Normal endometrial cells in cervical cytology: systematic review of prevalence and relation to significant endometrial pathology

J Med Screen 2008;15:188-198
doi:10.1258/jms.2008.008069
© 2008 Medical Screening Society

 

This Article
Right arrow
Abstract

Freely available
Right arrow
Figures Only
Right arrow

Full Text (PDF)

Right arrow
Alert me when this article is cited
Right arrow
Alert me if a correction is posted
Services
Right arrow
Email this article to a friend
Right arrow

Similar articles in this journal

Right arrow
Similar articles in PubMed
Right arrow
Alert me to new issues of the journal
Right arrow
Download to citation manager
Right arrow
Citing Articles
Right arrow

Citing Articles via Web of Science (1)

Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow
Articles by Canfell, K.
Right arrow
Articles by Beral, V.
Right arrow Search for Related Content
PubMed
Right arrow
PubMed Citation
Social Bookmarking

What’s this?


Original Articles


Karen CanfellDPhil, Sydney Rotary Research Fellow 
,


Cancer Epidemiology Research Unit, Cancer Council New South Wales, 153 Dowling Street, Woolloomooloo, New South Wales, Sydney, Australia and School of Public Health, The University of Sydney, Australia


Yoon Jung KangMPH (Hons), PhD Student
,


Cancer Epidemiology Research Unit, New South Wales, 153 Dowling Street, Woolloomooloo, New South Wales, Sydney, Australia and School of Public Health, The University of Sydney, Australia


Mark ClementsPhD, Research Fellow
,


The National Centre for Epidemiology and Population Health, The Australian National University, Canberra, Australia


Aye Myat MoaMPH, Research Assistant
,


Cancer Epidemiology Research Unit, New South Wales, 153 Dowling Street, Woolloomooloo, New South Wales, Sydney, Australia


Valerie BeralFRS, Director
,


Cancer Epidemiology Unit, The University of Oxford, Oxford, UK

Correspondence to: Dr Karen Canfell, Sydney Rotary Research Fellow, Cancer Epidemiology Research Unit, Cancer Council NSW, 153 Dowling Street, Woolloomooloo, New South Wales, Australia; karenc{at}nswcc.org.au


Objectives To estimate the prevalence of normal endometrial cells (NECs)and the proportion of NECs associated with significant endometrialpathology in conventional and liquid-based cytology (LBC) cervicalsmears; and to assess the association between NECs and clinicalsymptoms in women with endometrial hyperplasia or carcinoma.

Methods Systematic review of the literature and meta-analysis of prevalenceand proportion data. The review was confined to studies reportingon NECs in smears from postmenopausal women or women aged 40+.

Results A total of 22 relevant primary studies were identified from 1970 to 2007. The overall summary estimate for the prevalence of NECs in smears from postmenopausal women or women aged 40+ in all screening smears was 0.4% (95% CI 0.2–0.7%); this was 0.3% (95% CI 0.1–0.5%) and 0.9% (95% CI 0.5–1.4%) for conventional and LBC smears, respectively; P = 0.003 for difference. The overall estimate for the proportion of NECs associated with significant endometrial pathology was 7% (95% CI 4–10%); this was 11% (95% CI 8–14%) and 2% (95% CI 1–2%) for conventional and LBC smears, respectively; P < 0.001 for difference. In women with significant endometrialpathology, the presence of NECs in followed-up women was associatedwith abnormal uterine bleeding in 79% (95% CI 68–87%)of cases.

Conclusion Compared with conventional cytology, LBC may be associated witha higher prevalence of NECs but these are less likely to beassociated with endometrial pathology. This finding might beexplained by more consistent use of sampling instruments forLBC with better access to the endocervical canal or alternativelyby changes over time, broadly coincident with the introductionof LBC, in the population in which NECs are reported. In followed-upwomen with NECs, most endometrial pathology is accompanied bysymptoms, implying that a relatively smaller number of additionalcases are identified through follow-up of asymptomatic women.





INTRODUCTION

Go to previous sectionTop

 INTRODUCTION
Go to next sectionMETHODS

Go to next sectionRESULTS

Go to next sectionDISCUSSION

Go to next sectionACKNOWLEDGEMENTS

Go to next sectionREFERENCES

 

Endometrial carcinoma is one of the most common gynaecological malignancies, occurring predominantly in postmenopausal women.1 The established risk factors include obesity2 and use of oestrogen-only hormone replacement therapy (HRT) and tibolone.3 The established protective factors include use of oral contraceptives, high parity and late age at last birth.1 No widely accepted screening test for endometrial carcinoma exists, but cervical cytology has been found to be of some use in detecting endometrial disease. However, the sensitivity of cervical cytology for detection of endometrial pathology is relatively low, ranging from 25% to 55%.4,5 In one review, it was estimated that 48% of endometrial carcinomas are preceded by abnormal cervical cytology.6 It has been suggested that the likelihood of endometrial cell shedding is increased in higher tumour stage and endocervical involvement. Therefore, if atypical or malignant endometrial cells are identified in the cervical smear, reporting of the presence of such cells is warranted and further investigation is appropriate.4,710

