Intravasation is not only a complication but also a radiological finding for tubal pathologies in hysterosalpingography.

Despite the availability of advanced imaging modalities, hysterosalpingography (HSG) remains a widely accessible, cost-effective, and the most commonly preferred first-line imaging technique and is considered the most efficient method for visualizing and assessing the condition of the fallopian tubes. 1 According to the meta-analysis by Maheux-Lacroix et al, 2 HSG is considered an alternative to laparoscopy, which is the widely accepted standard for diagnosing tubal blockages. 3 The meta-analysis reports that the pooled estimates of sensitivity and specificity for HSG in detecting tubal occlusion are 0.94 (95% CI, 0.74-0.99) and 0.92 (95% CI, 0.87-0.95), respectively. 2 , 3 This indicates that HSG has a high level of accuracy in diagnosing tubal occlusion. Intravasation (IVZ) refers to the leakage of contrast media from the uterine cavity into myometrial vessels or surrounding venous or lymphatic structures, with reported rates ranging from 0.0% to 13%. 4 In this study, we aimed to accomplish 3 primary objectives: (1) to determine the prevalence of IVZ in the largest documented single-centre patient population in the literature, (2) to examine the relationship between IVZ and tubal occlusion, and (3) to evaluate the relevance and necessity of grading IVZ in HSG.

Before the HSG examinations, written informed consent was obtained from each patient for the procedure and for the use of their data in potential research. Ethical approval was granted by the hospital’s ethics committee (23.05.2024-140-7/19). The participants’ consent was documented, and the examination images were transferred to the Picture Archiving and Communication System (PACS) provided by Sectra AB, located at Teknikringen 20, SE-583 30 Linköping, Sweden. This was a cross-sectional study conducted at a tertiary hospital between January 2021 and May 2024. Inclusion criteria consisted of all patients who underwent HSG during this period. The study included 3032 HSG examinations conducted between January 2021 and May 2024. Exclusion criteria included cases with poor imaging quality ( n  = 49) or inadequate enface visualization of the uterus, as determined by the interpreting radiologist. Missing data were rare (<2%) and did not affect key outcomes significantly ( Figure 1 ). Flow diagram: patient selection. Radiological findings, including tubal occlusion and complications like venous IVZ, were prospectively categorized and stored in separate folders within the PACS system. The study recorded the number of patients with tubal patency, as well as cases of unilateral or bilateral occlusion. It also documented the presence of adhesions, indicated by the use of loculated water-soluble iodinated contrast agent (WSCA), and instances of IVZ, both with and without tubal occlusions. Additionally, the study noted the initiation and washout times for cases where IVZ was observed. HSG procedures were performed during the proliferative phase of the endometrium, typically between the 6th and 9th days after menstruation, utilizing only WSCA through manual infusion. Patients with irregular cycles or long-term amenorrhea underwent beta-HCG testing on the examination day. The patient lay supine on the X-ray table, with a speculum inserted into the vagina for viewing the cervix. A delicate catheter was then gently inserted through the cervix into the uterine cavity. Uterine positioning was initially assessed by administering a small quantity of WSCA after using a spasmolytic agent, such as Hyoscine butylbromide. If necessary, a flexible catheter with an inflated balloon, featuring a braided surface (ClinodeviceTR Co., Catheter OD/[FR] 5F, length 30 cm), was utilized to gently manipulate the uterus. The position of the uterus was re-evaluated after the manual injection of WSCA. In situations where the traction from the balloon catheter was inadequate, a tenaculum was used. After visualizing the fallopian tubes, the catheter balloon was deflated during the final phase of the examination to visualize the lower part of the uterus and cervix. The HSG examinations in this study were conducted using standard X-ray equipment available at the hospital. The examination team consisted of 2 nurses and 1 X-ray technician, all specifically trained in HSG and each with over 10 years of experience performing these procedures.

