Risks of Malignancy among 11,204 Patients with Endometrial Polyp: A Systematic Review and Meta-analysis

Abstract

Objectives:

To evaluate factors associated with malignancy in patients with endometrial polyps.

Materials and Methods:

We conducted electronic database research on PubMed, MEDLINE, EMBASE, COCHRANE, and Google Scholar from inception for all studies on endometrial polyp. After removing duplicates, and title and abstract screening, we had a total of 121 articles and 151 others from screening the reference list. Inclusion criteria included peri and postmenopausal women > 45 years diagnosed histopathologically with endometrial polyp(s). We excluded women with a history of endometrial cancer or hysterectomy.

Results:

Twenty studies were analyzed. Of 11204 patients with endometrial polyp, 287 had malignant polyps (2.75%), 182 (1.8%) had concomitant endometrial hyperplasia with atypia, and 520 (5.2%) had hyperplasia without atypia within the polyp. Menopausal women had a higher risk of pre-malignancy/malignancy than non-menopausal women (OR 5.63 (95CI 3.87, 8.20, I² = 0%, P < 0.001). Endometrial thickness on ultrasound in pre-malignancy/malignancy cases was significantly thicker than in the benign polyp (mean difference 4.2 mm, 95% CI 0.8 to 7.6 mm, I² = 18%, P = 0.02). Women who used tamoxifen or hormone replacement therapy (HRT) had a lower likelihood of endometrial pre-malignancy/malignancy, while women with abnormal uterine bleeding (AUB) had a higher probability of pre-malignancy/malignancy. The odd ratio of having pre-malignancy/malignancy among those who used tamoxifen was 0.50 (95% CI 0.26-0.94: I² 12%, p = 0.03).

Conclusion:

In women with endometrial polyp, menopausal age and thickened endometrium might increase the probability while tamoxifen or HRT use might lower the likelihood of endometrial pre-malignancy/malignancy; and the presence of AUB might signal endometrial pre-malignancy/malignancy.

INTRODUCTION

Endometrial polyp is a common uterine lesion in women.[1] Although most polyps are benign, some are hyperplastic and could progress to malignancy within the polyp itself or in the endometrium.[1] A meta-analysis of eight retrospective studies including 127 patients with the initial diagnosis of atypical endometrial polyp (atypia within an endometrial polyp) showed a 5.6% risk estimate for concurrent endometrial cancer.[2]

Several authors have evaluated factors that contribute to the development of malignancy associated with endometrial polyp. Although polyp size was considered a risk factor, the results of the studies were mixed. In a retrospective study of 472 postmenopausal women, the authors reported premalignant and malignant lesions in 11 (2.33%) cases; endometrial carcinoma in 4 cases (0.84%), and atypical endometrial hyperplasia in 7 others (1.49%). They could not determine any relationship between various polyp sizes (10, 15, 20 mm) and malignancy.[3] However, a retrospective study from Brazil with 1139 patients showed that endometrial polyps measuring more than 15 mm were significantly associated with hyperplasia (14.8%) compared to 7.7% in the group with smaller polyps.[4]

Advanced age has consistently emerged as a significant risk factor.[5] Compared to premenopausal women, postmenopausal women have a higher risk.[6] In addition, those with abnormal uterine bleeding (AUB) are at increased risk to harboring concealed cancer within the polyps than asymptomatic women.[7] Orvieto et al. reported that the use of hormone replacement therapy (HRT) was the only significant contributing factor predisposing women with endometrial polyp to abnormal histopathological changes.[8]

The purpose of our study was to provide a systematic review of the malignancy potential and the risk factors associated with endometrial malignancy in women with endometrial polyp. Specifically, we evaluated the use of tamoxifen, use of HRT, AUB, menopausal status, polyp size, and endometrial thickness as risk factors.

MATERIALS AND METHODS

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [Figure 1]. The protocol was registered in PROSPERO (CRD42022361378). Formal institutional review board approval was not required.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 flow diagram for systematic review of studies on endometrial polyp

Eligibility criteria, information sources, search strategy

We conducted an electronic database search on PubMed, Medline, Embase, Cochrane, and Google Scholar from inception to August 2022 for all studies on endometrial polyp and peri- and postmenopausal women. The following medical search terms and their combinations were used: Endometrial polyp* AND (Perimenopause* OR Postmenopausal* Women) AND (Endometrial Cancer OR [Premalignancy OR Premalignant Feature*]). We included studies in English or French language. Identified studies were exported into Rayyan, a web-based platform for systematic reviews.[9] In addition, we manually verified systematic reviews and meta-analyses on similar topics to identify additional studies. We removed duplicates before title and abstract screening.

