Prof WG McCluggage, Department of Pathology, Royal Group of Hospitals Trust, Grosvenor Road, Belfast BT12 6BL, Northern Ireland, UK. Email firstname.lastname@example.org
Objectives Endometrial polyps are a common cause of abnormal uterine bleeding. Rarely, a hyperplasia, either complex or atypical in type, is identified within a polyp in a biopsy or polypectomy specimen. Currently, it is not known whether the hyperplasia is likely to be confined to the polyp or also involve nonpolypoid endometrium. We aim to assess the likelihood of hyperplasia being confined to an endometrial polyp.
Design In this study, we identified 32 women from pathology archives in whom endometrial hyperplasia was present within a polyp. The number of endometrial polyps during the study period was 1031 and therefore 3.1% of all endometrial polyps diagnosed during the study period contained a hyperplasia.
Setting A major teaching hospital in the UK.
Methods The biopsies were retrieved from the pathology archives of Royal Group of Hospitals, Belfast, between 2000 and 2006. We traced any follow-up biopsy or hysterectomy specimens to evaluate the status of the surrounding endometrium.
Results The hyperplasias were complex (n= 23) or atypical (n= 9) in type. In 14 of 27 (52%) women in whom nonpolypoid endometrium was available for histological evaluation, either on the original biopsy or in a follow-up specimen, hyperplasia involved the nonpolypoid endometrium, and in three other women, hyperplasia was present in a polyp in follow-up specimens. Women with atypical hyperplasia in a polyp were slightly more likely to have hyperplasia in the surrounding endometrium than those with complex hyperplasia.
Conclusions Our study illustrates that the risk of endometrial hyperplasia in a polyp concurrently involving nonpolypoid endometrium is significant. We suggest a strategy for the management of women with hyperplasia identified within an endometrial polyp in a biopsy or polypectomy specimen.
Endometrial polyps are a common cause of abnormal uterine bleeding in both premenopausal and postmenopausal women.1 Occasionally, an adenocarcinoma, usually endometrioid or serous but sometimes clear cell or some other type, is identified within a polyp and in such circumstances a hysterectomy is usually performed since the carcinoma typically, although not always, involves the surrounding endometrium with or without myometrial or extrauterine involvement.2–5 An uncommon, but not rare, occurrence is the identification of a hyperplasia within an endometrial polyp in a polypectomy or biopsy specimen. As far as we are aware, there have been no studies, which have investigated the likelihood of an endometrial hyperplasia being confined to a polyp and, as such, the clinical significance of hyperplasia within a polyp is not known. In this study, we identified women in whom a hyperplasia was diagnosed within an endometrial polyp in a polypectomy or biopsy specimen. We aimed to ascertain the likelihood of the hyperplasia being confined to the polyp and, to investigate this, we evaluated any nonpolypoid endometrium in the original biopsy or in subsequent biopsy or in hysterectomy specimens. We also wished to correlate the likelihood of a clinically significant lesion being present in the surrounding endometrium with the type of hyperplasia and the menopausal status of the women.
Materials and methods
Women with endometrial hyperplasia within a polyp were identified from the pathology archives of the Royal Group of Hospitals Trust, Belfast, Northern Ireland, by performing a computer-based search using the SNOMED codes endometrium, uterus, hyperplasia, atypical hyperplasia (AH) and polyp. The search was performed between January 2000 and September 2006. All identified cases were reviewed by a gynaecological pathologist (W.G.M.) to verify the primary diagnosis. Most of the women had been reported by the same pathologist. Only those women in whom hyperplasia involved a polyp were included, i.e. cases where hyperplasia and a polyp coexisted but the hyperplasia did not involve the polyp were excluded. Using patient details, any follow-up biopsies or resection specimens were traced.
