased in studies with smaller sample sizes. Participants We included participants with infertility undergoing a visual tubal patency test and a diagnostic laparoscopy. We included women who have been trying to conceive for one year or more, both with low and high risk of tubal pathology, as well as unselected participants (i.e. women undergoing a visual tubal patency test without knowing about their risk of having tubal pathology). - High risk incorporates all women with a positive history of pelvic inflammatory disease, chlamydia or who are CAT‐positive, those with extensive abdominal or tubal surgery in the past, and those with abnormalities like endometriosis or possible hydrosalpinx discovered during physical examination. - Participants with low risk on tubal pathology are women with no previously mentioned conditions for high risk. Women who had previously had tubal testing were eligible only when the outcome of this test had not been used to select the study participants, as we did not include studies with a two‐gate design. We excluded women undergoing tubal testing after refertilisation (surgery to undo a tubal sterilisation). Index tests We included the following types of index tests. - Sono‐HSG (including HyFoSy and HyCoSy), used with commercially available foam, saline, saline and air or galactose, or combinations of these. We included studies that used two‐ or three‐dimensional modalities, with or without colour Doppler. We excluded studies conducted with contrast media that are no longer available (Echovist; galactose microparticles; Bayer Schering Pharma AG, Berlin, Germany) [40]. - HSG, with either oil‐based or water‐based contrast - THL, transvaginal endoscopy or fertiloscopy, conducted with reusable instruments or disposable trocars - MR‐HSG, including all techniques, MR protocols or contrast media used Target conditions We considered tubal occlusion as a dichotomous diagnosis for all tests, that is, occluded or patent (not occluded). As the unit of analysis was at the individual level, due to its clinical importance, we treated bilateral tubal occlusion and at least one‐sided tubal occlusion as two separate conditions, instead of a threshold. Similarly, we also considered hydrosalpinx as a dichotomous diagnosis.

standards Laparoscopy with methylene blue dye tubal patency testing was the reference standard. All participants in the studies included should have undergone this reference standard to avoid verification bias. We included only video‐assisted laparoscopy, as this is less operator‐dependent than the traditional monocular laparoscopy with direct visualisation only for the surgeon. We excluded studies on direct visualisation laparoscopy or those using CO2‐pertubation or indigo carmine dye for tubal testing during laparoscopy. In addition, we excluded studies with laparotomic tubal testing, and studies with another reference standard, for example, those using one of the index tests as the reference standard. Search methods for identification of studies We used a defined search strategy to search for all relevant published and unpublished studies on the diagnostic accuracy of visual tubal patency tests for tubal occlusion and hydrosalpinx. We carried out all searches in consultation with the Cochrane Gynaecology and Fertility Information Specialist. We checked Retraction Watch for retraction notices, errata or expressions of concern relating to each study and documented this in the characteristics table for each study [41]. Electronic searches We searched the following electronic databases on 6 November 2023: - Cochrane Central Register of Controlled Trials (CENTRAL; 2023, Issue 11) via the Cochrane Register of Studies Online (CRSO), web platform, searched from 1968 to 6 November 2023 (Supplementary material 1); - MEDLINE, Ovid platform, searched from 1968 to 6 November 2023 (Supplementary material 1); - Embase, Ovid platform, searched from 1980 to 6 November 2023 (Supplementary material 1); - CINAHL (Cumulative Index to Nursing and Allied Health Literature; EBSCO platform, searched from 1968 to 6 November 2023 (Supplementary material 1)). For each database, we used both index and free terms, and synonyms related to infertility, tubal pathology, (sono)‐hysterosalpingography, hydrolaparoscopy, and MR‐HSG. As the reference test, laparoscopy with methylene blue dye tubal patency testing, was first reported in 1968, we used this as the earliest search date for those databases with an inception date prior to 1968 [42]. We did not apply any language restrictions. We translated non‐English studies using an online translation tool, if necessary. The first search was performed on 1 November 2022 and a second one for additional new studies limited from 1 November 2022 was performed on 6 November 2023. Prior to publication, we searched for errata or retractions for reports of the included studies in PubMed (www.ncbi.nlm.nih.gov/pubmed) and Retraction Watch [41]. Searching other resources We also searched trials registers for studies that compared two or more index tests, and for other eligible observational studies. We searched ClinicalTrials.gov (clinicaltrials.gov/), International Standard Randomised Controlled Trial Number (ISRCTN) registry (www.isrctn.com/), and the World Health Organization (WHO) International Clinical Trials Platform (ICTRP) Search portal (apps.who.int/trialsearch/). We screened the reference lists of included studies and relevant systematic reviews for any additional studies. We also searched for ongoing and unpublished studies by approaching clinical experts and trialists in this field. Data collection and analysis Selection of studies Two authors (from RT, DK and KR) independently screened retrieved studies for eligibility on the basis of their titles and abstracts. If the study was potentially eligible, two authors (from RT, DK and KR) independently evaluated the full text for eligibility. We involved another author (RW) to solve any disagreements at both stages. If one of the authors was a co‐author on a potentially eligible study, that study was assessed by two other members of the team to avoid conflicts of interest. Where studies had multiple publications, we collated multiple reports of the same study under a single study ID with multiple references. We performed the study selection process in Covidence [43]. Data extraction and management Two review authors (RT and DK) performed the data extraction independently. When there was a disagreement between the two authors, we consulted a third author (RW). DK designed a data extraction form for this review and we pilot‐tested the form on three studies. We collected the following data from the included studies. - General information (first author, year of publication, country) - Participant characteristics (age, inclusion/exclusion criteria, numbers of participants) - Risk stratification (high/low risk for tubal pathology or unselected population) - Index test/reference standard details - 2 x 2 table for each outcome (true positives, true negatives, false positives, and false negatives) - Inconclusive tests - Adverse events When data for 2 x 2 tables were not available, we calculated these data from the test accuracy results (sensitivity, specificity, positive predictive value and negative predictive value). Next, we collected data on test‐specific related conditions found. We contacted study authors for information when needed. We used Review Manager for data input and to write the review [44]. Handling of inconclusive results We considered both valid inconclusive (intermediate or borderline) results and invalid inconclusive (indeterminate or uninterpretable) results in the analysis, as suggested by Shinkins and colleagues [45]. Participants with valid inconclusive results may receive further fertility treatment in clinical practice, but may also have another test in other settings. Therefore, we treated all valid inconclusive results as positive (i.e. occlusion) in the main analysis and as negative in a sensitivity analysis. Participants with invalid inconclusive results or procedure failures are more likely to have another test or a different index test in clinical practice, and some of these women may have conditions relevant to tubal pathology. Therefore, we excluded invalid inconclusive results from the main analysis and treated them as positive (i.e. occlusion) in a sensitivity analysis. We evaluated the robustness of the findings by using different methods to handle the inconclusive results. Please refer to Sensitivity analyses. Assessment of methodological quality We used the Quality Assessment of Diagnostic Accuracy Studies (QUADAS‐2) tool for the assessment of methodological quality of all included studies [46]. We evaluated the four domains in QUADAS‐2 (patient selection, index test, reference standard, and flow and timing) for risk of bias, and the first three domains for concerns regarding applicability. For comparative diagnostic test accuracy studies, we used the QUADAS‐C tool [47] (Supplementary material 7). Two review authors (RT and DK) evaluated the methodological quality of all the included studies independently. We settled disagreements by involving a third author (RW). Statistical analysis and data synthesis Diagnostic test accuracy for each index test We performed the analysis for each index test separately. We performed random‐effects