Full Text Review
Screening for Cervical Cancer
Prepared by Brenda J. Morrison, PhD, Associate
Professor, Department of Health Care and Epidemiology,
University of British Columbia, Vancouver,
BC
These recommendations were finalized by the Task Force in January 1992
Contents
Overview
In 1979
the Canadian Task Force on the Periodic Health Examination found that there
was fair justification for the view that the incidence and mortality of
invasive cancer of the cervix could be reduced by taking cervical smears
(B recommendation). The optimum age and frequency at which smears should
be taken was not known but was identified as a research priority. This
basic position has not changed but in 199
1
the
evidence was re-examined by the Task Force in response to the publication
of the Report of a National Workshop on Screening for Cancer of the Cervix.<
1>
Examining the ovaries at the time of cervical cancer screening should
also be considered for women at high-risk of ovarian cancer (see Chapter
72).
Burden
of Suffering
Cervical cancer is the eleventh most common cancer among
women in Canada. In 1993,
an estimated 1,300
women developed cervical cancer and about 400 died of the disease. In 1989,
an estimated 10,000
potential years of life were lost due to cervical cancer. The major factors
known to be associated with the occurrence of cervical cancer are age of
first sexual intercourse, number of consorts, number of previous partners
of consorts, smoking, low socio-economic status and possibly infections
such as human papillomavirus. However, note that screening for human papilloma-virus
(Chapter 63) is not recommended as part of cervical cancer screening. Rather
the factors associated with the highest risk are having many sexual partners,
a consort with many sexual partners and an early age of first intercourse.
Most studies show that for each additional partner the risk appears to
increase linearly by a factor ranging from 0.5 to 1,
up to a relative risk of at least 9; for early age of first intercourse
(before 17
or 18 years
of age), the relative risk has been reported to range between 2 and 3.
Depending upon the stage at which the disease is
detected, five-year survival rates range from 90% for stage 1
to
10-15%
for stage 4, suggesting that any screening measure which permits earlier
detection will improve survival.
Maneuver
At present only one test is suitable for general population
screening, the Papanicolaou smear test. False positive test results are
not a major concern because a second smear can be taken and colposcopy
done with minimal risk and relatively little expense; however some needless
anxiety will be generated. False negative test results, on the other hand,
allow the disease to escape detection until it reaches a less treatable
stage. Even in well run laboratories estimates for false negative error
rates run as high as 25%. The error can result from failure of the physician
to obtain malignant cells from the cervix or failure of the laboratory
process, including misinterpretation of the slide. Obtaining an adequate
smear can be difficult since cytological abnormalities usually arise at
the squamo-columnar junction and the site of this junction regresses up
the cervical canal as a woman ages. As for laboratory error, the false
negative rate appears to be lower in laboratories that process a large
volume of smears.
Effectiveness
of Screening
Although the etiology of cancer of the cervix has been
the subject of much debate and there are still uncertainties concerning
the progression of dysplasia and carcinoma in situ to invasive cancer,
there is evidence from both cohort and case-control studies that participation
in cervical cytology screening programs reduces the incidence of invasive
disease.<2-7> It has been estimated that somewhere between a quarter
and one half of carcinoma in situ progresses; for dysplasia, the proportion
is less. In the majority of cases progression appears to be slow. Models
of screening frequency indicate that even a ten year smear interval would
reduce the incidence of invasive cancer by about two-thirds. Because screening
for this disease is so widespread, a more precise measurement of its benefit,
as assessed by a randomized controlled trial, is not ethically possible.
Recommendations
of Others
It is not our purpose to review the cohort and case-control
evidence since this material has been presented in three major reports
produced by the Canadian government during the last 15
years<2,8-9> dealing exclusively with cervical screening policy. The
first, known as the Walton Report,<8> recommended that all women over
the age of 18
who have had sexual intercourse should be screened and that, if the initial
smear showed no atypia, a second smear be taken in a year’s time to affirm
that the first was not a false negative. If both these and later smears
were normal, screening should be carried out every three years, until the
age of 35 and every five years thereafter to age 60.