Simple hyperplasia (SH) was not diagnosed in an endometrial polyp since dilated glands exhibiting proliferative activity and resembling the histological features of SH are a normal finding within a polyp. Complex hyperplasia (CH) was diagnosed when there was significant crowding of glands with an increased gland to stroma ratio. This had to involve more than a few glands and the glands were required to exhibit proliferative activity. AH was diagnosed when, in addition to the features described in CH, there was cytological atypia, usually associated with loss of polarity, rounding of nuclei and prominent nucleoli. Figure 1A,B illustrate examples of CH and AH, respectively, in an endometrial polyp. Women with adenocarcinoma involving an endometrial polyp were excluded from the study.
During the study period, a total of 1031 endometrial polyps were diagnosed. Thirty-two women were identified, in whom endometrial hyperplasia was diagnosed within an endometrial polyp, corresponding to 3.1% of all endometrial polyps. In 30 women, this was in a biopsy (either a pipelle, curettage or polypectomy), and in two women, this was in a hysterectomy specimen. There were 23 women with CH and nine women with AH.
The women ranged from 42 to 81 years (mean 62.7 years). Only one woman was less than the age of 50 years. Thirty of 32 women had endometrial sampling to investigate abnormal uterine bleeding. Of these 30 women, 18 women had at least one further follow-up procedure; 10 women had a hysterectomy with or without unilateral or bilateral salpingo-oophorectomy, 2 women underwent transcervical resection of endometrium (TCRE) and 6 women had repeat endometrial biopsy. In the other two women, hyperplasia was identified within a polyp in a hysterectomy specimen. These women did not have a preceding biopsy.
Of the 21 women in whom CH was present within a polyp in a biopsy specimen, there was follow up in 12 women. In 13 women in whom surrounding endometrium was represented in the original specimen (this included the two hysterectomy women with no preceding biopsy), seven women had CH involving the nonpolypoid endometrium. Seven women with follow up had a hysterectomy and five underwent repeat endometrial sampling. Altogether in the nine hysterectomy specimens, there was no hyperplasia in three, there was CH involving nonpolypoid endometrium in five and there was AH involved nonpolypoid endometrium in one case. In five women with repeat biopsies, three women had no hyperplasia in the follow-up specimen, one woman had CH in polypoid and nonpolypoid endometrium and one woman had AH confined to a polyp. There were nine women with CH, with no follow-up biopsies. In six of these women, nonpolypoid endometrium was represented in the original biopsy; three of these women had CH in the nonpolypoid endometrium.
In summary, in 10 of 20 women (50%), hyperplasia (nine CH, one AH) was present in surrounding nonpolypoid endometrium, either in the initial specimen or at follow up or both.
The initial biopsy findings, follow-up procedures (where applicable) and histological findings of these procedures are summarised in Table 1.
Table 1. Summary of findings in women with CH
Women’s age (years)
Surrounding endometrium represented in original biopsy
Surrounding endometrium involved by hyperplasia in original biopsy
Of the nine women with AH in an endometrial polyp, six women had some form of follow up. In all six women, surrounding nonpolypoid endometrium was not sampled in the initial biopsy. Three women had a hysterectomy, two women underwent TCRE and one woman had a repeat endometrial biopsy. In the hysterectomy specimens, AH was confined to endometrial polyps in two cases, and in the other, there was CH involving the nonpolypoid endometrium. In both women who underwent TCRE, AH was present in nonpolypoid endometrium. In the woman who had follow-up biopsy, there was no hyperplasia. There were three women with AH, with no follow-up biopsies. In one of these, nonpolypoid endometrium was represented in the original biopsy and this exhibited SH.
In summary, six of seven women had hyperplasia at follow up or in nonpolypoid endometrium represented in the original biopsy (one SH, one CH and four AH). However, the hyperplasia only involved nonpolypoid endometrium in four of seven (57%) women.
The initial biopsy findings, follow-up procedures (where applicable) and histological findings of these procedures are summarised in Table 2.
