Fax: (413) 794–8166
Adherence to practice guidelines for atypical glandular cells on cervical cytology
Article first published online: 12 JUN 2012
Copyright © 2012 American Cancer Society
Volume 121, Issue 1, pages 47–53, January 2013
How to Cite
Sharpless, K. E., King, C. R. and Schnatz, P. F. (2013), Adherence to practice guidelines for atypical glandular cells on cervical cytology. Cancer Cytopathology, 121: 47–53. doi: 10.1002/cncy.21211
- Issue published online: 14 JAN 2013
- Article first published online: 12 JUN 2012
- Manuscript Accepted: 4 MAY 2012
- Manuscript Revised: 24 APR 2012
- Manuscript Received: 3 JAN 2012
Atypical glandular cells (AGC) on cervical cytology are high–risk, requiring an extensive evaluation. Compliance with practice guidelines for AGC, however, has been low. Some AGC cytology reports contain cytopathologist recommendations for evaluation. This study determines whether evaluation rates for AGC have improved over time, and whether cytopathologists' recommendations correlate with the types of evaluation women receive.
Evaluation rates from 284 women with AGC (2004-2007) were compared with findings from 1998-2001. Rates of evaluations were compared based on cytology report recommendations.
A total of 76.1% of the AGC cases had histologic sampling, and 58.8% had a comprehensive evaluation. These rates are higher than those from 1998-2001 (63.5% and 35.8%, respectively; P<.01). Rates of evaluations of women with AGC “favor neoplasia” did not increase between the 2 time periods. Between 2004-2005 and 2006-2007, rates of comprehensive initial evaluations and endometrial sampling in women ≥35 years of age did not increase. Of the AGC reports that did contain cytopathologist recommendations, 28% were consistent with practice guidelines, 26% recommended an incomplete histologic evaluation, and 46% recommended repeat cytology. Women whose AGC report recommended a comprehensive evaluation or any histologic evaluation were more likely to have a comprehensive work–up (79%) than those whose reports did not contain recommendations (55%, P <0.01) or recommended repeat cytology (51%, P<0.02).
Adherence to practice guidelines for the evaluation of women with AGC has improved but continues to be suboptimal. Our findings suggest that continuing education and including practice guidelines on AGC cytology reports may improve compliance. Cancer (Cancer Cytopathol) 2013;121:47–53 © 2012 American Cancer Society.
The finding of atypical glandular cells (AGC) on cervical cytology is rare (incidence <1%) and is associated with a high rate of disease (approximately 29%), including a 5% risk of malignancy.1 Diseases associated with AGC Paps originate from squamous or glandular tissue and from a variety of sites, including the cervix, endometrium, and ovaries, along with the urinary and gastrointestinal tracts. The types of disease associated with AGC vary by age. In one study, most women <35 years of age with an AGC Pap had cervical intraepithelial neoplasia (CIN) 2 or 3 (88% of the diagnoses); women ≥35 years of age had a greater variety of diseases, including CIN 2 or 3 (27% of the diagnoses), adenocarcinoma in situ (15%), and endometrial adenocarinoma (23%).2 Women ≥35 years of age have more endometrial disease and a higher rate of cancer than younger women.2
The 2001 Bethesda system subclassifies AGC Paps into 2 categories: AGC, either endocervical, endometrial, or not otherwise specified (AGC–NOS); or AGC “favor neoplasia,” either endocervical or not otherwise specified (AGC–FN).3 AGC–FN is associated with a higher rate of disease and malignancy than AGC–NOS.2
Because of the high rate and wide variety of diseases associated with AGC cytology, current practice guidelines recommend an extensive initial evaluation of women with AGC, based on age and AGC subclass; and a secondary evaluation (ie, cervical conization) in certain high–risk patients.4, 5 Multiple published reports have documented the low rates of histologic sampling in women with AGC. Twenty–six studies published between 1988 and 2004 had an average rate of histologic sampling of 57%.1 In a recent large retrospective study of AGC, only 65% of cases had sufficient tissue follow–up.6 Two published surveys of practitioners found few clinicians choose to manage women with AGC according to recommended practice guidelines.7, 8 A small study published in 2006 found that 77% of women with AGC were evaluated according to practice guidelines.9
A study of 477 AGC cases at Hartford Hospital (Hartford, CT) during the years 1998-200110 found that <40% of women with AGC had a comprehensive initial evaluation and <10% had an appropriate secondary evaluation. This current study was designed to determine whether adherence to practice guidelines for the management of AGC Paps has improved between 1998-2001 and 2004-2007. In addition, this study compares the written recommendations by cytopathologists' on the Pap reports with the type of evaluation performed.