The management of normal endometrial cells (NECs) in cervical cytological smears is less clearly established. An association between the presence of normal exfoliated endometrial cells and significant pathological outcomes was proposed in the 1960s and 1970s after analysis of samples obtained by endocervical aspiration.1113 Endometrial cells are a normal component of the smear in the first half of the menstrual cycle when endometrial breakdown and shedding occurs. Over 60% of smears obtained from women in the first few days of the cycle may have endometrial cells present14 and clusters of endometrial cells of stromal, glandular or mixed origin are observed.15 However, if NECs are observed in smears taken when physiological exfoliation is not expected – the second half of the menstrual cycle – the pathological significance may be greater, although cell shedding at this time can be associated with use of oral contraceptives, intrauterine contraceptive devices or use of HRT.7 Because therates of endometrial carcinoma are highest in postmenopausalwomen, NECs observed in smears from women aged 40+ may be ofgreater clinical significance.

Irrespective of whether or not NECs are observed in cervical cytology, a proportion of patients with significant endometrial disease are symptomatic, presenting with abnormal uterine bleeding and therefore warranting further investigation. Options for management include transvaginal ultrasound, hysteroscopy, endometrial biopsy, and dilatation and curettage. In the UK, transvaginal ultrasound is used as the first-line procedure, followed by further investigation if the endometrial thickness is greater than 4 mm.16,17 Most patients prefer to undergo endometrial assessment with hysteroscopy to rule out the risk of cancer.18 However, the procedure has some associated complications, including uterine perforation.1921 In one study of operative hysteroscopy, the overall complication rate was 3%, with a 1% rate of uterine perforation22 although the complication rate is lower for diagnostic hysteroscopy.19

In a proportion of women, the presence of NECs without symptomscould potentially signal the presence of significant endometrialpathology. The reporting of NECs and the associated managementstrategies must therefore take into consideration the numberof additional cases of endometrial disease detected by activefollow-up of asymptomatic women, beyond those that are detectedsymptomatically. In the UK, initial observations suggest anincrease in the prevalence of NECs since the transition fromconventional to liquid-based cytology (LBC) was initiated in2003 (Prof. Julietta Patnick, personal communication). It ispossible that the reported increase relates to the differencein sampling instrument associated with the transition to LBC.Conventional cytology in the UK prior to the 2003 technologychange was often-performed using the Ayer or (after 1987) theAylesbury wooden spatulas, whereas LBC samples are generallyobtained using a cervical broom. Therefore, more consistentsampling of cells in the endocervix – and potentiallyof the lower uterine segment – may have been a consequenceof the transition to LBC, and this could have increased theprevalence of endometrial cells in the cervical smear.

The objectives of the current study were, first, to performa systematic review of the literature in order to estimate theprevalence of NECs in conventional and LBC smears. The secondobjective was to assess the proportion of NECs associated withsignificant endometrial pathology and quantify any differencesin the underlying rate of significant pathological outcomesbetween conventional and LBC smears. The third objective wasto assess the association between NECs and uterine bleedingin women with significant endometrial pathology, in order toquantify the proportion of women who are diagnosed only throughactive follow-up of NECs, and not via investigation of symptoms.We also summarized the available information on factors thatcan modify the likelihood of observing NECs on the smear, includingthe effects of use of oral contraceptives, intrauterine contraceptivedevices, HRT and tamoxifen.





METHODS

Go to previous sectionTop

Go to previous sectionINTRODUCTION

 METHODS
Go to next sectionRESULTS

Go to next sectionDISCUSSION

Go to next sectionACKNOWLEDGEMENTS

Go to next sectionREFERENCES

 

Two investigators (YJK and AMM) searched Medline and EMBASEfor articles published between 1970 and 2007 relating to NECsin cervical cytology, with the final search performed on 19February 2008. The search terms used as specified for Medlinewere (Pap* smear* OR vaginal smear [MESH] OR cervical cytologyOR LBC OR uterine cervical dysplasia [MESH] OR uterine cervicalneoplasms [MESH] OR cervical intraepithelial neoplasia [MESH]OR CIN) AND (endometrial cells OR endometrial neoplasm* OR endometrialhyperplasia). We also searched the reference lists of identifiedpapers for further relevant source articles. The titles andabstracts were reviewed for inclusion according to the followingcriteria: reporting original data on NECs in smears from postmenopausalwomen or from women aged 40 or older; and presenting data onone or more of the following subjects: the prevalence of NECsin cervical cytology when the majority of smears were takenin the screening context and the total number of smears in thepopulation were precisely reported; the proportion of NECs withassociated endometrial pathology; or the association betweenthe presence of NECs and clinical symptoms in women with endometrialdisease. We excluded conference abstracts and early publicationsfrom multiple reports of the same study. The two investigatorsindependently conducted the search, reviewed titles and abstracts,and extracted data from included studies, with disagreementsand queries resolved at each stage via consultation with a thirdinvestigator (KC).