meta‐analysis in a bivariate model, in which we modelled sensitivity and specificity directly [48]. Test positivity is based on the judgement of the specialist who performed the index test or a radiologist, and is a binary decision regarding absence or presence of the diagnosis. Therefore, it is appropriate to focus on summary estimates for sensitivity and specificity, and thus we used a bivariate model. For each index test, we presented pairs of sensitivity and specificity with their 95% confidence intervals (CIs) for each study, as well as the pooled sensitivity and specificity, in a forest plot. We presented the summary receiver operating characteristic (SROC) plot with summary points, their confidence regions and prediction regions. Non‐convergence could be an issue in bivariate models, when the data were sparse or the number of studies was small. To facilitate convergence, we employed orthogonal‐triangular (QR) decomposition of the variance–covariance matrix in scenarios where the between‐study correlation was on the boundary of its parameter space [49]. We implemented this approach in the metadta package in Stata [38], and validated its outputs using datasets with sparse data. Unit of analysis The unit of analysis refers to the unit as the denominator in the 2 x 2 tables for the included studies. In the context of the current review, studies either report the number of women with tubal occlusion among all participants (i.e. the participant level), or report the number of fallopian tubes with occlusion among all fallopian tubes (i.e. the tube level). 'Participant‐level data' in this review does not refer to individual participant data, as we based the current systematic review on aggregate data only and did not obtain any individual participant data. We analysed the data at a participant level in the analysis, as this is more clinically relevant. However, because a significant number of studies reported outcomes by tube, rather than by participant, we additionally also performed tube‐level analysis. Comparative diagnostic test accuracy for different index tests Direct comparison We included studies that directly reported two or more index tests compared with the reference standard in the primary analysis of comparative diagnostic accuracy of different index tests. We planned to add a covariate for the type of test in the bivariate model to compare the differences in accuracy, and perform a likelihood ratio test to compare models [50]. We also presented SROC plots linking estimates of two different index tests from the same studies. Indirect comparison As indirect comparisons are prone to bias, we did not perform this analysis due to different outcome reporting of the only two direct comparative studies. Adverse events We reported adverse events for all index tests. We used the metadta command in Stata to perform all the analyses [38, 51]. We used Review Manager (Version 9.12.0) to produce the forest plots [44]. Investigations of heterogeneity We planned the following assessments of heterogeneity. - Population characteristics: population risk stratification (high risk, low risk and unselected risk for tubal pathology) - Index test characteristics: - sono‐HSG (2‐dimensional (D)/3D versus 2D; different contrast media; use of colour Doppler or not; operator skills); - HSG (oil‐based versus water‐based contrast media; operator skills); - MR‐HSG (different viscosity contrast media); - THL (operator skills). - All these covariates are categorical variables. We fitted the models separately in different subgroups and performed visual inspections of SROC. Sensitivity analyses We planned the following sensitivity analyses. - Different approaches to handling inconclusive results: - treating valid inconclusive results as negative (i.e. patent); - treating invalid inconclusive results as positive (i.e. occluded). - - Limiting to studies at low risk of bias in the index test and reference standard domains. Assessment of reporting bias We did not evaluate reporting bias in this systematic review because statistical investigation of publication and reporting bias is not routinely recommended in DTA systematic reviews, as stated in the Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy [52, 53]. Summary of findings and assessment of the certainty of the evidence For diagnostic accuracy of individual index tests, we assessed the certainty of the evidence according to the GRADE guidance [54, 55]. We evaluated risk of bias, indirectness, inconsistency and imprecision, but did not assess publication bias for the reasons mentioned in Assessment of reporting bias. We produced summary of findings tables for each index test for the outcome of bilateral tubal occlusion. We presented the number of studies and women, study design, summary of findings (numbers and 95% confidence intervals for the index test on pooled sensitivity and specificity), and certainty of evidence in the summary of findings tables. For comparative diagnostic accuracy of different index tests, we planned to evaluate the certainty of the evidence according to the GRADE guidance [56], involving the same four domains as mentioned above for diagnostic accuracy of individual index tests. We did not plan indirect comparisons (between‐study comparisons), given that evidence resulting from indirect comparisons is likely to be of low certainty and the methodological work in this area is under development [56]. We planned to produce summary of findings tables for comparative diagnostic accuracy if we included more than two studies for each comparison, for the outcome of bilateral tubal occlusion.

Results of the search The flow diagram of the study screening process is presented in Figure 1. We identified 11,787 records: 11,121 through database searching on 12 December 2022, and 666 during the search update on 6 November 2023. We did not include any additional records through searching trials registers or manual searching. After removing 799 duplicates, we screened the titles and abstracts of 10,998 records for eligibility. We retrieved and screened the full texts of 180 records, resulting in 21 studies being included in this review. We excluded 123 studies (see Supplementary material 3 for details of the excluded studies). We list reasons for excluding studies in Figure 1. The most common reasons for exclusion were: - not all participants underwent the reference test (32 studies); - the inclusion criteria for the index test or reference test were not met (26 studies); and - the time frame between both tests was beyond three months (19 studies). Two studies reported insufficient information [57, 58]. We contacted the study authors but without success, and therefore excluded these studies. Thirty‐six studies are awaiting classification due to lack of information on the time interval between the index and reference test (Supplementary material 4). We did not find any ongoing studies. An overview of synthesis and included studies is summarised in Table 1Table 2; Table 3; Table 4 and Supplementary material 5, and an overview of the characteristics of all included studies is presented in Supplementary material 2. | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of a single index test: sono‐HSG | |||||||| | Battaglia 1996 | Italy | NR | 60 | Infertile women Aged 22–39 years Duration of infertility 3–14 years. Exclusion: hydrosalpinx | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: saline Used modality: 2‐D Colour doppler used: no | NR | | Gao 2011 | China | NR | 95 | Infertile women Aged 21–43 years Primary and secondary subfertility | NR | Tubal occlusion per tube | Contrast: SonoVue Used modality: 2‐D Colour doppler used: no | Experienced | | He 2017 | China | The Medical Research Foundation of Guangdong Province (grant B2014266), the Research Initiative of Southern Medical University (grant PY2014N082), and Clinical Research of Southern Medical University (grant LC2016YM014) | 56 | Women visiting infertility clinics and undergoing elective surgery in 1 month Mean age 29.8 years | Unselected | Tubal occlusion per tube | Contrast: SonoVue Used modality: 4‐D Colour doppler used: no | NR | | Kaur 2015 | India | NR | 50 | Women presenting with infertility Mean age primary infertility: 25.8 years and mean age secondary infertility: 30.6 years | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Hydrosalpinx | Contrast: saline Used modality: 2‐D Colour doppler used: no | NR | | Ludwin 2017 | Poland | No external funding | 144 | Sexually active women Aged 20–41 years Primary or secondary infertility, no prior tubal testing; mean age of 32.3 years | Unselected | Tubal occlusion per tube | Contrast: ExEm Foam Used modality: 2‐ and 3‐D Colour doppler used: yes | Experienced | | Sharaf 2022 | Egypt | No funding | 122 | Women scheduled for DLS as a part of their infertility workup, no prior tubal testing Mean age 28 | Unselected | Tubal occlusion per tube | Contrast: lidocaine‐foam Used modality: 2‐D Colour doppler used: yes | Experienced | | Shi 2019 | China | NR | 112 | Women with infertility Aged 21–47 years | NR | Tubal occlusion per tube | Contrast: Sonovil powder Used modality: 4‐D Colour doppler used: yes | Experienced | | Soliman 2015 | Egypt | Internal research resources of the Department of Obstetrics and Gynecology, University of Alexandria, Egypt, and no external sources of funding were involved | 50 | Women having DLS because of primary or secondary infertility Mean age 29 years | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Tubal occlusion per tube | Contrast: saline Used modality: 2‐ and 3‐D Colour doppler used: yes | Experienced | | ‐D: dimensional; DLS: diagnostic laparoscopy; NR: not reported; sono‐HSG: sono‐hysterosalpingography | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of a single index test: HSG | |||||||| | Allam 2014 | Egypt | NR | 80 | Infertile women having DLS as a part of infertility workup | Unselected | Bilateral tubal occlusion | Contrast: water‐based | NR | | Foroozanfard 2013 | Iran | This study was supported by the Deputy of Research, Kashan University of Medical Sciences (KAUMS) | 62 | Infertile women examined by HSG as part of routine infertility evaluation Mean age primary infertility 26.25 years and secondary infertility 29.73 years | Low risk | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: water‐based | NR | | Igbodike 2022 | Nigeria | No funding | 128 | All primary and secondary infertile women, reproductive age, presented with utero‐tubal infertility and willing to undergo HSG and DLS Mean age 33.9 years | High risk | Tubal occlusion per tube Hydrosalpinx | Contrast: water‐based | NR | | Johnson 1994 | USA | NR | 50 | Women evaluated for infertility Mean age 33 years | NR | Tubal occlusion per tube | Contrast: water‐based | NR | | Kehila 2014 | Tunisia | NR | 120 | Women followed for infertility in whom HSG was performed followed by DLS Mean age 35.3 years | Unselected | Bilateral tubal occlusion | Contrast: not specified | NR | | Sakar 2008 | Turkey | NR | 82 | Women followed for infertility (primary and secondary) Mean age 29.3 years | NR | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: water‐based | NR | | Sharma 2023 | India | NR | 105 | Infertile women, primary and secondary, duration ≥ 1 year with no previous pelvic surgeries Mean age NR | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Hydrosalpinx | Contrast: water‐based | NR | | Tvarijonaviciene 2008 | Lithuania | NR | 153 | Consecutive infertile women who conform WHO classification; mean age 30.5 years | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: water‐based | NR | | DLS: diagnostic laparoscopy; HSG: hysterosalpingography; NR: not reported; WHO: World Health Organization | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of a single index test: THL | |||||||| | Casa 2002 | Italy | NR | 60 | Consecutive women with unexplained primary infertility for > 1 year Mean age 32.1 years | Low risk | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | NA | Experienced | | Darai 2000 | France | NR | 60 | Consecutive women referred for DLS due to infertility Mean age 31 (+/‐ 6.5) years | Low risk | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Hydrosalpinx | NA | Less experienced in THL and experienced in DLS | | Watrelot 2003 | France, Belgium, Tunisia | The study was supported by a grant provided by Soprane SA, France | 92 | Women scheduled for DLS as part of routine infertility assessment; mean age 32 years | Unselected | Bilateral tubal occlusion Tubal occlusion per tube | NA | Less experienced in THL and experienced in DLS | | DLS: diagnostic laparoscopy; NA: not applicable; NR: not reported; THL: transvaginal hydrolaparoscopy | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of index tests: sono‐HSG and HSG | |||||||| | Darwish 1999 | Egypt | NR | 84 | Infertile women with indication for DLS Mean age 23 years | NR | Hydrosalpinx per tube | Contrast sono‐HSG: saline Used modality: only 2‐D Colour doppler used: no Contrast HSG: not specified | NR | | Rezk 2015 | Egypt | NR | 104 | Unexplained infertility > 1 year Aged 20–40 years; mean age 27.47 years | Low risk | Bilateral tubal occlusion | Contrast sono‐HSG: saline Used modality: 2‐D Colour doppler used: no Contrast HSG: not specified | NR | | ‐D: dimensional; DLS: diagnostic laparoscopy; HSG: hysterosalpingography; NA: not applicable; NR: not reported; sono‐HSG: sono‐hysterosalpingography | - Sono‐HSG: 10 studies reported on sono‐HSG as an index test. Four of them reported on bilateral tubal occlusion as target condition (Battaglia 1996 [59]; Kaur 2015 [60]; Rezk 2015 [61]; Soliman 2015 [62]), three also on at least one‐sided tubal occlusion (Battaglia 1996; Kaur 2015; Soliman 2015), six on tubal occlusion by tube (Gao 2011 [63]; He 2017 [64]; Ludwin 2017 [65]; Sharaf 2022 [66]; Shi 2019 [67]; Soliman 2015), and two on hydrosalpinx (Kaur 2015; Darwish 1999 [68]). - HSG: 10 studies reported on HSG as an index test. Seven reported on bilateral tubal occlusion as the target condition (Allam 2014 [69]; Foroozanfard 2013 [70]; Kehila 2014 [71]; Rezk 2015; Sakar 2008 [72]; Sharma 2023 [73]; Tvarijonaviciene 2008 [74]), four on at least one‐sided tubal occlusion (Foroozanfard 2013; Sakar 2008; Sharma 2023; Tvarijonaviciene 2008), two on tubal occlusion by tube (Igbodike 2022 [75]; Johnson 1994 [76]), two on hydrosalpinx by tube (Igbodike 2022; Darwish 1999), and one on hydrosalpinx by woman (Sharma 2023; Igbodike 2022). - THL: three studies reported on THL as an index test, all of which had bilateral tubal occlusion as the target condition (Casa 2002 [77]; Darai 2000 [78]; Watrelot 2003 [79]). Two also reported on at least one‐sided tubal occlusion (Casa 2002; Darai 2000), one on tubal occlusion by tube (Watrelot 2003), and one on hydrosalpinx (Darai 2000). - MR‐HSG: no eligible studies reported on MR‐HSG as an index test. Of the 21 included studies, two studies directly compared both HSG and sono‐HSG to the reference test (Darwish 1999; Rezk 2015), while the other 19 reported on the diagnostic accuracy of one single index test. We included three non‐English‐language studies, which were translated online, or if in Chinese, translated by our Chinese co‐author (Gao 2011; Kehila 2014; Shi 2019). Furthermore, 13 studies did not specify whether the participants represented an unselected or well‐defined low‐ or high‐risk population. Seven studies did not report whether the index test was performed by a trained person. No studies reported on inconclusive test results, therefore, we could not perform a sensitivity analysis for different approaches to handling inconclusive results. Methodological quality of included studies The results of the methodological quality assessment of the included studies are summarised and presented in Figure 2 and Figure 3. Patient selection: risk of bias - Sono‐HSG: we judged four studies at low risk of bias in the patient selection domain, and the other six studies at unclear risk of bias, as they provided no information on participant selection. - HSG: we judged four studies at low risk of bias in the patient selection domain, and the other six studies at unclear risk of bias, as they provided no information on participant selection, or did not mention inclusion and exclusion criteria. - THL: we judged two studies at low risk of bias in the patient selection domain, and one at high risk because they did not enrol a consecutive or random sample, as participants were selected by the clinicians themselves (Watrelot 2003). Patient selection: applicability concerns - Sono‐HSG: we judged three studies at low concern for applicability to the patient selection domain; applicability was unclear for seven studies, as they provided no information on whether the participants represented an unselected or well‐defined high‐ or low‐risk population. - HSG: we judged four studies at low concern for applicability to the patient selection domain; applicability was unclear for six studies, as they provided no information on whether the participants represented an unselected or well‐defined high‐ or low‐risk population. - THL: we judged two studies at low concern for applicability for the patient selection domain, and the other study at high concern for applicability for this domain, because they included participants with confirmed tubal patency or proximal tubal occlusion on HSG (Watrelot 2003). Index test: risk of bias - Sono‐HSG: we judged eight studies at low risk of bias in the index test domain, and two studies at unclear risk of bias, as they did not provide sufficient information. - HSG: we judged four studies at low risk of bias for the index test domain, and six studies at unclear risk of bias, as we did not know if they assessed HSG results without knowledge of the results of the reference standard. - THL: we judged all three studies at low risk of bias in the index test domain, as different surgeons performed the THL and DLS. Index test: applicability concerns - Sono‐HSG: we judged six studies at low concern for applicability in the index test domain, and one study at high concern for applicability, because it assessed tubal patency by observation of contrast in the pouch of Douglas, while the tubes themselves were not observed (Rezk 2015). For three studies, concern about applicability was unclear as they provided no information on the experience of the clinicians performing sono‐HSG. - HSG: we judged five studies at low concern for applicability in the index test domain, and five studies at unclear concern, as they provided no information on the experience of the clinicians who performed HSG and DLS. - THL: we judged all three studies at low concern for applicability in the index test domain, as experienced surgeons performed THL and DLS.