MATERIALS AND METHODS
The study protocol was originally reviewed and approved by the Hartford Hospital Institutional Review Board (IRB). The current analysis has been conducted with the approval of The Reading Hospital and Medical Center's IRB. Cases of AGC on cervical cytology from the years 2004-2007 were identified from the Hartford Hospital pathology database. Cases with a prior AGC Pap, prior total hysterectomy, known disease accounting for the AGC Pap, and AGC Paps with concurrent squamous abnormalities were excluded. Only cases with follow–up data (ie, additional Paps or gynecologic histologic specimens) were included. There were at least 1.5 years of follow–up for each case. Data collected for each case included the age of the patient, provider, date of AGC Pap, all dates and results of subsequent cytology and histology samples, and the cytopathologist recommendation, if any, for further management. The subclass of AGC Pap (AGC–FN and AGC–NOS) was determined by the descriptive modifiers, in a similar fashion to the prior study from 1998-2001.10 Any Pap described as “favor reactive” was classified as AGC–NOS.
Disease was defined as high–grade dysplasia or worse. Cases of low grade dysplasia or CIN 1 were also noted. A delayed diagnosis was defined as any diagnosis occurring over 1 year from the initial AGC Pap. A complete initial evaluation was defined as an endocervical curettage (ECC) for women <35 years of age. For women ≥35 years of age, a complete initial evaluation was defined as an ECC plus an endometrial biopsy. It was assumed that women who had an ECC also underwent colposcopy simultaneously. These procedures were considered part of the complete initial evaluation if they took place within 1 year of the initial AGC Pap or within 6 months prior. High–risk human papillomavirus (HPV) testing was not included in the definition of a comprehensive initial evaluation, since the recommendation for HPV testing of AGC Paps was not published until 2007.4 Secondary evaluations were defined as any diagnostic cervical conization (cold–knife cone or loop electrosurgical excision procedures) or hysterectomy.
Cytopathologist recommendations were divided into 4 types of recommendations: a complete histologic evaluation, an incomplete histologic evaluation, repeat Pap, or no recommendation. Examples of recommendations for an incomplete histologic evaluation include the statement “recommend biopsy” and recommendations for only an endocervical evaluation or only an endometrial evaluation in women ≥35 years of age. Most recommendations for a repeat Pap stated “Recommend repeat.” Cytopathologist recommendations from a previously published case series of AGC from Hartford Hospital (1998-2001) were similarly recorded and analyzed.2
Statistical analysis was performed using a chi–square test with Yates' correction. 11 For age comparison, the Mann–Whitney test was used (Stata/IC version 10.1 for Windows, StataCorp LP, College Station TX). P ≤ 0.05 was considered statistically significant.
There were 234,828 Paps screened by the Hartford Hospital pathology department during the years 2004-2007. The rate of AGC was 0.17% (410/234,828) during this period. A total of 126 cases of AGC were excluded from the study: 20 with a prior AGC, 9 with prior total hysterectomy, 27 with known disease accounting for the AGC Pap, 12 with concurrent finding of atypical squamous cells, 1 with concurrent finding of squamous intraepithelial lesion, and 57 lacking follow–up data. These exclusion criteria are the same that were used in the prior study from years 1998-2001.10 The study sample after these exclusions consisted of 284 cases of AGC. The rate of disease was 15.8% (45/284), and the rate of malignancy was 5.3%. There were also 8 cases of CIN 1.
