Endometrial ablation is one of the less invasive surgical options for women with menorrhagia. The first-generation techniques included roller ball ablation, bipolar endometrial resection, and laser ablation, all of which required visualization of the cavity for the entirety of the procedure.1 Second generation, nonhysteroscopic techniques have since been developed that require less skill and can be performed in the office.2,3 These are devices that destroy the endometrium blindly by various methods, including heated water in a balloon, bipolar radiofrequency, cryoablation, microwave energy, and circulating hot fluid. These techniques offer an alternative to hysterectomy for women who are poor surgical candidates or who wish to avoid major surgery. Head-to-head comparisons of first- and second-generation ablation techniques with hysterectomy have found higher numbers of repeat operations with ablation, less patient satisfaction, and lower rates of amenorrhea. Although hysterectomy guarantees amenorrhea, it has a longer recovery time, greater cost, and much higher complication rate.4–6
Postablation pain has been described after endometrial ablation and variously attributed to hematometra and hematosalpinx.7–9 Postablation tubal sterilization syndrome has been described in patients with a history of tubal ligation in which pain develops as a result of remnant endometrium at the cornua.8,9 It has traditionally been thought that adenomyosis is responsible for pain after ablation, although several studies have not found an association between adenomyosis on preoperative imaging and hysterectomy.10–13 However, the sensitivity for identifying adenomyosis on ultrasound has been found to range from 53% to 85%.14
Although continued bleeding is a common cause for ablation failure, postablation pain is experienced by a substantial number of patients, and many of these patients request a hysterectomy.13 One prior study specifically looked at risk factors for pain after endometrial ablation and found that a history of dysmenorrhea, smoking, tubal ligation, and younger age were all associated with developing pain.13 The aim of this study was to identify prognostic factors that put women at risk of dissatisfaction as a result of pain and subsequent rates that these women request a hysterectomy.
MATERIALS AND METHODS
The institutional review boards of Drexel University and the University of Pennsylvania granted approval to conduct this study before data collection. A retrospective chart review investigating factors associated with pain and hysterectomy after endometrial ablation was undertaken. Data on consecutive patients were gathered for all endometrial ablations performed at Hahnemann University Hospital and the Hospital of the University of Pennsylvania hospitals from January 2006 to May 2013 and entered into the database by two of the authors (K.M.W. and J.P.). The earlier time point marks the advent of electronic medical records at Hahnemann University Hospital and was chosen for ease of data collection. At both sites, office endometrial ablations are captured in the same electronic medical record as operating room procedures. All patients older than 18 years of age were included. Patients were excluded if they had a history of coagulopathy, the procedure was miscoded, or the procedure was aborted. Ablation techniques included a ThermaChoice balloon, microwave, circulating hot water, and bipolar radiofrequency, First-generation techniques were not delineated as a result of their rarity. Patients were identified using the International Classification of Diseases, 9th Revision code for endometrial ablation (68.23) and Current Procedural Terminology codes for hysteroscopy with endometrial ablation, endometrial ablation without hysteroscopic guidance, and endometrial cryoablation with ultrasonic guidance (58563, 58353, and 58356, respectively). Data were abstracted from medical records including operative reports, all follow-up office visits, radiology, and pathology reports.
Our primary outcome was the development of new or worsening pain after endometrial ablation. Independent variables included type of ablation, date of ablation, age at the time of ablation, parity, self-reported race, number of prior cesarean deliveries, history of tubal sterilization, body mass index, development of pain after ablation, findings on radiologic imaging, and endometrial stripe thickness. The secondary outcome was hysterectomy after ablation. In cases in which a hysterectomy was performed, additional data collected included interval from ablation to hysterectomy, indication, and findings on surgical pathology. By abstracting data from every subsequent clinic visit after endometrial ablation, we identified postablation pain when the patient reported to the clinician pain that was new or worse after the procedure that lasted at least 2 months beyond the initial postoperative visit. Patients were defined as lost to follow-up if they did not return more than 6 weeks postoperatively.