standard: risk of bias - Sono‐HSG: we judged four studies at low risk of bias in the reference standard domain, and six studies at unclear risk of bias, as they provided no information on the type of dye they used during DLS (Battaglia 1996; He 2017; Kaur 2015), or it was unclear if they assessed the reference test results without knowledge of the index test (He 2017; Kaur 2015; Darwish 1999), or both. - HSG: we judged three studies at low risk of bias for the reference standard domain, and seven studies at unclear risk of bias, as there was no detailed description of the reference test or blinding of the surgeons. - THL: we judged one study at low risk of bias for the reference standard domain, and one study at high risk of bias for this domain, because they performed standard laparoscopy without dye testing (Darai 2000). The risk of bias was unclear for one study, as they provided no information on the type of dye they used (Casa 2002).

standard: applicability concerns We judged studies as low concern for applicability in the reference standard domain if they described the target condition as one‐sided tubal pathology or bilateral tubal patency by woman. We judged studies at high concern for applicability for this domain if the target condition was by tube instead of by participant. - Sono‐HSG: we judged three studies at low concern for applicability in the reference test domain, and four studies at high concern (Gao 2011; He 2017; Ludwin 2017; Sharaf 2022). - HSG: we judged seven studies at low concern for applicability in the reference test domain, and two studies at high concern (Igbodike 2022; Johnson 1994). - THL: we judged two studies at low concern for applicability in the reference test domain, and one study at high concern (Watrelot 2003). Flow and timing: risk of bias We included studies only if all participants received the reference standard, and if there was an appropriate interval between the index test and reference standard of a maximum of three months. - Sono‐HSG: we judged all 10 studies at low risk of bias in the flow and timing domain. - HSG: we judged seven studies at low risk of bias for the flow and timing domain, and three studies at high risk of bias, because not all participants were included in the analyses (Allam 2014; Tvarijonaviciene 2008), or not all participants had the same reference standard (Johnson 1994; used indigo carmine dye for some participants). - THL: we judged all three studies at high risk of bias in the flow and timing domain, because all three studies did not include all the participants in the analysis (Casa 2002; Darai 2000; Watrelot 2003). Findings Sono‐hysterosalpingography (sono‐HSG) For bilateral tubal occlusion, the pooled sensitivity was 0.98 (95% CI 0.19 to1.00; 4 studies, 259 women; moderate‐certainty evidence) and specificity was 0.99 (95% CI 0.93 to 1.00; 4 studies, 259 women; high‐certainty evidence; Figure 4; Supplement 8). The overall sensitivity and specificity for at least one‐sided tubal occlusion were 1.00 and 0.89 (95% CI 0.82 to 0.93; 3 studies, 157 women; Figure 5; Supplement 9). For occlusion by tube, the pooled sensitivity and specificity were 0.95 (95% CI 0.83 to 0.99) and 0.94 (95% CI 0.90 to 0.97; 6 studies, 1109 tubes), respectively (see forest plot Figure 6; Supplement 10; SROC Figure 7). Subgroup analysis showed that colour Doppler was associated with higher specificity than standard ultrasound for diagnosing occlusion by tube, but with similar sensitivity (Supplement 11). We did not find differences in diagnostic accuracy between the use of foam or saline as contrast media (Supplement 12), or the use of 3D/2D and 2D ultrasound (Supplement 13). We did not perform other subgroup analyses as data were not available. We were not able to perform sensitivity analysis on inconclusive findings due to lack of data. Sensitivity analysis of studies based on low risk of bias in the index test and reference standard domains showed a sensitivity of 0.96 (95% CI 0.84 to 0.99) and a specificity of 0.97 (0.95 to 0.98; 3 studies, 589 tubes; Supplement 14). We included two studies in this review for hydrosalpinx but could not conduct a meta‐analysis because one study reported on hydrosalpinx by tube (sensitivity 0.60 (95% CI 0.15 to 0.95); specificity 1.00 (95% CI 0.98 to 1.00) 168 tubes; Darwish 1999), and the other on hydrosalpinx by woman (sensitivity 1.00 (95% CI 0.03 to 1.00); specificity 1.00 (95% CI 0.93 to 1.00); 50 women; Kaur 2015). Hysterosalpingography (HSG) For bilateral tubal occlusion, the pooled sensitivity was 0.77 (95% CI 0.58 to 0.89; 7 studies, 670 women; very low‐certainty evidence) and the pooled specificity was 0.94 (95% CI 0.87 to 0.97; 7 studies, 670 women; moderate‐certainty evidence; see forest plot Figure 8; Supplement 15; SROC Figure 9). The pooled sensitivity and specificity for at least one‐sided tubal occlusion were 0.82 (95% CI 0.52 to 0.95) and 0.84 (95% CI 0.60 to 0.94; 4 studies, 394 women), respectively (see forest plot: Figure 10; Supplement 16; SROC Figure 11). For occlusion by tube, the overall sensitivity and specificity were 0.92 (95% CI 0.83 to 0.96; 2 studies, 290 tubes) and 0.90 (95% CI 0.86 to 0.94; 2 studies, 290 tubes), respectively (see forest plot Figure 12; Supplement 17). Subgroup analysis on population risk showed that HSG had similar sensitivity and specificity in both low‐risk and unselected populations (Supplement 18). Except for two studies that did not report details about contrast media, all the studies used water‐based contrast media. No studies reported operator skill levels. Therefore, we did not carry out these subgroup analyses. Sensitivity analysis on studies at low risk of bias in the index test and reference standard domains left only one study, showing a sensitivity of 0.89 (95% CI 0.67 to 0.99) and a specificity of 0.90 (95% CI 0.84 to 0.95) for bilateral tubal occlusion. We could not perform sensitivity analysis on inconclusive findings due to lack of data. For hydrosalpinx by tube, the pooled sensitivity was 1.00 and specificity was 0.96 (95% CI 0.96 to 0.98; 2 studies, 360 tubes; Figure 13; Supplement 19; Darwish 1999; Igbodike 2022). Transvaginal hydrolaparoscopy (THL) For bilateral tubal occlusion, the pooled sensitivity and specificity were 0.95 (95% CI 0.30 to 1.00; 3 studies, 172 women; low‐certainty evidence) and 0.99 (95% CI 0.84 to 1.00; 3 studies, 172 women; moderate‐certainty evidence), respectively (Figure 14; Supplement 20). For at least one‐sided tubal occlusion, the pooled specificity was 1.00 and the sensitivity was 0.87 (95% CI 0.66 to 0.96; 2 studies, 110 women; Figure 15; Supplement 21). We could not perform meta‐analysis for tubal occlusion by tube and hydrosalpinx, because only one study reported on these conditions. We did not perform subgroup analysis due to the limited number of included studies. We could not perform sensitivity analysis on inconclusive findings due to lack of data. The only study at low risk of bias in both the index test and reference standard domains showed a sensitivity of 0.88 (95% CI 0.47 to 1.00) and a specificity of 0.96 (95% CI 0.87 to 1.00; Watrelot 2003). Comparative diagnostic test accuracy Rezk and colleagues compared the diagnostic accuracy of HSG and sono‐HSG with the reference standard for bilateral tubal pathology. They showed a sensitivity of 0.48 (95% CI 0.29 to 0.62) for HSG and 0.66 (95% CI 0.46 to 0.82) for sono‐HSG, while the specificity was comparable for both modalities (1.00, 95% CI 0.95 to 1.00; Rezk 2015). Darwish and colleagues compared the diagnostic accuracy of HSG and sono‐HSG with the reference standard for hydrosalpinx by tube. For HSG, they showed both high sensitivity and specificity: 1.00 (95% CI 0.48 to 1.00) and 0.99 (95% CI 0.96 to 1.00), respectively. For sono‐HSG, the sensitivity was lower with a broader confidence interval than the specificity: 0.60 (95% CI 0.15 to 0.95) and 1.00 (95% CI 0.98 to 1.00), respectively (Darwish 1999). Adverse events - Sono‐HSG: six of the 10 studies reported adverse events. Post‐procedural fever occurred in 11 of the 607 women (1.8%), and cervical laceration in five of the 607 women (0.8%). - HSG: five of the 10 studies reported the incidence of adverse events. Post‐procedural fever occurred in 14 of the 546 women (2.6%), and vaginal bleeding or cervical laceration was reported in 39 women (7.1%). No other adverse events were reported. - THL: all three studies reported adverse events. Perforation occurred in four of the 202 women (2.0%). No other adverse events were reported.

