PREDICTing Post-Embolization Syndrome after uterine fibroid embolization: the PREDICT-PES study.

Uterine artery embolization (UAE) is a safe, effective, and durable treatment for women with symptomatic fibroids and is sometimes referred to as uterine fibroid embolization (UFE). 1 , 2 One of the side effects of the procedure is that a large proportion of women suffer from a phenomenon known as post-embolization syndrome (PES). The syndrome is characterized by a combination of the following: pelvic pain, nausea, low-grade fever, and malaise. 3 Biochemistry will reveal a leucocytosis and elevated C-reactive protein. 4 Symptoms can last for up to 2 weeks, but PES is usually only significant in the first 12 h after the embolization procedure. 3 While not completely understood, it is felt that the syndrome results from ischaemia-mediated release of pro-inflammatory cytokines including interlukin-6 and tumour necrosis factor-alpha. 3 , 5 There are few studies that have shown exactly which patients are at higher risk of suffering a more significant form of PES, herein referred to as clinically significant PES or csPES, by measuring post-operative pain. In their 2009 study, Ruuskanen et al. assessed 62 patients with significant pain after UAE and correlated this with technical factors and MRI imaging characteristics. They showed that a larger volume of embolic agent and more ischaemia on MRI correlated with pain. 4 It has been thus proposed that if ischaemia causes pain, then larger fibroids and/or smaller particles are more likely to result in csPES. 3 , 5–7 However, some studies have refuted this, suggesting the rate of significant pain in women with large fibroids is not increased. 8 We remain in a position where pre-procedure counselling of the patients’ risk of csPES is difficult to predict, resulting in a more generalized warning to all patients regarding the likelihood of pain. Unfortunately, it is not currently possible to prevent PES from occurring and thus management of PES is broadly supportive, not preventative. The exact choice of pain medications will vary according to the operator, but a combination of treatments has been shown to be effective. 9 , 10 Paracetamol, non-steroidal anti-inflammatory drugs (NSAIDs), and short-term opiate analgesics are effective. 9 , 11 , 12 NSAIDS have been shown to be particularly important in mediating the symptoms due to the hypothesized inflammatory-mediated activation of PES. 11 , 12 Some additional but not yet standardized adjuncts have also been devised to help reduce the burden of csPES, and include the use of glucocorticoids, muscle relaxants, superior hypogastric nerve blockade, and intra-arterial lidocaine. 10 , 13–15 Within 12-24 h of the procedure, symptoms will reduce temporally and be low enough that the patient is suitable for discharge home. In this study, we aimed to identify the rate of csPES in our cohort and identify risk factors that were associated with the development of csPES. We hope this will lead to an improved consent process through better pre-operative counselling.