Continuous variables were expressed as mean±standard deviation, and categorical variables were expressed as number of cases (n) and percentage of occurrence (%). Odds ratios (ORs) with 95% confidence intervals (CIs) for pain and hysterectomy were also calculated. Backward stepwise logistic regression was used to analyze the effect of categorical variables on the probability of pain or hysterectomy after ablation, and the final analysis included significant variables along with race, history of dysmenorrhea, and history of tubal sterilization. Fisher's exact tests and independent t tests were used where appropriate. Cox proportional hazards model was used to evaluate time to hysterectomy. Statistical significance was set at P<.05. Statistical analysis was performed using IBM SPSS Statistics for Windows 20.0.
A total of 388 patients were identified using International Classification of Diseases, 9th Revision and Current Procedural Terminology codes. Eighty-eight patients were excluded for the reasons listed previously (Fig. 1). Records were incomplete or unavailable for 71 patients, 10 patients had a history of coagulopathy, and in seven cases, the procedure was aborted or unable to be completed. There were 30 patients (10.0%) who were lost to follow-up, leaving a total of 270 for follow-up and analysis. Table 1 summarizes the overall demographics of the study population and those who were available for analysis. Differences between the populations were seen in BMI (P=.01), history of tubal sterilization (P=.01), and uterine mean size on imaging (P=.01).
Thermal balloon and bipolar radiofrequency were used most often (Table 1). Of those ablations done in the office, most were bipolar radiofrequency (n=27 [73%]) with the rest being thermal balloon (n=10 [27%]). Conversely, for procedures in the operating room, most were done by thermal balloon (n=117 [49.4%]) with bipolar radiofrequency for 99 cases (39.8%). This difference was statistically significant (P=.004). For all patients, the most common imaging finding was leiomyoma (n=104 [34.7%]) followed by no specific findings (n=91 [30.3%]), suspected adenomyosis (n=23 [7.7%]), both leiomyoma and adenomyosis (n=17 [5.7%]), thickened endometrial stripe (n=10 [3.3%]), polyp (n=8 [2.7%]), and leiomyoma and polyps (n=2 [0.7%]). For 45 patients (15%), either no imaging was performed preoperatively or a report was not available.
Overall, 62 (23.0%) patients developed worsening or new pain after ablation (Fig. 1). For six (9.7%) of these patients, either no imaging was performed as part of the preoperative workup or the report was not available. Of those patients who had imaging, the most common findings were leiomyomas in 31 (50.0%), no significant pathology in 14 (22.6%), adenomyosis in 13 (21.0%), and both leiomyomas and adenomyosis in six (9.7%). There were no significant differences in patient characteristics between patients who developed pain and those who did not (Table 2). The relationship between developing pain and several variables is shown in Table 3. A history of dysmenorrhea gave 74% higher risk of developing postablation pain (adjusted OR 1.74, 95% CI 1.06–2.87; P=.03) as well as a history of tubal sterilization, which more than doubled the risk (adjusted OR 2.06, 95% CI 1.14–3.70; P=.02). White women were 45% less likely to develop pain after ablation than nonwhites (adjusted OR 0.55, 95% CI 0.34–0.89; P=.014). There was no difference in rates of pain after thermal balloon (25.4%) and bipolar radiofrequency ablation methods (16.0%) (adjusted OR 1.27, 95% CI 0.59–1.69; P=.99).