Summary of main results Our search strategy led to 21 included studies. Two studies performed a direct comparison between two index tests. The other 19 studies reported on the diagnostic accuracy of one single index test. We found that sono‐HSG, HSG and THL for the detection of bilateral tubal occlusion have sensitivities and specificities ranging from 0.77 to 0.98 and from 0.94 to 0.99, respectively. We did not include any studies on the diagnostic accuracy of MR‐HSG. The overall certainty of evidence varied from very low to moderate for sensitivity and from low to high for specificity. For comparative diagnostic accuracy for detection of bilateral tubal occlusion, we included only one study that compared both sono‐HSG and HSG to the reference standard. We judged this study at low risk of bias and with low concerns for applicability. It found a sensitivity of 0.48 (95% CI 0.29 to 0.62) for HSG and 0.66 (95% CI 0.46 to 0.82) for sono‐HSG, while the specificity was comparable for both modalities: 1.00 (95% CI 0.95 to 1.00). For hydrosalpinx, we could not conduct meta‐analysis for THL and sono‐HSG, as the number of studies was limited. For HSG, the pooled sensitivity and specificity were both high for the detection of hydrosalpinx (1.00 and 0.96, respectively). One study performed a comparative analysis for sono‐HSG and HSG in detecting hydrosalpinx. We judged it at unclear risk of bias and unclear concerns for applicability. Both sensitivity and specificity were high for HSG: 1.00 (95% CI 0.48 to 1.00) and 0.99 (95% CI 0.96 to 1.00), respectively. For sono‐HSG, the sensitivity was lower with a broader confidence interval than the specificity: 0.60 (95% CI 0.15 to 0.95) and 1.00 (95% CI 0.98 to 1.00), respectively. We could not carry out subgroup analysis on population characteristics, only for sono‐HSG on some index test characteristics. Colour Doppler was found to have a higher specificity than standard ultrasound for diagnosing occlusion by tube, but with similar sensitivity. We found no differences in diagnostic accuracy between the use of foam or saline as contrast media, nor the use of 3D/2D and 2D ultrasound. We did not perform other subgroup analyses as data were not available. Strengths and weaknesses of the review Strengths This is the first Cochrane review that provides an overview of the accuracy of different visual tubal patency tests. We executed the review according to the protocol. We included sono‐HSG only if the contrast medium is still commercially available to ensure that the findings of the results can be applied in daily practice. Furthermore, we included only video‐assisted laparoscopy, as this is less operator‐dependent than traditional monocular laparoscopy. In this review, we considered tubal patency as the target condition both at the tubal level and the individual level. We emphasise that it is important to consider tubal patency at an individual level, as pregnancy chances are significantly lower in women with bilateral tubal pathology when compared to women with bilateral tubal patency. In cases of unilateral tubal pathology, the chances of natural conception are almost comparable to those in women without tubal pathology (10). Furthermore, we assessed heterogeneity by considering risk stratification for tubal pathology. If the chance for tubal pathology is higher in advance, the sensitivity of the index test becomes more important. We also considered different test characteristics such as contrast medium, operator skills, and for sono‐HSG, test modality when possible. These are all factors that might influence the accuracy of the different tests. Weaknesses A limitation is that we could not perform all the analyses published in the protocol due to lack of data [37]. We could only perform partial subgroup analysis for sono‐HSG and HSG but not for THL, and none of the studies took operator skills into account. - For sono‐HSG, we were able to assess the influence of test characteristics (usage of 3D imaging, different contrast media and adding colour Doppler) in a subgroup analysis. However, these results should be interpreted cautiously as numbers are still limited. - For HSG, we could only carry out subgroup analysis on population risk stratification, which showed no difference in sensitivity or specificity for low risk and unselected populations. Furthermore, we were able to carry out a sensitivity analysis on studies at low risk of bias in the index test and reference standard domains for only one study on HSG. We could not perform sensitivity analysis for the handling of inconclusive results due to the unavailability of data. However, inconclusive results are more relevant to blocked tubes, and this condition is relatively rare, so we expect that the prevalence of inconclusive results is low. We did not use MetaDTA, an interactive online application for meta‐analysis of DTA studies, to produce the figures as planned [36]. Instead, we used the metadta command in Stata [38], as it became available after the publication of the protocol [37]. One of the limitations of the review process was the exclusion of studies with a sample size of fewer than 50 women, resulting in the exclusion of 14 studies. This sample size threshold was predefined in the protocol as these were expected to be mainly older low‐quality studies or pilot studies, which are often not designed or reported in accordance with current diagnostic test accuracy study standards. Notably, of the 14 studies excluded due to insufficient sample size, 10 also met at least one additional exclusion criterion. Therefore, even if the sample size criterion had not been applied, only four of these studies would have been eligible for inclusion in the analysis. Of these four studies, two reported on sono‐HSG, one on HSG, and one on THL. - For sono‐HSG, the Schoubroeck study reported outcomes by tube and found both sensitivity and specificity to be 1.0 (Van 2013 [80]). In contrast, the Crequat study reported outcomes per participant, defining tubal pathology as bilateral tubal occlusion, and found a sensitivity of 0.66 and a specificity of 0.71; however, this study included only 10 women (Crequat 1993 [81]). - For HSG, Shah 2005 reported on bilateral tubal occlusion in 37 women and found a sensitivity of 1.0 and a specificity of 0.92. Although this sensitivity is higher than the pooled sensitivity observed in the included studies, the prevalence of tubal pathology in the study by Shah was notably high (91.8%) (Shah 2005 [82]). - For THL, Nawroth and colleagues reported tubal occlusion per tube in 80 tubes and found both sensitivity and specificity to be 1.0, which is comparable to the pooled estimates in our review (Nawroth 2001 [83]). Additionally, we did not include studies with a time frame beyond three months between the index test and the reference standard or if this time frame was not described, resulting in 19 studies excluded and 36 studies awaiting classification. These 36 studies are likely to have included women with a time frame between the index test and reference standard beyond three months. Although inclusion of these studies might have given more data, the tubal patency status for these individuals may have changed between the tests. Therefore, including them would have resulted in biased results. To overcome the statistical challenges caused by the inclusion of sparse data in various analyses, an individual participant data (IPD) meta‐analysis could be considered to allow a multilevel modelling‐based approach, although it would take years to get the IPD due to current increasing regulatory requirements globally. Another limitation of this review is that the literature search was not updated at the time of publication. The review was initially submitted in November 2024, at which time the most recent search was approximately one year old. Given that there has been a lack of filters for diagnostic test accuracy, an updated search would result in a high yield of studies to screen. After careful consideration, we decided not to update the