In 1982
the Walton report was revised<9> because of concerns that the sexual
freedom resulting from improvements in contraceptive technology might have
induced an "epidemic" of cervical cancer among younger women. The recommendation
for women under thiry-five was changed to annual screening. Since that
time, however, it has been concluded that for the majority of cases the
progression of dysplasia and carcinoma in situ is slow. As a consequence
the third report, The National Workshop Report, published in 1991,
returned to the original recommendation of screening every three years
and, in the interests of simplicity, suggested continuation of this frequency
for women over 35. Another change was extension of the surveillance period
to the age of 69, since most invasive disease occurs in post-menopausal
women, many of whom have never had a smear or have had one taken infrequently.
The report also dropped the recommendation for more frequent screening
for high-risk women. The Workshop urged the establishment of organized
cervical screening programs with a central laboratory in each province,
based on the assumption that programs with centralized laboratories would
be in place throughout the country.
In 1989,
the U.S. Preventive Services Task Force recommended regular papanicolaou
testing for all women who are or have been sexually active. Beginning with
the onset of sexual activity, the test should be repeated every one to
three years at the physicians discretion up to age 65 after which testing
may be discontinued if previous smears have been consistently normal. This
recommendation is currently under review.<10>
Frequency of
Routine Screening
Since no controlled trials have been carried out, no
well grounded scientific evidence exists on which to base recommendations
for the optimum frequency of screening. The recommendations of the most
recent National Workshop<1>
are largely based on the report from the International Agency for Research
on Cancer<11>
which, in turn, was based on case-control studies and computer models constructed
with parameters derived from cohort studies. However, Knox<12>
has recently published an analysis showing that estimates of the relative
protection of screening (and therapy) derived from such case-control and
cohort studies are potentially invalid. The estimate depends on the true
protective effect of the test and therapeutic procedures but is also determined
by the ratio of the "uptake rate" of the screening procedure (or proportion
being screened) in the case population and control population. If the uptake
is unequal in the two groups, then the estimate may be biased. The greater
the uptake of screening in the control group, relative to that for the
cases, the greater the apparent protection. Many reports have documented
that the women who are at highest risk of developing cervical cancer have
infrequent or no smears. Therefore estimates of relative risk will be strongly
confounded by the difference in uptake rates between the cases and controls.
Conclusions
and Recommendations
In light of the foregoing considerations, it is difficult
to formulate frequency recommendations that have solid scientific validity.
On an observational basis all that can be said is that women who are presently
being screened regularly have a very low rate of developing cervical cancer,
whereas those who are being screened infrequently or not at all, have a
much higher risk of developing the disease. In the British Columbia cohort
study only 4% of cases of carcinomas in situ or worse were detected by
smears subsequent to a third smear.<2>
There is fair evidence to include Papanicolaou smears
in the periodic health examination of women who have been sexually active
(B Recommendation).
The Task Force concurs with the Workshop’s recommendations concerning frequency
of smears once central laboratories are operational. Until such time, it
recommends that high-risk women be screened more frequently than every
3 years since the false negative rate may run well above 25%, and the disease
progression rate is known to vary. Physicians should keep in mind, however,
that these high-risk women constitute a small proportion of the population.
Because some social stigma may be attached to being
labelled as being at high risk, individual physicians should consider informing
a patient of the risk factors and allowing her to make the decision on
the appropriate frequency of screening.
The Task Force emphasizes that the largest group
of women at greatest risk of dying from cervical cancer are those who have
never been screened. The majority of these are women over 50, particularly
native women and women who have immigrated from poorer countries. It is
more important that these women be screened, even if only once, than that
average or low-risk women be screened frequently.
Unanswered
Questions (Research Agenda)
Research into methodologies to reach women who have
never been screened is critical. In future, most women reaching middle-age
will have had many cervical smears and may therefore be at a different
risk than the average middle-aged or older woman is at present. Most will
have had nothing but negative smears. It is important to determine if,
in well tested women, a history of numerous negative smears is associated
with a substantially reduced risk.
Evidence
The evidence used in this review was taken from the
collection of the author. This review was initiated in June 1989
and the recommendations were finalized by the Task Force in January, 1992.
A technical report with a full reference list dated February 1992
is available upon request.
Full Citation
Link to Structured Abstract of
this review
Link to Summary Table of this
review
Link to Selected References list of this review
Top of Page
CTFPHC
Home Page
Reprinted in modified format by the Canadian
Task Force on Preventive Health Care
with permission.
For any technical issues please contact: webmaster@ctfphc.org
Original Copyright
© 1994 Minister of Supply and Services Canada.
Last modified March 27, 1998.