JNCI | Articles 721
Organized cervical cytology screening programs in developed
countries have accompanied marked decreases in cervical cancer
incidence and mortality ( 1 ). In addition to the benefits of finding
cancers at an early stage, detection and treatment of cervical intra-
epithelial neoplasia (CIN) appear to account for the decline in
cervical cancer incidence ( 2 ). Although many women are treated
for CIN annually, the Cochrane Colposcopy and Cervical
Cytopathology Collaborative Group ( 3 ) recently highlighted the
lack of international consensus on optimal surveillance strategies
after treatment. Lack of data on long-term outcomes, including
risks of recurrence among women treated for CIN, is a critical
barrier to the formulation of evidence-based recommendations
for follow-up surveillance strategies. Findings ( 4 , 5 ) from ALTS
(ie, ASCUS-LSIL [atypical squamous cells of undetermined sig-
nificance – low-grade squamous intraepithelial lesion] Triage
Study) have been useful in providing data on outcomes for a
period of 8 – 24 months after treatment, but information on the
long-term risks of subsequent CIN or invasive cancer among
Affiliations of authors: Department of Family and Community Medicine,
Center for Healthcare Policy and Research, University of California,
Sacramento, CA (JM); Surveillance and Outcomes Unit (CM), Division of
Gynecologic Oncology (TE), and Department of Population Oncology (AC),
British Columbia Cancer Agency, Vancouver, BC, Canada; Department of
Obstetrics, Gynecology and Reproductive Sciences, Department of Epidemiology
and Biostatistics, University of California, San Francisco, CA (GFS) .
Correspondence to: Joy Melnikow, MD, MPH, Department of Family and
Community Medicine, Center for Healthcare Policy and Research, University
of California, Davis, 4860 Y St, Ste 2300, Sacramento, CA 95817 (e-mail:
See “Funding” and “Notes” following “References.”
© The Author 2009. Published by Oxford University Press. All rights reserved.
For Permissions, please e-mail: firstname.lastname@example.org.
Cervical Intraepithelial Neoplasia Outcomes After
Treatment: Long-term Follow-up From the British
Columbia Cohort Study
Joy Melnikow , Colleen McGahan , George F. Sawaya , Thomas Ehlen , Andrew Coldman
Background Information on the long-term risk of cervical intraepithelial neoplasia (CIN) recurrence among women
treated for CIN is limited yet critical for evidence-based surveillance recommendations.
Methods We retrospectively identified 37 142 women treated for CIN 1, 2, or 3 from January 1, 1986, through
December 31, 2000 (CIN cohort), from the British Columbia Cancer Agency cytology database and linked
their records with cancer registry and vital statistics data. Treatment included cryotherapy, loop electro-
surgical excision procedure, cone biopsy, and laser vaporization or excision. A comparison cohort con-
tained 71 213 women with normal cytology and no previous CIN diagnosis. Follow-up continued through
December 31, 2004. Among women in both cohorts under active surveillance, we compared rates of CIN
2 or 3 (CIN 2/3) and cervical cancer. Cumulative incidence rates of CIN 2/3 and 95% confidence intervals
(CIs) were estimated by a life table approach by using annual rates. Cumulative rates of invasive cancer
were examined by the person-years method.
Results Overall observed cumulative rates of CIN 2/3 in the first 6 years after treatment were 14.0% (95% CI =
13.84% to 14.15%) for women originally treated for CIN 3, 9.3% (95% CI = 9.09% to 9.42%) for CIN 2, and
5.6% (95% CI = 4.91% to 5.21%) for CIN 1. Annual rates of CIN 2/3 were less than 1% after 6 years. Initial
diagnosis, age, and treatment type were associated with a diagnosis of CIN 2/3 after treatment, with 6-year
adjusted rates for women aged 40 – 49 years ranging from 2.6% (95% CI = 1.9% to 3.4%) for treatment
of CIN 1 with the loop electrosurgical excision procedure to 34.0% (95% CI = 30.9% to 37.1%) for treatment
of CIN 3 with cryotherapy. Overall incidence of invasive cancer (per 100 000 woman-years) was higher in
the CIN cohort (37 invasive cancers, 95% CI = 30.6 to 42.5 cancers) than in the comparison cohort (six cancers,
95% CI = 4.3 to 7.7 cancers). Cryotherapy, compared with other treatments, was associated with the highest
rate of subsequent disease (adjusted odds ratio for invasive cancer = 2.98, 95% CI = 2.09 to 4.60).
Conclusion Risk of CIN 2/3 after treatment was associated with initial CIN grade, treatment type, and age. Long-term
risk of invasive cancer remained higher among women treated for CIN, particularly those treated with
J Natl Cancer Inst 2009;101: 721 – 728
722 Articles | JNCI Vol. 101, Issue 10 | May 20, 2009
women previously treated for CIN is limited. This information,
however, is crucial for the formulation of rational, evidence-based,
cost-effective strategies for treatment and follow-up.