The smear technology for each study was classified as conventional or LBC. Studies reporting that 85% or more of smears were prepared using a particular technology were classified on that basis. Studies not specifying a technology, or in which both conventional and LBC specimens were included without specifying the relative proportions, were further reviewed and classified according to study period and country. For European and Canadian studies we assumed that studies conducted before 2003 involved conventional cytology. In the USA, ThinPrep LBC (Cytyc Corp, Marlborough, MA, USA) received Food and Drug Administration approval in 1996; by the end of 2001, the market share for ThinPrep in the USA was over 50%, growing to 70% by 2005.23 Therefore, we assumedthat studies conducted in USA before 1997 involved conventionalcytology and studies conducted after 2001 predominantly involvedLBC; for studies conducted in the period 1997–2001, thesmear technology, if unspecified, was classified as predominantlyconventional cytology. Sensitivity analysis was performed toassess the impact of these assumptions. Most studies involvingLBC did not specify the manufacturer, and we were thereforeunable to classify our results on the basis of specific liquid-basedtechnologies.

We classified the population in which NECs were reported for each eligible study as one of the two types – Type 1 populations involved women who were postmenopausal and/or 50 years or older, whereas Type 2 populations involved women aged 40 or older.For studies providing data on NEC prevalence, we classifiedthe study according to whether the prevalence was reported asa percentage of all smears (including those from younger women),or as a percentage of smears taken from the population in whichNECs were reported; and separate analysis was performed forthe two prevalence estimates. For studies that included premenopausalwomen over 40 years of age and provided information on whetherNECs were observed in the first or second half of the menstrualcycle, we extracted results only for NECs in the second halfof the cycle.

A number of studies conducted in the USA reported on NEC prevalence in different smear series before and after the implementation of The Bethesda System (TBS) 2001 for cytological reporting. Prior to the implementation of TBS 2001, NECs in postmenopausal women were reported, whereas TBS 2001 recommended reporting of NECs in all women over 40 years of age.8 For some studies,relevant data could be extracted from both smear series. Becausethe year in which LBC became the dominant technology in theUSA (2001) broadly coincided with the change in the cytologyreporting to TBS 2001 (and hence a chance in the group of womenin which NECs were reported), we also performed stratified analysisaccording to the study population in which NECs were reported.Additionally, to account for potential confounding by studypopulation in the assessment of differences due to smear technology,we performed multivariate logistic regression to adjust forthe study population.

For the purposes of the review ‘significant endometrialpathology’ was defined as endometrial hyperplasia (generallyincluding simple endometrial hyperplasia with atypia and complexendometrial hyperplasia) or endometrial carcinoma. Because thevarious studies did not consistently report on the associationbetween NECs and benign endometrial polyps, we did not includethis outcome in our review.

Data were analysed using Stata 9 software (StataCorp, TX, USA) and R version 2.7.1 (R Foundation for Statistical Computing, Vienna, Austria). Because the data were highly heterogeneous, we used random effects meta-analysis to calculate pooled estimates of prevalence and proportions with 95% CI. In order to test for the significance of differences in prevalence and proportions by smear technology while adjusting for study population, we used generalized linear mixed regression models with binomially distributed outcomes, a logit link and a global random effect to calculate P values for univariate and multivariate analyses.





RESULTS

Go to previous sectionTop

Go to previous sectionINTRODUCTION

Go to previous sectionMETHODS

 RESULTS
Go to next sectionDISCUSSION

Go to next sectionACKNOWLEDGEMENTS

Go to next sectionREFERENCES

 

The initial search identified 411 studies, 37 of which were potentially eligible according to the title/abstract and 22 of which met the inclusion criteria after full review14,2444 (Table 1). Of these, 12 reported on the prevalence of NECs in cervical cytology, 19 reported on the proportion of women with NECs who had significant endometrial pathology and 10 reported on the association between the presence of NECs and clinical symptoms in women with significant endometrial pathology. Because the identified studies involved retrospective review of information from clinical smear series, they were all considered to be only of ‘low’ or ‘fair’ quality according to the QUADAS criteria for diagnostic accuracy studies (which specify that ‘high quality’ studies involve prospective recruitment and blind interpretation of the reference standard with respect to the test).45



View this table:
[in this window]
[in a new window]
Table 1 List of included studies


Prevalence of NECs in cervical cytology

Figure 1A and B detail the results for the prevalence of NECs in cervical smears taken in the screening context. Figure 1A provides results for the majority of the studies in which prevalence was reported as a percentage of all smears, including those from younger women; whereas Figure 1B provides resultsfor the smaller number of studies reporting on prevalence asa percentage of the population in which NECs were reported;either postmenopausal women (type 1) or women aged 40+ (type2).



