Endovascular reintervention versus conservative medical therapy for persistent or recurrent venous-origin chronic pelvic pain after primary ovarian vein embolization.

Conception and design: SX, XW, JH, XLu, CC, XLiu Analysis and interpretation: FL, TL Data collection: SX, XW Writing the article: SX, FL, XW Critical revision of the article: SX, TL, JH, XLu, CC, XLiu Final approval of the article: SX, FL, TL, XW, JH, XLu, CC, XLiu Statistical analysis: SX Obtained funding: XW Overall responsibility: XLiu SX and FL contributed equally to this article and share co-first authorship.

JYZZ243, Fundamental research program funding of Ninth People's Hospital affiliated to Shanghai Jiao Tong University School of Medicine.

This study was a single center, retrospective cohort study. Patients diagnosed with PeVD between January 1, 2018, and June 1, 2025, were retrospectively reviewed based on their outpatient and inpatient medical records at a single medical center that receives nationwide referrals. During the study period, a total of 190 hospitalized patients with PeVD who underwent endovascular treatment and 2677 outpatient visits were retrospectively screened. The study was conducted in accordance with the Declaration of Helsinki, and the institutional review board of the hospital approved the study protocol for this retrospective analysis (SH9H-2026-T37-1). The patients included in this study satisfied these criteria: (1) female patients, aged ≥18 years, with a confirmed diagnosis of PeVD; (2) underwent primary endovascular OV embolization, with or without additional embolization of internal iliac vein (IIV) tributaries or pelvic varices, at our institution or another medical center; (3) persistent CPP after initial embolization over 3-month follow-up or recurrent CPP during follow-up; (4) subsequently received either venoactive drugs (VADs) therapy or secondary endovascular reintervention; (5) computed tomography venography (CTV), magnetic resonance imaging, or transabdominal ultrasound study showing residual or recurrent pelvic varices with a diameter ≥5 mm. 14 For reintervention group, reflux in OV, IIV, or their tributaries, together with contrast delayed clearance in pelvic varices were required; (6) follow-up ≥ 6 months following the decision for secondary management. Exclusion criteria included (1) gynecological-related diseases (eg, endometriosis, adenomyosis, or pelvic inflammatory disease) and any identifiable abdominal or pelvic malignancies; (2) planned staged intervention, where the secondary interventional procedure was part of a multistep therapeutic strategy (eg, staged bilateral OV embolization); (3) adjunctive therapies during follow-up, such as acupuncture or pelvic floor electrical stimulation; (4) iliac stenting during reintervention. The decision to pursue endovascular reintervention or conservative medical therapy was made through shared decision-making between the treating physicians and patients, based on symptom severity, imaging findings, and patient preference. Patient data were extracted from the institutional electronic medical record system, encompassing both outpatient and inpatient encounters. Baseline characteristics at the decision-making point for secondary management were recorded. Extracted variables comprised demographic data (such as age and parity), clinical symptoms, procedural details of the first and the second intervention, Visual Analog Scale (VAS) scores, and anatomical findings from imaging examination. Regarding the “presenting complaint,” recurrence was defined as the return of CPP after at least 3 months of complete or significant symptom relief following the initial intervention. Patients who failed to achieve a ≥2-point reduction in VAS score at the 3-month follow-up following the initial intervention were classified as no relief, whereas those who achieved a ≥2-point reduction but continued to experience residual CPP symptoms were classified as partial relief. Preprocedural imaging was reviewed to identify residual or recurrent pelvic varices, assess their anatomical distribution, and evaluate potential communication with the OV or IIV systems, thereby guiding target catheterization and embolization. Bilateral IIV venography was not routinely undertaken in all patients. Pelvic varices, including the adnexal, parametrial, uterine, vesical, and rectal varices, and venous reflux pathways, including OV related reflux, and IIV tributary reflux were regarded as the primary treatment targets for embolization during reintervention. All endovascular procedures in the reintervention group were performed under local anesthesia without sedation. The common femoral vein was the preferred access site. Catheterization of the right OV used a Cobra (Cook Medical) or modified Pigtail catheter. IIV catheterization was typically performed via contralateral femoral access, with advancement of a Cobra catheter across the confluence of the common iliac veins. Pelvic varices were accessed coaxially using microguidewire and microcatheter (2.1-2.8F). The “sandwich technique” was generally applied during embolization which combines metallic embolization devices (0.018 and 0.035 detachable coils, Interlock; and Amplatzer, Abbott) with 1% to 3% sclerosant foam (sclerosant: air = 1:4; 2-5 mL foam for each target territory; Aethoxysklerol, Kreussler Pharma). Patients performed Valsalva maneuver during sclerosant injection to enhance dispersion. Liquid embolic agents, such as cyanoacrylate or ethylene vinyl alcohol copolymer were not applied in this set of patients. For the conservative medical group, the primary treatment consisted of VADs, specifically micronized purified flavonoid fraction (eg, Diosmin, 450 mg twice daily) or horse chestnut seed extract (eg, Aescuven, 300 mg twice daily) for a minimum duration of 3 months. Long-term maintenance dosages were adjusted according to clinical response during follow-up. Following reintervention, all patients received a standardized 3-month course of VAD as part of routine postoperative management to reduce postprocedural symptoms and facilitate recovery. 