Management recommendations for the treatment and follow-
up of CIN have evolved over time. Findings ( 4 , 5 ) from ALTS led
to recommendations for CIN treatment and follow-up by the
American Society for Colposcopy and Cervical Pathology (ASCCP)
( 6 ). Current ASCCP recommendations for follow-up after treat-
ment of CIN 2 or 3 (CIN 2/3) are a single human papillomavirus
(HPV) test 6 – 12 months after treatment, two consecutive cytology
tests or cytology with colposcopy 6 months apart, followed by
routine screening if tests are normal ( 7 ). The interval for routine
screening is unspecifi ed, but the guidelines note that elevated risk
of recurrent CIN or invasive cancer persists for many years after
treatment and that follow-up should continue for at least 20 years,
refl ecting the conclusion of a recent systematic review of observa-
tional studies ( 8 ). Guidelines from the British Columbia Cancer
Agency at present recommend colposcopy 4 – 6 months after treat-
ment of CIN 2/3. If results are normal, follow-up cytology is rec-
ommended 12 months after treatment ( 9 ). During the study period
(from January 1, 1986, through December 31, 2000), the recom-
mendation for CIN 2/3 was for follow-up colposcopy with cytol-
ogy at 3, 7, and 13 months after treatment and then annual
cytology. For CIN 1, the ASCCP recommends follow-up without
initial treatment, with HPV testing every 12 months or cytology
every 6 – 12 months and a return to routine screening if the HPV
test or two consecutive cytology tests are negative ( 7 ). In the 1980s
and 1990s, early treatment of CIN 1 was common practice in the
United States, whereas observation of women with CIN 1 with
repeat cytology was considered usual care in British Columbia.
British Columbia has had an organized cervical cytology
screening program in place since 1955. Since the 1980s, cervical
cytology and follow-up results have been included in a single
population-based database that links all cytology, colposcopy, and
histology reports. Screening recommendations during the period
covered by our study were that sexually active women be screened
every 2 – 3 years after two negative annual tests but that high-risk
women (ie, those with early onset of sexual activity or multiple
partners) be screened annually ( 10 ).
We identifi ed a retrospective cohort of women treated for CIN
(ie, the CIN cohort) and followed these women for up to 18 years
to examine the risks of subsequent CIN 2/3 and invasive cancer.
We compared subsequent rates of CIN 2/3 and of invasive cancer
in the CIN cohort with those of a low-risk cohort (ie, the compari-
son cohort) who had not previously been treated for CIN, and we
followed these women for the same time period. Linkage with the
British Columbia Cancer Registry identifi ed all women with inci-
dent cases of cervical cancer in each cohort.
Participants and Methods
This study was reviewed and approved by the University of British
Columbia Research Ethics Board and the University of California
Davis Committee for the Protection of Research Subjects. It was
considered exempt from written informed consent.
The CIN cohort consisted of 37 142 women retrospectively identi-
fied who were treated for CIN 1, 2, or 3 from January 1, 1986,
through December 31, 2000 (ie, the CIN cohort), from the British
Columbia Cancer Agency cytology database. Their records were
linked to cancer registry and vital statistics data. Follow-up contin-
ued through December 31, 2004. All women had a cytology test
within the year before the CIN diagnosis and had documented
treatment of CIN. Treatment included cryotherapy, a loop elec-
trosurgical excision procedure, cone biopsy, and laser vaporization
or excision. Women with no recorded treatment were excluded
from the analysis. Because the grade of CIN found on the initial
biopsy examination could differ from that found on excisional
specimen among women who were treated with loop electrosurgi-
cal excision procedure or cone biopsy, we defined the index CIN
diagnosis as the most severe pathological result occurring within a
6-month period. For example, a woman with an initial biopsy
result of CIN 2 and a subsequent cone biopsy within 6 months that
had a result of CIN 3 was considered to have CIN 3. Women who
had a histological diagnosis of invasive cancer within 6 months of
a biopsy examination with a result of CIN were considered to have
had missed prevalent invasive cancer and were excluded from the
study. Records of women treated with either cone biopsy or loop
electrosurgical excision procedure and found to have positive sur-
gical margins, those who were treated by hysterectomy or cervi-
cectomy, and those with incomplete records were also excluded.
CONTEXT AND CAVEATS
Only limited information on the long-term risk of cervical intraepi-
thelial neoplasia (CIN) recurrence among women treated for CIN is
A retrospective cohort of women who were treated for CIN 1, 2, or
3 and a comparison cohort of women with normal cytology and no
previous CIN diagnosis.
The risk of subsequent CIN 2 or 3 (CIN 2/3) was associated with
initial CIN grade, treatment type, and age. Two years after treat-
ment for CIN, there was a rapid decline in the rate of subsequent
diagnoses of CIN 2/3. However, 6 years after treatment the risk of
invasive cancer continued to be higher among women treated for
CIN than among those with no CIN diagnosis, although the risk of
CIN 2/3 remained low. Risk was highest for women treated for CIN
Findings of this study support the shift in recommendations for
screening of women with CIN from indefinite annual screening to
an initial period of 6 – 18 months of more intensive annual examina-
tion, followed by a return to routine screening.
Data were observational. Treatment patterns shifted during the
study period. Among women treated with cryotherapy or laser
ablation, neither satisfactory colposcopy results nor treatment
according to guidelines could be documented.
From the Editors
JNCI | Articles 723
In British Columbia from 1986 through 2000, women were
screened for cervical cancer by cytology (ie, Papanicolaou testing)
through an organized cervical cancer screening program. Guidelines
recommended follow-up of all abnormal results. A histological
diagnosis of CIN 2/3 was treated with one of the following proce-
dures: cone biopsy, loop electrosurgical excision procedure, laser
excision or vaporization, or cryotherapy. For CIN 1, treatment was
optional and was most common for women with persistent disease.