View larger version (32K):
[in this window]
[in a new window]
Figure 1 (A) Summary estimate for prevalence of normal endometrial cells (NECs) in older women as a percentage of all smears (from women of all ages). (B) Summary estimate for prevalence of NECs in older women in the postmenopausal period or women aged 40+ years

 

The reported prevalence of NECs in all smears varied between 0.1% and 1.4% (Figure 1A) and the overall summary estimate was 0.4% (95% CI 0.2–0.7%). For conventional smears, the corresponding prevalence was 0.3% (95% CI 0.1–0.5%), whereas for LBC the prevalence was 0.9% (95% CI 0.5–1.4%); this difference was significant in a univariate regression analysis (P = 0.003). For studies reporting on NEC prevalence only within type 1 or type 2 populations, the range reported was between 0.2% and 3.0% (Figure 1B) and the overall summary estimate was 0.7% (95% CI 0.4–1.4%); for conventional smears prevalence was 0.6% (95% CI 0.3–1.1%), whereas for LBC the prevalence was 1.2% (95% CI 0.3–4.2%); this difference was not significant (P = 0.23).

In the analysis of NEC prevalence, we were unable to perform multivariate analysis to adjust for the type of study population because studies reporting results for all combinations of smear technology and population were not available. However, if studies were classified by study population rather than smear technology, univariate analysis indicated a significant difference in prevalence (P < 0.001). Therefore, the study population appears to havean impact on NEC prevalence that may be independent of the smeartechnology.


Histological outcomes following NECs

Table 2 and Figure 2 summarize the results for thestudies reporting on histological outcomes following smear reportsof NECs. The majority of studies were retrospective audits ofcytology and pathology records for which referral practicesvaried, and the rate of follow-up of NECs varied between 18%and 79% (in studies in which this was specified).



View this table:
[in this window]
[in a new window]
Table 2 Summary of studies reporting endometrial histology outcomes after the identification of NECs in cytological smears



View larger version (28K):
[in this window]
[in a new window]
Figure 2 (A) Summary estimate for overall proportion of followed-up women with normal endometrial cells (NECs) found to have significant endometrial pathology. (B) Summary estimate for overall proportion of followed-up women with NECs found to have endometrial carcinoma

 

The proportion of followed-up women who were found to have any significant endometrial pathology ranged from 1% to 24% in the various studies. The overall estimate for the proportion was 7% (95% CI 4–10%) (Figure 2A); for conventional smears it was 11% (95% CI 8–14%) and for LBC it was 2% (95% CI 1–2%). The difference in proportions associated with smear technology was significant both in the univariate analysis and after adjusting for study population (P < 0.001). If only endometrial carcinoma outcomes were considered, the overall estimate of the proportion of NECs with underlying carcinoma was 3% (95% CI 2–5%) (Figure 2B). In the case of conventional smears, the summary estimate was 4% (95% CI 3–7%), whereas for liquid-based smears, it was 1% (95% CI 1–2%); the difference associated with the smear technology was significant in the univariate analysis (P = 0.001), but in this case was not significant after adjusting for population (P = 0.08).


Sensitivity analysis for smear technology

We performed sensitivity analysis to assess the impact on ourestimates of including post-1996 studies in which the test technologyhad been classified on the basis of study period and country(and not explicitly specified in the primary study). For bothestimates of NEC prevalence and of rates of significant endometrialpathology, the results did not change substantially when thesestudies were excluded. The sensitivity analysis was remarkablyrobust compared to all of the main study findings. The onlyqualitative change was weaker support for a difference in theprevalence of NECs in all smears between conventional smearsand LBC, with a p value of 0.05 compared to a p value of 0.003in the main analysis.


Association between NECs and symptoms

All eligible studies in the review of the association between NECs and clinical symptoms in women with significant endometrial pathology involved only postmenopausal women, and Table 3 and Figure 3 summarize the results. In women with NECs who also had significant pathology, the proportion with symptoms (abnormal uterine bleeding) varied from 43% to 100% and the summary estimate was 79% (95% CI 68–87%) (Figure 3A). For endometrial carcinoma, the summary estimate for the proportion with symptoms was 77% (95% CI 62–87%) (Figure 3B).It should be noted that several studies reported symptoms inall women with significant pathology or carcinoma, and thesehad to be excluded from the random effects meta-analysis, fortechnical reasons. Therefore, in this sense the summary estimatesfor the proportion of women with symptoms may be conservative.However, it is likely that the presence of symptoms influencedthe probability of follow-up in women with NECs, and it remainspossible that undetected endometrial pathology was present insome asymptomatic women with NECs who were not followed-up.Therefore, these findings can be considered only in relationto the existing management practices with NECs.



