[ACOG] Cervical Cytology 2009.pdf
[ACOG] Cervical Cytology 2009.pdf
CLINICAL MANAGEMENT GUIDELINES FOR OBSTETRICIAN–GYNECOLOGISTS
NUMBER 109, DECEMBER 2009
(Replaces Practice Bulletin Number 45, August 2003, Committee Opinion Number 300, October 2004,
and Committee Opinion Number 431, May 2009)
This Practice Bulletin was devel-
oped by the ACOG Committee on
Practice Bulletins—Gynecology with
the assistance of Alan Waxman,
MD. The information is designed to
aid practitioners in making deci-
sions about appropriate obstetric
and gynecologic care. These guide-
lines should not be construed as dic-
tating an exclusive course of
treatment or procedure. Variations
in practice may be warranted based
on the needs of the individual patient,
resources, and limitations unique to
the institution or type of practice.
Cervical Cytology
Screening
The incidence of cervical cancer has decreased more than 50% in the past 30
years because of widespread screening with cervical cytology. In 1975, the rate
was 14.8 per 100,000 women in the United States; by 2006, it had been reduced
to 6.5 per 100,000 women. Mortality from the disease has undergone a similar
decrease (1). The American Cancer Society estimates 11,270 new cases of cer-
vical cancer in the United States in 2009, with 4,070 deaths from the disease
(2). Recent estimates worldwide, however, are of almost 500,000 new cases and
240,000 deaths from the disease per year (3). When cervical cytology screening
programs have been introduced into communities, marked reductions in cervical
cancer incidence have followed (4–6).
New technology for performing cervical cancer screening is evolving rapidly,
as are recommendations for classifying and interpreting the results. The pur-
pose of this document is to provide a review of the best available evidence on
screening for cervical cancer. Specific equipment and techniques for perform-
ing cervical cytology and interpretation of the results are not discussed.
Background
Despite the demonstrated success of cervical cancer screening, it is estimated
that 50% of the women in whom cervical cancer is diagnosed each year have
never had cervical cytology testing. Another 10% had not been screened within
the 5 years before diagnosis (7). Thus, one approach to reducing the incidence
and mortality of cervical cancer would be to increase screening rates among
women who currently are not screened or who are screened infrequently.
Although rates of cervical cancer are on the decline in women born in the
United States, women who are immigrants to the United States from countries
where cervical cytology screening is not the norm are an especially high-risk
group (8).
ACOG
PRACTICE
BULLETIN
THE AMERICAN COLLEGE OF
OBSTETRICIANS AND
GYNECOLOGISTS
WOMEN’S HEALTH CARE PHYSICIANS
2 ACOG Practice Bulletin No. 109
Addressing Errors in Cervical Cytology
In some cases, cervical cancer is undetected despite a
recent screening test because of errors in sampling, inter-
pretation, or follow-up. Sampling errors occur when dys-
plastic cells on the cervix are not transferred to the slide;
errors of interpretation are attributed to lack of recogni-
tion of abnormal cells in the laboratory. These two
sources of false-negative test results are associated with
30% of the new cases of cervical cancer each year (7, 9).
The problem of errors in interpretation is compounded
by inconsistency among cytologists. When results of
monolayer cytology specimens were reviewed by quality
control pathologists, only negative and low-grade squa-
mous intraepithelial lesion (LSIL) readings had greater
than 50% consistency (10). Most revised results were
downgraded to lesser diagnoses. Of those reported as
atypical squamous cells of undetermined significance
(ASC-US), 39% were downgraded to negative on further
review. Of those originally interpreted as high-grade
squamous intraepithelial lesions (HSIL), 53% were rein-
terpreted as LSIL, ASC-US, or negative (11).
Natural History of Cervical Neoplasia
Infection with human papillomavirus (HPV) is a neces-
sary factor in the development of squamous cervical neo-
plasia; however, most HPV-infected women will not
develop significant cervical abnormalities (10, 12–15).
The infection is easily transmitted during sexual inter-
course. Most women, especially younger women, have
an effective immune response that clears the infection or
reduces the viral load to undetectable levels in an aver-
age of 8–24 months (12, 16–22). Factors that determine
which HPV infections will develop into squamous
intraepithelial lesions have been poorly identified. The
HPV type and the persistence of an HPV infection are
perhaps the most important determinants of progression
(12). Cigarette smoking may be a cofactor, and a com-
promised immune system appears to play a role in some
women (12).