Because we wished to understand the risks of persistence or recur-
rence of CIN after treatment, only women with documented treat-
ment in the 6-month window starting from the time of index
diagnosis were selected for the cohort. The most defi nitive treat-
ment within the 6-month window was identifi ed as the index treat-
ment. Defi nitive treatment was defi ned by the following hierarchy
(from most to least invasive): cone biopsy, loop electrosurgical
excision procedure, laser excision or ablation, and cryotherapy.
The comparison cohort was selected from a 10% random
sample of records of all women in the British Columbia cytology
database from January 1, 1985, through December 31, 2000. We
selected the comparison cohort to establish the baseline incidence
of disease in a low-risk, well-screened population. To defi ne a low-
risk study sample, women aged 21 years or older with three con-
secutive normal cervical cytology tests and no previous history of
CIN were included. We defi ned the index cytology as the third
normal test. Women with a documented history of hysterectomy
or cervicectomy before the study period and those with incomplete
records were excluded.
The end of the observation period for the two cohorts was
December 31, 2004. Records for both CIN and comparison
cohorts were linked to the British Columbia Cancer Registry fi les
to identify patients with invasive cancer who were not identifi ed
through the cytology database. If stage was not reported, pathol-
ogy reports were reviewed to defi ne the stage. Microinvasive can-
cers (International Federation of Gynecology and Obstetrics stage
IA1) were considered to be stage I.
Follow-up started from the treatment date of the index diagno-
sis for women in the CIN cohort and from the index cytology test
for women in the comparison cohort. Follow-up continued until a
woman had a diagnosis of invasive cancer by histology, was docu-
mented to have had a hysterectomy or cervicectomy, died, or until
December 31, 2004.
Event occurrence (ie, cervical cytology, colposcopy, or recurrent
disease) was measured by the elapsed time since index treatment.
Analysis of subsequent disease occurrence was performed sepa-
rately for a diagnosis of CIN 2/3 and for invasive cervical cancer.
The analysis of subsequent CIN 2/3 after treatment included
treated subjects remaining in active surveillance via periodic cytol-
ogy. Because CIN 2/3 is an asymptomatic state, only women
undergoing periodic cytology had the potential for diagnosis.
Rates of CIN 2/3 were computed by year since index treatment on
the basis of the number of women under active surveillance, with
the number of women in active surveillance diagnosed with CIN
2/3 in that year divided by the number in active surveillance
screened in that year. Women were considered to be in active
surveillance for a 3-year period after their last cytology, that is,
women were not considered as being at risk for CIN 2/3 for a
specifi c year if they had not also undergone cytology screening in
at least one of the preceding 2 years. Women not in active surveil-
lance were excluded from both the numerator and the denomina-
tor for the analysis of the years in question. Rates of CIN 2/3 were
examined over time by index diagnosis (CIN 1, 2, or 3), index
treatment (cryotherapy, laser excision or ablation, loop electrosur-
gical excision procedure, or cone biopsy), and age (21 – 29, 30 – 39,
40 – 49, or ≥ 50 years at index treatment for the CIN cohort and by
age at index cytology for the comparison cohort).
Logistic regression analysis was performed to examine the
independent effects of these factors on the odds of subsequent
CIN 2/3 after treatment. Coeffi cients from this analysis were then
applied to estimate annual incidence rates by time since treatment.
Cumulative incidence rates of CIN 2/3 after treatment and their
95% confi dence intervals (CIs) were estimated by a life table
approach by using the annual rates.
Cumulative rates of invasive cancer subsequent to treatment for
CIN were examined separately by the person-years method.
Because invasive cervical cancer may be present in unscreened
women with symptoms, we assumed that all women in the CIN
and comparison cohorts were at risk until the fi rst invasive cancer
diagnosis, hysterectomy or cervicectomy, death, or the end of the
observation period. Poisson regression analysis was used to exam-
ine the independent effects of index diagnosis, age, and treatment
type during the fi rst 10 years of cohort follow-up. Because of lim-
ited statistical power due to the relatively small number of women
who were diagnosed with invasive cancer, treatment type was
dichotomized as cryotherapy compared with any other treatment.
Coeffi cients were then applied to predict annual incidence rates.
Cumulative rates of invasive cancer were also examined over time
for women considered to be under active surveillance, which was
defi ned in the same way as for the estimates of CIN recurrence (ie,
having at least one cytology test with any result in the preceding
2 years before a cancer diagnosis). All analyses were conducted
with SAS/STAT software version 9.1.3 (SAS Institute, Cary, NC).
All statistical tests were two-sided.
We identified 63 722 women with a diagnosis of CIN between
January 1, 1986, and December 31, 2000. Because our objective
was to examine rates of CIN 2/3 and invasive cancer after treat-
ment, 22 615 women with no record of treatment were excluded.
In addition, 2776 women whose treatment was hysterectomy or
cervicectomy, 1052 women who had either a cone biopsy or a loop
electrosurgical excision procedure without clear margins, and
137 women with incomplete data were excluded. The final CIN
cohort, therefore, was composed of 37 142 women. In this group,
a total of 3013 women who were diagnosed with subsequent CIN
2/3 were identified. During the follow-up period, 809 women with
cytology that was interpreted as moderate dysplasia or worse and
no documented biopsy result within the next 2 years were censored
from the cohort. Among the 809 censored, 85 had a subsequent
diagnosis of CIN 2/3 that occurred more than 2 years after the
abnormal test, three were diagnosed with invasive cancer, and 721
had no record of a pathological diagnosis in the follow-up period.
