Noninvasive diagnostic tools for pelvic congestion syndrome: a systematic review.

The initial search yielded 588 articles; cross‐searching reference lists yielded another 21 records. After screening on title and abstract we selected 58 articles. These articles were assessed for eligibility based on the full text manuscript. Eventually, nine articles fully matched our eligibility criteria; six studies described ultrasonography 11 , 20 , 23 , 24 , 25 , 26 and three studies described MRI 12 , 14 , 27 . The process of study identification is displayed in Figure  2 , using a prisma flow diagram 28 . Two studies, describing 131 Xenon clearance and applied potential tomography, were excluded because the diagnostic tools are now considered obsolete (no published data since 1981 and 1991, respectively) 29 , 30 . PRISMA flow chart. [Color figure can be viewed at http://wileyonlinelibrary.com ]. The nine included studies were each assessed on 12 items. In total, 60 items were scored with an “L” (low risk of bias), eight items with an “H” (high risk of bias) and four items with a “U” (unclear risk of bias). Overall, seven of the included studies met most of the quality indicators of the QUADAS‐2 tool (Table  1 ). Regarding two studies the quality assessment could not be completed because of a lack of information, for this reason we concluded that the results could not be substantially reviewed 25 , 26 . Methodological quality assessment (QUADAS‐2 criteria) PCS, pelvic congestion syndrome; QUADAS‐2, Quality Assessment of Diagnostic Accuracy Studies. L, low risk; H, high risk; U, unclear risk. The characteristics of the study populations are presented in Table  2 and study characteristics are shown in Table  3 . Some form of venography was used in all studies as reference standard. A variety of diagnostic criteria and thresholds were used, as shown in Table  3 . Characteristics of the study population CPP, chronic pelvic pain; NM, not mentioned; PCS, pelvic congestion syndrome. Chronic pelvic pain: dull pelvic pain of variable intensity, persisting for at least 6 months. Venographic congestion score = Ovarian vein diameter 1–4 mm (1 point), 5–8 mm (2), > 9 mm 3 ; Contrast clearance  40 s 3 ; Congestion absent (1), moderate (2), severe 3 . Ovarian vein diameter > 5–10 mm, uterine vein engorgement, congestion of ovarian plexuses, filling of pelvic veins across the midline, filling of vulvovaginal and thigh varicosities. Characteristics of included studies MR PCVM, magnetic resonance phase‐contrast velocity mapping; MRV, magnetic resonance venography; PCS, pelvic congestion syndrome; TAUS, transabdominal ultrasonography; TR‐MRA, time‐resolved magnetic resonance angiography; TVUS, transvaginal ultrasonography. The main results are summarized in Table  4 . Campbell et al. 20 and Halligan et al. 23 studied the same ultrasound congestion score and power Doppler assessment of vascularity, but were not able to discriminate between women with PCS and control women. The occurrence of pelvic varicoceles on transvaginal ultrasonography had sensitivity and specificity of 100% and 83–100%, respectively, according to Giacchetto et al. 24 and Park et al. 11 . Aside from varicoceles, Park et al. 11 found a specificity of 91% when communication between bilateral pelvic varicosities via transuterine crossing veins, > 5 mm, was observed. Reversed caudal flow, seen with transabdominal ultrasonography, was found in all patients with PCS proven by venography, resulting in a sensitivity of 100%. An ovarian vein diameter > 5 or > 6 mm seen on transabdominal ultrasonography had a positive predictive value of 71.2% or 83.3%, respectively 11 . As secondary outcome, women with PCS had statistically significantly more and smaller ovarian follicles, smaller uterine volume and thinner endometrium in comparison with healthy control women 20 , 23 . Main outcomes MR PCVM, magnetic resonance phase‐contrast velocity mapping; MRV, magnetic resonance venography; PCS, pelvic congestion syndrome; TAUS, transabdominal ultrasound; TR‐MRA, time‐resolved magnetic resonance angiography; TVUS, transvaginal ultrasound; Sens: sensitivity; Spec: specificity, both followed by 95% confidence intervals between brackets. 95% CI, when not reported in literature, was calculated based upon reported results. Classification of correlation by Dancey and Reidy (2004) [31]. Three studies regarding MRI were included in this review; in each study a slightly different technique was applied. Asciutto et al. studied the occurrence of venous insufficiency in the pelvic plexus, ovarian vein and hypogastric vein, using magnetic resonance venography 12 . This resulted in sensitivities of 91%, 88% and 100%, respectively. The specificities were found to be 42%, 67% and 38%, respectively. Meneses et al. used a different kind of MRI technique, based upon phase‐contrast velocity mapping 27 . They found a sensitivity of 100% and specificity of 50%, based upon nine patients with suspected PCS, in whom both left and right side were evaluated 27 . Time‐resolved magnetic resonance angiography, as studied by Yang et al., resulted in a sensitivity of 100% among the 19 patients included 14 . As secondary outcome, they made a distinction between grade I and II, as demonstrated in Table  3 , which resulted in a specificity of 100%.