15 In addition, prophylactic anticoagulation with Rivaroxaban (15 mg once daily) was recommended for 1 month, particularly for those with left renal vein or iliac vein obstruction. Preprocedural imaging and reintervention venographic findings were retrospectively reviewed by two experienced authors to identify the anatomical distribution of pelvic varices and the potential reflux patterns associated with persistent or recurrent CPP. Any disagreements were resolved through consensus review. According to the anatomical distribution of pelvic varices within the V2 domain of the SVP classification, 2 we descriptively distinguished two pelvic venous plexus regions for procedural assessment: V2a (superior) predominantly comprised OV-drainage pathways, including the OVs, accessory OVs, ovarian venous plexus, and varices surrounding the uterine fundus; V2b (inferior) predominantly comprised IIV-related drainage pathways, including the uterine venous plexus surrounding the uterine body and cervix, as well as the vaginal, vesical, rectal venous plexuses and their tributaries 16 , 17 ( Fig 1 ). Fig 1 Schematic illustration of the pelvic venous reservoirs. V2a (superior) predominantly comprised ovarian vein-drainage pathways, including the ovarian veins, accessory ovarian veins, ovarian venous plexus, and varices surrounding the uterine fundus; V2b (inferior) predominantly comprised internal iliac vein-related drainage pathways, including the uterine venous plexus surrounding the uterine body and cervix, as well as the vaginal, vesical, rectal venous plexuses and their tributaries. Schematic illustration of the pelvic venous reservoirs. V2a (superior) predominantly comprised ovarian vein-drainage pathways, including the ovarian veins, accessory ovarian veins, ovarian venous plexus, and varices surrounding the uterine fundus; V2b (inferior) predominantly comprised internal iliac vein-related drainage pathways, including the uterine venous plexus surrounding the uterine body and cervix, as well as the vaginal, vesical, rectal venous plexuses and their tributaries. The primary endpoint was the 3-month CPP relief rate, which was defined as the percentage of patients achieving a reduction of ≥2 points in the VAS score at 3 months compared with pretreatment value. 18 The second endpoint was the longitudinal vas score trajectories over 12 months between the two groups. Outpatient visits were scheduled at 1, 3, 6, and 12 months postdischarge for the reintervention group, and every 1 to 3 months for the conservative group depending on symptomatic changes. VAS scores were collected at each follow-up visit. For patients with missing values at these intervals, telephone follow-ups were performed to retrospectively collect VAS scores at the corresponding time points. Transabdominal ultrasound study or computed tomography venography was routinely performed within the first 3 months postintervention to assess radiological changes. Incomplete embolization was defined as the persistence of residual pelvic varices within the target venous territory on follow-up imaging, indicating that the target venous region could not be reached and completely embolized during the procedure. Categorical variables were summarized using frequency and percentage. Continuous variables were summarized using mean (SD) values or median values and interquartile ranges as appropriate. The χ 2 test or Fisher exact test was used to compare categorical variables between groups (reintervention group vs medical therapy group), and Student t -test or the Mann-Whitney U test was used for continuous variables based on data distribution. Standardized mean differences were calculated to assess the magnitude of differences in baseline covariates between the two groups. The inverse probability of treatment weighting (IPTW) used a propensity score to weight the comparison for the imbalance of confounders observed at baseline between the two groups. Propensity scores were calculated via logistic regression, with adjustment for the following baseline covariates selected based on clinical relevance and group imbalances: age, baseline VAS score, varices diameter, presenting complaints, and left renal vein compression. The association between treatment and 3-month CPP relief rate was evaluated using IPTW-weighted logistic regression models. Results were presented as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). To assess the consistency of the therapeutic effect across subgroups, interaction terms (treatment × subgroup variable) were included in the models, and P values for interaction were calculated using the Wald test. The VAS score trajectories over 12 months between the two groups were analyzed using a weighted Linear Mixed-Effects Model. Post hoc pairwise comparisons were performed using estimated marginal means to evaluate the intergroup differences at each follow-up point (1, 3, 6, and 12 months, Supplementary Table , online only). To evaluate the consistency of the treatment effect difference at the 12-month follow-up, subgroup analyses were performed across predefined clinical strata. For each subgroup, the adjusted mean difference (AMD) in VAS scores between the two groups was estimated using the weighted models. Heterogeneity of the treatment effect across subgroups was evaluated by testing the significance of a multiplicative interaction term. All statistical tests were two-sided, and a P value < .05 was considered statistically significant. All analyses were conducted in R, version 4.5.3 (R Foundation for Statistical Computing).