724 Articles | JNCI Vol. 101, Issue 10 | May 20, 2009
For the comparison cohort, 72 323 women were initially selected;
999 women were excluded because of a previous hysterectomy or
cervicectomy, and 111 were excluded for incomplete data. The
final comparison cohort was composed of 71 213 women. Within
this group, there were 989 women with a first diagnosis of CIN 2/3
during the follow-up period.
The majority of women in the treatment cohort were treated for
CIN 3. Younger women and women with CIN 1 and 2 were
treated more often with cryotherapy, whereas women 50 years or
older and those with CIN 3 more often underwent cone biopsy
( Table 1 ). Women in the CIN cohort tended to be younger (82.6%
were <40 years) than women in the comparison cohort (64.4% were
<40 years) ( Table 2 ). Rates of subsequent CIN 2/3 up to 15 years
after treatment are shown in Figure 1 for the CIN and comparison
cohorts by index diagnosis for women under active surveillance.
Rates fell rapidly over the fi rst 4 years, and within 6 years, annual
rates of CIN 2/3 for all index diagnoses in the CIN cohort had
fallen to less than 1% and were comparable to incidence of CIN in
the comparison cohort. Rates were highest in the fi rst 2 years after
Table 1 . CIN cohort distribution by index diagnosis, index treatment (cone biopsy, LEEP, laser excision or ablation, or cryotherapy), and
age (n = 37 142) *
Index diagnosis Age, y
No. of women (% of age group)
CIN 1 (n = 6988 ) 21 – 29
30 – 39
40 – 49
CIN 2 (n = 10 823)21 – 29
30 – 39
40 – 49
CIN 3 (n = 19 331) 21 – 29
30 – 39
40 – 49
Total 987740669309 13 890
* CIN = cervical intraepithelial neoplasia; LEEP = loop electrosurgical excision procedure; laser = laser excision or ablation.
Table 2 . Age at initial treatment or at index cytology test for CIN
and comparison cohorts *
No. in CIN
No. in comparison
21 – 30
30 – 39
40 – 49
18 155 (48.9)
12 533 (33.7)
37 142 (100.0)
24 322 (34.2)
20 822 (29.2)
12 499 (17.6)
13 570 (19.1)
71 213 (100.0)
* The CIN cohort included women 21 years or older with a cytology test in
the previous year, a first histological diagnosis of CIN 1, 2, or 3 between
1986 – 2000, and documented treatment. The comparison cohort was
selected from a 10% random sample of women in the British Columbia
Cancer Agency cytology database from 1985 – 2000, aged 21 years or older,
who had three consecutive negative cytology tests and no history of CIN.
The age range presented is for initial treatment for the CIN cohort, and
for the index cytology test for the comparison cohort. CIN = cervical
Rate per 1000 women
2468 10 12 14 16
Time since index diagnosis (years)
Figure 1 . Incidence rates of cervical intraepithelial neoplasia 2 or 3 (CIN
2/3) per 1000 women by index diagnosis among women under active
surveillance in the CIN and comparison (COMP) cohorts. The index
event was treatment for the CIN cohort and normal cytology result at
entry for the comparison cohort. Error bars = 95% confi dence
treatment and increased monotonically with CIN grade. Overall
observed cumulative rates of CIN 2/3 for the fi rst 6 years after
treatment were 14.0% (95% CI = 13.84% to 14.15%) for women
originally treated for CIN 3, 9.3% (95% CI = 9.09% to 9.42%) for
CIN 2, and 5.6% (95% CI = 4.91% to 5.21%) for CIN 1.
Logistic regression analysis was used to estimate the overall
adjusted rates of subsequent CIN 2/3 in the CIN cohort under
active surveillance during the fi rst 6 years of follow-up. Six years
was chosen because of the low rates of CIN 2/3 after the fi rst 6 years
of follow-up (as shown in Figure 1 ). The analysis included age,
index diagnosis, index treatment, and year since index treatment as
independent variables. Age was categorized in decades. Interactions
between age, index diagnosis, and index treatment were explored by
use of forward selection, but year was restricted to be a main effect.
The resulting model included statistically signifi cant interactions
for age and index diagnosis and for index diagnosis and treatment,
JNCI | Articles 725
as well as main effects for age, diagnosis, treatment, and year (with
details in Supplementary Table 1 , available online).
In general, women 40 years or older and those with more severe
index disease had higher rates of CIN 2/3 after treatment. The
CIN 2/3 rates after treatment were lowest for cone biopsy and
highest for cryotherapy. Initial diagnosis, age, and treatment type
were all associated with a diagnosis of CIN 2/3 after treatment,
with the 6-year adjusted rates for women aged 40 – 49 years ranging
from 26.4 diagnoses per 1000 women (95% CI = 18.5 to 34.3 diag-
noses per 1000 women) for treatment of CIN 1 with loop electro-
surgical excision procedure to 340.0 diagnoses per 1000 women
(95% CI = 309.0 to 370.8 diagnoses per 1000 women) for treat-
ment of CIN 3 with cryotherapy ( Table 3 ). For the comparison
cohort, the 6-year cumulative rate of subsequent CIN 2/3 was
9.8 diagnoses per 1000 women (95% CI = 8.2 to 11.4 diagnoses per
1000 women) among women aged 40 – 49 years and was 54.8 diag-
noses per 1000 women (95% CI = 50.2 to 59.3 diagnoses per 1000
women) among women aged 20 – 29 years ( Table 4 ).