View this table:
[in this window]
[in a new window]
Table 3 Summary of studies reporting the proportion of symptomatic women, in women with normal endometrial cells and significant pathology



View larger version (35K):
[in this window]
[in a new window]
Figure 3 (A) Summary estimate for proportion of followed-up women with normal endometrial cells (NECs) found to have significant endometrial pathology who also present with symptoms. (B) Summary estimate for proportion of followed-up women with NECs found to have endometrial carcinoma who also present with symptoms


Effect of other factors

Use of oral contraceptives decreases the prevalence of endometrial cells observed in the first half of the cycle14 and may result in NECs or breakthrough bleeding in the latter half of the cycle.7 In women using an intrauterine contraceptive device, the prevalence of endometrial cells increases throughout the menstrual cycle, with an average prevalence of 26% in the first half of the cycle and 6% in the second half of the cycle.14 Use of tamoxifen may be associated with endometrial hyperplasia46 and it may also be associated with clusters of parabasal cells that can be mistaken for endometrial cells.4

Use of HRT may be associated with endometrial thickening, abnormal uterine bleeding and increased prevalence of NECs. However, in women using HRT, abnormal uterine bleeding does not appear to have a strong association with histological outcomes.47 Among women with NECs, rates of significant endometrial pathology in HRT users and non-users are not well-defined.4,29,33,34,37 Mount et al.33 reviewed results from 52,662 smears from womenwho were postmenopausal or over the age of 50, and found thatHRT users had a significantly higher prevalence of NECs (thestudy did not report on whether HRT-users were current or ever-usersat the time of cytological sampling, or on the type of HRT).The overall NEC prevalence was 1.1% in this population, witha prevalence of 1.52% in HRT-users and 0.097% in non-users.

Therefore, all four factors considered above may increase thelikelihood of observing NECs in the smear. Use of oral contraceptivesor HRT may increase the prevalence of NECs, but the impact onthe positive predictive value for endometrial disease is notclear, as these factors can modify the risk of developing endometrialcancer.





DISCUSSION

Go to previous sectionTop

Go to previous sectionINTRODUCTION

Go to previous sectionMETHODS

Go to previous sectionRESULTS

 DISCUSSION
Go to next sectionACKNOWLEDGEMENTS

Go to next sectionREFERENCES

 

The significance of NECs in cervical cytology depends on severalfactors, including the age and menopausal status of the womanand, if premenopausal, the phase of the menstrual cycle duringwhich sampling was performed. Based on data from multiple studies,we estimated the overall prevalence of NECs as a percentageof the entire screening population to be 0.4% of which 7% offollowed-up women are found to have significant pathology. Ifonly conventional smears are considered, the estimated prevalenceof NECs is 0.3%, of which 11% are associated with significantpathology. In contrast, for LBC smears, the estimated prevalenceof NECs is 0.9%, of which only 2% are associated with significantpathology. These results imply that, irrespective of the smeartechnology, the rate of underlying endometrial pathology detectableby NECs in screening smears is approximately 20–30 per100,000 on a population basis. The results also imply that thepositive predictive value of NECs for significant pathologyis reduced when LBC is used as the smear technology.

Our analysis has several limitations, the most important ofwhich is that the primary studies reporting on NEC prevalenceinvolved a retrospective review of smear records, and were conductedin different countries at different times. Data for conventionalcytology and in LBC were obtained from different studies, soin the comparison of smear technologies we were not able tocontrol for differences in histological follow-up rates or otherdifferences in management practice in various settings. Therefore,we cannot exclude the possibility that factors other than smeartechnology had an impact on NEC reporting rates. Potentially,the most important of these is the population in which NECswere reported. Although our finding that the rate of underlyingsignificant endometrial pathology in NECs from LBC smears issignificantly lower than for conventional smears was made inthe context of adjusting for study population, it remains possiblethat the lower positive predictive value of NECs in LBC smearsis related to a change in the population in which NECs are reported(i.e. in women aged 40+ rather than in postmenopausal women),which was broadly coincident in time with the introduction ofLBC technology. In addition, not all primary studies reportedon the type of smear technology used, and in those cases webased the classification of smear technology on the study periodand country. However, sensitivity analysis showed that our mainresults were not sensitive to this factor. Despite these limitations,the included primary studies do represent the best availableinformation on outcomes following a finding of NECs in cervicalcytology, and our review represents a synthesis of that information.