The mean age of the study group was 43 (standard deviation 35-50), with an age range of 17-91 years. There were 83 women <35 years of age and 201 women ≥35 years of age. Two hundred forty–five cases were subclassified as AGC–NOS and 39 were subclassified as AGC–FN. The providers were predominantly obstetrician/gynecologists. Other providers included internal medicine and family practice physicians (3.5%), advance practice nurse practitioners (8.8%), and certified nurse midwives (1.8%). The rate of AGC decreased between the 2 time periods, while the rate of disease associated with AGC increased (Table 1). The mean age, rate of each age group, subclass, and providers were similar between the 2 time periods (Table 1).
|Rate of AGC on cytology, %||0.27||0.17||<.01|
|Rate of clinically significant disease, %||8.8||15.8||<.01|
|Rate of malignancy, %||2.7||5.3||.11|
|Mean age, y||44||43||.74|
|Rate of women age <35 y, %||23.1||29.2||.07|
|Rate of women age ≥35 y, %||76.9||70.8||—|
|Rate of AGC-NOS, %||88.5||86.3||.44|
|Rate of AGC-FN, %||11.5||13.7||—|
|Rate of OB/GYN providers, %||88||86||.46|
|Rate of other providers, %||12||14||—|
The rate of histologic sampling in the 2004-2007 study sample was 76.1% (216/284); 23.9% (68/284) had follow–up cytology only. Between the 2 time periods, there were statistically significant increases in the rates of histologic sampling overall, in AGC–NOS and in women ≥35 years (Figure 1).
The rate of comprehensive initial evaluation was 58.8% (167/284) in the 2004-2007 study sample (Figure 2). During 2004-2007, women ≥35 years were less likely to have a comprehensive initial evaluation (53.7%) than women <35 years (71.1%; P <0.02), similar to findings from 1998-2001.10 A statistically significant increase in the rate of comprehensive initial evaluations was seen between the 2 time periods overall, for AGC–NOS, and in women ≥35 years (Figure 2).
The rates of ECC, endometrial sampling, and endometrial sampling in women ≥35 years increased between 1998-2001 and 2004-2007 (Figure 3). The rate of ECC was similar in women <35 years compared with women ≥35 years of age during 2004-2007 (68.7% versus 64.7%, respectively; P = 0.61). The rate of endometrial sampling was higher in women ≥35 years than in women <35 years during 2004-2007 (58.2% versus 21.7%, respectively; P <0.01).
Previous data from 1998-2001 indicate that the rates of comprehensive initial evaluations and rates of endometrial sampling in women ≥35 years increased between the first 2 years of the study, 1998-1999, and the last 2 years of the study, 2000-2001.10 The rates of comprehensive initial evaluations did not increase between 2004-2005 and 2006-2007 (P = 0.72, Figure 4). The rate of endometrial sampling in women ≥35 years also did not increase between 2004-2005 and 2006-2007 (P = 0.10, Figure 4).
Rates of secondary evaluations in women with persistent AGC and AGC–FN remained low (24% and 0% in 2004-2007 and 8% and 2% in 1998-2001, respectively). A similar rate of delayed diagnoses (occurring over 1 year from the AGC Pap) was found for the 2 time periods (4.6% for 2004-2007 and 3% for 1998-2001). In 2004-2007, 28.9% (13/45) of the cases of clinically significant disease were delayed diagnoses.
AGC cytology reports from 2004-2007 contained cytopathologist recommendations for further management in 31% (89/284) of the cases; 69% (195/284) of the cases did not contain recommendations. Of the reports that did contain cytopathologists' recommendations, 28% (25/89) were consistent with the current management guidelines, 26% (23/89) recommended an incomplete histologic evaluation, and 46% (41/89) recommended repeating the Pap. In comparison, 46% (219/477) of AGC reports from 1998-2001 contained cytopathologist recommendations. Seven percent (15/219) were consistent with practice guidelines, 8% (17/219) recommended an incomplete histologic evaluation, and 85% (187/219) recommended repeating the Pap.
