Uterine artery embolization with or without methotrexate followed by D&C versus D&C alone for cervical pregnancy: a retrospective cohort study.

A total of 67 patients were enrolled. Treatment allocation was based on shared decision-making after discussion of risks and benefits. Consequently, 43 patients opted for and received UAE with or without MTX followed by D&C, while 24 patients underwent D&C alone. Among the 43 patients receiving UAE-based therapy, 25 underwent UAE alone and 18 received adjunctive intra-arterial MTX (Fig.  1 ). Fig. 1 Patient enrollment flowchart Patient enrollment flowchart Table  1 presents baseline characteristics. The two groups were comparable in demographic and obstetric parameters ( p  > 0.05 for all). Clinical presentation symptoms—including vaginal bleeding, abdominal pain, or asymptomatic status—were similarly distributed between groups ( p  > 0.05). Table 1 Baseline Demographic and Clinical Characteristics of Women with CP Characteristic UAE (± MTX) + D&C group ( n  = 43) D&C Alone group ( n  = 24) p value Age (years), mean ± SD 33.39 ± 5.51 32.86 ± 5.62 0.709 BMI (kg/m²), mean ± SD 23.8 ± 4.2 24.7 ± 3.5 0.377 Gravidity, mean ± SD 3.42 ± 1.41 3.22 ± 1.52 0.590 Parity, mean ± SD 0.86 ± 0.75 0.66 ± 0.56 0.259 Number of prior miscarriages, mean ± SD 1.48 ± 1.17 1.66 ± 0.99 0.527 Number of prior cesarean deliveries, mean ± SD 1.52 ± 0.62 1.45 ± 0.55 0.647 Conceived by ART, n (%) 8 (18.6%) 4 (16.7%) 0.720 Gestational age at diagnosis (days), mean ± SD 54.7 ± 11.4 48.5 ± 9.6 0.048 Fetal cardiac activity present, n (%) 16 (37.2%) 5 (20.8%) 0.038 Preoperative β-hCG (mIU/mL), mean ± SD 25,521 ± 8,299 4,472 ± 1,891 < 0.001 Gestational sac diameter (cm), mean ± SD 2.97 ± 1.25 1.34 ± 0.64 < 0.001 Symptoms at presentation, n (%) Asymptomatic 6 (14.0%) 3 (12.5%) 1.000 Vaginal bleeding 26 (60.5%) 13 (54.2%) 1.000 Abdominal pain 5 (11.6%) 4 (16.7%) 0.697 Vaginal bleeding + abdominal pain 6 (14.0%) 4 (16.7%) 1.000 Abbreviations: CP cervical pregnancy, UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, SD standard deviation, BMI body mass index, ART assisted reproductive technology, β-hCG beta-human chorionic gonadotropin Data are presented as mean ± SD for continuous variables and n (%) for categorical variables. p values were calculated using Student’s t-test for continuous variables and Chi-squareor Fisher’s exact test for categorical variables Baseline Demographic and Clinical Characteristics of Women with CP Abbreviations: CP cervical pregnancy, UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, SD standard deviation, BMI body mass index, ART assisted reproductive technology, β-hCG beta-human chorionic gonadotropin Data are presented as mean ± SD for continuous variables and n (%) for categorical variables. p values were calculated using Student’s t-test for continuous variables and Chi-squareor Fisher’s exact test for categorical variables However, markers of baseline disease severity were significantly higher in the UAE (± MTX) + D&C group: elevated preoperative β-hCG levels (25,521 ± 8,299 vs. 4,472 ± 1,891 mIU/mL; p  < 0.001), larger gestational sacs (2.97 ± 1.25 vs. 1.34 ± 0.64 cm; p  < 0.001), and a higher prevalence of fetal cardiac activity (37.2% vs. 20.8%; p  = 0.038). Primary and perioperative outcomes are summarized in Table  2 . The UAE (± MTX) + D&C group achieved faster biochemical resolution with shorter time to undetectable β-hCG (24.6 ± 3.5 vs. 30.7 ± 4.4 days; p  < 0.001) and steeper β-hCG decline by postoperative day 3 (61.7% ± 13.5% vs. 54.1% ± 11.8%; p  = 0.024). Median intraoperative blood loss was substantially lower in the UAE (± MTX) + D&C group (7 mL [IQR 6–10] vs. 47 mL [IQR 35.5–57]; p  < 0.001). The UAE (± MTX) + D&C group required no transfusions, ICU admission, or re-intervention, whereas 8.3% of D&C-alone patients needed transfusion and 16.7% required uterine tamponade. Table 2 Primary Outcome and Key Perioperative Outcomes Outcome UAE (± MTX) + D&C group ( n  = 43) D&C Alone group ( n  = 24) p value Days to undetectable β-hCG (Primary outcome) 24.6 ± 3.5 30.7 ± 4.4 < 0.001* Intraoperative blood loss (mL), Median (IQR) 7 (6–10) 47 (35.5–57) < 0.001 β-hCG decline on postoperative day 3 (%), mean ± SD 61.7 ± 13.5 54.1 ± 11.8 0.024 Blood transfusion required, n (%) 0 (0.0%) 2 (8.3%) 0.165 ICU admission, n (%) 0 (0.0%) 1 (4.2%) 0.358 Retained products of conception, n (%) 1 (2.3%) 3 (12.5%) 0.297 Re-intervention required, n (%) 0 (0.0%) 1 (4.2%) 0.358 Composite treatment success, n (%) Defined as: no transfusion, no ICU admission, no retained products, no re-intervention 42 (97.7%) 19 (79.2%) 0.038 Hospitalization duration (days), mean ± SD 5.72 ± 1.31 5.11 ± 0.98 0.051 