Our findings are consistent with the hypothesis that smearsusing LBC technology are more likely to involve sampling instrumentswith better access to the endocervical canal (at least in thesettings in which the primary studies were conducted), and aretherefore more likely to sample NECs. However, such ‘directlysampled’ endometrial cells may show morphological distinctionfrom exfoliated endometrial cells – directly sampled endometrialcells are likely to be well-preserved and include endometrialstroma and epithelial cells with crush artefact, whereas exfoliatedendometrial cells may be degenerate single cells or balls ofcells (Dr Huw Llewellyn, personal communication). It is possiblethat the diagnostic implications of the two sampling scenariosare different.

In postmenopausal women with significant endometrial pathology, we found that the presence of NECs in followed-up women is associated with abnormal uterine bleeding in the majority of cases. In the context of a management system that routinely refers women with postmenopausal bleeding for further evaluation, this finding implies that a relatively small number of additional cases are identified using the existing practices for follow-up of asymptomatic women with NECs. However, it should be borne in mind that the estimates for this proportion are based on relatively small numbers and that other considerations are also relevant in determining whether NECs are reported in cervical cytology. For example, shed endometrial cells are a differential diagnosis of squamous cell carcinoma in situ, and in LBC it may be more difficult to distinguish between endometrial cells and carcinoma in situcells compared with conventional cytology (Dr Huw Llewellyn,personal communication).

Currently, different approaches to the reporting and management of asymptomatic women with NECs are taken in various countries. Reporting of NECs in women over 40 years of age is currently recommended in England, the USA and New Zealand. In England, it is recommended that NECs observed outside the first 12 days of the cycle in women over 40 years of age be recorded, but the decision to refer is a clinical one, as all the relevant information may not have been available on the pathology request form.7 In the USA, the 1991 Bethesda System for cytological classification required the recording of the presence of endometrial cells only in postmenopausal women. However, this was modified in the 2001 Bethesda System, which required the reporting of the presence of endometrial cells in women 40 years of age or older, regardless of menopausal status or date of last menstrual period.8 The rationale for the change in reporting structure was the inability of the laboratory to reliably report on menopausal status or menstrual phase and other risk factors for NECs, such as use of exogenous hormones.8 Therefore, it was intended that the clinician used the information on the presence of NECs to interpret its significance and recommend further management in the context of other information. In New Zealand, a modified version of TBS 2001 is used for cytology classification and the recommendations specify that NECs are reported for all women 40 years of age or older, accompanied by a clinical note requesting that the clinician correlate the finding with any symptomatology of uterine pathology and refer accordingly.9 In the case of Australia, the 2005 National Health and Medical Research Council (NHMRC) guidelines for the management of asymptomatic women with abnormal cervical cytology specify that NECs in asymptomatic premenopausal and postmenopausal women should not be reported.10

In summary, this review has found that compared with conventionalcytology, LBC may be associated with a higher prevalence ofNECs, but these are less likely than in conventional smearsto be associated with significant endometrial pathology. Mostcases of significant endometrial pathology in followed-up womenwith NECs are accompanied by clinical symptoms, implying thatunder existing management practices, a relatively small numberof additional cases are identified via the follow-up of asymptomaticwomen with NECs. These findings emphasize the importance ofassessing symptomatic status at the time of smear taking, andof following-up symptomatic women, whether or not NECs are observedin cervical cytology.

 

 

 

 

 

 




ACKNOWLEDGEMENTS

Go to previous sectionTop

Go to previous sectionINTRODUCTION

Go to previous sectionMETHODS

Go to previous sectionRESULTS

Go to previous sectionDISCUSSION

 ACKNOWLEDGEMENTS
Go to next sectionREFERENCES

 

We thank Prof. Julietta Patnick and Dr Huw Llewellyn for constructivereview of the manuscript.

Funding source: This project was funded by the National HealthService Cervical Screening Programme in England.

Accepted for publication October 6, 2008.