Table 2. Summary of findings in women with AH
Women’s age (years)
Surrounding endometrium represented in original biopsy
Surrounding endometrium involved by hyperplasia in original biopsy
Histological findings in follow up
NA, not applicable.
AH confined to polyp
CH in nonpolypoid endometrium
AH in nonpolypoid endometrium
AH in polypoid and nonpolypoid endometrium
AH confined to polyp
Overall, in 14 of 27 (52%) women in whom surrounding endometrium was sampled at some stage, either on the initial biopsy or in a follow-up specimen, there was hyperplasia in the nonpolypoid endometrium. Hyperplasia involved nonpolypoid endometrium in 10 of 20 (50%) women with CH and in 4 of 7 (57%) women with AH. There were three additional cases where hyperplasia was not present in the surrounding endometrium but which involved a polyp in repeat biopsy or hysterectomy. Adenocarcinoma was not identified in subsequent specimens in any case. Twelve of 21 (57%) women with CH (not including the two women in whom hysterectomy was performed without a preceding biopsy) had some form of follow-up specimen, as had six of nine (67%) women with AH.
Endometrial hyperplasias are proliferative lesions of the endometrial glands with an increased risk of the subsequent development of endometrioid-type (type 1) adenocarcinoma. They are associated with hyperoestrogenic states. Currently, the classification system of endometrial hyperplasia favoured by most pathologists and in widespread use is that which was introduced in 1994 and which is endorsed in the latest World Health Organization (WHO) publication.6,7 This system uses two morphological criteria, glandular architectural complexity and nuclear atypia, and results in four categories of hyperplasia, namely simple nonatypical, complex nonatypical, simple atypical and complex atypical.6,7 In practice, nuclear atypia almost always occurs in endometria with a complex architecture, and it is extremely uncommon to diagnose simple AH. As such, the WHO classification can be condensed into three groups, i.e. SH(simple nonatypical), CH (complex nonatypical) and AH (complex AH).6,7 All three categories of endometrial hyperplasia are associated with an increased risk of subsequent development of endometrioid-type adenocarcinoma. This is low in SH and CH (approximately 1 and 3%, respectively); many of these lesions are self-limiting and regress, especially following hormonal treatment.8,9 The risk of developing endometrioid adenocarcinoma following a diagnosis of AH is difficult to ascertain as treatment, most often surgical but sometimes hormonal, is almost invariably instigated following diagnosis. However, the risk is significant and thought to be in the region of 25–40%.8,10 In some women, the identification of endometrioid adenocarcinoma in a subsequent specimen may reflect undersampling in the original biopsy rather than progression or be a result of the significant inter-observer variation among pathologists in the distinction between AH at the upper end of the spectrum and grade 1 endometrioid adenocarcinoma.11–14
The exact pathogenesis of endometrial polyps is not fully elucidated, but they are thought to originate as a localised hyperplasia of the basalis, perhaps secondary to hormonal influences.15–18 Tamoxifen administration is associated with an increased risk of development of endometrial polyps.19,20 Since both endometrial hyperplasias and polyps are associated with hyperestrogenism, it is not surprising that they may coexist, and occasionally, an endometrial hyperplasia, either complex or atypical, is identified within a polyp in a biopsy or polypectomy specimen.21–23 The pathologist is often asked by the clinician to comment on the likelihood of the hyperplasia being confined to the polyp. As far as we are aware, there are no studies, which have specifically looked at this issue and, as such, no guidance can be given.
We identified 32 women over an approximately 7-year period in whom a hyperplasia was identified within an endometrial polyp. One of the advantages of this study is that almost all of the biopsies were reported by a single gynaecological pathologist. During this period, a total of 1031 endometrial polyps were reported, corresponding to an incidence of hyperplasia occurring in a polyp of 3.1%. Other studies have found the incidence of hyperplasia in endometrial polyps to range between 11 and 29%,21–23 a somewhat higher figure than we found. However, these studies included cases where a diagnosis of SH within a polyp was preferred, a diagnosis we do not make for reasons stated earlier. Additionally, the term hyperplastic polyp is often used loosely to denote proliferative activity, rather than a true hyperplasia, within a polyp.