A total of 79 patients presenting with persistent or recurrent symptoms following previous pelvic venous interventions were enrolled in this study. Of these, 40 (50.6%) underwent reintervention, whereas 39 (49.4%) received medical therapy. The baseline data, including demographics, medical history, clinical symptoms, imaging findings, and primary intervention details, are summarized in Table I . The mean age of the overall cohort was 37.6 ± 9.2 years, and the median baseline VAS score was 6.0 (5.0-7.0). Regarding the presenting complaint, 42 cases (53.2%) of patients sought treatment due to symptom recurrence, 21 cases (26.6%) of patients were refractory to their primary intervention (no relief), and 16 cases (20.3%) experienced partial relief with residual symptoms. Although all patients manifested CPP symptoms localized to S2 (SVP classification), concomitant involvement of S1 (renal symptoms of venous origin) and S3 (extrapelvic symptoms of venous origin) was observed in nine cases (11.4%) and 26 cases (32.9%) of the cohort, respectively. All the patients underwent left OV embolization, and 21 cases (26.6%) underwent right OV embolization during their initial intervention ( Table I ). No significant differences were observed between the two groups regarding the first intervention details, including the primary intervention center, embolization regions, and anatomical characteristics ( Table I ). Table I Baseline characteristics and procedural details in the initial intervention Variables Overall (N = 79) Reintervention group (N = 40) Medical therapy group (N = 39) P value Unadjusted SMD Adjusted SMD Baseline characteristics  Age, years b 37.61 ± 9.16 35.83 ± 7.94 39.44 ± 10.04 .08 0.399 0.015  BMI, kg/m 2 .548 0.250  Normal weight 73 (92.4) 36 (90.0) 37 (94.9)  Overweight 5 (6.3) 3 (7.5) 2 (5.1)  Underweight 1 (1.3) 1 (2.5) 0 (0)  VAS-baseline a , b 6.00 [5.0, 7.0] 6.5 [6.0, 7.0] 6.0 [5.0, 6.0] .002 0.707 0.123  Time to first intervention or recurrent time, minutes a 7.0 [4.5, 13.5] 6.0 [4.75, 11.25] 8.0 [4.5, 15.5] .951 0.014  Plexus diameter, mm a , b 6.0 [5.0, 7.0] 6.0 [5.0, 7.9] 6.0 [5.00, 6.9] .111 0.364 0.040  Pregnancy history, %  No 5 (6.3) 3 (7.5) 2 (5.1) 1.00 0.098  Yes 74 (93.7) 37 (92.5) 37 (94.9)  Presenting complaint, % b .493 0.270 0.041  Recurrence 42 (53.2) 23 (57.5) 19 (48.7)  No relief 21 (26.6) 11 (27.5) 10 (25.6)  Partial relief 16 (20.3) 6 (15.0) 10 (25.6)  S1, %  No 70 (88.6) 34 (85.0) 36 (92.3) .504 0.232  Yes 9 (11.4) 6 (15.0) 3 (7.7)  S3, %  No 53 (67.1) 28 (70.0) 25 (64.1) .750 0.126  Yes 26 (32.9) 12 (30.0) 14 (35.9)  Varices region, % .922 0.091  V2a 20 (25.3) 10 (25.0) 10 (25.6)  V2b 28 (35.4) 15 (37.5) 13 (33.3)  V2a + V2b 31 (39.2) 15 (37.5) 16 (41.0)  IV compression, %  No 31 (39.2) 16 (40.0) 15 (38.5) 1.00. 0.032  Yes 48 (60.8) 24 (60.0) 24 (61.5)  Left renal vein compression, % b  No 61 (77.2) 28 (70.0) 33 (84.6) .200 0.354 0.031  Yes 18 (22.8) 12 (30.0) 6 (15.4) Procedural details in initial intervention  The center of primary intervention, %  Our center 55 (69.6) 26 (65.0) 29 (74.4) .509 0.205  External center 24 (30.4) 14 (35.0) 10 (25.6)  Left OV embolization, % 1.00 0.00  No 0 0 0  Yes 79 (100.0) 40 (100.0) 39 (100.0)  Right OV embolization, %  No 58 (73.4) 29 (72.5) 29 (74.4) 1.000 0.042  Yes 21 (26.6) 11 (27.5) 10 (25.6)  IIV embolization, %  No 62 (78.5) 33 (82.5) 29 (74.4) .551 0.252  Unilateral 16 (20.3) 7 (17.5) 9 (23.1)  Bilateral 1 (1.3) 0 (0.0) 1 (2.6)  Escape point embolization, % .414 0.185  No 73 (92.4) 36(90.0) 37 (94.9)  Yes 6 (7.6) 4 (10.0) 2 (5.1)  Iliac stenting, %  No 75 (94.9) 37 (92.5) 38 (97.4) .626 0.227  Yes 4 (5.1) 3 (7.5) 1 (2.6) BMI, Body mass index; IIV, internal iliac vein; IV, iliac vein; OV, ovarian vein; SMD, standardized mean difference; VAS, Visual Analog Scale. a Non-normally distributed data with Mann-Whitney U test. b Variables included in the inverse probability of treatment weighting model. Baseline characteristics and procedural details in the initial intervention BMI, Body mass index; IIV, internal iliac vein; IV, iliac vein; OV, ovarian vein; SMD, standardized mean difference; VAS, Visual Analog Scale. Non-normally distributed data with Mann-Whitney U test. Variables included in the inverse probability of treatment weighting model. Before IPTW adjustment, the only statistically significant difference between the two groups was the baseline VAS score, which was higher in the reintervention group compared with the medical therapy group (median, 6.5 [6.0-7.0] vs 6.0 [5.0-6.0]; P = .002). Furthermore, compared with the medical therapy group, patients in the reintervention group tended to be younger. After IPTW adjustment, the baseline characteristics were well-balanced between the two cohorts. The standardized mean differences for all weighted variables—including age, baseline VAS score, varices diameter, presenting complaints, and the presence of left renal vein compression—were reduced to around the 0.1, indicating adequate covariate balance between the two groups ( Table I , Supplementary Fig , online only). Among the 40 patients who underwent reintervention, right OV reflux was the most frequently identified potential reflux pattern, occurring in 19 patients (47.5%). IIV-related reflux was also common, including left IIV tributary reflux in 11 patients (27.5%) and bilateral IIV tributary reflux in nine patients (22.5%). Accessory OV reflux, OV tributary reflux, and recanalization or incomplete embolization of the OV were less frequent, occurring in 10.0%, 5.0%, and 5.0% of patients, respectively ( Table II ). Notably, complete target lesion embolization was not achieved in all cases, even during reintervention: incomplete embolization was observed in 28.0% cases (7/25) of V2a lesions and 46.7% cases (14/30) of V2b lesions ( P = .177). Table II Potential reflux patterns associated with persistent or recurrent chronic pelvic pain (CPP) after initial embolization Total (N = 40) Recurrence (N = 23) No relief (N = 11) Partial relief (N = 6) OV-related reflux  Untreated right OV reflux 19 (47.5) 9 (39.1) 6 (54.5) 4 (66.7)  Accessory OV reflux 4 (10.0) 3 (13.0) 1 (9.1) 0  OV tributary reflux 2 (5.0) 1 (4.3) 1 (9.1) 0  Recanalization or incomplete embolization of OV 2 (5.0) 2 (8.7) 0 0 IIV-related Reflux  Left IIV tributary reflux 11 (27.5) 8 (34.8) 1 (9.1) 2 (33.3)  Right IIV tributary reflux 1 (2.5) 0 1 (9.1) 0  Bilateral IIV tributary reflux 9 (22.5) 6 (26.1) 1 (9.1) 2 (33.3) IIV, Internal iliac vein; OV, ovarian vein. Potential reflux patterns associated with persistent or recurrent chronic pelvic pain (CPP) after initial embolization IIV, Internal iliac vein; OV, ovarian vein. Twenty-seven cases (67.50%) in the reintervention group and 14 cases (35.90%) in the medical therapy group achieved a 3-month CPP relief (VAS score reduction ≥2). After adjusted analysis, reintervention demonstrated a significant overall therapeutic advantage over medical treatment, with a more than threefold increase in the odds of CPP relief compared with medical therapy (OR, 3.03; 95% CI, 1.19-7.68; P = .022, Fig 2 ). Subgroup analyses demonstrated generally consistent treatment effects across all predefined subgroups, with no significant treatment-by-subgroup interactions observed (all P for interaction > 