Placement of uterine Foley balloon, n (%) 0 (0.0%) 4 (16.7%) 0.058 Total hospitalization cost (CNY), mean ± SD 16,923 ± 909 5,933 ± 718 < 0.001 Abbreviations: UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, β-hCG beta-human chorionic gonadotropin, SD standard deviation, IQR interquartile range, ICU intensive care unit, CNY Chinese Yuan The composite success outcome requires all four criteria to be met simultaneously. Data are presented as mean ± SD, median (IQR), or n (%). The Mann-Whitney U test was used to compare intraoperative blood loss due to the non-normal distribution. *Days to undetectable β-hCG was analyzed using the Kaplan–Meier method; between-group comparison was performed using the log-rank test ( p  < 0.001). Other p values were calculated using Student’s t-test for continuous variables and Chi-squareor Fisher’s exact test for categorical variables. No hysterectomy occurred in either group Primary Outcome and Key Perioperative Outcomes Days to undetectable β-hCG (Primary outcome) Composite treatment success, n (%) Defined as: no transfusion, no ICU admission, no retained products, no re-intervention Abbreviations: UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, β-hCG beta-human chorionic gonadotropin, SD standard deviation, IQR interquartile range, ICU intensive care unit, CNY Chinese Yuan The composite success outcome requires all four criteria to be met simultaneously. Data are presented as mean ± SD, median (IQR), or n (%). The Mann-Whitney U test was used to compare intraoperative blood loss due to the non-normal distribution. *Days to undetectable β-hCG was analyzed using the Kaplan–Meier method; between-group comparison was performed using the log-rank test ( p  < 0.001). Other p values were calculated using Student’s t-test for continuous variables and Chi-squareor Fisher’s exact test for categorical variables. No hysterectomy occurred in either group The composite treatment success rate was significantly higher in the UAE-based therapy group than in the D&C alone group (97.7% vs. 79.2%, p  = 0.038). The individual components of this composite outcome are detailed in Supplementary Table 1. No hysterectomy occurred in either group. Although hospitalization duration was slightly longer (5.72 vs. 5.11 days, p  = 0.051), total hospitalization costs were substantially higher in the UAE (± MTX) + D&C group (CNY [Chinese Yuan] 16,923 ± 909 vs. 5,933 ± 718, p  < 0.001). Kaplan–Meier curves were constructed to compare time to undetectable serum β-hCG (< 5 mIU/mL) between the UAE (± MTX) + D&C and D&C-alone groups (Fig.  2 ). The unadjusted hazard ratio (HR) for achieving β-hCG normalization in the UAE-based therapy group was 4.63 (95% CI: 2.60–8.26; p  < 0.001). The log-rank test p -value is < 0.001, indicating a statistically significant difference in favor of the UAE (± MTX) + D&C group. Fig. 2 Kaplan-Meier estimates of the proportion of patients with persistently detectable β-hCG over time by treatment group Kaplan-Meier estimates of the proportion of patients with persistently detectable β-hCG over time by treatment group Multivariable Cox regression analysis confirmed that UAE-based therapy was independently associated with faster β-hCG decline (aHR = 4.67, 95% CI: 1.63–13.51, p  = 0.004), after adjusting for baseline prognostic factors (log₁₀-transformed preoperative β-hCG, gestational age, fetal cardiac activity, and sac diameter) (Table  3 ). Table 3 Multivariable Cox proportional hazards regression analysis for time to undetectable β-hCG, adjusted for baseline disease severity Variable Adjusted hazard ratio (aHR) 95% confidence interval (CI) p value Treatment group (UAE ± MTX + D&C vs. D&C alone) 4.67 1.63–13.51 0.004 Fetal cardiac activity present (yes vs. no) 1.06 0.62–1.81 0.84 Gestational age at diagnosis (per 1-day increase) 1.01 0.97–1.06 0.54 Gestational sac diameter (per 1-cm increase) 0.92 0.63–1.34 0.67 Preoperative β-hCG (per 10-fold increase) a 0.97 0.36–2.65 0.96 Abbreviations: aHR adjusted hazard ratio, CI confidence interval, β-hCG beta-human chorionic gonadotropin, UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage a Preoperative β-hCG was log₁₀-transformed. Adjusted for log₁₀-transformed preoperative β-hCG level, gestational age, presence of fetal cardiac activity, and sac diameter. Reference group for treatment: D&C alone. All variables were retained in the final model using the enter method Multivariable Cox proportional hazards regression analysis for time to undetectable β-hCG, adjusted for baseline disease severity Abbreviations: aHR adjusted hazard ratio, CI confidence interval, β-hCG beta-human chorionic gonadotropin, UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage a Preoperative β-hCG was log₁₀-transformed. Adjusted for log₁₀-transformed preoperative β-hCG level, gestational age, presence of fetal cardiac activity, and sac diameter. Reference group for treatment: D&C alone. All variables were retained in the final model using the enter method In a parsimonious Cox model adjusting only for log₁₀-transformed preoperative β-hCG, UAE-based therapy remained significantly associated with faster β-hCG normalization (aHR = 4.36, 95% CI: 1.90–10.00, p  < 0.001), confirming the robustness of the treatment effect under reduced model burden (Supplementary Table 2). After IPTW, all baseline covariates achieved adequate balance (all SMDs < 0.1; Supplementary Table 3). In the weighted Cox model, UAE-based therapy was associated with significantly accelerated β-hCG decline (weighted HR = 4.15, 95% CI: 1.38–12.49, p  = 0.011), consistent with the primary multivariable analysis (Supplementary Table 2). Short-term complications and mid-term reproductive outcomes are shown in Table  4 . Within 7 days postoperatively, hypogastric pain—consistent with post-embolization syndrome—was significantly more common in the UAE (± MTX) + D&C group than in the D&C alone group (62.8% vs. 12.5%, p   0.05). Table 4 Short-Term Complications and Mid-Term Reproductive Outcomes UAE (± MTX) + D&C group ( n  = 43) D&C Alone group ( n  = 24) p value Postoperative Complications (within 7 days) Hypogastric pain (post-embolization syndrome), n (%) 27 (62.8%) 3 (12.5%)  38 °C), n (%) 3 (7.0%) 2 (8.3%) 1.000 Gastrointestinal reaction, n (%) 4 (9.3%) 3 (12.5%) 1.000 Hypomenorrhea at first follow-up, n (%) 5 (11.6%) 1 (4.2%) 0.625 Reproductive Recovery (at 1–3 months follow-up) Days to resumption of normal menses, Median (IQR) 46 (32–63) 35 (31–44) 0.002 Days to complete regression of cervical mass, Median (IQR) 27 (22–37) 33 (29–48) 0.009 Intrauterine adhesions (diagnosed by ultrasound/hysteroscopy), n (%) 5 (11.6%) 1 (4.2%) 0.625 Abbreviations: UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, IQR interquartile range Data are presented as median (IQR) for continuous variables and n (%) for categorical variables. p values were calculated using the Mann-Whitney U test for continuous variables and Chi-squareor Fisher’s exact test for categorical variables Short-Term Complications and Mid-Term Reproductive Outcomes Abbreviations: UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, IQR interquartile range Data are presented as median (IQR) for continuous variables and n (%) for categorical variables. p values were calculated using the Mann-Whitney U test for continuous variables and Chi-squareor Fisher’s exact test for categorical variables Menstrual recovery was significantly delayed in the UAE (± MTX) + D&C group, with a mean time to resumption of normal menses of 46 (32–63) days compared to 35 (31–44) days in the D&C group ( p  = 0.002). Despite this delay, the incidence of hypomenorrhea at first follow-up (11.6% vs. 4.2%, p  = 0.625) and intrauterine adhesions (11.6% vs. 4.2%, p  = 0.625) did not differ significantly between groups. Notably, the cervical mass resolved more rapidly in the UAE (± MTX) + D&C group (27 [22–37] days vs. 33 [29–48] days, p  = 0.009). Among the 43 patients treated with UAE-based regimens, in a subgroup analysis, those who additionally received MTX (UAE + MTX, n  = 18) had significantly higher preoperative β-hCG levels than those receiving UAE alone ( n  = 25) (30,200 ± 8,500 vs. 22,300 ± 7,100 mIU/mL, p  = 0.002), reflecting more advanced disease (Table  5 ). Despite this higher baseline burden, the UAE + MTX group achieved faster resolution of pregnancy, with a shorter time to undetectable β-hCG (22.0 ± 3.0 vs. 26.5 ± 3.3 days, p  < 0.001), greater β-hCG decline on postoperative day 3 (69.6% ± 11.0% vs. 56.0% ± 12.0%, p  < 0.001), and more rapid regression of the cervical mass (26 [21–34] vs. 30 [25–40] days, p  = 0.067). There were no significant differences in intraoperative blood loss, post-embolization syndrome, or time to menstrual recovery. In subgroup analysis, composite treatment success was achieved in all 18 patients (100%) receiving UAE + MTX + D&C and in 24 of 25 patients (96.0%) receiving UAE + D&C alone ( p  = 0.419). Given the small subgroup sizes, this analysis should be considered exploratory. Table 5 Exploratory subgroup analysis of Outcomes by Adjunctive MTX Use Among Patients Treated with UAE-Based Regimens (Not powered for definitive conclusions) Outcome UAE alone group ( n  = 25) UAE + MTX group ( n  = 18) p value Fetal cardiac activity present, n (%) 8 (32.0%) 8 (44.4%) 0.402 Pre-operative β-hCG (mIU/mL), mean ± SD 22,300 ± 7,100 30,200 ± 8,500 0.002 Days to undetectable