We included studies assessing the risk of endometrial cancer in patients with endometrial polyp. Review articles, case reports, case series (n <10), meeting abstracts, posters, letters, opinion articles, abstracts without full text, and non-English or non-French language were excluded. Our inclusion criteria were perimenopausal women over 45 years of age with endometrial polyp(s) who had undergone polypectomy or hysterectomy. We included women from any region or country. Exclusion criteria included women younger than 45 years, previous history of endometrial cancer, women who had had hysterectomy, and pregnant women.

Study selection

Two authors independently screened articles based on titles and abstract (JR and MM). The final list of full texts was reviewed and selected based on inclusion criteria. A third author (SA) resolved any discrepancy. Figure 1 shows the PRISMA selection flow chart.

Data extraction

Once consensus was reached regarding studies to be included in the final analysis, both authors (SA, ES) independently extracted data from selected studies and compiled them into an Excel spreadsheet. We extracted data on study characteristics, including study design, population size, number of participants, diagnostic methods, duration of follow-up, and outcomes including the prevalence of endometrial hyperplasia, associated risk factors (use of tamoxifen, use of HRT, AUB, menopausal status, polyp size, and endometrial thickness), as well as patient characteristics.

Assessment of risk of bias

Quality assessment was performed using the National Heart, Lung and Blood Institute Study Quality Assessment Tool for case reports, case series, or chart reviews [Supplementary Table 1].

Supplementary Table 1.

Bias Evaluation with NEWCASTLE - OTTAWA QUALITY ASSESSMENT SCALE for Cohort studies

Study	Case-cohort representative	Selection of the non- exposed cohort	Ascertainment of exposure	Outcome negative at start	Comparability by design	Comparability by analysis	Outcome assessment	Score
Ciavattini et al.	*	*	*	*	*		*	6
Uglietti et al.	*	*	*	*	*	*	*	7
Baiocchi et al.		*	*	*	*	*	*	6
Ferrazzi et al.		*	*	*	*	*	*	6

Data synthesis

The pooled prevalence was calculated for endometrial carcinoma, hyperplasia with atypia, and hyperplasia without atypia within the polyp and/or the endometrium. The presence of hyperplasia with or without atypia within endometrial polyp or endometrium was considered as premalignancy, while endometrial carcinoma was considered as malignancy. To quantify the associations between various risk factors and endometrial premalignancy/malignancy, we conducted meta-analyses and calculated the odds ratios (ORs) with their 95% confidence intervals (CIs). Individual and pooled estimates with their 95% CI were presented in forest plots. We assessed the heterogeneity in the effect estimates study by calculating I², as well as by inspecting the forest plots. A random-effects model (DerSimonian and Laird) was used when there was substantial heterogeneity, defined as I² > 50. Sensitivity analyses was done to investigate the possible source of heterogeneity and to assess the robustness of the results. A two-sided P ≤ 0.05 was considered statistically significant. Meta-analyses were performed with RevMan web software.

RESULTS

Study selection and study characteristics

We screened 1,660 publications that evaluated the risk of uterine malignancy in patients with endometrial polyp. The study flow of these publications is shown in Figure 1. Of the 1,660 studies originally identified, 1,539 were excluded as they did not include the risk factors of malignancy, and the indication of polyp removal based on the size [Supplementary Table 2]. Of the 68 studies initially considered, we excluded 48 studies that did not fulfill the inclusion criteria. Twenty chart reviews (cross-sectional studies) were ultimately included in the analysis, encompassing 10 studies from the original database research, and 10 others[5,6,7,8,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25] from Lee et al.’s systematic review[26] Characteristics of the included studies are shown in Table 1. The list of excluded studies with the reason of exclusion is presented in Supplementary Table 2.

Supplementary Table 2.

Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies

Criteria	Saveli et al.	Hileeto et al.	Lasmar et al.	Akış et al.	Arslan et al.	Oriveto et al	Bel et al.	Antunes A et al.	Ghoubara et al.
1. Was the research question or objective in this paper clearly stated?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
2. Was the study population clearly specified and defined?	Yes	Yes	Yes	yes	No	Yes	No	Yes	yes
3. Was the participation rate of eligible persons at least 50%?	NM	NM	No	Yes	NM	No	NM	NM	NM
4. Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
5. Was a sample size justification, power description, or variance and effect estimates provided?	NA	NA	NA	NA	NA	NA	NA	NA	NA
6. For the analyses in this paper, were the exposure (s) of interest measured prior to the outcome (s) being measured?	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A
7. Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A
8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)?	N/A	N/A	yes	N/A	N/A	yes	N/A	N/A	N/A
9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?	Yes	Yes	No	No	Yes	Yes	Yes	Yes	Yes
10. Was the exposure (s) assessed more than once over time?	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A
11. Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?	Yes	No	No	No	yes	yes	Yes	Yes	Yes
12. Were the outcome assessors blinded to the exposure status of participants?	NA	NA	NA	NA	NA	NA	NA	NA	NA
13. Was loss to follow-up after baseline 20% or less?	N/A	N/A	NA	N/A	NA	NA	NA	NA	NA
14. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure (s) and outcome (s)?	yes	NM	NM	NM	NM	NM	NM	NM	NM
Criteria	Karakaya et al.		Fernandez- Parra et al.		Wang et al.	Gregoriur et al.	Shushan et al.	Ben-Arie et al.	Dominiques et al.
1. Was the research question or objective in this paper clearly stated?	Yes		Yes		yes	yes	yes	yes	yes
2. Was the study population clearly specified and defined?	Yes		Yes		Yes	Yes	Yes	No	No
3. Was the participation rate of eligible persons at least 50%?	NM		NM		NM	NM	NM	NM	NM
4. Were all the subjects selected or recruited from the same or similar populations (including the same time period)? Were inclusion and exclusion criteria for being in the study prespecified and applied uniformly to all participants?	Yes		Yes		yes	yes	yes	yes	yes
5. Was a sample size justification, power description, or variance and effect estimates provided?	NA		NA		N/A	N/A	N/A	N/A	N/A
6. For the analyses in this paper, were the exposure (s) of interest measured prior to the outcome (s) being measured?	N/A		NA		N/A	N/A	N/A	N/A	N/A
7. Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?	N/A		NA		N/A	N/A	N/A	N/A	N/A
8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)?	yes		N/A		N/A	N/A	N/A	N/A	N/A
9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?	No		Yes		Yes	yes	yes	yes	yes
10. Was the exposure (s) assessed more than once over time?	N/A		N/A		N/A	N/A	N/A	N/A	N/A
11. Were the outcome measures (dependent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?	Yes		Yes		Yes	Yes	Yes	Yes	Yes
12. Were the outcome assessors blinded to the exposure status of participants?	NA		N/A		N/A	N/A	N/A	N/A	N/A
13. Was loss to follow-up after baseline 20% or less?	NA		N/A		N/A	N/A	N/A	N/A	N/A
14. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure (s) and outcome (s)?	NM		1		NM	NM	NM	NM	NM

Table 1.