0.05). Fig 2 Subgroup analysis of 3-month chronic pelvic pain relief. The forest plot illustrates the odds ratios (ORs) and 95% confidence intervals ( CIs ) comparing 3-month chronic pelvic pain relief rates between the endovascular reintervention and medical therapy groups. VAS , Visual Analog Scale. Subgroup analysis of 3-month chronic pelvic pain relief. The forest plot illustrates the odds ratios (ORs) and 95% confidence intervals ( CIs ) comparing 3-month chronic pelvic pain relief rates between the endovascular reintervention and medical therapy groups. VAS , Visual Analog Scale. VAS score at the 12-month follow-up was missing for one case in each group. After IPTW adjustment, there was no statistically significant difference in VAS scores between the reintervention and medical therapy groups at baseline ( Fig 3 , A and B ). The Linear Mixed-Effects Model revealed a significant group × time interaction ( P < .001), indicating that the therapeutic trajectories of the two groups were significantly different over the 12-month follow-up period. Post hoc pairwise comparisons demonstrated that the reintervention group achieved a significantly greater reduction in pain compared with the medical therapy group starting from the first month ( Fig 3 , C ; Supplementary Table , online only). Specifically, the reintervention group showed lower VAS scores at 1 month (AMD, −1.09; 95% CI, −1.77 to −0.40; P = .002), with the maximum therapeutic divergence observed at 3 months (AMD, −1.37; 95% CI, −1.96 to −0.78; P < .001). This superior symptom control was sustained through 6 months and remained statistically significant at the 12-month final follow-up (AMD, −0.70; 95% CI, −1.28 to −0.12, P = .021). These results suggest that surgical reintervention-based strategy provides more rapid and durable pain relief than conservative medical management in the total cohort. No severe complications, such as embolic coil or stent migration, or deep vein thrombosis, were observed in this cohort during 1-year follow-up. Fig 3 Longitudinal trajectories of Visual Analog Scale ( VAS ) scores between reintervention and medical therapy groups. A, Unadjusted mean VAS scores over the 12-month follow-up period. B, Inverse probability of treatment weighting ( IPTW )-weighted mean VAS scores showing balanced baseline pain intensity. C, Estimated marginal means derived from the Linear Mixed-Effects Model ( P < .001). Longitudinal trajectories of Visual Analog Scale ( VAS ) scores between reintervention and medical therapy groups. A, Unadjusted mean VAS scores over the 12-month follow-up period. B, Inverse probability of treatment weighting ( IPTW )-weighted mean VAS scores showing balanced baseline pain intensity. C, Estimated marginal means derived from the Linear Mixed-Effects Model ( P < .001). To evaluate the consistency of treatment effects, we performed subgroup analyses at the 12-month follow-up. Significant heterogeneity in long-term treatment effect was observed according to presenting complaint ( P for interaction < .001). Reintervention was associated with greater reductions in VAS scores among patients presenting with recurrent symptoms (AMD, −1.26; 95% CI, −1.96 to −0.57; P < .001) and partial symptom relief (AMD, −1.84; 95% CI, −3.12 to −0.57; P = .005), whereas no significant benefit was observed in patients reporting no relief after the initial embolization (AMD, −0.56; 95% CI, −1.70 to 0.57; P = .326). Subgroup analysis by initial intervention center revealed a more pronounced therapeutic benefit for reintervention among patients from external institutions (AMD, −2.48; 95% CI, −3.45 to −1.50; P < .001).