β-hCG, mean ± SD 26.5 ± 3.3 22.0 ± 3.0 < 0.001 β-hCG decline on postoperative day 3 (%), mean ± SD 56.0 ± 12.0 69.6 ± 11.0 < 0.001 Intraoperative blood loss (mL), Median (IQR) 8 (6–11) 7.5 (3.5–9.5) 0.377 Hypogastric pain (post-embolization syndrome), n (%) 16 (64.0%) 11 (61.1%) 0.842 Days to resumption of normal menses, Median (IQR) 48 (36–68) 39 (28.5–58.7) 0.203 Days to complete regression of cervical mass, Median (IQR) 30 (25–40) 26 (21–34) 0.067 Composite treatment success*, n (%) 24(96.0%) 18(100.0%) 0.419 Abbreviations: UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, β-hCG beta-human chorionic gonadotropin, SD standard deviation, IQR interquartile range This analysis is exploratory and not powered for definitive conclusions. Data are presented as mean ± SD, median (IQR), or n (%). p values were calculated using Student’s t-test for normally distributed continuous variables, the Mann-Whitney U test for non-normally distributed continuous variables (Days to resumption of normal menses, Days to complete regression of cervical mass, and Intraoperative blood loss), and Chi-squareor Fisher’s exact test for categorical variables. *Composite treatment success was compared using Fisher’s exact test. Group comparisons for time to undetectable β-hCG were assessed by log-rank test across subgroups Exploratory subgroup analysis of Outcomes by Adjunctive MTX Use Among Patients Treated with UAE-Based Regimens (Not powered for definitive conclusions) Abbreviations: UAE uterine artery embolization, MTX methotrexate, D&C dilation and curettage, β-hCG beta-human chorionic gonadotropin, SD standard deviation, IQR interquartile range This analysis is exploratory and not powered for definitive conclusions. Data are presented as mean ± SD, median (IQR), or n (%). p values were calculated using Student’s t-test for normally distributed continuous variables, the Mann-Whitney U test for non-normally distributed continuous variables (Days to resumption of normal menses, Days to complete regression of cervical mass, and Intraoperative blood loss), and Chi-squareor Fisher’s exact test for categorical variables. *Composite treatment success was compared using Fisher’s exact test. Group comparisons for time to undetectable β-hCG were assessed by log-rank test across subgroups

This retrospective cohort study adhered to the Declaration of Helsinki guidelines and received approval from the Ethics Committee at Fujian Maternity and Child Health Hospital (Approval No. 2026KY007). The requirement for informed consent was waived due to the retrospective design and use of anonymized data. We enrolled women with first-trimester CP (≤ 12 weeks’ gestation) diagnosed between September 2020 and June 2025. CP diagnosis required: (1) clinical presentation consistent with ectopic pregnancy; (2) imaging confirmation via transvaginal ultrasound or MRI; and (3) postoperative histopathological identification of chorionic villi. Inclusion criteria were as follows: (1) amenorrhea with clinical indicators consistent with CP; (2) confirmation via imaging modalities and serum β-hCG levels; (3) histopathological confirmation postoperatively. Exclusion criteria included severe comorbidities and incomplete data. Clinical data were extracted from electronic medical records, supplemented by telephone follow-up when necessary. Two trained researchers independently performed data collection and cross-verification, with discrepancies resolved by consensus or a third reviewer. Treatment strategy (UAE with or without MTX followed by D&C versus D&C alone) was not assigned by randomization or fixed protocol. Instead, it was determined through shared decision-making involving the patient, gynecologist, and interventional radiologist, taking into account baseline disease severity (e.g., β-hCG level, fetal cardiac activity), patient preference, and institutional resources. This non-randomized allocation introduces potential confounding by indication, which we addressed through multivariable adjustment in the primary analysis. The final follow-up was completed in September 2025. The exposure variable was the initial treatment strategy, classified as UAE (± MTX) followed by D&C versus D&C alone, which was extracted from procedural records and coded as a binary variable. All women underwent serial monitoring of serum β-hCG levels at admission, on postoperative day 3 after D&C, and weekly thereafter until levels fell below the non‑pregnant threshold (< 5 mIU/mL). Concurrently, transvaginal ultrasound was performed weekly to assess the cervical mass until complete resolution was documented. Outcome variables included the primary endpoint—time to undetectable serum β-hCG—as well as secondary