The study sample consisted of 3032 participants. The mean age was 32.8 ± 6.6 years, with a median of 32.0 years (ranging from 21.0 to 55.0 years). Regarding tubal patency, 16.4% ( n  = 497) of participants had bilateral tubal obstruction, while 83.6% ( n  = 2535) had bilateral tubal patency. For tubal occlusion, 83.6% ( n  = 2535) had no occlusion, while 16.4% ( n  = 497) had occluded tubes. Among the 497 individuals with tubal occlusion, 73.4% ( n  = 365) had non-secondary/iatrogenic occlusion, and 26.6% ( n  = 132) had secondary or iatrogenic occlusion. Regarding IVZ, 97.4% ( n  = 2954) had no IVZ, while 2.6% ( n  = 78) had IVZ. The occurrence of IVZ combined with tubal occlusion was noted in 2.4% ( n  = 72) of participants. Lastly, 98.8% ( n  = 2996) had no loculated WSCA or fimbrial phimosis, whereas 1.2% ( n  = 36) exhibited these conditions ( Table 1 ). The descriptive analysis of patient age, tubal patency, unilateral or bilateral tubal occlusions. Secondary (or iatrogenic) tubal occlusions refer to blockages caused by medical interventions or conditions such as tubal ligation, ectopic pregnancy, salpingo-oophorectomy, and other surgical procedures.Numbers ( n ) and percentages in parentheses are shown in bold. Abbreviation: WSCA = water-soluble iodinated contrast agent. The descriptive data for the patients with IVZ in the study includes the following values: For age, the sample size was 76, with a minimum age of 21 years, a maximum age of 52 years, an average age of 35 years, and a SD of 6.58. Regarding the initiation of IVZ during HSG, the sample size was also 76, with a minimum of 0 minutes, a maximum of 9 minutes, an average duration of 3.47 minutes, and a SD of 2.90. For the duration of IVZ in HSG, 25 cases were considered, with a minimum of 1 minute, a maximum of 8 minutes, an average of 3.12 minutes, and a SD of 2.28. The duration of HSG examinations with IVZ was recorded for 75 participants, with a minimum of 1 minute, a maximum of 9 minutes, an average of 3.08 minutes, and a SD of 1.80. Finally, the number of X-ray exposures taken during the HSG was recorded for 78 participants, with a minimum of 4 exposures, a maximum of 12 exposures, an average of 6.42 exposures, and a SD of 1.75. The comparison between individuals without tubal occlusion ( n  = 2535) and those with occlusion ( n  = 497) reveals significant differences in various factors. The mean age for those without occlusion is 32.5 ± 6.6 years, while for those with occlusion, the mean age is 34.7 ± 6.0 years, with the difference being statistically significant ( P  < .001). In terms of age distribution, 42.2% of individuals without occlusion are 30 years or younger, compared to only 26.2% in the occlusion group, while 57.8% of those without occlusion and 73.8% of those with occlusion are older than 30 years ( P  < .001). Regarding IVZ, 99.8% of individuals without occlusion did not experience IVZ, while 85.5% of those with occlusion did not, and 14.5% had IVZ ( P  < .001). Additionally, loculation of WSCA or fimbrial phimosis was rare in the group without occlusion, with 99.8% having no such conditions, whereas 93.6% of individuals with occlusion did not exhibit this condition, but 6.4% did ( P  < .001). The analysis shows the following results for individuals with and without tubal occlusion in relation to IVZ ( Table 2 ). Among the 78 individuals with IVZ, 72 had tubal occlusion, yielding a PPV of 92.3% (72/78). Among the 2954 individuals without IVZ, 2529 did not have tubal occlusion, yielding a NPV of 85.6% (2529/2954). The diagnostic value of intravasation as a radiological indicator for tubal occlusion is shown. A DOR of 71.4 indicates that the likelihood of observing occlusion is 71.4× higher in patients with intravasation compared to those without it. However, this is the