Characteristics of included studies of patients with endometrial polyp

Author, Year of Publication	Country	Study Design	Study Period	No. patients	Population	Aim
Akış et al. 2022[18]	Turkey	Retrospective cross-sectional	March 2011-March 2016	229	Patients who underwent hysteroscopy after a pre-diagnosis endometrial polyp.	Evaluating polyp size as measured by sonohysterography in predicting premalignant/malignant lesions.
Antunes et al. 2007[12]	Brazil	Retrospective observational Study	January 1998-December 2005	475	Women who underwent hysteroscopy polypectomy.	Evaluate the prevalence of premalignant and malignant polyps, menopausal status, hormone therapy and clinical characteristics in perimenopausal and postmenopausal women
Arslan et al. 2003[21]	Turkey	Retrospective cross-sectional	March 2001-December 2001	181	Postmenopausal women with endometrial polyp.	Characterize postmenopausal women with endometrial polyp and to evaluate their significance.
Baiocchi et al. 2009[13]	Italy	Cohort	January 1995-December 2006	1242	Patients with endometrial polyps and/or endometrial thickness of (≥5 mm) and/or abnormal uterine bleeding who underwent hysteroscopy.	Evaluate the risk of premalignant and malignant changes in endometrial polyps and to investigate whether clinical parameters could predict the histopathological features of some lesions
Bel et al. 2017[23]	France	Retrospective cross-sectional	October 2007- December 2014	631	Patients who underwent hysteroscopy for suspected polyp.	Estimate the prevalence of lesions in menopausal patients with a pre-operative diagnosis of endometrial polyp and evaluate the risk factors for malignancy.
Ben-Arie et al. 2004[19]	Israel	Retrospective observational Study	January 1996 to July 2001.	420	Women where endometrial polyp was suspected by diagnostic hysteroscopy	Determine the pre-malignant and malignant potential of endometrial polyps and to assess whether different clinical parameters are associated with malignancy in the polyps
Ciavattini et al. 2015[22]	Italy	Cohort	January 2010-May 2013	146	Women who underwent transvaginal ultrasound then and hysteroscopy.	Evaluating preperitoneal fat thickness and the risk of premalignant and malignant changes of endometrial polyp in overweight and obese women.
Domingues et al. 2009[20]	Portugal	Retrospective cross-sectional	2004-2007	481	Women who underwent a diagnostic hysteroscopy	Evaluate the risk of malignancy of endometrial poly in postmenopausal women.
Ferrazzi et al. 2009[14]	Italy	Cohort	January 2005-December 2006	1152	Patients treated for endometrial polyps	Evaluate the prevalence of cancer and premalignant lesions in polyps on atrophic endometrium in asymptomatic postmenopausal women
Fernandez-Parra et al. 2006[11]	Spain	Retrospective observational study	May 2002-July 2004	637	Women who underwent hysteroscopy.	Establish the validity of hysteroscopy for predicting cancer in endometrial polyp based on the number, size and hysteroscopic appearance
Ghoubara, et al. 2018[24]	United Kingdom	Retrospective cross-sectional	January 2011-December 2015	431	Women with postmenopausal bleeding	Quantify the prevalence and identify the predictors of hyperplasia and cancer in polyps.
Gregoriou et al. 2009[16]	Greece	Retrospective observational Study	January 2002-December 2006	425	Women who underwent hysteroscopic polypectomy.	Investigate the association of clinical parameters with the histological diagnosis and the prevalence of premalignant and malignant endometrial polyp.
Hileeto et al. 2005[6]	USA	Retrospective cross-sectional	1986-1995	513	Cases with endometrial polyp were retrieved from the pathology database	Investigate the age-group of cases with endometrial polyp and those with malignancy, as well as histological subtype.
Karakaya et al. 2018[25]	Turkey	Retrospective cross-sectional	2007-2016	133	Women aged>65 years with endometrial polyp.	Prevalence of malignant endometrial polyp in a population of geriatric women.
Lasmar et al. 2013[5]	Brazil	Retrospective cross-sectional	January 1, 1999-December 31, 2012	1136	Patients who underwent outpatient hysteroscopy .	Assess the correlation between the size of the polyp and histopathologic diagnosis of hyperplasia or cancer.
Orvieto et al. 1999[8]	Israel	Retrospective cross-sectional	January 1996- December 1997	146	All postmenopausal women referred to the ambulatory gynecological unit.	Characterize endometrial polyp in postmenopausal women and to determine risk factors for concomitant endometrial pathology.
Savelli et al. 2003[10]	Italy	Retrospective cross-sectional	January 1998 -December 2001	509	Patients who consecutively underwent hysteroscopic polypectomy	Determine the rate of benign, hyperplastic, and malignant endometrial polyp and whether clinical data can predict histopathologic outcome
Shushan et al. 2004[17]	Israel	Retrospective observational Study	January 1998-May 2003	300	Women who underwent hysteroscopic polypectomy.	Investigate the frequency of malignant endometrial polyps, and to characterize the hysteroscopic image of these polyps.
Uglietti et al. 2014[27]	Italy	Cohort	1 January 2007-31 December 2010	481	Patients who underwent outpatient hysteroscopy .	Assess frequency of malignancy of malignancy within polyps and identify risk factors for thus condition
Wang al. 2010[15]	China	Retrospective observational Study	January 2000 to February 2006	766	Patients who underwent hysteroscopic polypectomy.	Estimate the prevalence and risk factors of benign, premalignant, and malignancy .

