Materials and Methods
This retrospective cohort study included all patients who had their first endometrial ablation from January 2007 through June 2009 at Women and Infants Hospital. We excluded women who had had a prior endometrial ablation. This study was approved by the Institutional Review Board of Women and Infants Hospital (no. 12-0051).
We collected demographic information, medical history, preoperative bleeding pattern, and surgical information from hospital and office records. The main independent variable, preoperative bleeding pattern (heavy and irregular or heavy and irregular), was determined by a review of documentation in the hospital medical record and in history and physical examinations for the operating room. Office charts were reviewed when the bleeding pattern could not be sufficiently classified with the information available in the hospital medical record.
Because of confusion and misuse of terms such as menorrhagia, menometrorrhagia, and dysfunctional uterine bleeding, we did not rely on these terms to determine the regularity of the bleeding. Examples of terms that designated the bleeding such as regular included regular, monthly, predictable, cyclic, or a statement that bleeding occurred every X to Y days. Examples of terms that designated the bleeding such as irregular included irregular, unpredictable, erratic, ovulatory dysfunction, or anovulation. The main dependent variable, treatment failure, was defined as hysterectomy for any benign indication or repeat ablation within 36 months after the endometrial ablation. Our secondary dependent variable, subsequent gynecological procedures, was defined as endometrial biopsy, dilation and curettage, hysteroscopy, repeat ablation, or hysterectomy within 36 months after the endometrial ablation.
For our sample size calculations, we assumed an alpha = 0.05 and beta = 0.2. We estimated sample size needed for a 1:1 to 3:1 ratio of women with heavy and regular to women with heavy and irregular bleeding. We estimated a treatment failure rate of 15% at 3 years after endometrial ablation in women with heavy and regular bleeding based on previous studies. We set the minimal detectable difference in treatment failure between groups at 10%. Based on these assumptions, we needed a minimum of 560 patients with analyzable data. To account for the missing data in this retrospective chart review, we planned to review at least 960 medical records.
Categorical variables were compared by χ 2 or Fisher exact test. Continuous variables were compared between groups by Student t test or Wilcoxon rank-sum test. Multiple logistic regression was used to estimate the adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for the association of several variables with the outcome, treatment failure. All P values presented are 2 tailed, with P < .05 considered statistically significant.
Results
There were 968 records of women with endometrial ablations performed during the study period who were eligible for this review. Nine hundred sixty-one of the records (99.3%) were radiofrequency bipolar endometrial ablations. Two hundred ninety-three women (30.3%) were classified as having heavy and regular uterine bleeding and 352 (36.4%) as having heavy and irregular uterine bleeding. The bleeding pattern of 286 women (29.5%) could not be classified more specifically and was called heavy not otherwise specified. There were no significant differences in demographic or clinical characteristics between women with the different bleeding patterns ( Table 1 ).
| Characteristic | Total a |
|---|---|
| Age, mean in years (SD) | 42.7 (5.6) |
| Parity, median (range) | 2.0 (0.0–6.0) |
| Obesity (BMI ≥30 kg/m 2 ) | 312 (33.3) |
| Race/ethnicity | |
| White | 832 (87.9) |
| Black | 35 (3.7) |
| Hispanic | 49 (5.2) |
| Other | 30 (3.2) |
| Insurance | |
| Private | 785 (81.6) |
| Public | 146 (15.2) |
| None | 31 (3.2) |
| Tobacco use | 197 (21.1) |
| Preoperative bleeding pattern | |
| Heavy/regular | 293 (30.3) |
| Heavy/irregular | 352 (36.4) |
| Heavy NOS b | 286 (29.5) |
| Other | 37 (3.8) |
| Polyps on preoperative ultrasound | 221 (23.5) |
| Fibroids on preoperative ultrasound | 187 (19.9) |
| Previous tubal ligation | 391 (40.4) |
| Preoperative dysmenorrhea or pelvic pain | 110 (11.4) |
| Subsequent gynecological procedures c | |
| Any intervention | 158 (16.3) |
| Hysterectomy | 104 (10.7) |
| Repeat ablation | 12 (1.2) |
| Endometrial biopsy | 25 (2.6) |
| Dilation and curettage | 21 (2.2) |
| Hysteroscopy | 18 (1.9) |
a Data are n (column percentage) unless otherwise noted
b Heavy uterine bleeding, regularity/predictability not otherwise specified
c Gynecological procedures performed in the 36 months following endometrial ablation.
During the 36 months following endometrial ablation, 16.3% of the women (n = 158) had a gynecological procedure, 1.2% had a repeat ablation (n = 12), and 10.7% had a hysterectomy (n = 104). We found no difference in the proportion of women who had a subsequent gynecological procedure or experienced treatment failure between women with heavy and regular uterine bleeding and women with heavy and irregular uterine bleeding (16.4% vs 17.6%, P = .7 and 12.6% vs 12.2% P = .9, respectively) ( Tables 2 and 3 ).