One hundred forty-fi ve women were diagnosed with invasive
cancer in the entire CIN cohort during the 18-year follow-up
period, a rate of 37 cancers per 100 000 woman-years (95% CI =
30.6 to 42.5 cancers per 100 000 woman-years ), compared with 49
women in the comparison cohort, a rate of six cancers per 100 000
woman-years (95% CI = 4.3 to 7.7 cancers per 100 000 woman-
years). Among women in the CIN cohort who were under active
surveillance and were considered at risk for up to 3 years after their
last cytology test (with the same denominator that was used for the
analysis of CIN 2/3), 49 cases of invasive cancer occurred during
the follow-up period. The stage distribution for invasive cancers
was similar for the CIN cohort and the comparison group, with
most invasive cancers in early stages at diagnosis ( Table 5 ). The
cumulative rates of cancer in the CIN cohort and the CIN cohort
under active surveillance increased steadily relative to the compari-
son group for the fi rst 8 years of follow-up. After 8 years, the rates
for the cohort under surveillance diverged from the overall CIN
cohort, with higher rates for the overall CIN cohort were not sig-
nifi cant different from rates for the CIN though under surveillance
( Figure 2 ). From a Poisson regression analysis, independent risk
factors for invasive cancer were identifi ed as an initial diagnosis of
CIN 3, treatment with cryotherapy (compared with all other treat-
ments), and being 40 years or older ( Table 6 ). Cryotherapy, com-
pared with other treatments, was associated with the highest rate
of subsequent disease (adjusted odds ratio for invasive cancer =
2.98, 95% CI = 2.09 to 4.60).
This large, population-based cohort study with more than 300 000
woman-years of observation in the CIN cohort provided impor-
tant information that could contribute to evidence-based guide-
lines for follow-up of women treated for CIN. For women under
surveillance after treatment, risks for subsequent CIN 2/3 declined
to that of incident CIN in low-risk women by 6 years after treat-
ment. Rates of subsequent CIN 2/3 increased with age and initial
CIN grade and varied by treatment, being highest for women
older than 40 years who were treated for either CIN 2 or CIN 3
with cryotherapy. Previously, large population-based cohort stud-
ies from Sweden in 1989 ( 11 ) and 2007 ( 12 ) reported on invasive
cancer rates after treatment for CIN but did not have information
on the recurrence of CIN or whether women were under surveil-
lance after treatment. In our study, we found that invasive cancer
risk was markedly higher in women after treatment of CIN, even
among those undergoing active surveillance, compared with a
cohort of women without previous CIN.
Overall rates of CIN 2/3 declined rapidly for the fi rst 2 years
after treatment; however, during the fi rst 6 years of follow-up,
these rates ranged from 5% for women initially treated for CIN 1
to 14% for women treated for CIN 3. In ALTS, women with initial
low-grade squamous epithelial lesions who were referred for early
colposcopy had rates of subsequent CIN 2/3 of 8% – 13% during a
24-month follow-up ( 13 ). Others ( 7 , 14 ) have estimated rates of
subsequent CIN after treatment that ranged from 1% to 21%.
In this observational study, rates of subsequent CIN 2/3 could be
measured only for women who chose to obtain follow-up cytology.
We focused on estimating the risk of subsequent CIN in the
women under active surveillance. This group of women included a
Table 3 . Estimated rates of CIN 2/3 per 1000 women in the initial 6-year period after treatment for CIN by index diagnosis, age group,
and index treatment (cone biopsy, LEEP, laser excision or ablation, or cryotherapy) *
Index diagnosisAge, y
Estimated rate of CIN 2/3 after treatment, No. per 1000 women (95% CI)
ConeLEEP Laser Cryotherapy
CIN 121 – 29
30 – 39
40 – 49
28.1 (20.2 to 36.1)
33.1 (23.9 to 42.3)
31.9 (22.8 to 41.0)
22.4 (14.7 to 30.0)
23.3 (16.4 to 30.1)
27.4 (19.4 to 35.3)
26.4 (18.5 to 34.3)
18.5 (11.5 to 25.5)
45.6 (39.0 to 52.1)
53.5 (45.1 to 61.8)
51.6 (42.0 to 61.3)
36.3 (25.1 to 47.5)
54.7 (48.6 to 60.9)
64.1 (56.2 to 72.0)
61.9 (52.3 to 71.6)
43.7 (31.1 to 56.2)
CIN 221 – 29
30 – 39
40 – 49
34.8 (28.2 to 41.5)
35.0 (28.3 to 41.7)
35.7 (28.4 to 43.0)
29.5 (21.4 to 37.5)
41.9 (34.8 to 49.0)
42.2 (34.9 to 49.5)
43.0 (34.6 to 51.4)
35.5 (25.3 to 45.8)
73.0 (66.6 to 79.4)
73.0 (66.6 to 79.4)
74.8 (64.4 to 85.3)
62.1 (45.9 to 78.3)
133.3 (124.5 to 142.2)
134.1 (123.3 to 144.9)
136.5 (119.2 to 153.8)
114.1 (85.3 to 142.9)
CIN 321 – 29
30 – 39
40 – 49
56.3 (52.0 to 60.6)
62.9 (58.1 to 67.6)
85.3 (77.5 to 93.2)
90.4 (77.3 to 103.4)
86.1 (78.3 to 94.0)
95.8 (87.0 to 104.7)
129.0 (114.7 to 143.4)
136.4 (113.7 to 159.0)
117.2 (109.6 to 124.8)
130.1 (121.3 to 138.9)
173.6 (156.9 to 190.2)
183.0 (154.1 to 212.0)
241.6 (228.7 to 254.5)
265.1 (249.3 to 281.0)
340.0 (309.3 to 370.8)
355.4 (300.1 to 410.7)
* Rates are based on logistic regression analysis. CIN = cervical intraepithelial neoplasia; LEEP = loop electrosurgical excision procedure; laser = laser excision or
ablation; CI = confidence interval.