Evidence guiding the management of patients with persistent or recurrent venous-origin CPP after initial endovascular intervention is lacking. This single center retrospective study with 79 cases found that secondary pelvic endovascular reintervention provides more rapid CPP relief and superior efficacy within the 1-year follow-up compared with conservative medical therapy in patients with persistent or recurrent CPP after initial OV embolization. Patients undergoing reintervention demonstrated a more than threefold increase in the odds of achieving CPP relief within 3 months (OR, 3.03; 95% CI, 1.19-7.68; P = .022). Furthermore, the longitudinal analysis confirmed that this therapeutic advantage was sustained throughout the 12-month follow-up. The persistence or recurrence of symptoms following endovascular embolization in venous-origin CPP is closely related to its complex pathophysiology. (1) Technically, incomplete primary occlusion or residual reflux remains the most prevalent driver of failure. Given that 24% to 40% of patients possess multiple OV trunks, traditional coil-based strategies may “miss” accessory branches. 19 , 20 This hypothesis could be further supported by the inclusion criteria of this study, as all enrolled cases exhibited a pelvic varices diameter no less than 5 mm. (2) Physiologically, late-stage recurrence often results from vascular adaptation, including recanalization or the recruitment of alternative collateral pathways. 21 Conversely, early-stage nonresolution may be attributed to localized thrombophlebitis or inflammatory responses. 22 (3) Central sensitization offers a compelling explanation for persistent nonresponders. 23 , 24 (4) Finally, diagnostic ambiguity of venous-origin CPP must be acknowledged. Since pelvic varices is frequently observed in asymptomatic women, it is possible that in certain patients, radiologically identified venous reflux and varices are merely a concomitant finding rather than the primary etiology of the patient's CPP. 25 In this cohort, the 3-month CPP relief rate in the reintervention group was 67.50%, leaving 32.50% of patients without significant symptomatic improvement. This rate is notably lower than the 80.4% primary treatment CPP control rate previously reported by our center. 26 Several factors may account for this discrepancy. First, retrospective review of imaging and venography demonstrated that persistent or recurrent symptoms were frequently associated with complex residual reflux pathways, including untreated right OV reflux, accessory OVs, and internal iliac tributary reflux ( Table II ), suggesting more extensive venous disease than initially appreciated. Second, incomplete embolization of pelvic venous reservoirs remained relatively common, particularly in the V2b region, where some target varices and reflux sites were inaccessible to catheterization and therefore could not be completely embolized. Finally, although major gynecological and malignant causes were excluded, a nonvenous component of CPP cannot be completely ruled out and may contribute to persistent symptoms despite technically successful reintervention. To better characterize procedural findings, we descriptively divided the V2 from PeVD system into superior (V2a) and inferior (V2b) regions ( Fig 1 ) based on their anatomical distribution. The V2a region is typically more accessible during initial procedures via catheterization of the OVs. Due to extensive pelvic venous communications, V2a can be accessed from either side and may serve as a conduit to reach the V2b region. In contrast, the V2b region, located in the lower pelvis, drains predominantly through the IIV. 16 , 27 Embolizing dilated plexuses in V2b presents a challenge due to the complex branching of the IIV system. V2b plexuses are distributed more broadly across the anteroposterior pelvic planes, 16 creating significant hurdles for stable catheter positioning and effective target-zone navigation. Even in this reintervention series, successful embolization was achieved in only 53.3% (16/30) of V2b lesions, a rate numerically lower than the 72.0% (18/25) observed in V2a. This disparity reflects a technical difficulty inherent in accessing the distal pelvic venous plexus, particularly after primary interventions have altered the original hemodynamics. Although the symptom relief rate in the reintervention group was lower than that of primary treatments, it demonstrated a significant short-term advantage over the medical therapy group after adjustment (OR, 3.03; 95% CI, 1.19-7.68; P = .022). Notably, the relief rate in our medical therapy cohort was only 35.9% (14/39), which is lower than the success rates typically reported in the literature. 11 This discrepancy likely stems from the unique nature of our postprocedural population, where conservative management may be less effective for patients with persistent anatomical triggers. 