Fifty-one patients (18.9%) underwent hysterectomy after ablation (Fig. 1; Table 4). Pain was the most common indication for hysterectomy (n=16 [31.4%]). Other indications included continued bleeding among 13 (25.5%) patients, both pain and bleeding for 10 patients (19.6%), uterine prolapse (n=6 [11.8%]), hyperplasia (n=4 [7.8%]), postmenopausal bleeding (n=1 [2%]), and malignancy (n=1 [2%]). Regarding treatment after a failed ablation, seven (30.4%) patients with continued bleeding tried hormone therapy before hysterectomy, whereas 16 (69.6%) declined medical management. For those patients with pain postablation, five (19.2%) tried analgesics (nonsteroidal antiinflammatory drugs) without relief. The most common radiologic imaging findings for those patients who had a hysterectomy were as follows: leiomyomas, 24 (48.0%); adenomyosis, nine (18.0%); and no specific findings for 13 (26.0%). Imaging reports were not available for eight (15.6%) patients, and pathology reports were available for all but one patient. The most common histopathologic diagnosis was leiomyomas in nine (17.6%) patients, adenomyosis in six (11.8%) patients, and both leiomyoma and adenomyosis in 23 (45.1%) patients. No specific findings were found in six specimens (11.8%), hyperplasia in three (5.9%), polyps in two (3.9%), smooth muscle tumor of unknown malignant potential in one (2.0%), and malignancy in one specimen (2.0%). Of the patients with hyperplasias discovered on histologic analysis, two had endometrial biopsy by Pipelle sampling before hysterectomy, which showed benign pathology. The other patient had no sampling before hysterectomy. Malignancy was identified by Pipelle sampling preoperatively for one patient.
Of ablations performed in the operating room, the most common reasons for failure were bleeding and pain with nine cases each (25.0%). The next most common reason was both bleeding and pain for six (16.7%). Those done in the office failed mostly because of pain (n=5 [56%]), then bleeding (n=2 [22.2%]), and bleeding and pain (n=1 [1.1%]). The difference between ablation location and indication for hysterectomy was not significant (P=.61).
Time to hysterectomy for all indications ranged from 32 to 3,122 days (mean [standard deviation] 786 , median 619 days). The shortest time was for malignancy at 32 days. For the indication of pain, hysterectomies were performed at a range of 161–1,795 days (803 , median 745 days). Hysterectomies for both continued bleeding and pain were performed at a mean of 848 days (standard deviation 602, median 698, range 193–179). Time to hysterectomy for the indication of bleeding ranged from 55 to 3,122 days (824 , median 776 days). A Kaplan-Meier curve showing the time to hysterectomy for those patients who and did not develop pain reveals the time to hysterectomy as significantly shorter in patients who developed pain (P<.001; Fig. 2). Additionally, a history of cesarean delivery was significant for a shorter time to hysterectomy (hazard ratio 2.28, 95% CI 1.19–4.36; P=.013; Table 5).
Table 6 shows the relationship between different variables and the risk of hysterectomy after ablation. Patients with a history of cesarean delivery were more than twice as likely to have a failed ablation (adjusted OR 2.33, 95% CI 1.05–5.16; P=.037). Any abnormal uterine findings on radiologic imaging, including leiomyoma, adenomyosis, thickened endometrial stripe, and polyps before ablation, conferred an almost four times higher risk for hysterectomy (adjusted OR 3.96, 95% CI 1.25–12.56; P=.02). However, there were too few cases among the hysterectomies to analyze specific imaging findings (n=27). We also identified ablations performed in the operating room as 76% less likely to lead to hysterectomy (adjusted OR 0.24, 95% CI 0.07–0.77; P=.016). The rate of hysterectomy for thermal balloon ablation was no different than that for bipolar radiofrequency (16.7% compared with 13.4%, adjusted OR 1.14, 95% CI 0.56–2.30; P=.71). Nonwhite race was not a risk factor for hysterectomy.
We also compared the indications for hysterectomy for those patients with a history of cesarean delivery. For the indication of continued bleeding, seven (53.8%) patients had a history of cesarean delivery, whereas for patients with both bleeding and pain, 11 (47.8%) had a history of cesarean delivery. Comparing this with any other indication in which 10 had a history of cesarean delivery (35.7%), statistical significance was not reached (P=.56).