726 Articles | JNCI Vol. 101, Issue 10 | May 20, 2009
mean of 32 314 (87%) of the 37 142 women in the CIN cohort and
48 424 (68%) of the 71 213 women in the comparison cohort over
the fi rst 10 years. Our rates of CIN 2/3 cannot be applied to
women who were lost to follow-up after treatment. Because of the
linkage to the British Columbia Cancer Registry, however, we
were able to identify all women who were diagnosed with invasive
cervical cancer in British Columbia and thus to estimate the risk of
subsequent invasive cancer for the entire cohort, with the excep-
tion of those who migrated out of the province. The lower rates of
invasive cancers after the fi rst 8 years of follow-up that appeared in
women under active surveillance compared with women the over-
all CIN cohort indicate the importance of long-term surveillance
of women after treatment for CIN.
It is biologically plausible that long-term cryotherapy effi cacy
may be lower than that of excisional procedures because removing
rather than destroying tissue may provide greater protection.
Higher cancer rates after cryotherapy may also be related to
altered ability to adequately sample the transformation zone after
treatment, perhaps because structural changes of the cervix after
treatment for CIN differ between excisional and ablative proce-
dures. Alternatively, our exclusion of women with incomplete
excisional procedures (ie, those without clear surgical margins)
may have underestimated the subsequent recurrence risk in women
undergoing loop electrosurgical excision procedure and cone
biopsy. The effect of treatment type on subsequent CIN has been
inconsistent across studies. A meta-analysis of randomized con-
trolled trials of CIN treatment outcomes by Nuovo et al. ( 15 )
found that the median follow-up time for these trials was only
12 months, with CIN rates of 5% – 15%, and no statistically signifi -
cant outcome differences between treatment modalities. No inva-
sive cancers were reported during the follow-up of 3811 women in
these studies. Two more recent randomized controlled trials con-
Table 4 . Estimated rates of CIN 2/3 per 1000 women in the initial
6-year period for the comparison cohort by age group *
Estimated rate of CIN
2/3 in comparison group,
No. per 1000 women (95% CI)
21 – 29
30 – 39
40 – 49
54.8 (50.2 to 59.3)
26.3 (23.6 to 28.9)
9.8 (8.2 to 11.4)
10.9 (9.1 to 12.7)
* CIN 2/3 = cervical intraepithelial neoplasia 2 or 3; CI = confidence interval.
Table 6 . Association between risk factors and risk of invasive
cancer after treatment for CIN over the first 10 years of follow-up
(n = 37 142) *
Factor (comparison) Adjusted OR (95% CI)
Initial treatment (cryotherapy vs other)
Initial diagnosis (CIN 3 vs CIN 1 and 2)
Age ( ≥ 40 vs <40 y)
2.98 (2.09 to 4.26)
4.10 (2.70 to 6.22)
1.75 (1.12 to 2.74)
* Data are based on a Poisson regression analysis of the CIN cohort. Odds
ratios were adjusted for initial treatment, initial diagnosis, and age,
as appropriate. CIN = cervical intraepithelial neoplasia; OR = odds ratio;
CI = confidence interval.
CIN - under surveillance
CIN - all women
COMP - under surveillance
COMP - all women
95% CI for cumulative rate at 18 yrs: CIN - all women
95% CI for cumulative rate at 18 yrs: COMP - all women
Rate per 1000 women
02468 1012 1416 18 20
Time since initial treatment (years)
Figure 2 . Cumulative rates of invasive cervical cancer over the follow-up
period for all women in the cervical intraepithelial neoplasia (CIN) cohort,
for women in the CIN cohort under active surveillance, for the compari-
son cohort (COMP), and for the comparison cohort under surveillance
(COMP — under surveillance). We used the same defi nition as in the CIN
cohort (ie, must have had a cervical cytology within the previous 2 years
to be included for that year). Error bars = 95% confi dence intervals.
Table 5 . Stage distribution of invasive cervical cancer for all
women in the CIN and comparison cohorts, by stage of cancer *
No. in CIN
No. in comparison
* All women in the cohorts who developed cancer were included in this
analysis. CIN = cervical intraepithelial neoplasia.
cluded that women treated by loop electrosurgical excision proce-
dure had a lower rate of CIN after treatment than those treated
with cryotherapy ( 16 ) or laser vaporization ( 17 ). A report ( 18 ) of
follow-up for 2116 women treated by cryotherapy, laser excision
or ablation, or loop electrosurgical excision procedure for all
grades of CIN in the United Kingdom found that the rate of inva-
sive cervical cancer after treatment was 5.8 cancers per 1000
women for an 8-year period, whereas a long-term follow-up study
( 19 ) of 4417 women treated for CIN 3 with cone biopsy and whose
excised specimen had clear margins found no woman with invasive
cervical cancer during a median follow-up of 8.9 years. A recent
international systematic review of invasive cancer risk after treat-
ment for CIN found no statistically signifi cant differences in the
risk of invasive disease across treatment types ( 3 ).
