Benign uterine lesions, such as endometrial polyps and leiomyomata, are common in women, especially those over age 35 years.1–4 Submucosal leiomyomata or polyps are found in approximately 50% of premenopausal women referred for ultrasound evaluation of abnormal uterine bleeding.1,5–7 However, the prevalence of both lesions is lower in nonselected populations of perimenopausal women being evaluated for their initial complaint of abnormal bleeding.8 These lesions are also found in asymptomatic women.1 Endometrial polyps and leiomyomata are frequently removed surgically because of their association with abnormal uterine bleeding.1,9 Despite the high prevalence of these conditions and the morbidity attributed to them, little is known about their natural history.
In 1999, using saline infusion sonography, we compared the prevalence of benign uterine lesions in 80 women with abnormal uterine bleeding to a cohort of 100 asymptomatic premenopausal women over age 30 years.1 We detected polyps in 10% and leiomyomata in 13% of the asymptomatic controls. In our current study, we add to the understanding of the natural history of uterine polyps and leiomyomata. By reevaluating this group of initially asymptomatic women, we aimed to estimate the incidence rate and regression rate of uterine leiomyomata and polyps, as determined by saline infusion sonography.
MATERIALS AND METHODS
This study was approved by the University of Iowa Human Subjects Committee. All 100 premenopausal women who participated as controls in our initial study were invited to participate. The mean time between participation in the two studies was 2.6 years. We chose this time interval to allow time for growth or regression in the lesions while minimizing the number of women who would become postmenopausal during the interval. We contacted women by letter and, if we received no response, by telephone calls.
In the initial study, women were recruited by local advertisement. To qualify as a control, women were determined to be free of any complaints of abnormal uterine bleeding for 3 months before the evaluation, as assessed by a standardized questionnaire.1 To determine whether abnormal bleeding had developed in the interim, women were asked to complete a second standardized questionnaire about their bleeding patterns in the previous 2.5 years, any treatments for bleeding, and surgical procedures, or hormonal medication use. For the determination of bleeding symptoms, the only difference between the first and second questionnaire was the time frame. For this study, we defined abnormal bleeding as a positive response to one of the following complaints: any spotting between periods, heavy menstrual bleeding for most of their menses, or seeking medical treatment for abnormal bleeding in the previous 2.5 years. It should be emphasized that characterization of abnormal bleeding was the subjective assessment of the patient, and answers were not validated by measuring blood loss by menstrual pad collection. We defined postmenopausal status as a positive response to the question, “Have you gone through menopause (stopped having natural periods)?”
A transvaginal ultrasound with saline infusion sonography was then performed as follows. A speculum was inserted, and the cervix was cleaned with an iodine-based solution. A 5.3-French Soules intrauterine insemination catheter was inserted until the fundus of the uterus was reached. The speculum was then removed, a standard transvaginal probe (5–8 MHz) was placed in the vagina, and approximately 2–10 mL of saline was introduced into the uterine cavity through the catheter to better visualize the presence of endometrial abnormalities. Saline infusion sonography examinations were scheduled shortly after menses had ceased if a woman was having menstrual periods. If she was not having periods, then saline infusion sonography examinations were scheduled at the woman's convenience.
Two sonographers performed all ultrasounds using the same machine, and one physician interpreted the films. Sonographers and physicians involved were unaware of the results of the first ultrasound and the answers on the subjects' questionnaires. Polyps were identified as generally hyperechoic sessile or pedunculated masses that were present in the endometrial cavity. Leiomyomata were identified as hypoechoic masses found in the myometrium, in the uterine cavity, or were subserosal in location. All uterine lesions were identified and measured in three diameters.
We calculated the point prevalence (proportion of women with lesions at a given time), cumulative incidence (proportion of women with polyps or leiomyomata at follow-up who were free of such lesions at the first ultrasound), spontaneous regression rate (proportion of women with polyps or leiomyomata on the first ultrasound but no such lesions on follow-up), and growth of the lesions seen on first ultrasound.
Statistical analysis was performed utilizing paired t tests for comparing continuous variables in the same individual and Student t test for comparing continuous variables between groups after conditions of normality of data were met. χ2 and Fisher exact test were used for comparing the categoric variables between groups. For all tests a P value of less than .05 was considered statistically significant.
Of 100 women that we attempted to contact for this study, 64 agreed and were eligible to participate. Among nonparticipants, we were unable to contact nine women, nine had moved out of state, 16 refused participation most commonly because of the inconvenience, one had a hysterectomy for endometriosis, and one was pregnant. There were no significant differences in age or findings on the initial ultrasound when participants and nonparticipants were compared (Table 1).
