Abstract
Background: The main objective was to investigate if occurrence of hyperplastic polyps was reduced by use of the levonorgestrel-impregnated intrauterine system (LNG-IUS, Mirena®; Bayer) and if the LNG-IUS was more effective compared to oral medroxyprogesterone acetate (MPA) or observation-only. Patients and Methods: Patients (N=59) with hyperplastic polyps were given LNG-IUS, 10 mg oral MPA taken 10 days per cycle, or had observation-only for six months. Diagnosis of histological specimens was performed by light microscopy according to the WHO classification and D-score prior to and after six months therapy. Results: No polyps were found in women treated with LNG-IUS (18/18). Five women treated with cyclic MPA (5/20, 25%) and two (2/21, 9%) with observation had normal endometrium without polyps after six months. Conclusion: No former study has shown that LNG-IUS is effective at reducing the occurrence of hyperplastic endometrial polyps. The effect is superior to that of oral progestin and observation-only.
Endometrial polyps, solitary or multiple, may be detected by conventional transvaginal ultrasound. However, ultrasonographic findings are not specific for these structures, so small polyps may easily be overlooked in daily routine (1, 2). Fragments of small hyperplastic endometrial polyps are frequent incidental findings in Pipelle® (endometrial suction curette, Cooper Surgical, UK) biopsy on routine microscopy. Usually such biopsies have been taken in outpatient routine due to irregular bleeding to exclude pre-malignant or malignant disease. Only a minority of these polyps are diagnosed prior to biopsy according to patient records.
The prevalence of endometrial polyps shows great differences in comparable studies, ranging between 10% and 35% in different populations (3-5). Most endometrial polyps are benign; however, studies have reported pre-malignant and malignant changes to occur in up to 23 and 12%, respectively (3, 6, 7). Comparably, malignancy in endometrial polyps associated with tamoxifen use has been reported to be considerably more frequent (8). Thus, hyperplastic endometrial polyps may be considered as risk factors for endometrial carcinoma, which necessitates removal or alternative therapeutic procedures (7, 9).
Resection by hysteroscopy has been recommended as a safe and effective therapy for endometrial polyps and allows for evaluation by histology. However, the requirement for general anesthesia is a disadvantage in outpatient practice (2). Lieng and co-workers reported spontaneous regression of small endometrial polyps in 27% of cases after one year (10). Thus, for smaller polyps, conservative treatment might be considered a viable option in an outpatient setting, depending on patient preference and risk factors. For elderly and seriously-ill women, medical therapy might be the only preferable option. Oral progestin has been used for decades as routine therapy for endometrial hyperplasia (11, 12). Recently, use of a levonorgestrel-impregnated intrauterine device (LNG-IUS) was found to lead to a higher regression rate compared to oral therapy for endometrial hyperplasia independent of polyps (13, 14). Only a few studies have demonstrated that the LNG-IUS can prevent development of benign endometrial polyps in patients on tamoxifen (15-18). Thus, it remains to be shown if the LNG-IUS is an effective therapy for hyperplastic polyps in women independent of tamoxifen use.
In the present population-based observational cohort study, the main objective was to investigate the occurrence of hyperplastic endometrial polyps after six months in LNG-IUS users. We also intended to compare the effect of LNG-IUS to low-dose oral progestin therapy and observation only. Furthermore, we intended to evaluate if the WHO classification and the prognostic D-score algorithm was applicable to hyperplastic polyps (19-23).