crude OR value. The adjusted OR value is calculated using multivariate logistic regression analysis. Percentages are shown in bold. Abbreviations: NPV = negative predictive value, PPV = positive predictive value. For the overall sample of 3032 individuals, the sensitivity, which measures the proportion of those with occlusion correctly identified, is 14.5% (72/497). The specificity, which measures the proportion of those without occlusion correctly identified, is 99.8% (2529/2535). The logistic regression analysis of factors associated with tubal occlusion reveals both univariate and multivariate results. In the univariate analysis, the crude ORs and 95% CIs for different factors were as follows: for women older than 30 years (compared to those 30 years or younger), the OR was 2.1 (95% CI, 1.7-2.6, P  < .001); for the presence of IVZ (compared to none), the OR was 71.4 (95% CI, 30.8-165.3, P  < .001); and for the presence of loculated WSCA or fimbrial phimosis (compared to none), the OR was 43.5 (95% CI, 15.3-123.7, P  < .001). In the multivariate analysis, the adjusted ORs for these factors were 2.0 (95% CI, 1.6-2.5, P  < .001) for age over 30 years, 54.6 (95% CI, 23.3-127.7, P  < .001) for IVZ, and 29.6 (95% CI, 9.8-88.9, P  < .001) for loculated WSCA or fimbrial phimosis. These findings indicate that each of these factors are significantly associated with tubal occlusion. In the group of 72 participants with IVZ and tubal occlusion, the mean age was 35.6 ± 6.3 years, with a median age of 35.0 years (ranging from 21.0 to 52.0 years). Regarding the type of occlusion, 45.8% ( n  = 33) had unilateral occlusion, with 29.2% ( n  = 21) on the right side, 16.7% ( n  = 12) on the left side, and 54.2% ( n  = 39) had bilateral occlusion. In terms of the site of IVZ compared to the site of occlusion, 41.7% ( n  = 30) had ipsilateral IVZ, 45.8% ( n  = 33) had bilateral IVZ, 1.4% ( n  = 1) had contralateral IVZ on the right, and 6.9% ( n  = 5) had contralateral IVZ on the left. Secondary tubal occlusion was present in 11.1% ( n  = 8) of cases, while 88.9% ( n  = 64) had non-secondary occlusion. Regarding loculated WSCA or fimbrial phimosis, 83.3% ( n  = 60) had no such conditions, while 16.7% ( n  = 12) exhibited these conditions. Regarding infertility, 26.4% ( n  = 19) had primary infertility, while 73.6% ( n  = 53) had secondary infertility. The mean duration of infertility was 4.3 ± 4.0 years, with a median duration of 2.0 years (ranging from 0.1 to 16.0 years). The distribution of IVZ and occlusion sites in the 72 participants with tubal occlusion is summarized ( Table 3 ). In total, there were 21 cases with right-side IVZ, 12 with left-side IVZ, and 39 with bilateral IVZ, summing to 72 cases. The table illustrates the details of intravasation (IVZ) and tubal occlusion, highlighting site-specific features. Among 39 cases of bilateral tubal occlusion, IVZs were observed bilaterally in 31 HSGs and unilaterally in 8 cases; specifically, 3 cases were on the left side and 5 cases on the right side. The volume of contrast agent administered did not exceed 10 mL in any case. The Kappa value for inter-observer agreement was almost perfect, with a value of 0.97.

This study underscores IVZ as not merely a complication but a significant radiological finding with diagnostic implications in HSG. With the largest documented patient cohort, we demonstrated that IVZ, when present, is a strong predictor of tubal occlusion, offering high specificity and PPV. Moreover, the association between IVZ and tubal occlusion highlights the need to redefine its role in HSG evaluations. The findings also challenge the adequacy of the existing classification system, suggesting the need for a more comprehensive grading approach that incorporates initiation timing and anatomical characteristics.