outcomes such as intraoperative blood loss, composite treatment success, time to resumption of menses, and incidence of intrauterine adhesions. Serum β-hCG was quantified using a chemiluminescent immunoassay (CLIA; Roche Diagnostics) in the hospital laboratory. Time to undetectable serum β-hCG was pre-specified as the primary endpoint because it provides an objective, quantitative, and universally accepted measure of complete trophoblastic resolution in ectopic pregnancy. While hard clinical outcomes such as hemorrhage and hysterectomy are critically important, their expected rarity in modern conservative management limits statistical power. Composite treatment success was defined as the absence of all the following events: blood transfusion, ICU admission, retained products of conception, and need for any re‑intervention. Each component was considered equally important because all represent clinically significant failures of primary therapy requiring escalation of care (transfusion, ICU admission, or re-intervention) or incomplete resolution of ectopic tissue (retained products), all of which compromise the goal of safe, single-stage, fertility-sparing treatment. Placement of a uterine Foley balloon for hemostasis was considered supportive care, not re-intervention. Covariates considered for adjustment included age, body mass index (BMI), gravidity, parity, history of miscarriage or cesarean delivery, use of assisted reproductive technology (ART), gestational age at diagnosis, presence of fetal cardiac activity, preoperative β‑hCG level, and gestational sac diameter. All covariates were assessed within 24 h of admission. Missing data accounted for less than 5% of cases (only one missing value for menstrual recovery time). Given the low rate of missingness, a complete‑case analysis was performed, and multiple imputation was not applied. The UAE procedure was conducted by an interventional radiologist under local anesthesia. Selective catheterization of the uterine arteries was performed, followed by embolization using gelatin sponge particles (560–710 μm) until near‑stasis of flow was achieved. For patients with elevated preoperative β‑hCG levels—suggesting more active trophoblastic disease—1 mg/kg MTX was administered intra‑arterially prior to embolization. D&C was carried out by a specialist gynecologist one day after UAE in the combined treatment group, or as the standalone procedure in the D&C‑alone group. The operative process was meticulously documented, intraoperative blood loss was quantified using standardized gravimetric methods, and all extracted tissue was sent for histopathological examination. Normality was assessed using the Shapiro-Wilk test and Q-Q plots. Normally distributed continuous variables are presented as mean ± standard deviation (SD) and compared using Student’s t-test, while non-normally distributed variables are reported as median (interquartile range [IQR]) and compared using the Mann-Whitney U test. Categorical variables were presented as numbers (percentages) and compared using the Chi-square test or Fisher’s exact test for categorical variables. Days to undetectable β-hCG (defined as < 5 mIU/mL) was analyzed as a time-to-event outcome. Kaplan–Meier curves were constructed to estimate the probability of achieving undetectable β-hCG over time, and differences between the two treatment groups were compared using the log-rank test. To address potential confounding by indication arising from baseline imbalances, a multivariable Cox proportional hazards model was fitted to evaluate the independent association between treatment strategy and time to undetectable β-hCG. The model was adjusted for preoperative β-hCG (log₁₀-transformed), gestational age at diagnosis, presence of fetal cardiac activity, and gestational sac diameter, based on their clinical relevance and association with treatment allocation. All covariates were entered into the final model using the enter method. Adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) were reported. We conducted a series of sensitivity analyses. First, a parsimonious Cox model was fitted adjusting only for log 10 -transformed preoperative β-hCG—the single strongest confounder between treatment groups. Second, to provide a more robust approach to confounding adjustment, we applied inverse probability of treatment weighting (IPTW). A propensity score was estimated using logistic regression with covariates including log 10 -transformed preoperative β-hCG, gestational age at diagnosis, presence of fetal cardiac activity, and gestational sac diameter. Stabilized weights were calculated. The distribution of weights was inspected, and extreme values were truncated at the 