Patients and Methods
Patients. During the period from 1st of January 1999 until the 31st of December 2004 a total of 68 consecutive patients in our health region (North Norway) were given the initial histopathological diagnosis of hyperplastic endometrial polyp and considered for the study. The endometrial biopsies had been taken by Pipelle after the patients had seen their gynecologist, mostly due to irregular or post-menopausal bleeding. Ultrasound was performed prior to Pipelle biopsy. Endometrial specimens were immediately sent for histopathological investigation. Among the 68 consecutive patients, 6 were excluded from the study due to lack of follow-up biopsy. Three women had immediate hysterectomy due to strong suspicion of co-existing malignancy. Ultimately, 59 women with hyperplastic polyps were included. Irregular or post-menopausal bleeding were the main reasons for consultation for 53 out of the 59 women. Prior to Pipelle biopsy, only six out of the 59 women were diagnosed with endometrial polyps and five had thickened endometrium at ultrasonograhpy. All the patients were treated and followed at outpatient clinics at the Gynaecological Department, University Hospital of North Norway or in one of the seven local hospitals within our health region. Choice of treatment modality was left to the gynecologist. The patient records showed that one of three different therapy strategies had been chosen: LNG-IUS, 10 mg oral medroxyprogesterone acetate (MPA) taken 10 days per cycle for six months, or observation-only. Control biopsy and ultrasound investigation had been performed after six months of such strategy for all 59 patients. In spite of the histological diagnosis, none of the patients had been referred for hysteroscopy. Patients' records showed that none of the included women had developed endometrial carcinoma as of 1st of June 2013 after 126-174 months of follow-up.
Histological specimens. The histological material, defined as index and follow-up biopsies obtained by Pipelle, had been sent to the Department of Pathology, University Hospital of North Norway for routine assessment. The specimens were processed in the laboratory, fixed in buffered formaldehyde, embedded in paraffin before standard histological sections stained by haematoxylin and eosin were made. Microscopic assessment was performed by a trained gynecologic pathologist (AO) and one additional routine pathologist who were blinded to each other's diagnosis. Agreement after discordant results was always obtained after discussion at a multi-headed microscope. A diagnosis of hyperplastic endometrial polyp was given when microscopy revealed polyps or fragments of polyps (diameter of polyps or fragments were always >5 mm and <12 mm). Hyperplastic polyps and fragments were identified by the characteristic shape covered by surface epithelium, and by fibrous stroma with thick-walled or enlarged vessels. Another mandatory criterion was the presence of hyperplastic glands organized with or without complexity and atypical changes similar to endometrial hyperplasia (21, 24). Hyperplastic glandular changes in index biopsies as well as follow-up biopsies were divided into one of three groups: simple hyperplasia, complex hyperplasia or atypical hyperplasia, according to the WHO classification, which is still considered the gold standard for evaluation of endometrial hyperplasia (20, 21). To increase the diagnostic reproducibility for the hyperplastic lesions, the D-score analyses always followed the WHO classification (see below). The D-score analyses were always carried out by an experienced engineer. Normal endometrium in the follow-up biopsies after therapy or observation only, was defined as ordinary proliferative endometrium or endometrium with progestin effect lacking the characteristic features of polyp and with no suspicion of polyp at ultrasound examination. All clinical and histopathological information were registered and maintained in a separate database and finally supplemented by information from hospital records.
Morphometric analysis D-score. As reproducibility of the WHO classification for endometrial hyperplasia performed by light microscopy is considered rather poor, we introduced the morphometric image analysis algorithm D-score in order to improve the selection of risk groups for hyperplastic lesions. Thus, hyperplastic lesions with D-score >1 or D-score 0-1 are considered to be at low or medium risk of cancer development; patients with D-score <0 are assessed as having a high risk of co-existent or future carcinoma (19, 22, 23). In the original computerized morphometric analysis study on endometrial hyperplasia, a total of 10 nuclear features and 12 architectural features were analyzed. Using a linear stepwise regression analysis and discriminant analysis, three of these quantitative features were selected as having significant independent prognostic value and were combined into the formula called the D-score. The measurements were performed with a Q-PRODIT image analysis system (version 6.1; Leica, Cambridge, UK). The method of the D-score is described in detail in former studies (19; 23). The D-score method has been implemented as routine analysis in our health region and includes recommendation for patient follow-up (22, 23).
Ethics. The study was recommended by the local Ethical Committee (approval number 23/2004) and the national data management authorities.
Statistics. Fisher's exact test with two-sided p-value were used for comparison of demographic data using GraphPad InStat version 3.06 (GraphPad Software, San Diego, CA, USA). A p-value of less than 0.05 was considered significant.