The primary predisposing factors for IVZ include cavitary surgical procedures (such as curettages, hysteroscopy, adhesiolysis, septoplasty, etc.) that result in endometrial injury, and high intra-cavitary pressures caused by the rapid infusion of contrast media, whether administered manually or via an infusion pump. Unilateral or bilateral tubal occlusion and endometrial inflammation are also significant factors. 3 , 5 Additionally, potential predisposing causes include the early follicular and late luteal phases of the menstrual cycle, menometrorrhagia, secondary infertility, endometriosis, and Müllerian anomalies. 6 IVZ is recognized as a complication during HSG, with prevalence rates varying depending on the type of contrast media used. It presents a potential diagnostic challenge for inexperienced reviewers in distinguishing it from free peritoneal contrast spillage. In a recent nationwide survey of 1876 HSGs using WSCM, IVZ occurred in 25 cases (1.8%). 7 In our single-centre study, which included 3032 HSGs, IVZ was observed in 78 cases (2.6%). The imaging technique for HSG at our centre differed from that employed in nationwide survey by Roest et al. 7 We used basic X-ray equipment because of limitations in our facility, while the survey employed fluoroscopy. This methodological difference may account for the higher prevalence rate observed in our study. Despite appearing disadvantageous, our approach provides certain benefits, such as enhanced imaging capabilities and reduced X-ray exposure for patients. The X-ray technique produces static images, offering high-definition still images of the body part being examined. In contrast, fluoroscopy generates live video-format images by capturing numerous consecutive still images. Consequently, fluoroscopy employs a lower dose for live imaging, resulting in lower-resolution images compared to plain X-ray. When focusing on visualizing tubal patency, the human eye might overlook the initiation of IVZ. Additionally, X-rays present a lower risk of radiation-related complications compared to fluoroscopy. Although no universal gold standard exists for confirming tubal occlusion in HSG findings, the presence of IVZ was analysed using specific diagnostic metrics, as described. This study acknowledges the absence of surgical validation as a limitation. However, HSG is regarded as an alternative to laparoscopy, which is the widely accepted standard for diagnosing tubal blockages. 2 , 3 One of the limitations in the present study is the absence of oil-soluble contrast agent, attributed to 2 primary reasons: cost constraints and the potential risk of oil embolism, especially when using basic X-ray equipment. Another limitation that may affect the applicability of the findings is the lack of a pressure manometer during manual WSCM infusion into the uterine cavity. IVZs do not occur in all cases of tubal occlusions; however, when present, they serve as a significant indicator of tubal occlusion, rather than being considered just a complication. It is observed that occlusion occurs approximately 55× more frequently in cases with IVZ compared to cases where IVZ is not observed. Among 39 cases of bilateral tubal occlusion, IVZs were observed bilaterally in 31 HSGs and unilaterally in 8 cases ( Table 3 ). The ovarian veins stem from the ovarian plexuses, which interconnect with the adnexal and uterine venous plexuses supplied by the bilateral uterine veins. The left ovarian vein empties into the left renal vein, while the right ovarian vein drains directly into the inferior vena cava. In as many as 40% of patients, multiple gonadal veins can be observed. 8–10 The path of contrast agent IVZ and washout aligns with this anatomical venous pathway. When IVZ occurs, X-ray imaging captures the contrast agent along this route, with visibility dependent on the exposure time. The contrast media can be observed along the entire pathway simultaneously, irrespective of the underlying physiopathology causing the IVZ, which holds significance ( Figures 2-8 ). In cases of mild or slight pathology, a minimal quantity of contrast media infiltrates the veins, potentially leading to the observation of the contrast agent in sequential locations such as the myometrium, parametrium, and ovarian veins during the same HSG examination ( Figures 2 and 5-8 ). This suggests that the classification presented by Dusak et al 6 could be deceptive, and there is a need to explore a new system. We believe that incorporating the initiation time of IVZ into Dusak et al’s system, an aspect currently absent could enhance the grading to better represent the level of pathology in the uterine venous circulation system. Bilateral proximal tubal occlusion and venous intravasation