Synthesis of results

In those included 20 chart-review studies, all 11204 patients had their polyp removed with histopathological confirmation. Two hundred and eighty-seven polyps were malignant (2.75%), 182 (1.8%) had hyperplasia with atypia, and 520 (5.2%) had hyperplasia without atypia within the polyp and/or the endometrium [Tables 2 and 3].

Table 2.

Prevalence of endometrial precarcinoma (hyperplasia with and without atypia) and carcinoma among women with endometrial polyp

	Sample size	n (%)
All
Endometrial precarcinoma*	11,204	809 (7.2)
EMP with hyperplasia without atypia**	9922	520 (5.2)
EMP with atypical hyperplasia**	9922	182 (1.8)
Endometrial carcinoma	11,204	287 (2.6)
1999–2010
Endometrial precarcinoma*	8017	555 (6.9)
EMP with hyperplasia without atypia	8017	426 (5.3)
EMP with atypical hyperplasia	8017	129 (1.6)
Endometrial carcinoma	8017	207 (2.6)
2011–2022
Endometrial precarcinoma*	3187	254 (8.0)
EMP with hyperplasia without atypia**	1905	94 (4.9)
EMP with atypical hyperplasia**	1905	53 (2.8)
Endometrial carcinoma	3187	80 (2.5)

*Combined atypical and nonatypical hyperplasia, **Excluding studies by Lasmar et al.[5] and Ciavattini et al.[22] EMP: Endometrial polyp

Table 3.

Summary of prevalence of benign: endometrial polyp, or hyperplasia without atypia in polyp premalignant/malignant: hyperplasia with atypia, or endometrial carcinoma for included studies

Author(s)	Population (Inclusion criteria)	Polyp with endometrial hyperplasia without atypia	Polyp with endometrial hyperplasia with atypia	Polyp with endometrial carcinoma
Akış et al. 2022[18]	229 patients who underwent hysteroscopic polypectomy after pre-diagnosis on saline infusion sonohysterography	13 (3.65)	4 (1.1%)	2 (0.9%)
Antunes A et al. 2007[12]	475 postmenopausal women who underwent hysteroscopic polypectomy	64 (13.5%)	5 (1.1%)	13 (2.7%)
Arslan et al. 2003[21]	181 postmenopausal women with endometrial polyp	30 (16.6%)	4 (2.2%)	None
Baiocchi et al. 2009[13]	1242 postmenopausal women with endometrial polyp	None	15 (1.3%)	45 (3.5%)
Bel et al. 2017[23]	631 menopausal patients over 45 years who underwent hysteroscopic polypectomy	11 (1.74%)	11 (1.74%)	30 (4.75%)
Ben-Arie et al. 2004[19]	420 patients with endometrial polyp (symptomatic and asymptomatic)	48 (11.5%)	14 (3.3%)	13 (3.0%)
Ciavattini et al. 2015[22]*	146 overweight and obese women with endometrial polyp diagnosed at transvaginal ultrasound and then underwent hysteroscopy polypectomy	5 (3.4%)*		7 (4.8%)
Domingues et al. 2009[20]	481 postmenopausal women over 40 years with and without abnormal uterine bleeding	11 (2.3%)	9 (1.8%)	5 (1.0%)
Ferrazzi et al. 2009[14]	1152 asymptomatic postmenopausal women and 770 patients with abnormal uterine bleeding	103 (5.4%)	31 (1.6%)	33 (1.7%)
Fernandez-Parra et al. 2006[11]	637 women with endometrial polyp (hysteroscopic polypectomy or hysterectomy)	None	None	10 (1.5%)
Ghoubara, et al. 2018[24]	431 women with endometrial polyps and postmenopausal bleeding	10 (2.3%)	20 (4.7%)	4 (1.0%)
Gregoriou et al. 2009[16]	425 women with abnormal uterine bleeding or asymptomatic women with endometrial polyp.	75 (14.5%)	6 (1.2%)	10 (1.9%)
Hileeto et al. 2005[6]	513 cases with endometrial polyp	None	None	66 (13%)
Karakaya et al. 2018[25]	133 women over >65 years with pathological diagnosis of endometrial polyp	7 (5.3%)	7 (5.3%)	12 (9.0%)
Lasmar et al. 2013[5]*	1136 asymptomatic patients aged 15–52 years with a hysteroscopic diagnosis of endometrial polyp.	102 (9.0%)*		None
Orvieto et al. 1999[8]	146 postmenopausal women with polyp	11 (7.5%)	4 (2.7%)	None
Savelli et al. 2003[10]	509 patients who underwent hysteroscopic polypectomy over a 48-month period	131 (25.74)	16 (3.14)	4 (0.79)
Shushan et al. 2004[17]	300 patients who underwent hysteroscopic polypectomy	None	None	4 (1.3%)
Uglietti et al. 2014[27]	481 postmenopausal with endometrial polyp on ultrasound. Those diagnosed at hysteroscopy but not suspected at ultrasound were excluded.	53 (4%)	11 (1%)	25 (2%)
Wang al. 2010[15]	766 women who underwent hysteroscopic polypectomy with exclusion criteria of endometrial carcinoma with polypoid features	84 (11%)	25 (3.3%)	4 (0.5%)