Of the 64 participants, 57 were still premenopausal, whereas seven women had become menopausal since the initial ultrasound. The mean age of women participating in this study was 44 years (range 33–56 years). More than 90% of participants were white, which is consistent with the population of Iowa. Fifteen women (23%) reported using hormonal medications in the intervening 2.5 years, including birth control pills, Depo-Provera (Pharmacia & Upjohn Co., Kalamazoo, MI), or hormone replacement therapy in the case of menopausal women. However, only three women reported that they were using hormones as medical treatment for abnormal bleeding. No women had undergone any diagnostic or therapeutic surgical procedures involving the uterus, and none had been pregnant since the first ultrasound. The average interval between the first ultrasound (ultrasound 1) and the second ultrasound (ultrasound 2) was 2.6 years (range 2.4–2.7 years).
Endometrial polyps were detected in seven of 64 (11%) women at ultrasound 1 (Table 1). Five women had solitary polyps, and two women had two polyps each. Four of these seven women no longer had polyps at ultrasound 2. Polyps that regressed tended to be smaller than polyps that persisted. The mean diameter of polyps that regressed was 0.7 cm (n = 6, range 0.4–1.1 cm) compared with 1.3 cm (n = 3, range 0.8–1.8 cm) for polyps that persisted (P = .16). There was no difference in mean diameter of polyps that persisted from ultrasound 1 to ultrasound 2 (1.3 cm versus 1.6 cm, P = .44). Of women who had polyps on ultrasound 1, two took hormonal medications (oral contraceptive pills or hormone replacement therapy) in the intervening years and both had regression of their polyps. Five women did not use any hormonal medications, and two of these women had regression of their polyps whereas three did not. Only one woman whose polyp regressed had become menopausal in the intervening years.
Of women with no evidence of uterine polyps by saline infusion sonography on ultrasound 1, seven women had endometrial polyps present at ultrasound 2, yielding a cumulative incidence of 12%. The 2.5-year cumulative incidence rate of polyps was not significantly different in women who reported hormone use since the first ultrasound versus those who did not (one of 13 [8%] versus six of 44 [14%], respectively; P > .99).
In total, ten women had endometrial polyps detected by saline infusion sonography at ultrasound 2 (Table 1). Seven women had new polyps and three women had persistent polyps yielding a point prevalence rate of 16%. The mean diameter of these polyps was 1.2 cm (range 0.6–2.2 cm). The mean age of women with polyps was not statistically different from women without endometrial polyps (44.6 versus 43.9 years, P = .38). The percentage of women reporting hormone use among those with polyps was not different from women without polyps (10% versus 26%, P = .26).
Eleven of 64 women (16%) had a total of 18 uterine leiomyomata detected at ultrasound 1. One woman had a single leiomyoma on ultrasound 1 and had a normal uterus with no leiomyoma detected on ultrasound 2. Among three other women with multiple leiomyomata on the first ultrasound, a total of five leiomyomata were not detected on ultrasound 2. Women who had at least one leiomyoma that was no longer detectable tended to be older than women whose leiomyomata remained (50.5 versus 46.8 years, P = .15) although none of these women reported becoming menopausal in the intervening years. The mean diameter of leiomyomata that were no longer detected was significantly smaller than that of leiomyomata that persisted (1.1 cm versus 2.2 cm, P = .008). The leiomyomata that persisted tended to be larger in ultrasound 2 compared with ultrasound 1 (mean diameter 3.4 versus 2.2 cm, P = .13). The average growth of leiomyomata in mean diameter was 1.2 cm per 2.5 years with a range of −0.9–6.8 cm per 2.5 years. Of women with no leiomyomata on ultrasound 1, seven women developed a total of eight leiomyomata (2.5 year cumulative incidence rate of 13%). Two women, each with one leiomyoma on ultrasound 1, developed a second leiomyoma in the interval.
Seventeen women had a total of 22 leiomyomata detected on ultrasound 2, yielding a point prevalence of 27% (Table 1). There was no significant difference in age or reported hormone use between women with and without leiomyomata.
Using our definition of abnormal bleeding, 25 women (39%) reported abnormal uterine bleeding at the time of the second ultrasound. Five women reported seeking medical advice for abnormal uterine bleeding. No woman had a surgical procedure or endometrial biopsy for abnormal uterine bleeding.
Women with uterine polyps detected on ultrasound 2 reported abnormal uterine bleeding more often than women with no uterine abnormalities (70% versus 32%; P = .035). Women with leiomyomata also reported abnormal bleeding more often than women with no uterine abnormalities, but this difference did not reach statistical significance (47% versus 33%; P = .36). Women with leiomyomata and abnormal bleeding tended to have larger leiomyomata than women with leiomyomata who did not have that complaint (34.2 versus 20.4 mm, P = .13). All leiomyomata were intramural in location, and none distorted the endometrial cavity.