The invasive cervical cancer rate in the CIN treatment cohort
of 37 cancers per 100 000 woman-years fell between the rate in a
Finnish cohort of 23 cancers per 100 000 woman-years ( 20 ) and the
rate in an international systematic review of cohort studies of
56 cancers per 100 000 woman-years ( 8 ). The ongoing higher rate
of invasive cancer among women in the CIN cohort, despite the
decline in the rates of CIN 2/3, was also observed in the systematic
review ( 8 ) and supports an underlying increased risk in this group.
However, the optimal length and intensity of surveillance for
this group remains unclear. Among the women who developed
invasive cancers in our study, the relatively high proportion of
stage I cancers (77%) attests to the benefi ts of ongoing surveillance,
JNCI | Articles 727
as compared with only 53% of cervical cancers that were diagnosed
at stage I in the US Surveillance, Epidemiology and End Results
cancer registry for the period from 1988 through 2003 ( 2 ).
Our study had several limitations. Interpretation of our results
is limited by the nature of observational cohort data. Treatment
patterns shifted in British Columbia over the study period, so that
loop electrosurgical excision procedure became more common and
laser excision and ablation became less common. Most treatment
was provided in the provincial colposcopy clinics and hence was
more likely to be relatively uniform and in accordance with guide-
lines for the study period. However, we were not able to document
the presence of satisfactory colposcopy results among the women
who were treated with cryotherapy or laser ablation. Although
guidelines in British Columbia endorsed cryotherapy or laser abla-
tion for women with satisfactory colposcopy only, we did not know
how many women were treated in accordance with guidelines.
We have limited information on regional differences in surveil-
lance practices or changes over time in British Columbia because
most surveillance was provided by general practitioners. To evalu-
ate further whether variation in treatment protocols or surveil-
lance practices may have contributed to our fi ndings, we examined
the outcomes of women who were initially treated at a single urban
tertiary hospital that treated and followed the highest volume of
women with CIN in British Columbia and that also used consis-
tent protocols. These outcomes were similar to those of overall
dataset (data not shown). Testing for HPV was not performed in
British Columbia during the study period, so we were limited to
evaluating outcomes on the basis of surveillance with cervical
cytology. Our analysis included only variables available in the data-
base and could not examine the associations of socioeconomic
status or race or ethnicity with outcomes.
The higher risk of CIN 2/3 and invasive cancer after treatment
for women who were treated with cryotherapy, particularly for
those treated for CIN 3, was substantial and differs from outcomes
reported in some randomized trials ( 15 ). If confi rmed, this fi nding
creates a dilemma for providers and their patients when making
decisions about treatment of CIN. In contrast to cryotherapy,
excisional methods of treatment have been associated with an
increased risk of both early complications of hemorrhage in ran-
domized trials ( 15 ) and with an increased risk of preterm delivery
and low birth weight in subsequent pregnancies in retrospective
studies ( 21 – 24 ). Further, cryotherapy is less costly to provide and
the technique is easier to learn, making it more readily available in
low-resource settings in which most women needing treatment for
CIN reside ( 23 ). Future randomized trials will need longer term
follow-up to defi ne the impact of treatment choice on subsequent
CIN and invasive cancer.
These fi ndings support the recent shift in the ASCCP guide-
lines for women who have been treated for CIN from indefi nite
annual screening to a return to routine screening after an initial
period of more intensive follow-up that may take the form of cytol-
ogy, HPV testing, or cytology with colposcopy during the fi rst
6 – 18 months. Given the rapid decline in subsequent diagnoses of
CIN 2/3 after the fi rst 2 years but the ongoing elevated risk of
invasive cervical cancer, it appears that consistency of follow-up for
10 – 20 years may be important for detecting disease after treatment.
More intensive follow-up strategies are likely to be important for
those women older than 40 years who were treated for CIN 3,
particularly if they were treated with cryotherapy. Cost-effectiveness
studies are needed to defi ne optimal surveillance strategies that
may differ by CIN grade, treatment type, and age.
1. Woolf SH . Screening for cervical cancer . In: Goldbloom RB , Lawrence RS ,
eds. Preventing Disease: Beyond the Rhetoric . New York, NY : Springer
Verlag ; 1990 : 319 – 323 .
2. Surveillance, Epidemiology, and End Results (SEER) Program . SEER*Stat
Database: Incidence – SEER 17 Regs Limited-Use, Nov 2006 Sub (1973 –
2004 varying) . National Cancer Institute, DCCPS, Surveillance Research
Program, Cancer Statistics Branch . Released April 2007, based on the
November 2006 submission. Available at: www.seer.cancer.gov .
3. Kyrgiou M , Tsoumpou I , Vredoussis T , et al . Management of minor
cervical abnormalities: a systematic review and a meta-analysis of the lit-
erature . Cancer Treat Rev. 2006 ; 32 ( 6 ): 515 – 523 .
4. The ASCUS-LSIL Triage Study (ALTS) Group . Results of a randomized
trial on the management of cytology interpretations of atypical squamous
cells of undetermined signifi cance . Am J Obstet Gynecol . 2003 ; 188 ( 6 ):
1383 – 1392 .
5. The ASCUS-LSIL Triage Study (ALTS) Group . A randomized trial on
the management of low-grade squamous intraepithelial lesion cytology
interpretations . Am J Obstet Gynecol . 2003 ; 188 ( 6 ): 1393 – 1400 .
6. Wright TC Jr , Cox JT , Massad LS , Carlson J , Twiggs LB , Wilkinson EJ .
2001 consensus guidelines for the management of women with cervical
intraepithelial neoplasia . Am J Obstet Gynecol . 2003 ; 189 ( 1 ): 295 – 304 .