REFERENCES

Go to previous sectionTop

Go to previous sectionINTRODUCTION

Go to previous sectionMETHODS

Go to previous sectionRESULTS

Go to previous sectionDISCUSSION

Go to previous sectionACKNOWLEDGEMENTS

 REFERENCES


  1. Bray F, dos SSI, Moller H, et al. Endometrial cancer incidence trends in Europe: underlying determinants and prospects for prevention. Cancer Epidemiol Biomarkers Prev 2005;14:1132–42[Abstract/Free Full Text]
  2. Reeves GK, Pirie K, Beral V, et al. Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study. Br Med J 2007;335:1134[Abstract/Free Full Text]
  3. Beral V, Bull D, Reeves G. Endometrial cancer and hormone-replacement therapy in the Million Women Study. Lancet 2005;365:1543–51[Medline]
  4. Fadare O, Ghofrani M, Chacho MS, et al. The significance of benign endometrial cells in cervicovaginal smears. Adv Anat Pathol 2005;12:274–87[Medline]
  5. Mitchell H, Giles G, Medley G. Accuracy and survival benefit of cytological prediction of endometrial carcinoma on routine cervical smears. Int J Gynecol Pathol 1993;12:34–40[Medline]
  6. Fukuda K, Mori M, Uchiyama M, et al. Preoperative cervical cytology in endometrial carcinoma and its clinicopathologic relevance. Gynecol Oncol 1999;72:273–7[Medline]
  7. Johnson EJ, Patnick J. Achievable standards, benchmarks for reporting, and criteria for evaluating cervical cytopathology. Second edition including revised performance indicators. Cytopathology 2000;11:212–41[Medline]
  8. Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 2002;287:2114–9[Abstract/Free Full Text]
  9. National Cervical Screening Programme (NZ). Bethesda 2001: New Zealand Modified. Codes, Descriptors and Assessment of Sample Adequacy for Cytology Laboratories. Wellington, New Zealand: Ministry of Health; 2005
  10. National Health and Medical Research Council. Screening to Prevent Cervical Cancer: Guidelines for the Management of Asymptomatic Women with Screen Detected Abnormalities. See http://www.nhmrc.gov.au/publications/synopses/wh39syn.htm (last checked 15 February 2006)
  11. Liu W, Barrow MJ, Spitler MF, et al. Normal exfoliation of endometrial cells in premenopausal women. Acta Cytol 1963;7:211–4[Medline]
  12. Ng AB. The cellular detection of endometrial carcinoma and its precursors. Gynecol Oncol 1974;2:162–79[Medline]
  13. Ng AB, Reagan JW, Hawliczek S, et al. Significance of endometrial cells in the detection of endometrial carcinoma and its precursors. Acta Cytol 1974;18:356–61[Medline]
  14. Vooijs GP, van der GY, Vooijs MA. The presence of endometrial cells in cervical smears in relation to the day of the menstrual cycle and the method of contraception. Acta Cytol 1987;31:427–33[Medline]
  15. Chang BS, Pinkus GS, Cibas ES. Exfoliated endometrial cell clusters in cervical cytologic preparations are derived from endometrial stroma and glands. Am J Clin Pathol 2006;125:77–81[Abstract/Free Full Text]
  16. Guruwadayarhalli B, Jones SE, Srinivasan V. Hysteroscopy in the diagnosis of postmenopausal bleeding. Menopause Int 2007;13:132–4[Abstract/Free Full Text]
  17. Williams SC, Lopez C, Yoong A, et al. Developing a robust and efficient pathway for the referral and investigation of women with post-menopausal bleeding using a cut-off of < or =4 mm for normal thickness. Br J Radiol 2007;80:719–23[Abstract/Free Full Text]
  18. Timmermans A, Opmeer BC, Veersema S, et al. Patients’ preferences in the evaluation of postmenopausal bleeding. BJOG 2007;114:1146–9[Medline]
  19. Paschopoulos M, Polyzos NP, Lavasidis LG, et al. Safety issues of hysteroscopic surgery. Ann N Y Acad Sci 2006;1092:229–34[Medline]
  20. Nappi C, Di Spiezio SA, Greco E, et al. Prevention of adhesions in gynaecological endoscopy. Hum Reprod Update 2007;13:379–94[Abstract/Free Full Text]
  21. Stamatellos I, Koutsougeras G, Karamanidis D, et al. Results after hysteroscopic management of premenopausal patients with dysfunctional uterine bleeding or intrauterine lesions. Clin Exp Obstet Gynecol 2007;34:35–8[Medline]
  22. Shveiky D, Rojansky N, Revel A, et al. Complications of hysteroscopic surgery: “Beyond the learning curve”. J Minim Invasive Gynecol 2007;14:218–22[Medline]
  23. Dobosz J. Testing Positive for Growth. MA, USA: Cytyc Corporation, 2008. See http://www.forbes.com/2005/03/30/cz_jd_0330gurusow_inl.html (accessed May 2008)
  24. Cherkis RC, Patten SFJr., Andrews TJ, et al. Significance of normal endometrial cells detected by cervical cytology. Obstet Gynecol 1988;71:242–4[Medline]
  25. Yancey M, Magelssen D, Demaurez A, et al. Classification of endometrial cells on cervical cytology. Obstet Gynecol 1990;76:1000–5[Medline]
  26. Gomez-Fernandez CR, Ganjei-Azar P, Capote-Dishaw J, et al. Reporting normal endometrial cells in Pap smears: an outcome appraisal. Gynecol Oncol 1999;74:381–4[Medline]