search, as doing so would have further delayed publication. Although updating the search may have resulted in the inclusion of more recently published studies, we believe that this is unlikely to have materially affected the main findings or conclusions of the review. Applicability of findings to the review question Almost half of the included studies did not report on the risk stratification for tubal pathology within their study populations. This lack of information limits our ability to determine whether the findings are equally applicable to women with low and high risk for tubal pathology and the generalisability to specific clinical settings remains uncertain. Furthermore, since sono‐HSG is an operator‐dependent test, the absence of information on operator skills in one‐third of the included studies raises concerns about how these results translate to real‐world practice, where variation in expertise may directly influence test accuracy. For THL, a learning curve also applies, which could impact complications and failure rates. However, all three included studies reported that procedures were performed by experienced operators, which may not reflect outcomes in less experienced operators. These limitations should be considered when applying our findings to broader clinical practice. We aimed to analyse the data at a participant level, as this is more clinically relevant than tubal level. However, for sono‐HSG, the majority of the studies reported on tubal occlusion by tube, instead of by woman. We found comparable diagnostic accuracy for double‐sided tubal occlusion and tubal occlusion per tube. For the diagnosis of hydrosalpinx, we could perform only one meta‐analysis, including two studies on hydrosalpinx per tube for HSG. For the other index tests and target conditions, insufficient studies were eligible for inclusion in our review. We previously discussed that we could not evaluate heterogeneity concerning population risk stratification and operator skills for all index tests due to poorly described patient characteristics in the included studies. We were able to evaluate test characteristics for sono‐HSG, however, the number of studies reporting on this was still limited and therefore these results should be interpreted with caution. This review aims to guide clinical decisions on fertility treatment based on tubal patency instead of surgical intervention to treat tubal patency, or to identify the cause of tubal pathology. It should be acknowledged that sono‐HSG and HSG are limited to determining tubal patency and cannot distinguish between intraluminal causes (e.g. mucosal fibrosis, mucus plugs) and extrinsic pathology (e.g. peritubal adhesions). In contrast, THL and laparoscopy allow identification of extrinsic causes, which are most often related to pelvic infections or endometriosis. Falloposcopy may further contribute by enabling direct visualisation of intraluminal pathology, and although falloposcopic tuboplasty has been explored as a therapeutic intervention, its role in clinical practice remains limited. Comparison with previous systematic reviews Laparoscopy for tubal occlusion Sono‐HSG To date, seven meta‐analyses have evaluated the diagnostic accuracy of sono‐HSG with diagnostic laparoscopy for assessing tubal occlusion. Of these, two examined 2D HyCoSy [18, 84], three 3D/4D HyCoSy [85, 86, 87], one 3D/4D HyCoSy and HyFoSy [88], and one 2D/3D/4D HyCoSy [89]. These meta‐analyses reported sensitivities between 0.86 and 0.98 and specificities between 0.90 and 0.95. They observed high heterogeneity between the studies, mostly because of small sample sizes and the use of Echovist. The majority of their included studies do not overlap with ours. Maheux and colleagues included nine studies that directly compared sono‐HSG and HSG with diagnostic laparoscopy. The pooled sensitivity for sono‐HSG and HSG were 0.95 (95% CI 0.78 to 0.99) and 0.94 (95% CI 0.74 to 0.99), respectively. The pooled specificity was 0.96 (95% CI 0.89 to 0.96) for sono‐HSG and 0.92 (95% CI 0.87 to 0.95) for HSG. Contrary to our findings on the direct comparison of both index tests, they found that HSG had a comparable sensitivity for tubal occlusion as sono‐HSG. We excluded from our review all the studies included in Maheux's review, mainly because the time frame between index test and reference standard was beyond three months or undefined, or the sample size was below 50 women [18]. HSG One systematic review with meta‐analysis has previously been published on HSG versus diagnostic laparoscopy, which included 19 studies that evaluated tubal pathology as an outcome [90]. An overall sensitivity and specificity could not be calculated due to high heterogeneity, and subgroup analysis showed that homogeneity was only shown for the three studies in which HSG and laparoscopy were performed by different clinicians without knowledge of the other test. The overall sensitivity and specificity for these studies were 0.65 (95% CI 0.50 to 0.78) and 0.83 (95% CI 0.77 to 0.88), respectively. This systematic review was published in 1991 and therefore did not include any of the studies included in our review. However, their difference in sensitivity and specificity is in line with our findings, indicating that tubal patency detected on HSG does not always rule out tubal pathology [90]. When considering the effect of different patient characteristics, such as the woman's age, duration of subfertility and low risk for tubal pathology on the accuracy of HSG, it seems that age and duration of subfertility do not influence the accuracy of HSG, while for women at low risk for tubal pathology the sensitivity was lower [91]. Within our review, subgroup analysis on population risk showed that HSG had similar sensitivity and specificity in both low‐risk and unselected populations. THL No systematic reviews have compared THL with diagnostic laparoscopy. Laparoscopy for hydrosalpinx Sono‐HSG, HSG and THL None of the above‐mentioned systematic reviews reported on hydrosalpinx as an outcome. Alternative reference standard Sono‐HSG One systematic review evaluated the diagnostic accuracy of HyFoSy with either HSG or diagnostic laparoscopy. In total, 622 tubes were included, showing a pooled sensitivity of 0.99 (95% CI 0.89 to 0.99) and specificity of 0.91 (95% CI 0.53 to 0.98). HSG and THL We found no systematic reviews that compared THL or HSG with one of the other index tests. MR‐HSG Although we could not include eligible studies on the diagnostic performance of MR‐HSG, there are two published systematic reviews that compared the diagnostic accuracy of MR‐HSG. - Li and colleagues included six studies and 101 women using HSG as the gold standard. They found a pooled sensitivity of 0.91 (95% CI 0.48 to 0.99) and specificity of 1.00 (95% CI 0.87 to 1.00) for the diagnosis of tubal occlusion (26). - Chen and colleagues evaluated the accuracy of HyCoSy versus MR‐HSG with either HSG or laparoscopy [92]. They found that HyCoSy had a lower sensitivity than MR‐HSG (0.89 vs 1.00) but higher specificity (0.93 vs. 0.82). The heterogeneity for both modalities was high. Subgroup analysis for HyCoSy showed that the use of 3D or 4D imaging resulted in higher diagnostic accuracy. The studies included in the meta‐analysis of Chen and colleagues overlap with those included in the meta‐analysis of Li and colleagues, with the addition of studies that compared MR‐HSG with laparoscopy. Since laparoscopy is considered a reference standard, this might explain the differences in diagnostic accuracy found between both meta‐analyses. Additional information