28 Reintervention group maintained a significant therapeutic advantage over medical therapy throughout the 12-month follow-up ( P < .001) as the longitudinal analysis demonstrates ( Fig 3 , C ). However, the longitudinal analysis revealed a gradual convergence of clinical outcomes between the two cohorts. This narrowing gap was mainly due to a slight increase in VAS scores in the reintervention group during the later follow-up period. A factor contributing to this phenomenon may be the presence of underaddressed venous outflow obstruction within this cohort. Despite the high prevalence of iliac vein compression (60.8%) and left renal vein compression (22.8%), no stenting was used. The unsolved outflow obstruction may be a potential cause of the VAS score increase observed during later follow-up. 29 , 30 In addition, given the extensive interconnections within the pelvic venous network, incompletely embolized pelvic venous reservoirs may be re-exposed to venous hypertension through collateral redistribution of reflux and thereby contribute to symptom recurrence and the gradual attenuation of treatment benefit. 31 The 12-month subgroup analysis suggested that patients reporting no relief after the initial embolization derived substantially less benefit from reintervention than the other subgroups ( Fig 4 ). This trend could also be identified at 3 months, although no significance ( Fig 2 ). Notably, Table II did not reveal substantial differences in residual reflux patterns between the no-relief group and the other symptom categories. These findings raise the possibility that, in some patients, the identified pelvic varices and reflux pathways may represent concomitant anatomical findings rather than the primary drivers of CPP, suggesting the coexistence of a nonvenous pain reason. 32 Therefore, in patients who experience no CPP relief after the initial embolization, the indication for repeat venous intervention should be considered cautiously, and alternative causes of CPP should be actively investigated. Fig 4 Subgroup analysis of adjusted mean differences ( AMDs ) in Visual Analog Scale ( VAS ) at 12-month Follow-up. The forest plot illustrates the AMD and 95% confidence intervals ( CIs ) in VAS scores between the reintervention and medical therapy groups with weighted Linear Mixed-Effects Model. Subgroup analysis of adjusted mean differences ( AMDs ) in Visual Analog Scale ( VAS ) at 12-month Follow-up. The forest plot illustrates the AMD and 95% confidence intervals ( CIs ) in VAS scores between the reintervention and medical therapy groups with weighted Linear Mixed-Effects Model. The institution of the primary intervention emerged as another potential driver of heterogeneity. A greater benefit was observed in patients referred from external institutions compared to those initially treated at our own center ( Fig 4 ). A plausible explanation is the technical and procedural variability between institutions. A secondary intervention at a specialized institution may overcome previous blind spots, such as the incomplete occlusion of accessory OV trunks or the overlooked distal pelvic venous reservoirs, as a new diagnostic perspective and diverse access strategies may lead to the successful management of previously untreated venous segments. The limitations of this study should be acknowledged. This was a single-center retrospective cohort study with a relatively small sample size, which may limit the generalizability of our findings. Although IPTW was applied to reduce baseline imbalances, and two complementary statistical models were used to evaluate treatment effects, residual confounding from unmeasured variables cannot be completely excluded, particularly regarding VAD management, as information on preintervention VAD exposure, response to VAD therapy, and treatment compliance was not available for analysis. Therefore, the overall level of evidence remains limited. Moreover, in the conservative treatment group, the venous origin of CPP was inferred primarily from pelvic varices ≥5 mm without consistent hemodynamic confirmation of venous reflux, which may have introduced diagnostic uncertainty compared with the reintervention group. The sample size was further reduced after stratification, particularly in subgroup analyses, which may have limited statistical power and increased the risk of type II error. Therefore, these subgroup findings should be interpreted with caution and regarded as hypothesis-generating rather than definitive.

In this single-center retrospective cohort study, secondary pelvic endovascular reintervention-based strategy demonstrated significantly superior effectiveness compared with conservative medical therapy alone for patients with persistent or recurrent CPP after initial endovascular pelvic venous embolization. The benefit was evident within 3 months and persisted over 12 months, with no major procedure-related complications observed during follow-up. However, treatment effects appeared to be heterogeneous at 12 months, with patients who had achieved an initial symptomatic response and those initially treated at external institutions appearing to derive greater benefit from reintervention. These findings suggest that reintervention is a reasonable option for selected patients, but larger prospective studies are needed to refine patient selection and optimize management strategies.

None.