This retrospective case-control study was performed according to the Strengthening The Reporting of OBservational studies in Epidemiology (STROBE) checklist and approved by our Human Research and Ethics Committee (HREC). The Alfred hospital is a quaternary referral hospital in Melbourne, Australia. Our cohort included patients who underwent UAE for the purpose of treating symptomatic uterine fibroids between the 18-month study dates of March 1, 2022 and September 1, 2023. The study dates corresponded with the introduction of an updated protocol for premedication, detailed below. All patients were treated after pre-operative assessment in a dedicated interventional radiology outpatient clinic. This included history taking, physical examination, pre-operative imaging, and biochemistry. Patients were treated on a morning operating list, and all patients remained admitted for an overnight hospital stay, with discharge approximately 24-h post-procedure. Prior to the procedure, all patients received the following pre-medications: Analgesic: Paracetamol 1 g intravenous infusion Parecoxib 40 mg intravenous Morphine 5 mg intravenous Antiemetic: Ondansetron 8 mg intravenous Droperidol 0.625 mg intravenous Prophylactic antibiotic: Cefazolin 2 g intravenous Metronidazole 500 mg intravenous infusion Procedures were performed by 1 of 6 fellowship-trained interventional radiologists, or an advanced training fellow under direct supervision. All procedures were performed under conscious sedation using titrated 1 mg boluses of midazolam, via transfemoral access using a standard technique as previously documented. 16 Our embolic agent was polyvinyl alcohol (PVA) particles (Merit Medical Bearing nsPVA or Boston Scientific Contour PVA). The choice of particle size was according to operator preference and embolization endpoint was diminished forward flow. After the procedure, all patients were given a patient-controlled analgesia (PCA) with each button press delivering a 1 mg morphine bolus with a 5-min lockout, alongside 6-hourly 1 g oral paracetamol. At the same time, patients have available additional antiemetic medications pro re nata (PRN), including metoclopramide 10 mg, ondansetron 4 mg, dexamethasone 4 mg, and droperidol 0.625 mg. Patients are nursed in a dedicated short-stay ward familiar with the management of patients after the UAE procedure. During an afternoon interventional radiology ward round, patients are assessed and typically suitable for conversion from PCA to oral opiate analgesic which are given PRN, either oxycodone 5 mg 2-hourly or tapentadol 50 mg 4-hourly. On discharge, patients are advised to take regular NSAID and paracetamol for 7 days, or until their symptoms abate. This PES protocol was developed in conjunction with a senior pain specialist, and the hospital acute pain service was available for consultation in the event of any concern. UAE for symptomatic fibroids. Age over 18 years. UAE performed for an indication other than fibroids. Data were collected from a combination of the electronic medical record (EMR) and the Radiology Information System. The authors acknowledge that almost all patients have some form of PES after the procedure, usually “mild PES”. As such, the definition of csPES was made to distinguish those who required additional management outside routine care. In our cohort, csPES was defined as maximum pain on visual analogue scale (VAS) of greater than 5 out of 10 , plus at least 1 of the following 3 criteria: Morphine PCA total dose greater than or equal to 10 mg . This did not include the bolus 5 mg dose given as part of the controlled premedications. Post-operative fever , defined as temperature greater than 37.8 degree Celsius. The use of 2 or more antiemetic agents in the post-operative period. This did not include the agents given as part of the controlled premedications. Maximum pain was measured on a Likert VAS and reported between a value of 0 and 10. This value was documented in the EMR at regular intervals after the procedure. The highest documented post-operative VAS score was recorded for this study. Particle size was determined to be the smallest size of PVA particles used in the treatment. For example, if the physician began with 355-500 micrometre particles, but also gave 500-710 micrometre particles towards the end of the procedure, the smallest size was documented given smaller particle size has traditionally been associated with post-operative pain. 5–7 Fibroid number was recorded as the number of fibroids on pre-procedure imaging. This was documented on an ordinal scale of 1 to 6, with the final group consisting of those with 6 or more fibroids. Patients were also included as “6 or more” when the imaging reported that the uterus was multifibroid and there were too many to individually measure. Prior pelvic surgery included any previous open, laparoscopic, or hysteroscopic surgery to the pelvic region. Long-term analgesic use was defined as any patient who was taking regular opiate analgesic at the time of initial consultation for the management of any form of chronic pain. This pain was not limited to the pelvic region. Statistical analysis was performed using Stata (Version 17.0-BE, StataCorp, Texas, USA). Data were assessed for normalization and according to the data type presented as number (percentage), mean (standard deviation or SD), or median (range). Summary statistics were compared using student’s t -test or Mann-Whitney U test. Covariates were controlled for confounding by multivariate logistic regression using the binary outcome of PES. A 2-tailed probability value of less than 0.05 was deemed significant.

A total of 79 patients were treated with UAE during the study period. After applying exclusion criteria to 10 patients due to underlying adenomyosis or endometriosis as the primary pathology, 69 patients were enrolled in the study. The rate of csPES in our cohort was 47.8% (33 patients). Enrolled patients were of mean age 46.2 years (SD 5.6). The overall median uterine volume was 393 mL (range 80-2288), and most women had a single fibroid (22, 31.9%) and fibroid disease that was mostly intramural in location (37, 53.6%). The remaining summary statistics are shown in Table 1 , including the entire list of covariates. Summary statistics of the cohort including covariates and univariate analysis. 30: 8 (13.3%) 30: 5 (17.4%) 30: 3 (9.7%) 1: 22 (31.9%) 2: 8 (11.6%) 3: 12 (17.4%) 4: 3 (4.4%) 5: 9 (13.0%) 6 or more: 15 (21.8%) 1: 8 (24.2%) 2: 4 (12.1%) 3: 7 (21.2%) 4: 2 (6.1%) 5: 5 (15.1%) 6 or more: 7 (21.2%) 1: 14 (38.9%) 2: 4 (11.1%) 3: 5 (13.9%) 4: 1 (2.8%) 5: 4 (11.1%) 6 or more: 8 (22.2%) Submucosal: 5 (7.3%) Intramural: 37 (53.6%) Subserosal: 7 (10.1%) Combination: 20 (29.0%) Submucosal: 3 (9.1%) Intramural: 15 (45.4%) Subserosal: 3 (9.1%) Combination: 12 (36.4%) Submucosal: 2 (5.6%) Intramural: 22 (61.1%) Subserosal: 4 (1.1%) Combination: 8 (22.2%) 355-500: 16 (23.9%) 500-710: 53 (76.8%) 355-500: 7 (21.2%) 500-710: 26 (78.9%) 355-500: 9 (25.0%) 500-710: 27 (75.0%) Student’s t -test or Mann-Whitney U test, according to the type of data. Abbreviations: PES = post embolization syndrome, csPES = clinically significant post embolization syndrome. Due to the definition of csPES in our study, the group with csPES had both a higher median pain score (8 vs 4.5, P  < .001) and a higher mean dose of morphine PCA (23.2 mg vs 10.3 mg, P  < .001). These are shown in more detail in Table 2 and Figures 1 and 2 . Box plot comparing the range of maximal pain scores for patients on a Likert visual analogue scale between the groups. The group with csPES had a higher median pain score. Abbreviations: PES = post-embolization syndrome, csPES = clinically significant PES. Box plot comparing the range of total morphine PCA dose between the groups. The group with csPES had a higher mean morphine dose. Abbreviations: PES = post-embolization syndrome, csPES = clinically significant PES. Summary of parameters used to define clinically significant post-embolization syndrome. Student’s t -test or Mann-Whitney U test, according to the type of data. On Likert visual analogue scale between 0 and 10. Abbreviations: PES = post embolization syndrome, csPES = clinically significant post embolization syndrome. On univariate analysis, the group with csPES had a higher median uterine volume (519 mL vs 284 mL), P  = .019 ( Figure 3 ). In addition, the group with csPES had a higher proportion of nulliparity (66.7% vs 38.9%), P  = .021. The remaining covariates were not significantly different on univariate analysis. Box plot comparing the overall uterine volume between the groups. The csPES group has a higher median uterine volume on univariate analysis. Abbreviations: PES = post-embolization syndrome, csPES = clinically significant PES. After adjusting for confounding using multiparametric logistic regression ( Table 3 ), uterine volume was no longer associated with the development of csPES (OR: 1.00, 95% CI: 0.999-1.002, P  = .770). However, a negative trend was shown between increasing age and the development of csPES (OR: 0.87, 95% CI: 0.748-1.002, P  = .054). In addition, the association between nulliparity and csPES remained significant after logistic regression (OR: 5.51, 95% CI: 1.297-23.410, P  = .021). Further patient details of the nulliparous cohort are shown in Supplement S1 . Multiparametric logistic regression analysis using the binary outcome of clinically significant post-embolization syndrome (csPES), and including a range of studied covariates. Excluded by the statistical model due to small sample.