We would like to thank Dr Marian Showell, the Information Specialist from Cochrane Gynaecology and Fertility, for developing the search strategy, and Dr Elena Kostova, the Managing Editor of Cochrane Gynaecology and Fertility, for assisting with the protocol preparation. We acknowledge the late Professor Marlies Y Bongers for her contributions to women’s health research and for serving as an inspiring role model to many co‐authors. This review honours her memory and reflects a small part of her enduring impact on advancing women's health. Editorial team and peer‐reviewer contributions Cochrane Gynaecology and Fertility supported the authors in the development of this review. The following people conducted the editorial process for this article: - DTA Sign‐off Editor (final editorial decision): Danielle van der Windt, School of Medicine, Keele University, UK; - Clinical Sign‐off Editor (final editorial decision): Assoc. Prof Vanessa Jordan, University of Auckland; - Managing Editor (selected peer reviewers, provided editorial guidance to authors, edited the article): Leanne Jones, Cochrane Editorial Service; - Editorial Assistant (selected peer reviewers, conducted editorial policy checks, collated peer‐reviewer comments and supported the editorial team): Andrew Savage, Cochrane Editorial Service; - Copy Editor (copy editing and production): Denise Mitchell, Cochrane Central Production Service; - Peer‐reviewers (provided comments and recommended an editorial decision): Kazunori Nagasaka, Department of Obstetrics and Gynecology, Teikyo University, Tokyo (clinical/content review); Kavita Mandrelle (patient and public review); Nia Wyn Roberts, Bodleian Health Care Libraries, University of Oxford (search review); - Three additional peer reviewers provided clinical, statistical and methodological peer review but chose not to be publicly acknowledged. Contributions of authors RT, CK, BWM and RW conceived and designed the review. KR, MYB and VM contributed to the development of the protocol. RT, KR, RW, DK were involved in screening and/or data extraction of this review and assessed the methodological quality. RW was involved in data analysis. All authors contributed to the interpretation of the data. RT, RW, DK drafted the review. All authors revised the review critically for important intellectual content. MYB was deceased as of October 2025, after R1 of the review was submitted, and there have been no substantive changes to the review since the deceased author's contribution. All authors except for MYB approved the final version. Declarations of interest RT wrote a thesis Transvaginal hydrolaparoscopy in the diagnosis of tubal pathology. RT was involved with studies that may be eligible for inclusion in this review. DK is a PhD student focusing on diagnostic and therapeutic aspects of tubal patency testing. DK reports receiving a travel fee from Guerbet LLC. KR is a PhD student focusing on hysterosalpingography. CK is co‐promotor for RT's thesis and co‐author of several articles on visual tubal patency tests. VM reports receiving travel and speakers’ fees as well as research grants from Guerbet LLC, Merck and Ferring. MYB: author deceased; was promotor of RT's thesis and had no relevant conflicts of interest. BWM is supported by an NHMRC (National Health and Medical Research Council) Investigator grant (GNT1176437) for a study eligible for inclusion in this review. BWM reports consultancy, travel support and research funding from Merck and consultancy for Organon, Ferring and Norgine. BWM was previously an editor for Cochrane Pregnancy and Childbirth (ended in 2023), but had no involvement in the editorial processing of this review. BWM was involved in studies that may be eligible for inclusion in this review. RW is supported by an NHMRC Investigator grant (GNT2009767). RW is an editorial board member of Cochrane Gynaecology and Fertility but had no involvement in the editorial processing of this review. Sources of support Internal sources - No sources of support provided External sources - NHMRC, Australia RW is supported by an NHMRC Emerging Leadership Investigator Grant (2009767). Registration and protocol Protocol (2022) https://doi.org/10.1002/14651858.CD014968 Data, code and other materials As part of the published Cochrane review, the following are made available for download for users of the Cochrane Library. Full search strategies for each database (Supplementary material 1); full citations of each unique report for all included studies (Supplementary material 2), studies excluded at the full‐text screen (Supplementary material 3), and studies awaiting classification (Supplementary material 4); analysis data (Supplementary material 5) including overall estimates and settings, subgroup estimates, and individual data rows; data package (Supplementary material 6); and tailored QUADAS‐2 and QUADAS‐C tools (Supplementary material 7). Appropriate permissions have been obtained for such use. Data management was partly conducted within Cochrane’s authoring tool, Review Manager, using the inbuilt computation methods. Supplements 8‐21, Stata data files and Stata syntax for statistical analysis can be accessible via the following link. https://doi.org/10.17605/OSF.IO/DYG3R History Protocol first published: Issue 6, 2022 To view the full reference list, organized by study, for included studies, excluded studies, studies awaiting classification and/or ongoing studies, see the characteristics of studies. Alternatively, this grouping of references is available via the downloadable data package.

The content of this article (as distinct from the content of the ‘Supplementary materials’, ‘Related’ and ‘About this Review’ sections) has been peer reviewed and prepared for publication in accordance with Cochrane’s editorial policies . Though the aim of these processes is, among other things, to check the accuracy of such content, responsibility for the content remains that of the author(s) and Cochrane gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, the content of the article. | Population: women presenting with infertility Prior testing: medical history, physical examination, with or without serological testing (chlamydia antibody testing) Setting: fertility clinics or outpatient Index test: sono‐HSG Reference standard: diagnostic laparoscopy (video‐assisted) with tubal testing with methylene blue Target condition: bilateral tubal occlusion Studies: one direct comparison study, comparing sono‐HSG and HSG (Rezk 2015), and three single‐test studies (Battaglia 1996, Kaur 2015, Soliman 2015) | ||||| | Number of women | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Prevalencea | Certainty of evidence | What do the results mean? | |---|---|---|---|---|---| | 259 women | 0.98 (0.19 to 1.00) | 0.99 (0.93 to 1.00) | 18.2% (182/1000) | Sensitivity: moderateb Specificity: high | With a prevalence of 18.2%, 182 out of 1000 women have bilateral tubal occlusion. - Out of 186 women with bilateral tubal occlusion on sono‐HSG, this is correct in 178 women (96%) and incorrect in 8 women (4%). - Out of 814 women without bilateral tubal occlusion on sono‐HSG, this is correct in 810 women (99.5%), and incorrect in 4 women (0.5%). | | CI: confidence interval; HSG: hysterosalpingography; sono‐HSG: sono‐hysterosalpingography | ||||| | aThis prevalence is based on the actual prevalence observed across the included studies. | | Population: women presenting with infertility Prior testing: medical history, physical examination, with or without serological testing (chlamydia antibody testing) Setting: fertility clinics or outpatient Index test: HSG Reference standards: diagnostic laparoscopy (video‐assisted) with tubal testing with methylene blue Target condition: bilateral tubal occlusion Studies: one study directly compared sono‐HSG and HSG (Rezk 2015), six single‐test studies (Allam 2014; Foroozanfard 2013; Kehila 2014; Sakar 2008; Sharma 2023; Tvarijonaviciene 2008) | ||||| | Number of women | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Prevalencea | Certainty of evidence | What do the results mean? | |---|---|---|---|---|---| | 670 women | 0.77 (0.58 to 0.89) | 0.94 (0.87 to 0.97) | 18.2% (182/1000) | Sensitivity: very lowb,c Specificity: moderateb | With a prevalence of 18.2%, 182 out of 1000 women have bilateral tubal occlusion. - Out of 189 women with bilateral tubal occlusion on HSG, this is correct in 140 women (74%) and incorrect in 49 women (26%). - Out of 811 women without bilateral