*Combined atypical and non-atypical hyperplasia

The odd ratio of having premalignancy/malignancy among those who used tamoxifen was 0.50 (95% CI 0.26–0.94: I² 12%, P = 0.03), [Figure 2a]. In the sensitivity analysis, we excluded the largest weight study by Baiocchi et al.[13] The results showed an OR of 0.74 (95% CI 0.38, 1.44; I² = 0%, P = 0.38). This suggests that tamoxifen use could reduce the probability of endometrial premalignancy/malignancy, although the statistical significance was driven by the largest study [Figure 2b].

Figure 2.

Forest plots for tamoxifen (a); Forest plots for tamoxifen after excluding a study by Baiocchi et al. (b)

Concerning HRT, the OR of having premalignancy/malignancy among those who used HRT was 0.35 (95% CI 0.20–0.61; I² 12%) [Figure 3a]. However, a sensitivity analysis by removing Arslan et al.,[21] the largest weight study, showed an OR of 0.56 (95% CI 0.30, 1.06; I² 0%, P = 0.07) [Figure 3b]. This also suggests that the role of HRT in reducing the probability of endometrial premalignancy/malignancy was driven by the largest study. Among 200 subjects diagnosed with premalignant or malignant conditions, 159 were menopausal. Menopausal women had a likelihood of having premalignancy/malignancy 5.63 (95% CI 3.87–8.20, I² = 0%, P < 0.001) higher than the nonmenopausal women [Figure 4]. A sensitivity analysis by excluding the study by Baiocchi yielded similar results: OR 5.76 (95CI 3.65, 9.08, I² = 0%, P < 0.001) [Figure 5], which confirmed that menopause was a risk factor for endometrial premalignancy/malignancy.

Figure 3.

Forest plots for hormone replacement therapy (HRT) (a); Forest plots for HRT after excluding a study by Arslan et al. (b)

Figure 4.

Forest plots for menopause

Figure 5.

Forest plots for polyp diameter random effect

There were two conflicting studies regarding polyp diameter and the pooled mean difference was inconclusive with a very wide CI and high heterogeneity (–0.49 mm, 95% CI –8.70–7.72 mm, I² = 89%, P = 0.06) [Figure 5]. The endometrial thickness was a risk factor contributing to malignancy potential within polyps. The pooled estimate of endometrial thickness showed that the endometrium in premalignancy/malignancy cases was significantly thicker than in the benign polyp (mean difference 4.21 mm, 95% CI 0.77–7.64 mm, I² = 18%, P = 0.02) [Figure 6].

Figure 6.

Forest plots for endometrial thickness

Women with AUB had a likelihood of having premalignancy/malignancy 1.89 (95% CI 1.06–3.35, I² = 78%, P = 0.03) higher than those without AUB [Figure 7a]. A sensitivity analysis by excluding the largest study by Baiocchi showed an OR of 1.77 (95% CI 0.92, 3.38: I² 79%, P = 0.09) [Figure 7b]. It suggests that endometrial premalignancy/malignancy is more likely with the presence of AUB, although the statistical significance was driven by the largest study.