Despite decades of study, the natural history of cer-
vical intraepithelial lesions is still not completely under-
stood. The oncogenic agent is well established to be one
of 15–18 “high-risk” types of HPV (23). The once widely
held concept that low-grade lesions are necessary pre-
cursors to the high-grade lesions and subsequent invasive
cancer has been questioned (10, 12, 15, 24).
Human papillomavirus infections are most common
in teenagers and women in their early 20s, with preva-
lence decreasing as women age (25–28). In adolescents
and young women, HPV infections and dysplasia are
likely to resolve spontaneously (16, 17, 19, 29–31). This
suggests that HPV infections found in older women are
more likely to reflect persistent infections acquired in the
past, and correlates well with increasing rates of HSIL
with increasing age.
The recent introduction of a vaccine targeting HPV-16
and HPV-18, the two most common cancer causing HPV
types, has advanced the promise of primary prevention
of cervical cancer. The vaccine does not protect women
against approximately 30% of cervical cancer caused by
HPV types other than HPV-16 and HPV-18. Further-
more, women already exposed to HPV-16 and HPV-18
can expect a lower level of protection from the vaccine
than the nearly 100% protection demonstrated in clinical
trials involving women not exposed to the virus (32, 33).
If immunization is widely implemented, it has been pro-
posed that the impact in terms of reduction in cervical
cancer will not begin to be realized for another 15–20
years (34). In the meantime, secondary prevention,
through a screening regimen of cervical cytology with or
without concomitant HPV DNA testing remains the best
approach to protecting women from cervical cancer.
Women who have been immunized against HPV-16 and
HPV-18 should be screened by the same regimen as non-
immunized women. Understanding the natural history of
HPV infection is important to establishing a balance
between sufficient testing to prevent cancer, while avoid-
ing overtesting with its increased cost and morbidity.
Techniques of Cervical Cytology
Both liquid-based and conventional methods of cervical
cytology screening are acceptable for screening. The
majority of cervical cytology screening performed in the
United States uses a liquid-based process. According to
a 2003 survey, nearly 90% of obstetrician—gynecolo-
gists use liquid-based cytology (35). Exfoliated cells are
collected from the transformation zone of the cervix and
may be transferred to a vial of liquid preservative that
is processed in the laboratory to produce a slide for inter-
pretation—the liquid-based technique—or may be
transferred directly to the slide and fixed using the con-
ventional technique. Performance of conventional cervi-
cal cytology requires avoidance of contaminating blood,
discharge, and lubricant. The liquid-based technology
will filter out most contaminating blood and inflammatory
cells and debris. A small amount of lubricant may be used
on the speculum and will remain on the vaginal walls
prior to reaching the cervix. Lubricant on the cervix itself
will interfere with the transfer of cells. Even with the liq-
uid-based technique, heavy menstrual blood may limit
the number of squamous cells available for interpretation.
Prompt suspension of the cells in the liquid eliminates the
problem of air drying artifact, which may limit the inter-
pretation of conventional cervical cytology.
ACOG Practice Bulletin No. 109 3
The use of liquid-based cytology has advantages
and disadvantages compared with conventional cervical
cytology screening. The principal disadvantages are the
higher cost and a decreased specificity. The advantages
are in the convenience of being able to test for HPV,
gonorrhea, and chlamydial infection directly from the
residual sample after the cells have been extracted for
cytology. In addition, cytotechnologists find liquid-based
tests easier to read. Some studies have found fewer
“unsatisfactory” results, although this claim has not been
consistent (36). Whether the liquid-based cytology tests
are more sensitive or specific is unclear. A meta-analysis
of eight studies identified by the authors to be method-
ologically sound found no significant difference in sen-
sitivity or specificity between the two technologies in their
ability to diagnose cervical intraepithelial neoplasia 2
(CIN 2) or higher using a cytology threshold of LSIL or
HSIL. If the threshold for colposcopy was lowered to
ASC-US, however, the liquid-based cytology had a sig-
nificantly lower specificity (37).