| Characteristic | Heavy and regular (n = 293) a | Heavy and irregular (n = 352) | P value |
|---|---|---|---|
| Age, mean in years (SD) | 42.6 (5.2) | 43.2 (5.7) | .1 |
| Parity, median (range) | 2.0 (0.0–6.0) | 2.0 (0.0–5.0) | .6 |
| BMI | |||
| Nonobese (<30 kg/m 2 ) | 189 (68.0) | 218 (63.4) | .2 |
| Obese (≥30 kg/m 2 ) | 89 (32.0) | 126 (36.6) | |
| Race/ethnicity | |||
| White | 252 (88.4) | 302 (87.3) | .3 |
| Black | 13 (4.6) | 11 (3.2) | |
| Hispanic | 10 (3.5) | 23 (6.6) | |
| Other | 10 (3.5) | 10 (2.9) | |
| Insurance | |||
| Private | 237 (81.4) | 286 (81.7) | .5 |
| Public | 42 (14.4) | 55 (15.7) | |
| None | 12 (4.1) | 9 (2.6) | |
| Tobacco use | 51 (18.1) | 80 (23.4) | .1 |
| Polyps on preoperative ultrasound | 76 (26.3) | 83 (24.1) | .6 |
| Fibroids on preoperative ultrasound | 69 (23.9) | 65 (18.9) | .1 |
| Previous tubal ligation | 133 (45.4) | 134 (38.1) | .07 |
| Preoperative dysmenorrhea or pelvic pain | 43 (14.7) | 42 (11.9) | .3 |
| Subsequent gynecological procedures b | |||
| Any intervention | 48 (16.4) | 62 (17.6) | .7 |
| Hysterectomy | 35 (11.9) | 39 (11.1) | .8 |
| Repeat ablation | 2 (0.7) | 4 (1.1) | .7 |
| Endometrial biopsy | 7 (2.4) | 11 (3.1) | .6 |
| Dilation and curettage | 7 (2.4) | 8 (2.3) | 1.00 |
| Hysteroscopy | 5 (1.7) | 8 (2.3) | .8 |
a Data are n (column percentage) unless otherwise noted
b Gynecological procedures performed in the 36 months following endometrial ablation.
| Preoperative characteristic | Any subsequent gynecological procedure c | Treatment failure | ||||
|---|---|---|---|---|---|---|
| Yes | No | P value | Yes | No | P value | |
| Preoperative bleeding pattern d | ||||||
| Heavy and regular | 48 (16.4) | 245 (83.6) | .7 | 37 (12.6) | 256 (87.4) | .9 |
| Heavy and irregular | 62 (17.6) | 290 (82.4) | 43 (12.2) | 309 (87.8) | ||
| Previous tubal ligation | ||||||
| Yes | 82 (21.0) | 309 (79.0) | .001 | 64 (16.4) | 327 (83.6) | .0008 |
| No | 76 (13.2) | 501 (86.8) | 52 (9.0) | 525 (91.0) | ||
| Dysmenorrhea or pelvic pain | ||||||
| Yes | 28 (25.5) | 82 (74.5) | .009 | 24 (21.8) | 86 (78.2) | .002 |
| No | 130 (15.2) | 728 (84.8) | 92 (10.7) | 766 (89.3) | ||
| BMI | ||||||
| Nonobese (<30 kg/m 2 ) | 86 (13.8) | 539 (86.2) | .001 | 61 (9.8) | 564 (90.2) | .003 |
| Obese (≥30 kg/m 2 ) | 69 (22.1) | 243 (77.9) | 52 (16.7) | 260 (83.3) | ||
| Polyps on ultrasound | ||||||
| Yes | 36 (16.3) | 185 (83.7) | 1.00 | 24 (10.9) | 197 (89.1) | .6 |
| No | 116 (16.1) | 604 (83.9) | 88 (12.2) | 632 (87.8) | ||
| Fibroids on ultrasound | ||||||
| Yes | 37 (19.8) | 150 (80.2) | .1 | 28 (15.0) | 159 (85.0) | .2 |
| No | 115 (15.2) | 640 (84.8) | 84 (11.1) | 671 (88.9) | ||
a Gynecological procedures (endometrial biopsy, dilation and curettage, hysteroscopy, repeat ablation, or hysterectomy) were performed in the 36 months following endometrial ablation
b Treatment failure was defined as repeat ablation or hysterectomy in the 36 months following endometrial ablation
Women with a history of tubal ligation were more likely to experience treatment failure after endometrial ablation compared with women without a history of tubal ligation (16.4% vs 9.0%, P = .0008) ( Table 3 ). Similarly, compared with their counterparts who did not have preoperative dysmenorrhea or pelvic pain or obesity, women with dysmenorrhea or pelvic pain, and women who were obese were more likely to experience treatment failure after ablation (21.8% vs 10.7%, P = .002 and 16.7% vs 9.8%, P = .003, respectively). The incidence of having a gynecological procedure in the 36 months following endometrial ablation was also greater in patients with prior tubal ligation, preoperative pelvic pain, and obesity ( Table 3 ).
Multiple logistic regression was performed to estimate odds of treatment failure and subsequent gynecological procedures. We adjusted for preoperative uterine bleeding pattern and factors found to be associated with our dependent variables in the univariate analyses (previous tubal ligation, dysmenorrhea or pelvic pain, and obesity). The odds of treatment failure for women with heavy and irregular uterine bleeding was not significantly increased compared with women with heavy and regular uterine bleeding (adjusted OR [aOR], 1.07; 95% CI, 0.65–1.74). However, consistent with our univariate analysis, we did identify several other factors that were associated with treatment failure.
Compared with women without a previous tubal ligation, women with a tubal ligation were at increased odds of both treatment failure and subsequent gynecological procedures (aOR, 1.94; 95% CI, 1.30–2.91, aOR, 1.71; 95% CI, 1.20–2.43, respectively). Women with preoperative pelvic pain or obesity were also at increased odds of treatment failure and gynecological procedures (aOR, 2.42; 95% CI, 1.44–4.06 and aOR, 1.93; 95% CI, 1.20–3.13 for pain, aOR, 1.82; 95% CI, 1.21–2.73 and aOR, 1.75; 95% CI, 1.22–2.50 for obesity, respectively) ( Table 4 ).