Women whose AGC report from 2004-2007 recommended a comprehensive evaluation or any histologic evaluation were more likely to have a comprehensive work–up (79%) than those whose reports did not contain recommendations (55%; P<0.01) or recommended repeating the Pap (51%; P<0.02). No difference in rates of comprehensive evaluation or any histologic evaluation was seen between those whose reports did not contain recommendations and those whose reports recommended repeating the Pap (P = 0.75). Similarly, during 1998-2001, women with an AGC report recommending a comprehensive evaluation or any histologic evaluation were more likely to have a comprehensive work–up (72%) than those whose reports did not contain recommendations (38%; P<0.01) or recommended repeating the Pap (29%; P<0.01). During the 1998-2001 time period, there was also no difference seen between those whose reports did not contain recommendations and those whose reports recommended repeating the Pap (P = 0.07).
While rates of histologic sampling and comprehensive initial evaluations have improved over time, women with AGC continue to be undermanaged. Fewer than 60% of women with AGC are managed appropriately in the initial evaluation, and fewer than 25% of women who require a secondary evaluation receive it. The low rates of comprehensive evaluations in women ≥35 are mainly attributable to the low rate of endometrial sampling. These findings suggest that clinicians may be unaware of the increased risk of endometrial pathology associated with AGC in older women. The continued low rates of comprehensive initial evaluations and secondary evaluations in women with AGC–FN also may indicate a lack of knowledge of the high–risk nature of this subclassification.
The overall rate of AGC decreased in the study sample, whereas the disease rate increased between 1998-2001 and 2004-2007, indicating that women with AGC may be at higher risk for disease than in the past. It is unclear whether the decrease in AGC rate is due to changes in the cytopathologic criteria for the diagnosis or changes in the population over time. The higher disease rate may reflect the lower rate of AGC. Alternatively, more disease may have been identified during 2004-2007 than 1998-2001 because more comprehensive evaluations took place during the later period. These findings support the even greater importance of following practice guidelines in women with AGC.
The drawbacks of the current study include its retrospective design and reliance on data from a pathology database. Women who lacked follow–up in the pathology database were specifically excluded because these women may have changed providers or been referred elsewhere. There may have been women who had part of their management at another unaffiliated practice or patients who were noncompliant with part of their care (ie, they come in for Paps, but not colposcopy), which would have underestimated the number of women appropriately managed.
Clinical practice guidelines are “systematically developed statements to assist practitioner and patient decisions about appropriate health care for specific clinical circumstances.”12 The American Society for Colposcopy and Cervical Pathology (ASCCP), along with other professional societies and federal and international organizations, jointly develop and publish comprehensive evidence–based guidelines called consensus guidelines to help manage women with abnormal Paps. The 2006 consensus guidelines recommend that the initial evaluation for AGC (except for atypical endometrial cells) include colposcopy with directed biopsies and ECC.4 In addition, endometrial sampling is recommended in women ≥35 years of age. Endometrial sampling is also recommended in women <35 years of age if they have clinical risk factors for endometrial cancer. The recommendation for women with atypical endometrial cells is that they may undergo an endometrial biopsy and ECC without colposcopy. If no disease is identified, then the recommendation is to perform colposcopy. Current guidelines also state that high–risk HPV testing is preferred for all AGC Paps (except AGC–FN) to determine subsequent follow–up in the management of women with AGC, after a comprehensive evaluation.4 High–risk HPV testing may also help guide initial evaluations in women with AGC.13-15
Subsequent management is determined based on the initial evaluation, AGC subclassification, and HPV status.4 If initial testing in women with AGC–FN is negative, a secondary evaluation consisting of a diagnostic excisional procedure is recommended. A diagnostic excisional procedure may be necessary for women with an initial negative evaluation and persistent AGC (having more than 1 AGC).4