Cytologic Reporting
The nomenclature for reporting cervical cytology results
has undergone several changes since the publication of
the original Papanicolaou system. The Bethesda System
of reporting is the most widely used system in the United
States. First proposed in 1988, it was revised in 1991 and
again in 2001 (38–40). Highlights of the 2001 Bethesda
System classification are summarized as follows (40):
• Specimen adequacy—Slides are to be reported as “sat-
isfactory” or “unsatisfactory” for interpretation. The
presence or absence of an endocervical or transforma-
tion zone component is described in the narrative por-
tion of the laboratory report, as are other quality
indicators, such as partly obscuring inflammation or
blood. If a slide is categorized as unsatisfactory, the
reason should be specified. If abnormalities are found
on an otherwise unsatisfactory slide, it will, by defini-
tion, be considered satisfactory for interpretation.
• Negative for intraepithelial lesion or malignancy—
This designation should be used for slides with no
cytologic evidence of neoplasia. When specific
organisms are identified (eg, Trichomonas vaginalis,
Candida species, shift in flora suggestive of bacter-
ial vaginosis, bacteria consistent with Actinomyces
species, and cellular changes consistent with herpes
simplex virus), they are reported and categorized as
“negative for intraepithelial lesion or malignancy.”
Other nonneoplastic findings, including reactive
cellular changes associated with inflammation, radi-
ation, or an intrauterine device, as well as glandu-
lar cells posthysterectomy or atrophy, also may be
included in this category. Endometrial cells found in
a woman aged 40 years or older will be listed under
this category, but the finding of endometrial cells
will not be reported routinely if noted in a woman
younger than 40 years. A finding of endometrial cells
on cytology in asymptomatic premenopausal women
is rarely associated with significant pathology (41).
• Atypical squamous cells (ASC)—The epithelial
abnormality ASC is diagnosed when the degree of
nuclear atypia is not sufficient to warrant a diag-
nosis of squamous intraepithelial lesion. It is sub-
categorized into “atypical squamous cells of
undetermined significance” (ASC-US) and “atypi-
cal squamous cells cannot exclude HSIL” (ASC-H).
The category ASC-H includes those cytologic
changes suggestive of HSIL but lacking sufficient
criteria for definitive interpretation. The literature
suggests ASC-H should represent 5–15% of the total
pool of ASC but would have a significantly higher
predictive value for diagnosing CIN 2 or CIN 3 than
ASC-US (42, 43).
• Atypical glandular cells—This term designates cells
exhibiting atypia that are of glandular rather than
squamous origin and replaces the term “atypical
glandular cells of undetermined significance.” The
finding of atypical glandular cells on cytology is
more likely to be associated with both squamous
and glandular abnormalities than is ASC-US, and
the workup required of atypical glandular cells is
more aggressive (44, 45).
• The 2001 terminology subdivides atypical glandular
cells by cell type, ie, atypical endocervical cells,
atypical endometrial cells, or atypical glandular
cells not otherwise specified. The subdivision of
“favor neoplastic” is maintained in the 2001 report-
ing system. Because sufficient cytologic criteria exist
to designate endocervical adenocarcinoma and ade-
nocarcinoma in situ, these two findings are reported
when identified.
• Low-grade squamous intraepithelial lesions—As in
the original terminology, the 2001 nomenclature
combines cytologic findings of CIN 1 (mild dyspla-
sia) and those consistent with HPV infections into
the category LSIL.
• High-grade squamous intraepithelial lesions—The
category of HSIL combines CIN 2 and CIN 3 (mod-
erate dysplasia, severe dysplasia, and carcinoma in
situ).
• Squamous cell carcinoma
• The absence of endocervical cells or a transforma-
tion zone component may reflect that the transfor-
4 ACOG Practice Bulletin No. 109
mation zone was not well sampled. This finding is
common in pregnant women and in postmenopausal
women in whom the transformation zone has receded
onto the canal. Data conflict as to whether the lack
of these cells is associated with an increase in squa-
mous intraepithelial lesions. Women with this find-
ing whose recent cervical cytology test results have
been normal without intervening findings of ASC-US
or worse may be monitored by repeat cervical cytol-
ogy screening in 1 year. Others, including those with
incompletely evaluated abnormal test results, incom-
pletely visualized cervix, immunocompromised status,
and poor prior screening, should have repeat cervi-
cal cytology screening within 6 months. Pregnant
women lacking endocervical cells or transformation
zone component should have repeat cervical cytol-
ogy screening postpartum (45, 46).