We conducted this prospective cohort study to determine the natural history of untreated endometrial polyps and leiomyomata. Although transvaginal ultrasound and saline infusion sonography do not provide histologic proof of the presence of lesions, we previously found that saline infusion sonography has a sensitivity of 97% and a specificity of 86% compared with surgical and pathologic findings as the gold standard.1 Others have also noted the high sensitivity and specificity of saline infusion sonography with lesions as small as 2 mm seen reliably.10 Transvaginal sonography has been shown to have high reproducibility and low interobserver variability for the detection of uterine abnormalities.11 Sonograms were performed by experienced sonographers and physicians, and when possible, we performed the saline infusion sonography shortly after menses had stopped to avoid false-positive saline infusion sonography findings that can occur because of blood clots or tufts of thickened late proliferative or secretory endometrium.5 Therefore, transvaginal ultrasonography and saline infusion sonography provide an accurate means of detecting and following up even small benign uterine lesions.
We confirmed our previous observation that small endometrial polyps are commonly found by saline infusion sonography in perimenopausal women who are asymptomatic or who have minor complaints of abnormal uterine bleeding for which they are not seeking treatment. We found the cumulative prevalence of polyps to be 16%, which is slightly higher than the 10% prevalence rate found in our initial study. This increased prevalence over time is consistent with our observation that polyps develop more often in older premenopausal women and are uncommon in women under age 35 years.1,3 The prevalence of polyps in these relatively asymptomatic women is less than what has been reported in several series of premenopausal women who have been evaluated for abnormal uterine bleeding by ultrasound and saline infusion sonography.1,5–7 However, our prevalence rate is higher than the 1% prevalence of polyps reported in a series of asymptomatic women (age range 19–44 years, no mean age reported) who had elective hysteroscopic sterilization.12 The higher prevalence in our study could be explained by several factors. First, we are studying an older population, and polyps are more common among perimenopausal women. The peak prevalence of polyps occurs in women between 40 and 50 years.3 Second, more patients who had sterilization could have been using hormonal contraception before the procedure, a practice that might influence polyp formation. Finally, it is possible that some small polyps detected by saline infusion sonography are not polyps but are focal irregularities of the endometrium that had a polypoid appearance on ultrasound. We took precautions against this possibility. The saline infusion sonography was performed in the early follicular phase when possible, and we required that the lesion be present in several planes.
We have shown that polyps can regress without medical or surgical treatment. Four of seven women with polyps detected on ultrasound 1 no longer had polyps 2.5 years later. We believe that this represents true regression and not a detection error, given the high sensitivity and specificity of saline infusion sonography for small lesions. The mechanism by which these polyps regressed is not known. Two women who had polyps that regressed reported isolated episodes of heavy menses associated with cramping and uterine pain followed by resumption of normal menses. It is possible that those episodes were associated with the passage of an endometrial polyp. These findings suggest that asymptomatic polyps, particularly small polyps, will have a reasonable likelihood of spontaneous regression if followed up over time. Further studies are needed to determine whether small, symptomatic polyps also regress and do so in an acceptable time frame for the patient. Our data suggest that larger polyps (greater than 1 cm in mean diameter) are less likely to regress.
The prevalence of leiomyomata also increased in this cohort of women. This finding is expected because previous studies have shown that older age before menopause is a risk factor for the development of leiomyomata. We calculated a cumulative incidence rate of 13% over 2.5 years. This rate is higher than the incidence rate of 1.25% per year reported for white women in this age range in the Nurse's Health Study, a large questionnaire-based study of clinically detected leiomyomata.4 The higher incidence rate in our study was expected because we used a more sensitive technique for the detection of leiomyomata, and many of these leiomyomata were asymptomatic and not clinically apparent.
The growth patterns of leiomyomata varied greatly during the study period. Although most had a greater mean diameter at ultrasound 2 than at ultrasound 1, the change in diameter ranged from a decrease of 0.9 cm to an increase of 6.8 cm over the 2.5-year period. Four relatively older premenopausal women had six small leiomyomata that were not detected on the second ultrasound. Leiomyomata can regress after menopause, but we are unaware of this being described in premenopausal women. We cannot completely rule out the possibility that this finding is spurious and is attributable to detection errors. However, the leiomyomata that regressed had a mean diameter of 1.1 cm, a diameter well within the limits of detection using modern transvaginal ultrasound machines.