A computer‐aided search in PubMed and EMBASE was performed, in collaboration with a medical librarian. The databases were searched from inception until 6 May 2017, the search strategy can be found in the Supplementary material (Appendix S1). To identify additional suitable studies, the reference lists of reviews and included studies were cross‐searched manually. Titles and abstracts were independently screened by two reviewers (TN and MS) to identify suitable articles for the first selection. Full manuscripts of the remaining citations were obtained and reviewed by TN and MS. We excluded articles from the search if they were not relevant to the goals of the review, not published in the English language, included fewer than four patients, did not include a reference test, did not involve human participants or were reviews, letters or conference abstracts. Selection disagreements were resolved by consensus in cooperation with one of the co‐authors. Data were extracted from manuscripts using a data extraction form. On these data extraction forms information on study characteristics were noted. These included study design, sample size, inclusion and exclusion criteria, and baseline characteristics of included patients, such as age, ethnicity, parity and menopausal status. Information about the setting, parameters, thresholds and outcomes of the index and reference test was also extracted. Sensitivity and specificity of pelvic venous dilatation associated with CPP were set as main outcomes and, when necessary, these were calculated based on available data. A p ‐value < 0.05 was considered statistically significant. The methodological quality of included studies was assessed systematically using the Second Quality Assessment of Diagnostic Accuracy Studies (QUADAS‐2) tool. This 12‐item scale, applicable to diagnostic accuracy studies, is designed to assess the risk of bias and possible concerns regarding applicability 22 .

In this systematic review, we aimed to identify the value of noninvasive diagnostic tools in the work‐up of patients suspected of PCS. In ultrasonography, a vein > 5 mm crossing the uterine body, pelvic varicoceles and reversed caudal flow appeared to be the most indicative of PCS. Due to the limited number of patients in MRI studies, no firm conclusions could be drawn from these studies. Being a noninvasive diagnostic tool used in the work‐up of numerous gynecological symptoms, transvaginal ultrasonography would theoretically be the ideal first step in the diagnostics of PCS. The ultrasound congestion score however, based on vein diameter, number of veins in the sector and a subjective assessment of congestion, appeared to be unable to discriminate patients with PCS from healthy control women 20 , 23 . Park et al. found a high specificity (91%) when a vein > 5 mm was crossing the uterine body, yet the low sensitivity (25%) will subsequently result in a high chance of false‐negative results 11 . Pelvic varicoceles on transvaginal ultrasound appeared to have a high sensitivity (100%), assuming a good ability to rule out PCS when pelvic varicoceles are not identified 11 , 24 . Park et al. also studied the role of transabdominal ultrasonography next to the transvaginal approach, and concluded that an ovarian vein diameter of > 6 mm accounted for a positive predictive value of 83.3% 11 . However, the positive predictive value is only a relevant outcome when the study population reflects the real prevalence of disease and due to the case–control design used in the Park et al. study, this outcome may not be relevant. Reversed caudal flow, seen in the ovarian vein, appeared to be highly sensitive (100%) in detecting PCS with a conventional ultrasound Doppler technique 11 . The flow in congested adnexal veins is typically low; power Doppler assigns different color tones and brightness to the total energy of the Doppler signal, consequently making it more sensitive to motion 31 . Despite this hypothesis, the two studies included in this review that investigated power Doppler assessment of adnexal veins, were unable to discriminate women with PCS from healthy control women 20 , 23 . Furthermore, Malgor et al. demonstrated a compensatory right ovarian vein dilatation in the case of left ovarian vein reflux 32 . This is clinically relevant regarding the fact that treatment should be based not only on the veins’ size but also the amount of reflux. Magnetic resonance imaging allows the demonstration of the ovarian and gonadal veins in a complete examination of the pelvic anatomy, due to the multiplanar imaging capability. Three studies regarding MRI‐based