Clinical Considerations and
Recommendations
When should screening begin?
Cervical cancer screening should begin at age 21 years.
Cervical neoplasia develops in susceptible individuals in
response to a sexually transmitted infection with a high-
risk type of HPV (12–14, 20, 21). Human papillomavirus
causes carcinogenesis in the transformation zone of the
cervix, where the process of squamous metaplasia
replaces columnar with squamous epithelium (12). Squa-
mous metaplasia is active in the cervix during adolescence
and early adulthood. Human papillomavirus infections are
commonly acquired by young women shortly after the ini-
tiation of vaginal intercourse (16–20) but, in most, they
are cleared by the immune system within 1–2 years with-
out producing neoplastic changes (12, 16, 17, 22). The
risk of neoplastic transformation increases in those
women whose infections persist (12, 47, 48).
Further evidence for this model comes from studies
of age-specific prevalence of HPV infections, which
consistently show a high prevalence of infection in
teenagers, peaking in the third decade of life with a sub-
sequent decrease (25–28). In a report of 10,090 Pap tests
in females aged 12–18 years, 422 (5.7%) were reported
as LSIL and only 55 (0.7%) were HSIL (49). Moreover,
most dysplasia in adolescents regresses spontaneously. A
prospective study of 187 women aged 18–22 years with
LSIL found that 61% and 91% had reverted to negative
after 1 and 3 years of follow-up respectively. Only 3%
progressed to CIN 3 (29). Two smaller studies of adoles-
cents with biopsy confirmed CIN 2 showed 65% and
75% regression to negative after 18 months and 3 years
respectively (30, 31).
In contrast to the high rate of infection with HPV in
sexually active adolescents, invasive cervical cancer is
very rare in women younger than age 21 years. Only
0.1% of cases of cervical cancer occur before age 21
years (50). In a recent analysis of national data from
1998 through 2003, researchers from the Centers for
Disease Control and Prevention identified an average of
only 14 cases of invasive cancer each year in females
aged 15–19 years. Cancer cases in adolescents younger
than 15 years were too few to report. Based on this report
and Surveillance Epidemiology and End Results (SEER)
data from 2002–2006, this translates to an incidence rate
of 1–2 cases of cervical cancer per 1,000,000 females
aged 15–19 years (1, 50).
The recommendation to start screening at age 21
years regardless of the age of onset of sexual intercourse
is based in part on the very low incidence of cancer in
younger women. It is also based on the potential for
adverse effects associated with follow-up of young women
with abnormal cytology screening results.
The American College of Obstetricians and Gyne-
cologists endorsed the 2006 recommendations of the
American Society for Colposcopy and Cervical Path-
ology regarding management of adolescents with abnor-
mal cytology and cervical biopsy results (51). These
guidelines stress a conservative approach to women
younger than 21 years found to have ASC-US or LSIL
on cytology and most with histology findings less than
CIN 3. Delaying the onset of screening until age 21 years
is a logical incremental step in practice guidelines, con-
sistent with this conservative approach to management
of adolescents with cervical test result abnormalities.
Earlier onset of screening may increase anxiety,
morbidity, and expense from the test itself and overuse of
follow-up procedures. The emotional impact of labeling
an adolescent with both a sexually transmitted infection
and a potential precancer must be considered because
adolescence is a time of heightened concern for self-
image and emerging sexuality.
Although cancer is rare in adolescents, dysplasia is not
uncommon. An abnormal cervical cytology screening test
leads to a sequence of additional tests designed to identify
those with CIN 2 or worse. However, recent studies have
documented a significant increase in premature births in
women previously treated with excisional procedures for
dysplasia (52). Because adolescents have most or all of
their childbearing years ahead of them, it is important to
avoid unnecessary excision or ablation of the cervix.
Sexually active adolescents, ie, females younger than
21 years, should be counseled and tested for sexually
transmitted infections, and should be counseled regard-
ACOG Practice Bulletin No. 109 5
ing safe sex and contraception. These measures may be
carried out without cervical cytology screening and, in
the asymptomatic patient, without the use of a speculum.