Figure 7.

Forest plots for abnormal uterine bleeding (AUB) (a); Forest plots for AUB, Sensitivity analysis excluding a study by Baiocchi et al. (b)

Bias assessment for the included studies

Quality assessment

Of 20 studies, all articles explicitly demonstrated their objectives and research questions. With regards to study population, 16 (80%) studies accurately defined their chosen study populations. Several studies did not specify their population.[19,20,21 ,23] Concerning participation rates, only 5% of the articles achieved a participation rate of at least 50%. All articles selected and recruited subjects from the same or a comparable population. In addition, inclusion and exclusion criteria were clearly and systematically clarified in every article. 85% of the articles accurately validated exposure measures consistently and reliably across all study participants. However, 15% failed to maintain consistency in their approach [Supplementary Table 2].

DISCUSSION

The results of our study are similar to those of a systematic review published in 2010[26] suggesting that premalignant and malignant endometrial lesions are more likely to occur in postmenopausal women with endometrial polyp than in premenopausal women.[26] Menopausal women had a 5.6-fold higher likelihood of having premalignancy/malignancy than the nonmenopausal women. The sensitivity analysis yielded similar results (OR 5.76, P < 0.001). This could be related to the fact that advancing age is associated with a lack of progesterone in postmenopausal women. This age-related hormonal shift creates an imbalance between estrogen and progesterone leading to decreased regression as seen in young individuals.[6]

Of ten chart-reviews evaluating the association between body mass index (BMI) and the risks of malignancy,[10,12,14,15,16 ,22,23,24,25,27] only two reported significant associations between BMI and the likelihood of developing cancer[16,17] Similarly, in the previous review, only one study included in their meta-analysis found a statistically significant association between obesity and malignancy. The established association between obesity and endometrial cancer, attributed to the estrogenic environment may not extend to the malignancy potential of endometrial polyp.[16]

Six of eleven chart reviews reported increased malignancy risk in those with abnormal bleeding,[11,12,13,14,15,17,21,23,24,25 ,27] and our pooled result showed that AUB presented more in patients with premalignant/malignant lesions. Patients experiencing bleeding symptoms are more likely to seek medical attention leading to earlier diagnosis.

It has been reported that women with thickened endometrium are at increased risk of harboring cancer cells compared to their counterparts. The pooled estimate of three studies examining this risk[13,24,25] demonstrated a significant association with a mean difference of 4.2 mm (P = 0.02) thicker endometrium in the premalignant/malignant subjects. This could be due to the underlying factors that increase endometrial thickness. A thick endometrium might impede visualization of endometrial polyp on imaging studies, potentially delaying their detection, and providing more time for malignant transformation to occur within the polyp. In addition, without sonohysterography, the presence of a polyp inside the uterine cavity might give an impression of thick endometrium on transvaginal ultrasound.

Eight chart reviews[5,11,14,15,16,17,18,19] explored the association between polyp size and malignancy within endometrial polyps. Five of them found a significant association between polyp size and malignant potential within the polyp.[5,14,15,18,19] They reported a cutoff size of 10–22.5 mm with a size exceeding their cutoff point carried a significant risk of malignant transformation within the polyp.[14] Each study used a different cutoff and hence, we could not calculate the pooled estimate. It has been assumed that there seems to be a direct relationship between the quantity of cells and the mutation rate that leads to the development of malignancy.[28] It is also possible that larger polyps have more blood supply, facilitating the delivery of nutrients and oxygen to rapidly dividing cells facilitating malignant transformation. However, regardless of the presence or absence of a polyp, perimenopausal patients with AUB should undergo endometrial sampling. Empirically, polypectomy is indicated in asymptomatic women with endometrial polyp of ≥10 mm. In a retrospective study, Garcia et al. reported that benign endometrial of over 15 mm tends to recur.[29]

Hypertension and diabetes mellitus, although previously being considered as risk factors for endometrial carcinoma, were not associated with malignant transformation of endometrial polyp in several studies.[8,12,14,15,17,20,22,24,25] Likewise, in agreement with Lee et al.’s study, only one of five studies investigating hypertension in relation to endometrial malignancy risk found a significant association.[10] Biaocchi et al.[13] identified hypertension as a significant factor, while Gregoriou et al. underscored diabetes as an important contributor.[16]