1st and 99th percentiles to improve precision. Covariate balance after weighting was assessed using standardized mean differences (SMDs), with a threshold of < 0.1 indicating adequate balance. A weighted Cox proportional hazards model with a robust (sandwich) variance estimator was then fitted to estimate the treatment effect on time to undetectable β-hCG. These sensitivity analyses were designed to evaluate the robustness of the primary findings under reduced model complexity and alternative confounding-adjustment strategies. A two-sided p-value < 0.05 was considered statistically significant. All primary analyses were performed using SPSS version 27.0 (IBM Corp., Armonk, NY, USA). Sensitivity analyses including IPTW were conducted using R version 4.5.1 (R Foundation for Statistical Computing, Vienna, Austria) with the ‘survey’, ‘survival’, and ‘cobalt’ packages.

In this retrospective cohort study of women with CP, UAE-based therapy (with or without adjunctive MTX) followed by D&C was associated with significantly faster biochemical resolution, markedly reduced intraoperative blood loss, and higher composite treatment success rates compared to D&C alone—even among patients with more advanced disease. While UAE was associated with transient post-embolization syndrome and delayed menstrual recovery, surrogate reproductive outcomes—including menstrual recovery and intrauterine adhesion rates—were comparable between groups; however, long-term fertility was not directly assessed. Prospective, multicenter cohort studies or registries are urgently needed to validate these findings and reduce selection bias. Future research should prioritize long-term follow-up to evaluate reproductive outcomes and obstetric safety after fertility-sparing interventions.

Our study demonstrates that combined UAE (± MTX) + D&C was associated with faster biochemical resolution and improved procedural safety compared with D&C alone in first-trimester CP. Despite higher baseline disease burden, the UAE (± MTX) + D&C group achieved faster β-hCG decline, lower intraoperative blood loss (7 [6–10] vs. 47 [35.5–57] mL), and higher composite success rates. Importantly, after adjusting for key baseline prognostic factors—preoperative β-hCG, gestational age, fetal cardiac activity, and gestational sac diameter—UAE-based therapy remained independently associated with accelerated β-hCG normalization (aHR = 4.67, 95% CI: 1.63–13.51, p  = 0.004). This finding suggests that the observed benefit of UAE may not be attributable solely to selection bias and may reflect a therapeutic effect in mitigating trophoblastic viability and promoting resolution; however, given the non-randomized design, residual confounding cannot be excluded. These advantages were offset by a transient delay in menstrual recovery and substantially higher hospitalization costs. Our findings are consistent with and extend prior literature on the use of UAE in CP management. The pronounced reduction in intraoperative blood loss (7 [6–10] vs. 47 [35.5–57] mL) supports the core principle of UAE: preemptive occlusion of the cervical blood supply—mainly derived from the descending branches of the uterine arteries—thereby mitigating the risk of catastrophic hemorrhage during surgical evacuation [ 15 ]. Multiple studies demonstrate that UAE reduces intraoperative bleeding during curettage, consistent with our reduction in blood loss [ 17 , 18 ]. Moreover, the accelerated β-hCG decline and earlier anatomical resolution of the cervical mass provide quantitative evidence that devascularization promotes trophoblast necrosis [ 19 ]. Our study uniquely incorporates a composite treatment success reflecting multiple adverse events (transfusion, ICU admission, retained products, re-intervention). This patient-centered metric provides a more comprehensive assessment of treatment effectiveness than traditional isolated endpoints. The clinical benefits of this combined approach are further supported by multiple studies. Patients treated with UAE plus intra-arterial MTX infusion showed significantly better outcomes compared to those receiving systemic MTX followed by curettage. These benefits included a faster regression of serum β-hCG levels, a return to normal menstruation, and reduced hospital stays [ 20 ]. One study reported a 96.8% success rate for the UAE + MTX group, along with the shortest duration of vaginal bleeding and fastest normalization of β-hCG levels [ 21 ]. Notably, similar favorable outcomes have been reported in cesarean scar pregnancy (CSP), a condition with shared pathophysiology [ 22 ]. This cross-condition consistency reinforces the biological plausibility of our results. UAE induces rapid ischemia-mediated