Our definition of abnormal bleeding was rather liberal (see Methods), and this led to a high percentage of women reporting some bleeding abnormalities in the preceding 2.5 years. Polyps detected by saline infusion sonography were associated with the development of complaints of abnormal uterine bleeding. All women were asymptomatic upon entry to the study, and yet 70% of women with polyps on the second ultrasound had complaints of abnormal uterine bleeding, albeit minor complaints as few women considered the problem significant enough to seek medical advice or treatment. Many leiomyomata remained asymptomatic despite continued growth. There was a tendency for larger leiomyomata to be associated with more bleeding symptoms. None of the leiomyomata that we detected was intracavitary or distorted the endometrial cavity. We speculate that location, rather than size or growth of the leiomyomata, plays the most important role for the development of abnormal uterine bleeding. Because of the retrospective nature of our questionnaire in eliciting symptoms of abnormal uterine bleeding, we cannot draw firm conclusions about the relationship between ultrasound findings and specific bleeding patterns.
Ultrasound with saline infusion sonography allows for relatively painless detection of endometrial polyps.7 We invite future clinical trials to further study the natural history of benign lesions in larger and more ethnically varied populations of women. Further studies of potential medical treatment of women for either therapy or prevention of polyps will be possible using saline infusion sonography screening. Possible agents to consider might include hormones or nonsteroidal anti-inflammatory drugs. Progesterone-releasing intrauterine devices have been shown to reduce the development of polyps in women taking tamoxifen for breast cancer prevention.13 Nonsteroidal anti-inflammatory drugs have been shown to reduce the incidence of colonic polyps,14 but have never been investigated for their effect on endometrial polyps.
In conclusion, polyps and leiomyomata are common in asymptomatic premenopausal women. Small polyps have a reasonable chance of regressing spontaneously. This should be kept in mind when recommending therapy for women. This fact also serves to emphasize the importance of an untreated control group in randomized clinical trials designed to assess the therapeutic efficacy of interventions for polyps.
1. Clevenger-Hoeft M, Syrop CH, Stovall DW, Van Voorhis BJ. Sonohysterography in premenopausal women with and without abnormal bleeding. Obstet Gynecol 1999;94:516–20.
2. Syrop CH, Sahakian V. Transvaginal sonographic detection of endometrial polyps with fluid contrast augmentation. Obstet Gynecol 1992;79:1041–3.
3. Peterson WF, Novak ER. Endometrial polyps. Obstet Gynecol 1956;8:40–9.
4. Marshall L, Spiegelman D, Barbieri R, Goldman M, Manson J, Colditz G; et al. Variation in the incidence of uterine leiomyoma among premenopausal women by age and race. Obstet Gynecol 1997;90:967–73.
5. Goldstein SR. Use of ultrasonography for triage of perimenopausal patients with unexplained uterine bleeding. Am J Obstet Gynecol 1994;170:565–70.
6. De Vries L, Dijkhuizen F, Mol B, Brolmann H, Moret E, Heintz A. Comparison of transvaginal sonography, saline infusion sonography, and hysteroscopy in premenopausal women with abnormal uterine bleeding. J Clin Ultrasound 2000;28:217–23.
7. Widrich T, Bradley L, Mitchinson A, Collins R. Comparison of saline infused sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:1327–34.
8. Goldstein S, Zeltser I, Horan C, Snyder J, Schwartz L. Ultrasonography-based triage for perimenopausal patients with abnormal uterine bleeding. Am J Obstet Gynecol 1997;177:102–8.
9. Tjarks M, Van Voorhis B. Treatment of endometrial polyps. Obstet Gynecol 2000;96:886–9.
10. Widrich T, Bradley LD, Mitchinson AR, Collins RL. Comparison of saline infusion sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:1327–34.
11. Emanual MH, Aknum WM, Verdel MJC, Hart AAM. The reproducibility of the results of transvaginal sonography of the uterus in patients with abnormal uterine bleeding. Ultrasound Obstet Gynecol 1996;8:346–9.
12. Cooper J, Houck R, Rigberg H. The incidence of intrauterine abnormalities found at hysteroscopy in patients undergoing elective hysteroscopic sterilization. J Reprod Med 1983;28:659–61.
13. Gardner F, Konje J, Abrams K, Brown L, Khanna S, Al-Azzawi F, et al. Endometrial protection from tamoxifen-stimulated changes by a levonorgestrel-releasing intrauterine system: A randomized controlled trial. Lancet 2000; 356:1711–7.
14. Steinbach G, Lynch P, Phillips R, Wallace M, Hawk E, Gordon G, et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000;342:1946–52.