methods were included in this review; in each study a slightly different technique of MRI was applied 12 , 14 , 26 . Yang et al. investigated the feasibility of time‐resolved magnetic resonance angiography 14 . Time‐resolved magnetic resonance angiography is a quick and noninvasive technique to visualize physiological blood flow dynamics 14 . It is widely used and proven to be highly sensitive, when compared with conventional angiography for detecting pathology in a variety of blood vessels. High sensitivity was also demonstrated in the study of Yang et al. when reflux was detected in the ovarian veins of patients diagnosed with PCS 14 . The specificity could not be assessed due to the absence of controls (without pelvic pathology on venography) 14 . Asciutto et al. studied the application of magnetic resonance venography in the assessment of women with PCS 12 . Magnetic resonance venography appeared to be highly sensitive for insufficiency in pelvic plexus, ovarian or hypogastric veins. Especially hypogastric vein insufficiency accounted for a high sensitivity (100%), but the specificity was low, which results in a high prevalence of false‐positive results 12 . Computed tomography might account for a less expensive alternative to MRI, but at this moment there are no studies available describing the value of CT in the work‐up of patients suspected of PCS. A case report and a retrospective study performed in asymptomatic patients assume that it may be possible to diagnose PCS when incompetence or dilatation is visualized using CT 33 , 34 . CT and MRI are both operator dependent, but CT is easier to read and so less operator dependent. The downside is the limited amount of information that CT provides; MRI has the possibility to provide information on the differential diagnosis of PCS, which includes endometriosis and adenomyosis. Meneses et al. investigated the role of magnetic resonance phase‐contrast velocity mapping (MR PCVM) in patients with suspected PCS 27 . MR PCVM is a modality used to assess anomalous venous flow, and not solely anatomical changes. Hypothetically, PCS is an entity caused by dynamic disturbance of venous flow, the anatomical changes are the result of this anomalous flow. Meneses et al. found a high sensitivity, but the specificity could not reliably be determined due to the wide 95% confidence interval 27 . It is possible that MR PCVM can detect an earlier stage of PCS, in which the slow venous flow has not yet caused any dilatation of the ovarian veins. On the other hand, slow venous flow may not result in dilatation, varicosities, or pain in all patients. This review has some potential limitations. First, the included studies yielded different parameters and thresholds, which made a formal meta‐analysis impossible. Due to publication bias, relevant data may be missing. The studies of Park et al. and Halligan et al. were based upon a case–control design, which is a possible source for bias in results 11 , 23 . For example, a bias in the selection of patient population and verification may occur because the healthy control women did not undergo the reference standard. Consequently, it is important to note that the reference standard, venography, solely used in the studies included in this review is not 100% accurate. Based on the study of Beard et al., the sensitivity and specificity of venography were found to be 91% and 89%, respectively 19 . This represents a fundamental flaw in the test accuracy design of included studies. The index test is unable to perform better than the reference standard, and hence its value may be underestimated. Finally, all index tests described in this review are performed with the patients in supine position. It is likely that ovarian and pelvic varices may not be as prominent on images in this position. Especially in early stages, with mild changes in venous anatomy, ultrasonography could reach a higher diagnostic accuracy when the patient is examined in reverse Trendelenburg or when Valsalva's maneuver is applied. Furthermore, to enhance discriminative ability it could be of added value to examine patients later in the afternoon, which is also common practice in the work‐up of patients with pelvic organ prolapse. Labropoulos et al. described a more complete protocol on how patients can be examined with ultrasonography, including the position of the patient with a head elevation of 30 degrees 35 . Their technique aids in the overall visibility of the relevant veins in the lower abdomen 35 . However, this technique requires that patients have been fasting overnight, which is of course not a usual preparation