Regarding hormonal use, eight authors studied the relationship between tamoxifen use and polyp malignancy,[10,13,15,16,18,22,23,24] and seven included the use of HRT.[8,13,14,16,19,21,22] All those studies demonstrated that tamoxifen use reduced the likelihood of endometrial premalignancy/malignancy. We also found a lower probability of malignancy in endometrial polyps among HRT users. Because tamoxifen is associated with a reduction in estrogen receptors (ER) and an elevation in progesterone receptors within the polyp, McGurgan et al. postulated that a decrease in apoptosis facilitates the growth of polyp.[30]

It is possible that the lower prevalence of endometrial polyps in women treated with HRT could be due to the use of progestins in HRT. According to a recent systematic review by Tempfer et al., encompassing 8 case–control studies, 2 randomized control studies, and 9 cohort studies on the use of HRT in postmenopause women, the use of HRT, specifically continuous-combined menopausal hormone therapy with synthetic progestins), 9/19 studies showed reduction of the endometrial cancer risk with ORs/HR ranged between 0.24 and 0.71).[31] Studies showed that menopausal hormone therapy (continuous-combined or single progestins) could reduce the endometrial cancer risk in women with higher BMI index.[29]

The prevalence of premalignant endometrial lesions was 6.9% in studies published before 2010[26] and 8% in more recent studies [Table 2]. Similarly, an upward trend in the incidence of endometrial atypical hyperplasia was observed. The prevalence was 1.6% in studies before 2010, and 2.8% in those conducted after 2010. In general, this increasing trend could be due to several factors including obesity, estrogen replacement therapy, delayed childbearing, nulliparity, poor lifestyles including poor diet, sedentary lifestyles or smoking, environmental toxins, or increased awareness and early detection.

Strengths and limitations

The strengths of our study include the comprehensive search strategy adopted encompassing multiple databases and manual searches. The process of study selection, involving independent screening and resolution of discrepancies, further strengthens the reliability of the included literature. In addition, our inclusion criteria were rigorous, restricting the enrollment of patients with endometrial polyp/s or those who underwent endometrial polyp removal with histopathology diagnosis by either biopsy or hysterectomy, ensuring a more reliable diagnostic framework.

Our meta-analysis provides valuable insights into both the prevalence of malignancy in endometrial polyps and specific risk factors that elevate the likelihood of hyperplasia occurring within endometrial polyp, being one of the few observational-based studies in the medical literature. However, despite the thoroughness of our search strategy, we only found 20 chart-review studies, which is considered low-quality evidence and potentially limits the generalizability and interpretation of findings.

Given the nature of the design, we could not establish temporal associations between polyp, risk factors, and the occurrence of malignancy. We were also unable to further explore the types, doses, and frequency of HRT and tamoxifen use and their impact on the progress of malignancy. Due to the low incidence of malignancy, prospective studies require a large number of patients and will take many years to complete. We propose creation of a longitudinal data base with a large number of patients.

CONCLUSIONS

In women with endometrial polyps, menopausal age and thickened endometrium might increase the probability while tamoxifen or HRT use might lower the likelihood of endometrial premalignancy/malignancy; and the presence of AUB might signal endometrial premalignancy/malignancy.

Author contributions

• SA: Data Collection, Literature review search, Analysis, Writing, Review and Editing

• MM: Data Collection, Literature review search, Analysis, Writing, Review and Editing

• HK: Analysis, Writing, Review and Editing

• ES: Conceptualization, Methodology, Data Collection, Analysis, Writing, Review, supervision, and Editing

• JR: Data Collection, Literature review search, Writing, Review and Editing

• TT: Conceptualization, Methodology, Data Collection, Analysis, Writing, Review, supervision, and Editing.

All authors have read and agreed to the final version of the manuscript.

Data availability statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author.

Conflicts of interest

Dr. Togas Tulandi, an editorial board member at Gynecology and Minimally Invasive Therapy, had no role in the peer review process of or decision to publish this article. The other authors declared no conflicts of interest in writing this paper.

Funding Statement

Nil.