In this study, csPES occurred in 47.8% of our cohort, and nulliparity was strongly associated with the development of csPES. We can use this information to better counsel our patients regarding the odds of csPES when these risks are present, leading to an improved consent process, and the ability to plan more targeted pain management for women who are known to be nulliparous.

This study showed that the incidence rate of csPES in our cohort was 47.8%. 2 The results also show that nulliparity was significantly associated with the development of csPES, with patients at 5.5 times increased odds compared to patients who are primiparous or multiparous ( P  = .021). This risk factor has not previously been shown to be associated with csPES. In addition, after adjusting for confounding, a trend was seen between increasing age and a reduced likelihood of csPES ( P  = .054), where for every year increase in age, there was 13% reduced odds of developing csPES. Previous studies have suggested that the volume of infarcted tissue may relate to the development of csPES; however, in the small number of studies to date, post-operative pain did not show an association with larger uterine volumes. 8 , 17 While the univariate analysis in this study did show an association between median uterine volume and csPES, after adjusting for confounders, this was no longer significant and suggests that the effect in our cohort was due to the influence of other variables. This is important as, in the experience of the authors, many women with very large fibroids do not get referred for embolization and are also often not considered suitable for myomectomy. Thus, they may miss out on an opportunity for a uterine-spring treatment. In their 2001 study, Bilhim et al. compared post-procedural pain after randomization of 160 women into embolization with PVA particles of a range of sizes: 350-500, 500-700, and 700-900 micrometres. They showed a significantly higher mean post-embolization pain score in the group with predominantly smaller particles. 6 However, the results in this study did not support their findings. Those treated with smaller particle size had an increased odds of csPES (OR 4.66), but this was not significant ( P  = .122). These findings suggest that Interventional Radiologists (IRs) may be able to use a smaller particle if they wish, without worrying that the incidence of csPES will significantly increase. However, the discordance with previous studies does warrant further investigation into the topic. It may be that our pre- and post-embolization management technique helped to obviate the effect of different particle sizes. Practically, the results of this study will allow us to improve our informed discussion process, and thus we can specifically counsel younger and nulliparous patients about the higher likelihood of PES. We may also be able to proactively modify our management of nulliparous patients knowing that their risk of csPES is higher. As a result of this study, the authors have made iterative changes to csPES management moving forward and look forward to evaluating this as a case-control study (using these results as a control group) in the future. There are several limitations to acknowledge in this study. The demographic of patients reflects current referral patters in Australia, influenced by awareness and attitudes towards UFE, leading to a relatively older patient cohort. As a retrospective study, some study data points are limited to free-text data attainable in the EMR, and potential data like reason for preferring UFE over surgery, all treatments tried to date, and ethnicity are not universally available unless specifically documented. This is also a small sample based on patients treated in an 18-month time period. The reason for this is that the start of the study dates corresponded to a time where our premedication protocol changed from oral to intravenous, and the study conclusion dates correspond with the HREC application. This time period thus allowed us to have a controlled group without external influences. The study was also limited by our own premedication protocol and definitions that are based on our practice. This includes an ability to admit patients overnight under the interventional radiology bed card that is not universally available to some colleagues. In addition to this, the choice to use PVA particles is also not universal where many clinicians choose to use calibrated spheres, even though the latter has not been shown to improve clinical efficacy of the procedure. 18 Finally, it should be acknowledged that while all operators aim for a similar embolization endpoint, this is subjective and under- or over-embolization can occur due to natural variations, which may affect the degree of csPES.