trophoblast necrosis before surgical intervention. Among patients with elevated β-hCG or fetal cardiac activity, adjunctive intra-arterial MTX enhanced biochemical and anatomical resolution ( p  < 0.001), suggesting a synergistic effect. MTX exerts direct chemotoxic effects on proliferating trophoblasts, while UAE enhances this effect by prolonging drug retention and achieving localized high concentrations [ 23 ]. The delayed menstrual recovery in the UAE (± MTX) + D&C group (46 vs. 35 days) reflects transient endometrial ischemia. However, this did not increase the risk of permanent intrauterine damage, as evidenced by similar intrauterine adhesion rates between groups (11.6% vs. 4.2%; p  = 0.625). Concerns about endometrial damage were not supported by long-term data. In cesarean scar pregnancy cohorts, live birth rates were comparable between patients with reduced menstrual flow post-UAE and those with normal menstruation (63% vs. 38%; p  = 0.191) [ 24 ]. It should be noted that menstrual recovery and intrauterine adhesion rates are surrogate markers and do not directly measure fertility outcomes. Large-scale studies directly assessing long-term fertility after CP treatment remain limited; indirect support comes from related applications of UAE. For example, a series of 23 CSP patients seeking pregnancy showed that 19 (82.6%) achieved successful pregnancies, with 15 (78.9%) delivering healthy infants after UAE [ 25 ]. Reproductive outcomes after UAE for uterine fibroids and adenomyosis also demonstrate trends consistent with the general population of similar age groups [ 26 , 27 ]. The subgroup analysis revealed a composite treatment success rate of 100% in the UAE + MTX group compared with 96.0% in the UAE alone group. Although this difference did not reach statistical significance ( p  = 0.419)—likely due to limited power from small subgroup sizes—it may suggest a potential clinical benefit of adjunctive MTX, particularly considering that the UAE + MTX group tended to have higher baseline disease burden. Although the UAE + MTX subgroup showed numerically better outcomes, the small sample size ( n  = 18) limits statistical power. This finding is hypothesis-generating and requires validation in larger prospective studies. Importantly, our results support the emerging conceptual role of UAE as a “gateway” to conservative surgery, rather than a standalone therapy. This strategy aligns with recent findings in fibroid management, where preoperative UAE facilitated uterus-preserving surgery by minimizing vascularity [ 28 ]. Similarly, in CP, our findings suggest that UAE may serve as a bridging intervention that converts a high-risk, potentially hemorrhagic procedure into a controlled, fertility-sparing operation. (1) the largest comparative sample for CP; (2) multidimensional outcome assessment incorporating efficacy, safety, reproductive recovery, and economic burden; and (3) a novel patient-centered composite treatment success. The use of a pre-specified composite treatment success—reflecting a “no adverse event” clinical ideal—is a novel and patient-centered contribution to the literature on CP. Although the absolute difference in time to β-hCG normalization was approximately 6 days (24.6 vs. 30.7 days), this reduction may be clinically relevant, potentially translating into reduced follow-up burden, decreased patient anxiety associated with persistent trophoblastic activity, and earlier psychological closure—factors that are integral to patient-centered care in high-risk ectopic pregnancies. (1) High-risk patients (detectable fetal cardiac activity, with high β-hCG levels, or a gestational sac > 3 cm): UAE with or without MTX followed by D&C is recommended. (2) Low-risk patients (absence of above features): D&C alone may be considered only in centers with immediate access to interventional radiology and blood products. Multidisciplinary collaboration among gynecologic, interventional radiology, and ultrasound teams is essential for timely risk stratification and management decisions. (1) The retrospective, non-randomized design introduces an inherent risk of confounding by indication. While propensity score matching was not pursued due to potential sample loss, we employed IPTW to leverage the full cohort while balancing baseline covariates. Additionally, a series of sensitivity analyses—including a parsimonious Cox model adjusting only for the single strongest confounder (log₁₀ β-hCG)—confirmed that the treatment effect remained significant and directionally consistent across all modeling approaches (Supplementary Table 2). Nevertheless, the wide confidence intervals in the full multivariable model reflect the inherent