in a 30-year-old woman experiencing infertility for 2 years; gravida 1, parity 1, abortus 0. The X-ray exposure times, along with the initiation and washout times, have been inserted. Fluoroscopy-guided tubal recanalization was performed for the same patient shown in the previous figure, with no evidence of venous intravasation. Tubal patency and the absence of intravasation around the uterus were observed in the follow-up HSG exam of the same patients shown in Figure 1 , following fluoroscopy-guided tubal recanalization performed 9 months ago. Bilateral proximal tubal occlusion and venous intravasation in a 37-year-old woman with infertility for 1.5 years; gravida 2, parity 1, abortus 1, caesarean section 1. Venous intravasation begins 6 minutes and 8 seconds after the infusion of WSCM into the uterine cavity. Intravasation occurs within the myometrial, parametrial, and peripheral veins, extending to the right ovarian vein, and it is nearly undetectable 5 minutes and 10 seconds after the initiation of WSCM. Right tubal occlusion and ipsilateral parametrial and ovarian venous intravasation in a 42-year-old woman with infertility for 8 years; gravida 4, parity 3, abortus 1, caesarean section 1. Venous intravasation starts 47 seconds after the infusion of WSCM into the uterine cavity, specifically in the parametrial area. It extends to the right ovarian vein in the follow-up X-ray exposure. The uterine cavity exhibits a fundal indentation of 11 mm, verified with 3D transvaginal ultrasonography in the coronal plane. During the examination, a tenaculum was used to apply traction to the uterus. Bilateral proximal tubal occlusion and venous intravasation in a 35-year-old woman experiencing infertility for 2 years; gravida 0, parity 0, abortus 0. The X-ray exposure times, along with the initiation and washout times, have been inserted. Venous intravasation starts 4 minutes and 3 seconds after the infusion of WSCM into the uterine cavity, specifically in the parametrial area, and extends to the right ovarian vein. It is nearly undetectable 5 minutes and 3 seconds after the initiation of WSCM in the follow-up X-ray exposure. Tubal occlusion and venous intravasation in a 23-year-old woman with a unicornuate uterus experiencing infertility for a year; gravida 0, parity 0, abortus 0. The class of uterine anomalies is confirmed using 3D transvaginal ultrasonography in the coronal plane. Venous intravasation occurs 1 minute and 31 seconds after the infusion of WSCM into the uterine cavity, specifically in the parametrial area, and extends to the ovarian veins on both sides. Intravasation becomes completely washed out 6 minutes and 29 seconds after WSCM is initiated in the follow-up X-ray exposure. Despite its proximity to the genital tract’s microbial flora, the outer endometrial layer of the uterus lacks substantial lymphatic circulation. 11 Although IVZ can occur via the lymphatic system, none of the cases in this study were of lymphatic origin. This conclusion is supported by the rapid washout times and the mean duration of WSCA IVZ, which was 3 minutes and 12 seconds ( Table 2 ).

Statistical analysis was conducted using IBM SPSS Statistics for Windows, Version 25 (IBM Corp., Armonk, NY, United States). Descriptive statistics are presented as frequency, percentage, mean ± SD, and median (with minimum and maximum values). The normality of continuous variables was evaluated using visual methods, such as histograms and probability plots, along with analytical methods like the Kolmogorov-Smirnov test. The independent samples t -test was conducted to compare normally distributed data between 2 independent groups. The mean ± SD for the independent samples t -test is shown in the tables. Chi-square tests were employed to compare categorical variables among independent groups. To evaluate the potential of IVZ in predicting occlusion, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and diagnostic odds ratio (DOR) were calculated. Furthermore, factors associated with occlusion were assessed using logistic regression analysis. Significant variables identified in the univariate analysis ( P  < .05) were included in the multivariate logistic regression analysis, and adjusted OR were computed. In this study, a statistical significance level of P  < .05 was considered. The anonymized DICOM files from 22 HSGs with tubal occlusion and 18 HSGs with normal tubal patency were randomly mixed and exported from the PACS into a single folder. These files were then provided to the reviewers, excluding the initial observer (C.G.), with no patient identification or information regarding potential findings. Fleiss Kappa was used to calculate the inter-observer agreement among multiple observers.