In 52% of women in whom nonpolypoid endometrium was sampled, either at the time of the initial biopsy or on follow up, hyperplasia was present in the surrounding endometrium. Hyperplasia was slightly more likely to involve the surrounding endometrium in cases where the initial pathology was AH rather than CH. Women in whom AH was diagnosed in an endometrial polyp were more likely to undergo follow up. In our series, only one woman was aged less than 50 years, so we cannot comment on any association between the likelihood of hyperplasia being present in the surrounding endometrium and the menopausal status of the woman. It is perhaps surprising that no carcinomas were identified in the surrounding endometrium during the study. This is of some reassurance and illustrates that, although there is a significant likelihood of hyperplasia being present in the surrounding endometrium, there is little risk of adenocarcinoma. In three of our women, hyperplasia was confined to a polyp during follow up. We have not included these in the final numbers, as it is likely that only part of the polyp was removed during the initial procedure. Alternatively, the hyperplasia may have involved more than one polyp.
An unexpected finding was the discovery of a significant proportion of women (12 of 30) for whom there was no apparent follow up in the form of repeat endometrial sampling. It is possible that, as some of the cases are recent, follow-up biopsy has not yet been performed. In the remaining women, clinical follow up may have been carried out but no biopsy was performed, either because no lesion was identified at hysteroscopy or on scanning or a sample could not be obtained since the endometrium was atrophic. We did not contact clinicians regarding follow up. Other women may have been commenced on hormonal therapy and some may have been lost to follow up.
As far as we are aware, there are no standard recommendations for the management and follow up of women with hyperplasia identified in a polyp in an endometrial biopsy or polypectomy specimen. Given the significant likelihood of hyperplasia involving the nonpolypoid endometrium and the relative risks of CH and AH evolving into an endometrioid adenocarcinoma, we suggest a strategy, which takes into account both the type of hyperplasia and other parameters, such as the desirability to retain fertility and co-morbid factors. With a diagnosis of AH in a polyp, hysterectomy is the preferred treatment. This may be preceded by complete polypectomy (if not already performed) and curettage of the surrounding endometrium to exclude an adenocarcinoma since if diagnosed this would necessitate preoperative imaging and full surgical staging. However, the likelihood of an adenocarcinoma being present seems low. In selected cases of AH in a polyp, where preservation of fertility is desired or where the patient is a significant operative risk due to co-morbid factors, polypectomy and sampling of the surrounding endometrium followed by hormonal therapy may be instigated with close follow up, including regular repeat biopsies. Alternatively, TCRE may be performed, especially if hysterectomy is contraindicated due to co-morbidity. In women with CH within a polyp, polypectomy (if not already performed) and sampling of the surrounding endometrium should be undertaken. If the surrounding endometrium is normal or contains CH, hormonal therapy may be instigated with careful follow up and regular repeat biopsies. Alternatively, a hysterectomy could be performed if the patient wishes. If AH is diagnosed in the surrounding endometrium, management should be as described for AH. Figure 2 illustrates this algorithm for the management of endometrial hyperplasia within a polyp.
Most endometrial biopsy specimens are obtained at outpatient clinics by pipelle or other techniques during the investigation of abnormal uterine bleeding. It is not uncommon for the histopathologist to diagnose an endometrial polyp, which may or may not have been detected by the gynaecologist. On occasions, a diagnosis of hyperplasia within a polyp will be made, the significance of which has not been previously investigated. We have shown that there is a significant likelihood of hyperplasia also involving the surrounding endometrium, although the risk of adenocarcinoma is low. We have suggested a strategy for the follow up and management of women with hyperplasia diagnosed within an endometrial polyp.