for a visit to the outpatient gynecological department 35 . In patients with CPP and symptoms directing towards PCS, a good diagnostic work‐up must be readily available. This systematic review emphasizes the current gap in literature and lack of standardized criteria regarding PCS. With this review, we add important data on the accuracy of noninvasive diagnostic tools in the daily practice of gynecologists in the work‐up of patients with CPP. Pelvic veins > 5 mm, a vein crossing the uterine body from left to right and communicating with both ovarian plexus and reversed caudal flow seem to be the most indicative for PCS. Therefore, gynecologists should pay extra attention to these parameters and the possibility of PCS. Based on this systematic review, a validated noninvasive diagnostic tool is currently not available. Ultrasonography and MRI are the most investigated modalities. A vein > 5 mm crossing the uterine body, pelvic varicoceles and reversed caudal flow shown with ultrasonography seem to indicate PCS. Additionally, this review highlights important gaps in the available literature. Future studies should ideally investigate the role of transvaginal ultrasonography with the patient in different positions, such as reverse Trendelenburg and using the Valsalva's maneuver and correlate these with the typical complaints of PCS and the current reference standard for diagnosis, being venography. There is an urgent need for methodologically adequate diagnostic accuracy studies in patients with a suspicion of PCS.

Chronic pelvic pain (CPP) is a common yet underestimated condition in women of reproductive age, with a prevalence of 5.7–26.6% 1 , 2 , 3 , 4 , 5 , 6 , 7 . By the International Association for the Study of Pain CPP is defined as pain, perceived in structures related to the pelvis, which has been continuous or recurrent for at least 6 months. In gynecology, CPP has several differential diagnoses, such as endometriosis, pelvic inflammatory disease, pelvic adhesions, ovarian pathology and the often‐neglected pelvic congestion syndrome (PCS) 1 , 2 , 3 , 4 , 5 , 8 . Pelvic congestion syndrome is a clinical entity, first described in 1857, in which varicose veins are related to CPP 9 . The prevalence of PCS among patients with CPP is found to be 12–33% 1 , 2 , 3 , 4 , 5 , 10 . Patients with PCS are mostly premenopausal, multiparous women who have complaints of CPP accompanied with dysmenorrhea, and exacerbating symptoms during and after intercourse or prolonged standing 11 , 12 , 13 . Pelvic varices are caused by incompetence of the ovarian veins. These veins arise from the ovarian venous plexus and communicate with the uterine plexus in the broad ligament. Incompetence leads to retrograde venous flow, progressive development of pelvic varicosities and dilatation 14 , 15 . The origin of PCS is most likely to be multifactorial 15 . Two important sources of pelvic vein insufficiency are described in the literature. First, valvular insufficiency, due to congenital absent or incompetent valves, plays a possible role in the etiology 15 . Second, during pregnancy the vascular capacity of ovarian veins enlarges by up to 60 times their normal value 16 . This increased capacity causes mechanical pressure and can eventually contribute to persistent venous reflux; explaining why PCS is mainly seen in multiparous women 17 . Venography is considered the reference standard test for the diagnosis of pelvic venous disorders 18 , 19 . The congestion is defined as extensive when wide veins are tortuous, with great variation in caliber, and when individual veins are obscured by a pool of contrast medium 18 , 19 . Venography is a valid method for diagnosing PCS, but is invasive, time‐consuming and exposes the pelvis of women of childbearing age to radiation 18 , 19 . If noninvasive diagnostic tools were found to be accurate, venography could possibly be avoided in a large number of women. Several noninvasive diagnostic tools are used in the work‐up of patients with CPP and suspected pelvic venous insufficiency, for example ultrasonography (Figure  1 ), computed tomography (CT) and magnetic resonance imaging (MRI) 14 , 20 , 21 . However, the diagnostic value of these tests remains unclear. In this review, we aim to evaluate studies on noninvasive diagnostic tests in the work‐up of patients with suspected PCS. An example of tortuous veins visualized with pulsed Doppler during transvaginal ultrasonography, a possible sign of pelvic congestion syndrome. [Color figure can be viewed at http://wileyonlinelibrary.com ].

Appendix S1. Search strategy. Click here for additional data file.