  27. Gray JA, Nguyen GK. Cytologic detection of endometrial pathology by Pap smears. Diagn Cytopathol 1999;20:181–2[Medline]
  28. Chang A, Sandweiss L, Bose S. Cytologically benign endometrial cells in the papanicolaou smears of postmenopausal women. Gynecol Oncol 2001;80:37–43[Medline]
  29. Ashfaq R, Sharma S, Dulley T, et al. Clinical relevance of benign endometrial cells in postmenopausal women. Diagn Cytopathol 2001;25:235–8[Medline]
  30. Montz FJ. Significance of ‘normal’ endometrial cells in cervical cytology from asymptomatic postmenopausal women receiving hormone replacement therapy. Gynecol Oncol 2001;81:33–9[Medline]
  31. Sarode VR, Rader AE, Rose PG, et al. Significance of cytologically normal endometrial cells in cervical smears from postmenopausal women. Acta Cytol 2001;45:153–6[Medline]
  32. Wu HH, Schuetz MJIII, Cramer H. Significance of benign endometrial cells in Pap smears from postmenopausal women. J Reprod Med 2001;46:795–8[Medline]
  33. Mount SL, Wegner EK, Eltabbakh GH, et al. Significant increase of benign endometrial cells on Papanicolaou smears in women using hormone replacement therapy. Obstet Gynecol 2002;100:445–50[Medline]
  34. Brogi E, Tambouret R, Bell DA. Classification of benign endometrial glandular cells in cervical smears from postmenopausal women. Cancer 2002;96:60–6[Medline]
  35. Cai N, Spitzer M, Wasserman PG. Endometrial cells in pap smears of postmenopausal women: cytohistologic correlation. J Low Genit Tract Dis 2003;7:250–3[Medline]
  36. Browne TJ, Genest DR, Cibas ES. The clinical significance of benign-appearing endometrial cells on a Papanicolaou test in women 40 years or older. Am J Clin Pathol 2005;124:834–7[Abstract/Free Full Text]
  37. Simsir A, Carter W, Elgert P, et al. Reporting endometrial cells in women 40 years and older: assessing the clinical usefulness of Bethesda 2001. Am J Clin Pathol 2005;123:571–5[Abstract/Free Full Text]
  38. Thrall MJ, Kjeldahl KS, Savik K, et al. Significance of benign endometrial cells in papanicolaou tests from women aged >or = 40 years. Cancer 2005;105:207–16[Medline]
  39. Bean SM, Connolly K, Roberson J, et al. Incidence and clinical significance of morphologically benign-appearing endometrial cells in patients age 40 years or older: the impact of the 2001 Bethesda System. Cancer 2006;108:39–44[Medline]
  40. Siebers AG, Verbeek AL, Massuger LF, et al. Normal appearing endometrial cells in cervical smears of asymptomatic postmenopausal women have predictive value for significant endometrial pathology. Int J Gynecol Cancer 2006;16:1069–74[Medline]
  41. Aslan DL, Crapanzano JP, Harshan M, et al. The Bethesda System 2001 recommendation for reporting of benign appearing endometrial cells in Pap tests of women age 40 years and older leads to unwarranted surveillance when followed without clinical qualifiers. Gynecol Oncol 2007;107:86–93[Medline]
  42. Beal HN, Stone J, Beckmann MJ, et al. Endometrial cells identified in cervical cytology in women > or = 40 years of age: criteria for appropriate endometrial evaluation. Am J Obstet Gynecol 2007;196:568–5
  43. Kapali M, Agaram NP, Dabbs D, et al. Routine endometrial sampling of asymptomatic premenopausal women shedding normal endometrial cells in Papanicolaou tests is not cost effective. Cancer 2007;111:26–33[Medline]
  44. Moroney JW, Zahn CM, Heaton RB, et al. Normal endometrial cells in liquid-based cervical cytology specimens in women aged 40 or older. Gynecol Oncol 2007;105:672–6[Medline]
  45. Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25[Medline]
  46. d’Arailh AS, Michy T, Pioud R, et al. [Uterine abnormalities in non-menopausal women received tamoxifen for breast cancer adjuvant therapy.]. Gynecol Obstet Fertil 2007;35:1215–9[Medline]
  47. Hickey M. Bleeding with menopausal hormone therapy: physiological or pathological? Menopause Int 2007;13:188–90[Abstract/Free Full Text]


CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    What’s this?






This Article
Right arrow
Abstract

Freely available
Right arrow
Figures Only
Right arrow

Full Text (PDF)

Right arrow
Alert me when this article is cited
Right arrow
Alert me if a correction is posted
Services
Right arrow
Email this article to a friend
Right arrow

Similar articles in this journal

Right arrow
Similar articles in PubMed
Right arrow
Alert me to new issues of the journal
Right arrow
Download to citation manager
Right arrow
Citing Articles
Right arrow

Citing Articles via Web of Science (1)

Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow
Articles by Canfell, K.
Right arrow
Articles by Beral, V.
Right arrow Search for Related Content
PubMed
Right arrow
PubMed Citation
Social Bookmarking

What’s this?

 

Leave a Comment