What is the optimal frequency of cervical
cytology screening?
Cervical cytology screening is recommended every 2
years for women aged 21–29 years, with either conven-
tional or liquid-based cytology. Women aged 30 years
and older who have had three consecutive cervical cytol-
ogy test results that are negative for intraepithelial
lesions and malignancy may be screened every 3 years.
Certain risk factors have been associated with CIN in
observational studies; women with any of the following
risk factors may require more frequent cervical cytology
screening:
• Women who are infected with human immunodefi-
ciency virus (HIV)
• Women who are immunosuppressed (such as those
who have received renal transplants)
• Women who were exposed to diethylstilbestrol in
utero
• Women previously treated for CIN 2, CIN 3, or cancer
Women infected with HIV should have cervical cyto-
logy screening twice in the first year after diagnosis and
annually thereafter (53). Women treated in the past for
CIN 2, CIN 3, or cancer remain at risk for persistent or
recurrent disease for at least 20 years after treatment and
after initial posttreatment surveillance and should con-
tinue to have annual screening for at least 20 years
(54–58).
The optimal number of negative cervical cytology
test results needed to reduce the false-negative rate to
a minimum has not been determined (59, 60). It has
been demonstrated, however, that the rate of dysplasia
decreases as the number of sequential negative Pap test
results increases (61). Studies over the past several decades
have shown that in an organized program of cervical can-
cer screening, annual cytology examinations offer little
advantage over screening performed at 2- or 3-year inter-
vals (62–65). One study that did show an increase in
relative risk of cancer with screening at a 3- versus 1-year
intervals, found no significant difference between
screening at 2- versus 3-years. The absolute risk in this
well-screened population, however, was very low (66).
An evaluation of 31,728 women aged 30–64 years
screened in the National Breast and Cervical Cancer
Early Detection Program, found a prevalence of CIN 2
and CIN 3 of 0.028% and 0.019%, respectively among
those who had three or more negative Pap test results in
a row. There were no cases of invasive cancer in this
group. Using a computer model, they calculated the risk
of these women developing invasive cancer and estimated
4 women with cancer per 100,000 women over the next
3 years with annual Pap screening and 8 women with
cancer per 100,000 women with triennial screening.
Although this represents a doubling of cases with pro-
longing the interval to 3 years, the absolute number of
cases, 4 women with cancer per 100,000 women, is very
small, and the estimated cost of finding each additional
case of cancer was large (61).
Formal cost-effective analysis of data from this
national program showed that the most cost-effective
strategy for cervical cancer screening is cytology testing
no more often than every 3 years in women with prior
normal screening test results (67). Moreover, regardless
of age, annual Pap testing was never found to be cost-
effective (67).
In several studies, age was shown to play a role in
the sensitivity of screening. A negative cervical cytology
screening result confers less protection on women
younger than 30 years than in older women (61, 65, 68).
A recent British study of 4,012 women aged 20–69 years
with invasive cancer showed that whereas cytology
screening in 3 years prior to diagnosis offered a 60% and
80% reduction in the incidence of cervical cancer at ages
40 years and 64 years respectively, screening between
age 20 years and 24 years provided no significant reduc-
tion in invasive cancer in women younger than 30 years
(69).
In a woman aged 30 years or older who is known to
have multiple recent consecutive negative cervical cytol-
ogy test results, the risk of developing CIN 3 or cancer
is low, and screening at 3-year intervals is a safe, cost-
effective approach, with either conventional or liquid-
based cytology (37, 70).
Published studies have assumed a program of cervi-
cal cancer screening and follow-up. Most women in the
United States get opportunistic screening as their insur-
ance carriers and providers change. Patients are fre-
quently inaccurate in recalling the timing and results of
recent screening, more often underestimating the time
elapsed and incorrectly recalling abnormal results as
normal (71–74). Therefore, it is important for the physi-
cian to assess a new patient’s screening history—ie, the
date of her most recent cervical cytology test, frequency
and results of her prior tests, or prior abnormal test
results and management.
It is important to educate patients about the nature of
cervical cytology, its limitations, and the rationale for
prolonging the screening interval beyond every year. In
addition, regardless of the frequency of cervical cytology
screening, physicians also should inform their patients
...