Guidelines for the inclusion of educational notes and recommendations accompanying cervical cytology reports have been published.16 These guidelines state that recommendations are optional and at the discretion of the cytopathology laboratory and, if used, they should be concise and evidence–based. Notably, this study found that AGC cytology reports that contain no specific recommendations have rates of comprehensive initial evaluations similar to reports that recommend repeat cytology. The inclusion of recommendations that are consistent with practice guidelines or recommend any type of biopsy correlates with higher rates of appropriate evaluations. A previous study also found that women with abnormal cytology whose reports have a recommendation for a colposcopy or biopsy were more likely to receive biopsies than those whose reports lack the recommendation.17, 18 These findings suggest that adherence to practice guidelines by clinicians may be influenced by cytopathologist recommendations. Instituting standardized management guidelines on AGC cytology reports could increase compliance and decrease litigation, which is more common with Paps containing glandular abnormalities.19 For example, the statements: “Recommend cervical, endocervical and endometrial evaluation” for women ≥35 years and “Recommend cervical and endocervical evaluation, and, if clinically indicated, endometrial evaluation” for women <35 years could be included on AGC cytology reports. The recommendation for repeat cytology for AGC is not a standard of care. The 2006 consensus guidelines state that the use of repeat cytology for AGC is “unacceptable” for the initial evaluation of AGC.4
The observed increase in adherence to practice guidelines for AGC from 1998-2001 to 2004-2007 may be attributable to local and national educational efforts. The topic was presented at the Department of Obstetricians and Gynecologists Grand Rounds at Hartford Hospital in 2005 and 2006. The findings from 1998-2001 were published in 2 articles in 2005, 1 of which was designated an American Board of Obstetrics and Gynecology Maintenance of Certification article.2 Changes in the terminology for AGC may have also improved compliance. The 2001 Bethesda system changed the name of the finding from AGUS (atypical glandular cells of undetermined significance) to AGC, removing the phrase “of undetermined significance” to avoid possible confusion with the less risky finding ASCUS (atypical squamous cells of undetermined significance).3 The misleading subclassification of AGC “favor reactive” was also eliminated at that time. In addition, consensus guidelines were published, including the ASCCP 2001 Consensus Guidelines for the Management of Women with Cervical Cytological Abnormalities, and the American College of Obstetricians and Gynecologists (ACOG) Practice Bulletin for Management of Abnormal Cervical Cytology and Histology.20, 21
Further improvements in the adherence to practice guidelines for AGC may be seen following the publication of the 2006 ASCCP Consensus Guidelines for the Management of Women with Abnormal Cervical Screening Tests and the 2008 ACOG Practice Bulletin for the Management of Abnormal Cervical Cytology and Histology.4, 22 These publications and other educational efforts may increase both clinicians' and cytopathologists' awareness of current practice guidelines for AGC Paps. Future studies should explore the implementation of the use of standardized guidelines on cytology reports with this rare and high–risk finding.
No specific funding was disclosed.
CONFLICT OF INTEREST
The authors made no disclosures.
- 11http://www.quantpsy.org Accessed April 14, 2010.. Calculation for the chi–square test: an interactive calculation tool for chi–square tests of goodness of fit and independence.
- 12Field M, Lohr KN, eds. Clinical Practice Guidelines: Directions for a New Program. Washington, DC: National Academy Press; 1990.
- 16Papanicolaou Society of Cytopathology Practice Guidelines Task Force. Papanicolaou Society of Cytopathology guidelines for educational notes, disclaimers, and similar comments on reports of cervical cytology specimens. Diagn Cytopathol. 2003; 28: 282-285.
- 19Observations from Pap litigation consultations. Path Case Rev. 2011; 16: 73-82., .
- 21American College of Obstetricians and Gynecologists Practice Bulletin No. 66. Management of abnormal cervical cytology and histology. Obstet Gynecol. 2005; 106: 645-664.