7. Wright TC Jr , Massad LS , Dunton CJ , Spitzer M , Wilkinson EJ ,
Solomon D ; for the 2006 ASCCP-Sponsored Consensus Conference .
2006 consensus guidelines for the management of women with abnormal
cervical screening tests . J Low Genit Tract Dis . 2007 ; 11 ( 4 ): 201 – 222 .
8. Soutter WP , Sasieni P , Panoskaltsis . Long-term risk of invasive cervical
cancer after treatment of squamous cervical intraepithelial neoplasia . Int J
Cancer . 2006 ; 118 ( 8 ): 2048 – 2055 .
9. Cervical Cancer Screening Program . Screening for Cancer of the Cervix .
Vancouver, Canada : BC Cancer Agency ; 2007 .
10. Anderson GH , Boyes DA , Benedet JL , et al . Organisation and results of
the cervical cytology screening programme in British Columbia, 1955–85 .
BMJ . 1988 ; 296 ( 6627 ): 975 – 978 .
11. Petterssen F , Malker B . Invasive carcinoma of the uterine cervix following
diagnosis and treatment of in-situ carcinoma . Radiother Oncol . 1989 ; 16 ( 2 ):
115 – 120 .
12. Strander B , Andersson-Ellstrom A , Misom I , et al . Long term risk of inva-
sive cancer after treatment for cervical intraepithelial neoplasia grade 3:
population based cohort study . BMJ . 2007 ; 335 ( 7629 ): 1077 .
13. Cox JT , Schiffman M , Solomon D . Prospective follow-up suggests simi-
lar risk of subsequent cervical intraepithelial neoplasia grad 2 or 3
among women with cervical intraepithelial neoplasia grade 1 or negative
colposcopy and directed biopsy . Am J Obstet Gynecol . 2003 ; 188 ( 6 ):
1406 – 1412 .
14. Chan BKS , Melnikow J , Slee CA , Arellanes R , Sawaya GF . Post-treatment
human papillomavirus testing for recurrent cervical intraepithelial neopla-
sia: a systematic review [published online ahead of print January 22, 2009] .
Am J Obstet Gynecol . 2009 ; 200 ( 4 ): 422e1 – 422.e9 .
15. Nuovo J , Melnikow J , Willan AR , Chan BK . Treatment outcomes for
squamous intraepithelial lesions . Intl J Gynecol Obstet . 2000 ; 68 ( 1 ): 25 – 33 .
16. Chirenje ZM , Rusakaniko S , Akino V , Mlingo M . A randomized clinical
trial of loop electrosurgical excision procedure (LEEP) versus cryotherapy
in the treatment of cervical intraepithelial neoplasia . J Obstet Gynaecol .
2001 ; 21 ( 6 ): 617 – 621 .
17. Dey P , Gibbs A , Arnold DF , Saleh N , Hirsch PJ , Woodman CBJ . Loop
diathermy excision compared with cervical laser vaporization for the treat-
ment of intraepithelial neoplasia: a randomized controlled trial . BJOG .
2002 ; 109 ( 4 ): 381 – 385 .
18. Soutter WP , de Barros Lopes A , Fletcher A , et al . Invasive cervical cancer
after conservative therapy for cervical intraepithelial neoplasia . Lancet .
1997 ; 349 ( 9057 ): 978 – 980 .
728 Articles | JNCI Vol. 101, Issue 10 | May 20, 2009
19. Reich O , Pickel H , Lahousen M , Tamussino K , Winter R . Cervical intra-
epithelial neoplasia III: long-term outcome after cold-knife conization
with clear margins . Obstet Gynecol . 2001 ; 97 ( 3 ): 428 – 430 .
20. Kalliala I , Anttila A , Pukkala E , Nieminen P . Risk of cervical and other
cancers after treatment of cervical intraepithelial neoplasia: retrospective
cohort study . BMJ . 2005 ; 331 ( 7526 ): 1183 – 1185 .
21. Kyrgiou M , Koliopoulos G , Martin-Hirsch P , Arbyn M , Prendiville W ,
Paraskevaidis E . Obstetric outcomes after conservative treatment for
intraepithelial or early invasive cervical lesions: systematic review and
meta-analysis . Lancet . 2006 ; 367 ( 9509 ): 489 – 498 .
22. Bruinsma F , Lumley J , Tan J , Quinn M . Precancerous changes in
the cervix and risk of subsequent preterm birth . BJOG . 2007 ; 114 ( 1 ):
70 – 80 .
23. Albrechtsen S , Rasmussen S , Thoresen S , Irgens LM , Iversen OE .
Pregnancy outcome in women before and after cervical conisation:
population based cohort study . BMJ . 2008 ; 337 : a1343 . doi:10.1136/
24. Arbyn M , Kyrgiou M , Simoens C , et al . Perinatal mortality and other severe
adverse pregnancy outcomes associated with treatment of cervical intraepithe-
lial neoplasia: meta-analysis . BMJ . 2008 ; 337 : a1284 . doi:10.1136/bmj.A1284 .
National Cancer Institute (grant 1R01CA109142 ).
The authors had full responsibility for the design of the study, the collection
of the data, the analysis and interpretation of the data, the decision to sub-
mit the manuscript for publication, and the writing of the manuscript The
authors thank Catarina Kiefe, MD, PhD, for critical review of the manuscript
and Mairin Rooney and Galen Sanderlin for assistance with the fi gures and
Manuscript received July 17 , 2008 ; revised March 3 , 2009 ; accepted
March 12 , 2009 .