Results
Patients. A total of 59 women with microscopically-confirmed hyperplastic endometrial polyps at the index biopsy were included. The results are reported according to the occurrence of polyps or not in the follow-up biopsies taken after six months on LNG-IUS, oral progestin therapy or observation only. An absence of polyps is regarded as therapy response. When comparing demographic baseline data for the two groups, with and without polyps, there was no difference in distribution of age, although, the majority of patients were considered pre- and peri-menopausal (Table I). Mean body-mass index showed that the majority of the patients were slightly overweight. Smoking habits, menopausal status, and uterine size were equally distributed among the two patient groups (Table I). Patients with the highest parity showed the most favorable therapy response (Table I). Two of the patients in the no-response group (after oral MPA) had a BMI over 40, both were nulliparous. Six patients in the response group and eight in the no-response groups had been using hormone replacement therapy. One patient in the response group and four in the no-response groups had a history of cervical intra-epithelial neoplasia. None of the patients included in the study had been receiving tamoxifen.
Baseline demographic data of the included women according to the follow-up biopsy after six months.
Response to therapy. Table II clearly shows that oral treatment was less effective compared to LNG-IUS regarding the occurrence of endometrial polyps after six months. Observation-alone was the least effective therapy option. All the patients in the LNG-IUS group were cured independent of their primary diagnosis, whereas only 25% patients in the oral progestin group had response to therapy. Nineteen out of the 21 patients in the observational group, had remaining hyperplastic polyps at follow-up biopsy. However, there was no significant difference in this regard between oral MPA and observation-only (p=0.12). Table III shows that the LNG-IUS is effective for all D-score groups. In women treated with low-dose oral progestin, only 25% of the low- or medium-risk cases showed response and none of the high-risk cases had regressed. A very low spontaneous regression rate was demonstrated for the observation group (Table III). No endometrial carcinomas were diagnosed in patients with hyperplastic polyps after 6 months. After the follow-up biopsy, a total of 17 women had simple hysterectomy, three in the cyclic MPA-group and 14 in the observation-group, respectively. The reason for surgery was atypical hyperplasia with D-score <0 for nine women and persisting vaginal bleeding for the remaining. Two patients with atypical hyperplasia and D-score <0 had LNG-IUS and showed response. None of the patients in the LNG-IUS therapy group underwent subsequent hysterectomy. The hysterectomy specimens after surgery all confirmed histological diagnosis of hyperplastic polyps.
WHO classification versus D-score. Table IV shows the distribution of D-score according to WHO classification of the index biopsies in the current study. In 34 (61%) patients, WHO classification (SH and CH) and the D-score (D-score >1 and D-score 0-1) were perfectly consistent, demonstrating low or modest risk of malignant development. In 9 women (15%), the WHO diagnosis and the D-score were discordant, with a WHO diagnosis of atypical hyperplasia and a D-score >1 (Table IV). When all these nine women underwent hysterectomy after six months, microscopy of the endometrium revealed only simple hyperplasia, indicating that the sensitivity of the D-score is superior to that of the WHO classification (22).
Discussion
As far as we are aware of, this is the first study to demonstrate the efficacy of the LNG-IUS as therapy for small endometrial hyperplastic polyps found in routine biopsies independent of tamoxifen therapy. On the other hand, for the groups receiving low-dose oral progestin or no therapy, only 25% and 9% had normal histology after six months, respectively. Although, the number of patients in the present study is limited, our results strongly indicate that therapy response for small hyperplastic polyps may be dependent on the higher progestin concentration provided by the levonorgestrel impregnated intrauterine device.
Diagnosis according to WHO classification of the index biopsy of all included women before start of therapy or observation only, according to result of therapy.
The LNG-IUS used (Mirena®, Bayer) has been shown to release 20 μg of levonorgestrel per 24 h to the uterine mucosa (25, 26). Thus, the endometrial progestin concentration observed in women treated with the LNG-IUS was one hundred-fold higher compared to that in the endometrial mucosa after oral therapy (27). Over the past decades, although not included in the official list of indications, the LNG-IUS has been introduced as an alternative therapy for endometrial hyperplasia. Cohort studies, as well as case series, have shown promising results (28, 29). In a recent prospective multi-center randomized controlled trial comparing LNG-IUS and oral progestin in patients with endometrial hyperplasia without polyps, we found that the LNG-IUS was therapeutic for all patients with endometrial hyperplasia after six months' therapy (13). Comparable studies remain to be performed for endometrial polyps with hyperplastic changes.