In this study, we aimed to determine risk factors for pain and for hysterectomy after ablation. Pain developed in 23.0% of patients, similar to the 20.8% seen by Thomassee et al.13 Risk factors for postablation pain in the current study included nonwhite race, history of tubal sterilization, and history of dysmenorrhea. The percentage of procedures leading to hysterectomy in our cohort (18.9%) is similar to other studies.15,16
Thomassee et al13 found tubal sterilization as a risk factor for pain. Despite the finding that a history of tubal ligation conferred a higher risk of postablation pain, we did not find that it led to an increased risk of hysterectomy. The pain may not be severe enough to require definitive treatment or it may resolve over time. We also saw that a history of dysmenorrhea was associated with postablation pain. Other studies have shown that dysmenorrhea is associated with higher rates of treatment failure.12,13,17 Women may persist with the same level of dysmenorrhea as preablation and be dissatisfied that the procedure does not eliminate it. Patients who developed postablation pain requested a hysterectomy within a significantly shorter time than those without pain (Fig. 2; Table 5). Bleeding after ablation is often treated first with other methods such as hormones that may lengthen the time to hysterectomy. It is also possible that health care providers do not feel they can treat chronic pelvic pain effectively and offer definitive management more quickly.
The finding that white race decreased the risk for pain was surprising, and the cause for this is unknown. In our population, most nonwhite patients are African American who as a population has higher rates of leiomyomatous uteri.18
Consistent with Shavell et al,19 we found that a history of cesarean delivery is a risk factor for hysterectomy. This has been postulated to be the result of abnormal bleeding after cesarean delivery from a distorted lower uterine segment, although we did not find statistical significance when comparing history of cesarean delivery by indication.20 The presence of leiomyoma or adenomyosis has been suggested as a cause for ablation failure, and we found that any findings on imaging are associated with hysterectomy.10,11 Because these are the most common imaging findings, an analysis of each specific type of pathology may identify one as a risk factor.
A procedure done in the operating room decreased the risk for hysterectomy. Ablations done in the office may not be as thorough as those done in the operating room out of concern for patient comfort. Patient selection for location is unlikely to affect this result because patients with lower pain tolerance would opt for an operating room procedure.
The type of ablation did not affect the rates of pain or hysterectomy despite the difference between the proportions of thermal balloon and bipolar radiofrequency used in the operating room and office. This suggests ablation technique did not affect the outcome. Few studies have made direct comparisons among second-generation techniques, and results are mixed regarding risk of treatment failure and ablation type.21
Two cases of Pipelle sampling failed to diagnose hyperplasia preoperatively. Assessment of the endometrium after ablation may be compromised.22 There is evidence that the feasibility of Pipelle sampling is decreased, and the reliability of postablation is not well defined.23 Given this, patients with persistent postmenopausal bleeding or a high risk for malignancy should have a more thorough assessment.
The strengths of our study were the large sample size, the diversity of the patient population, and the inclusion of two large urban hospitals. The follow-up time for most patients was significant, and careful review of each follow-up visit was performed to assess procedure outcome. The main limitation of our study is its retrospective design, preventing the use of objective measures of pain such as an analog pain scale. However, the patients reported increased pain after the procedure, and patient satisfaction with the procedure is clinically relevant even if it is not corroborated with an analog scale. The percentage of patients lost to follow-up is another limitation of the study. There were significant differences between these patients and the patients with follow-up in regards to BMI, history of tubal sterilization, and uterine size, though this is not likely to negatively affect the final analysis (Table 1). Uterine size and BMI were not significant risk factors for pain or for hysterectomy (Table 3, Table 4). The rate of postablation pain and hysterectomy may actually be underestimated since the group lost to follow-up had a higher percentage of tubal sterilization.
1. Papadopoulos NP, Magos A. First-generation endometrial ablation: roller-ball vs loop vs laser. Best Pract Res Clin Obstet Gynaecol 2007;21:915–29.
2. Madhu CK, Nattey J, Naeem T. Second generation endometrial ablation techniques: an audit of clinical practice. Arch Gynecol Obstet 2009;280:599–602.
3. Reid PC. Endometrial ablation in England—coming of age? An examination of hospital episode statistics 1989/1990 to 2004/2005. Eur J Obstet Gynecol Reprod Biol 2007;135:191–4.
4. Dickersin K, Munro MG, Clark M, Langenberg P, Scherer RW, Frick KD, et al.. Hysterectomy compared with endometrial ablation for dysfunctional uterine bleeding: a randomized controlled trial. Obstet Gynecol 2007;110:1279–89.