tubal occlusion on HSG, this is correct in 769 women (95%), and incorrect in 42 women (5%). | | CI: confidence interval; HSG: hysterosalpingography; sono‐HSG: sono‐hysterosalpingography | ||||| | aThis prevalence is based on the actual prevalence observed across the included studies. | | Population: women presenting with infertility Prior testing: medical history, physical examination, with or without serological testing (chlamydia antibody testing) Setting: fertility clinics or outpatient Index test: THL Reference standards: diagnostic laparoscopy (video‐assisted) with tubal testing with methylene blue Target condition: bilateral tubal occlusion Studies: three single‐test studies (Casa 2002; Darai 2000; Watrelot 2003) | ||||| | Number of women | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Prevalencea | Certainty of evidence | What do the results mean? | |---|---|---|---|---|---| | 172 women | 0.95 (0.30 to 1.00) | 0.99 (0.84 to 1.00) | 18.2% (182/1000) | Sensitivity: lowb,c Specificity: moderateb | With a prevalence of 18.2%, 182 out of 1000 women have bilateral tubal occlusion. - Out of 181 women with bilateral tubal occlusion on THL, this is correct in 173 women (96%) and incorrect in 8 women (4%). - Out of 819 women without bilateral tubal occlusion on THL, this is correct in 810 women (99%), and incorrect in 9 women (1%). | | CI: confidence interval; THL: transvaginal hydrolaparoscopy | ||||| | aThis prevalence is based on the actual prevalence observed across the included studies. | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of a single index test: sono‐HSG | |||||||| | Battaglia 1996 | Italy | NR | 60 | Infertile women Aged 22–39 years Duration of infertility 3–14 years. Exclusion: hydrosalpinx | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: saline Used modality: 2‐D Colour doppler used: no | NR | | Gao 2011 | China | NR | 95 | Infertile women Aged 21–43 years Primary and secondary subfertility | NR | Tubal occlusion per tube | Contrast: SonoVue Used modality: 2‐D Colour doppler used: no | Experienced | | He 2017 | China | The Medical Research Foundation of Guangdong Province (grant B2014266), the Research Initiative of Southern Medical University (grant PY2014N082), and Clinical Research of Southern Medical University (grant LC2016YM014) | 56 | Women visiting infertility clinics and undergoing elective surgery in 1 month Mean age 29.8 years | Unselected | Tubal occlusion per tube | Contrast: SonoVue Used modality: 4‐D Colour doppler used: no | NR | | Kaur 2015 | India | NR | 50 | Women presenting with infertility Mean age primary infertility: 25.8 years and mean age secondary infertility: 30.6 years | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Hydrosalpinx | Contrast: saline Used modality: 2‐D Colour doppler used: no | NR | | Ludwin 2017 | Poland | No external funding | 144 | Sexually active women Aged 20–41 years Primary or secondary infertility, no prior tubal testing; mean age of 32.3 years | Unselected | Tubal occlusion per tube | Contrast: ExEm Foam Used modality: 2‐ and 3‐D Colour doppler used: yes | Experienced | | Sharaf 2022 | Egypt | No funding | 122 | Women scheduled for DLS as a part of their infertility workup, no prior tubal testing Mean age 28 | Unselected | Tubal occlusion per tube | Contrast: lidocaine‐foam Used modality: 2‐D Colour doppler used: yes | Experienced | | Shi 2019 | China | NR | 112 | Women with infertility Aged 21–47 years | NR | Tubal occlusion per tube | Contrast: Sonovil powder Used modality: 4‐D Colour doppler used: yes | Experienced | | Soliman 2015 | Egypt | Internal research resources of the Department of Obstetrics and Gynecology, University of Alexandria, Egypt, and no external sources of funding were involved | 50 | Women having DLS because of primary or secondary infertility Mean age 29 years | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Tubal occlusion per tube | Contrast: saline Used modality: 2‐ and 3‐D Colour doppler used: yes | Experienced | | ‐D: dimensional; DLS: diagnostic laparoscopy; NR: not reported; sono‐HSG: sono‐hysterosalpingography | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of a single index test: HSG | |||||||| | Allam 2014 | Egypt | NR | 80 | Infertile women having DLS as a part of infertility workup | Unselected | Bilateral tubal occlusion | Contrast: water‐based | NR | | Foroozanfard 2013 | Iran | This study was supported by the Deputy of Research, Kashan University of Medical Sciences (KAUMS) | 62 | Infertile women examined by HSG as part of routine infertility evaluation Mean age primary infertility 26.25 years and secondary infertility 29.73 years | Low risk | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: water‐based | NR | | Igbodike 2022 | Nigeria | No funding | 128 | All primary and secondary infertile women, reproductive age, presented with utero‐tubal infertility and willing to undergo HSG and DLS Mean age 33.9 years | High risk | Tubal occlusion per tube Hydrosalpinx | Contrast: water‐based | NR | | Johnson 1994 | USA | NR | 50 | Women evaluated for infertility Mean age 33 years | NR | Tubal occlusion per tube | Contrast: water‐based | NR | | Kehila 2014 | Tunisia | NR | 120 | Women followed for infertility in whom HSG was performed followed by DLS Mean age 35.3 years | Unselected | Bilateral tubal occlusion | Contrast: not specified | NR | | Sakar 2008 | Turkey | NR | 82 | Women followed for infertility (primary and secondary) Mean age 29.3 years | NR | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: water‐based | NR | | Sharma 2023 | India | NR | 105 | Infertile women, primary and secondary, duration ≥ 1 year with no previous pelvic surgeries Mean age NR | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Hydrosalpinx | Contrast: water‐based | NR | | Tvarijonaviciene 2008 | Lithuania | NR | 153 | Consecutive infertile women who conform WHO classification; mean age 30.5 years | Unselected | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | Contrast: water‐based | NR | | DLS: diagnostic laparoscopy; HSG: hysterosalpingography; NR: not reported; WHO: World Health Organization | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of a single index test: THL | |||||||| | Casa 2002 | Italy | NR | 60 | Consecutive women with unexplained primary infertility for > 1 year Mean age 32.1 years | Low risk | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman | NA | Experienced | | Darai 2000 | France | NR | 60 | Consecutive women referred for DLS due to infertility Mean age 31 (+/‐ 6.5) years | Low risk | Bilateral tubal occlusion At least one‐sided tubal occlusion per woman Hydrosalpinx | NA | Less experienced in THL and experienced in DLS | | Watrelot 2003 | France, Belgium, Tunisia | The study was supported by a grant provided by Soprane SA, France | 92 | Women scheduled for DLS as part of routine infertility assessment; mean age 32 years | Unselected | Bilateral tubal occlusion Tubal occlusion per tube | NA | Less experienced in THL and experienced in DLS | | DLS: diagnostic laparoscopy; NA: not applicable; NR: not reported; THL: transvaginal hydrolaparoscopy | | Study name (year) | Country of conduct | Funding | No. participants included | Population description | Risk stratif ication | Target condition | Test characteristics | Operator skills | |---|---|---|---|---|---|---|---|---| | Diagnostic test accuracy study of index tests: sono‐HSG and HSG | |||||||| | Darwish 1999 | Egypt | NR | 84 | Infertile women with indication for DLS Mean age 23 years | NR | Hydrosalpinx per tube | Contrast sono‐HSG: saline Used modality: only 2‐D Colour doppler used: no Contrast HSG: not specified | NR | | Rezk 2015 | Egypt | NR | 104 | Unexplained infertility > 1 year Aged 20–40 years; mean age 27.47 years | Low risk | Bilateral tubal occlusion | Contrast sono‐HSG: saline Used modality: 2‐D Colour doppler used: no Contrast HSG: not specified | NR | | ‐D: dimensional; DLS: diagnostic laparoscopy; HSG: hysterosalpingography; NA: not applicable; NR: not reported; sono‐HSG: sono‐hysterosalpingography |