imprecision associated with 67 events and 5 parameters, and the findings require validation in larger cohorts. Consequently, our findings should be interpreted as reflecting associations rather than causal effects. The observed benefit, while robust to adjustment, should be interpreted cautiously. (2) Although 67 patients represent the largest CP comparison cohort to date, sample size limits statistical power for detecting subtle differences in subgroup analyses. (3) The 1–3 month follow-up period precludes evaluation of long-term fertility outcomes (subsequent pregnancy rates, live birth rates, obstetric outcomes). Future prospective studies with extended follow-up are needed. (4) Single-center data may have limited generalizability regarding cost and procedural protocols. Multicenter prospective studies are needed to validate these findings across diverse healthcare settings. Furthermore, while our composite endpoint captures short-term procedural success, it does not account for patient-reported outcomes such as pain or quality of life—domains increasingly recognized as essential in fertility-sparing interventions.

Cervical pregnancy (CP), a rare form of ectopic gestation with a reported incidence of 1 in 2,500–18,000 pregnancies, poses substantial maternal risks [ 1 – 3 ]. The cervix, composed primarily of fibrous tissue, lacks the smooth muscle contractility necessary to control hemorrhage. This anatomical feature renders pregnancy interruption prone to severe, refractory hemorrhage. This situation is life-threatening and may necessitate emergency hysterectomy, resulting in permanent loss of fertility. Such an outcome is undeniably catastrophic for women of childbearing age [ 4 – 7 ]. The preservation of reproductive function while ensuring maternal safety remains the central therapeutic challenge. CP management varies considerably among institutions, reflecting limited high-quality evidence. Treatment options range from expectant management to pharmacological therapy (systemic methotrexate [MTX]) to surgical approaches (dilation and curettage [D&C], uterine artery embolization [UAE], or hysterectomy) [ 4 , 8 , 9 ]. Notably, major professional guidelines—including those from the Royal College of Obstetricians and Gynaecologists (RCOG) and the American College of Obstetricians and Gynecologists (ACOG)—do not provide strong recommendations due to the scarcity of high-quality comparative evidence [ 10 , 11 ]. Consequently, clinicians must navigate this complexity, often relying on limited case series or institutional protocols rather than robust comparative evidence. In conservative treatment approaches, both pharmacological therapy and surgical intervention have their limitations. MTX, a folate antagonist, acts by inhibiting dihydrofolate reductase, thereby impeding DNA synthesis in trophoblastic tissue. This makes it a classic medication for treating ectopic pregnancies [ 12 ]. However, the efficacy of MTX when administered systemically or through local injection may be influenced by various factors, including gestational age, serum β-human chorionic gonadotropin (β-hCG) levels, and the patient’s sensitivity, leading to potential treatment failure and subsequent bleeding [ 13 , 14 ]. On the other hand, surgical procedures such as curettage carry a significant risk of fatal hemorrhage due to the rich vascularization of the cervix and the lack of contractility in the myometrial layer. UAE, frequently combined with D&C, has emerged as a promising fertility-sparing approach by preemptively reducing vascularity in the highly perfused cervical region [ 6 , 15 , 16 ]. While several retrospective studies suggest UAE may decrease intraoperative bleeding and hysterectomy rates, critical questions remain unanswered: Does UAE truly accelerate biochemical resolution? Does it improve composite treatment success without compromising short- or mid-term reproductive recovery? And importantly, does adjunctive MTX confer additional benefit in more active disease? No study has comprehensively compared UAE-based therapy with D&C alone across efficacy, safety, reproductive outcomes, and cost. We conducted a retrospective cohort study comparing UAE (± MTX) + D&C with D&C alone in first-trimester CP across multiple dimensions of outcome. We evaluated efficacy, safety, composite treatment success, reproductive recovery, and economic burden, thereby generating comprehensive evidence to guide clinical practice and future research. We hypothesized that UAE-based therapy would lead to faster biochemical resolution, higher composite success rates, and acceptable reproductive recovery compared to D&C alone.

Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3.