According to existing literature, different hormonal approaches have been tested as therapy for endometrial polyps. Vercellini and co-workers reported that gonadotropin-releasing hormone agonist gave a short-term symptomatic relief for endometrial polyps (30). Another study compared the significance of different hormone replacement therapy regimes in endometrial polyp formation (31). This study concluded that progestin with high anti-estrogenic activity was most effective at preventing development of endometrial polyps (31).
Index biopsy by WHO classification according to D-score group.
Little is known about the natural history of endometrial polyps. In the present study, only 2 out of the 21 patients in the observational group had spontaneous regression assessed by Pipelle. When hysteroscopy was used as a diagnostic method, as many as 27% of polyps <1 cm had disappeared after one year of observation (10). That study strongly indicates that the majority of polyps does persist and need to be treated. In the current study, 9 out of the 19 cases with hyperplastic polyps diagnosed after six months in the observational group had atypical hyperplasia, as assessed by WHO classification of the index biopsy. This fact might have contributed to the low frequency of spontaneous regression for these patients.
Recurrence of endometrial polyps after surgical removal has been demonstrated to occur in 15% to 60% after surgical removal (32, 33). On the other hand, in a long-term follow-up study of women with endometrial hyperplasia without polyps, recurrence occurred for half of the cases when the LNG-IUS was removed. In contrast, women who kept the LNG-IUS in situ had no recurrence of hyperplasia (23). In the study of Vercellini and co-workers, symptoms were shown to recur when therapy was terminated (30). Thus, there may be a sustained need for hormonal treatment in order to prevent recurrence of hyperplastic polyps, and therapy with the LNG-IUS may persist for years.
Blind biopsy as performed in the present study is considered as an inaccurate procedure, with high specificity but rather low sensitivity, and can certainly miss pendunculated, as well as sessile, polyps (34). Even ultrasound may miss the diagnosis of endometrial polyps, however, this procedure can be further refined by use of contrast sonography with saline, or hysteroscopy (35). Nevertheless, incidental finding of endometrial polyps is common in routine pathology and there is an obvious need for immediate therapeutic recommendations.
The need for stratification of malignancy risk in patients with hyperplastic polyps cannot be neglected (3, 6, 7). The WHO classification system used as a gold standard for endometrial hyperplasia has been seriously criticized for its lack of reproducibility (36). Whether this system is applicable to polyps has never been determined. Although the D-score classification system demonstrates increased diagnostic accuracy in endometrial hyperplasia compared to the WHO classification (22), we do not know if this system is safe for use far hyperplastic polyps. However, the present result suggests that the method is more sensitive. Even if ultrasonography may contribute to diagnosis of pre-malignant and malignant changes within endometrial polyps, with a sensitivity ranging from 67% to 100% and specificity from 71% to 89%, the need for histological diagnosis will always be mandatory (37, 38). The WHO classification supplemented by the D-score might be important to predict risk of malignant transformation in polyps before therapy (22, 23, 39).
Our results strongly indicate that the LNG-IUS may be effective as therapy for small endometrial hyperplastic polyps and may reduce the need for hysterectomy in these patients. The effect is superior to oral progestin and observation only.
A main weakness of the current study is a lack of diagnostic accuracy due to the fact that hysteroscopy was not performed in order to confirm the presence or absence of polyps after therapy. Future prospective randomized controlled trials including optimal histological and clinical follow-up methods should be conducted to compare efficacy of different therapy regimes for small hyperplastic polyps.
Acknowledgements
We would like to thank Bjørn T.G. Moe, M.Sc., Inger Pettersen B.Sc., and Lena M. M. Lyså B.Sc. for skilled image analyses performed on endometrial specimens. We are grateful for grants from HelseNord.
Footnotes
-
This article is freely accessible online.
- Received February 4, 2014.
- Revision received February 21, 2014.
- Accepted February 25, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