5. Munro MG, Dickersin K, Clark M, Langenberg P, Scherer RW, Frick KD, et al.. The surgical treatments outcomes project for dysfunctional uterine bleeding: summary of an Agency for Health Research and Quality-sponsored randomized trial of endometrial ablation versus hysterectomy for women with heavy menstrual bleeding. Menopause 2011;18:445–52.
6. Pinion SB, Parkin DE, Abramovich DR, Naji A, Alexander DA, Russell IT, et al.. Randomised trial of hysterectomy, endometrial laser ablation, and transcervical endometrial resection for dysfunctional uterine bleeding. BMJ 1994;309:979–83.
7. Hubert SR, Marcus PS, Rothenberg JM, Schilder JM, Hurd WW. Hematometra after thermal balloon endometrial ablation in a patient with cervical incompetence. J Laparoendosc Adv Surg Tech A 2001;11:311–3.
8. Webb JC, Bush MR, Wood MD, Park GS. Hematosalpinx with pelvic pain after endometrial ablation confirms the postablation-tubal sterilization syndrome. J Am Assoc Gynecol Laparosc 1996;3:419–21.
9. Bae IH, Pagedas AC, Perkins HE, Bae DS. Postablation-tubal sterilization syndrome. J Am Assoc Gynecol Laparosc 1996;3:435–8.
10. Tresserra F, Grases P, Ubeda A, Pascual MA, Grases PJ, Labastida R. Morphological changes in hysterectomies after endometrial ablation. Hum Reprod 1999;14:1473–7.
11. McCausland AM, McCausland VM. Long-term complications of endometrial ablation: cause, diagnosis, treatment, and prevention. J Minim Invasive Gynecol 2007;14:399–406.
12. El-Nashar SA, Hopkins MR, Creedon DJ, St Sauver JL, Weaver AL, McGree ME, et al.. Prediction of treatment outcomes after global endometrial ablation. Obstet Gynecol 2009;113:97–106.
13. Thomassee MS, Curlin H, Yunker A, Anderson TL. Predicting pelvic pain after endometrial ablation: which preoperative patient characteristics are associated? J Minim Invasive Gynecol 2013;20:642–7.
14. Hanafi M. Ultrasound diagnosis of adenomyosis, leiomyoma, or combined with histopathological correlation. J Hum Reprod Sci 2013;6:189–93.
15. Longinotti MK, Jacobson GF, Hung YY, Learman LA. Probability of hysterectomy after endometrial ablation. Obstet Gynecol 2008;112:1214–20.
16. Fürst SN, Philipsen T, Joergensen JC. Ten-year follow-up of endometrial ablation. Acta Obstet Gynecol Scand 2007;86:334–8.
17. Peeters JA, Penninx JP, Mol BW, Bongers MY. Prognostic factors for the success of endometrial ablation in the treatment of menorrhagia with special reference to previous cesarean section. Eur J Obstet Gynecol Reprod Biol 2013;167:100–3.
18. Richard-Davis G. Uterine fibroid: the burden borne by African American women. J Womens Health (Larchmt) 2013;22:793–4.
19. Shavell VI, Diamond MP, Senter JP, Kruger ML, Johns DA. Hysterectomy subsequent to endometrial ablation. J Minim Invasive Gynecol 2012;19:459–64.
20. Tower AM, Frishman GN. Cesarean scar defects: an underrecognized cause of abnormal uterine bleeding and other gynecologic complications. J Minim Invasive Gynecol 2013;20:562–72.
21. Daniels JP, Middleton LJ, Champaneria R, Khan KS, Cooper K, Mol BW, et al.. Second generation endometrial ablation techniques for heavy menstrual bleeding: network meta-analysis. BMJ 2012;344:e2564.
22. Luo XL, Lim CE, Li L, Wong WS. Hysteroscopic appearance of endometrial cavity after microwave endometrial ablation. J Minim Invasive Gynecol 2010;17:30–6.
23. Ahonkallio SJ, Liakka AK, Martikainen HK, Santala MJ. Feasibility of endometrial assessment after thermal ablation. Eur J Obstet Gynecol Reprod Biol 2009;147:69–71.