Journal of the National Cancer Institute Advance Access published online on April 29, 2008
JNCI Journal of the National Cancer Institute, doi:10.1093/jnci/djn099
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Author 2008. Published by Oxford University Press.
ARTICLES |
Screening-Preventable Cervical Cancer Risks: Evidence From a Nationwide Audit in Sweden
Affiliations of authors: Department of Obstetrics and Gynecology, Gävle Hospital, Gävle, Sweden (BA); Cancer Screening Unit, Oncologic Center M8, Karolinska University Hospital, Stockholm, Sweden (LK, ST); Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden (PS); Department of Obstetrics and Gynecology, Umeå University Hospital, Umeå, Sweden (LS); Oncologic center, Western Region, Sahlgrenska University Hospital, Göteborg, Sweden (BS); Department of Pathology and Clinical Cytology, Sahlgrenska University Hospital, Göteborg, Sweden (WR); Department of Medical Microbiology, Lund University, University Hospital, Malmö, Sweden (JD)
Correspondence to: Bengt Andrae, MD, Department of Obstetrics and Gynecology, Gävle Hospital, S-801 87 Gävle, Sweden (e-mail: bengt.andrae{at}gmail.com).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Background: The effectiveness of cervical cancer screening programs differs widely in different populations. The reasons for these differences are unclear. Routine and comprehensive audits have been proposed as an ethically required component of screening. We performed a nationwide audit of the effectiveness of the Swedish cervical cancer screening program.
Methods: We identified all invasive cervical cancer cases that were diagnosed in Sweden from January 1, 1999, through December 31, 2001, and had been reported to the Swedish Cancer Registry (n = 1230 cases). We verified the diagnoses by histopathologic rereview and matched each case subject to five (population-based) age-matched control subjects who were identified from the National Population Register. The Pap smear screening histories for case and control subjects were reviewed for a 6-year period using the National Cervical Cancer Screening Register, which contains data on essentially all relevant cytological and histological diagnoses in Sweden. Odds ratios (ORs), and their 95% confidence intervals (CIs), of cervical cancer according to screening history were calculated in conditional logistic regression models. All statistical tests were two-sided.
Results: Women who had not had a Pap smear within the recommended screening interval had higher risk of cervical cancer than women who had been screened (OR = 2.52, 95% CI = 2.19 to 2.91). This risk was similarly increased for all age groups (Phomogeneity = .96). The risk for nonsquamous cell cervical cancers (OR = 1.59, 95% CI = 1.20 to 2.11) was also increased. Women who had not had a Pap smear within the recommended screening interval had a particularly high risk of advanced cancers (OR = 4.82, 95% CI = 3.61 to 6.44). Among women who had been screened within the recommended interval, those with abnormal Pap smears had a higher risk of cervical cancer than those with normal smears (OR = 7.55, 95% CI = 5.88 to 9.69) and constituted 11.5% of all women with cervical cancer.
Conclusions: Nonadherence to screening intervals was the major reason for cervical cancer morbidity. The screening program was equally effective for women of all ages and was also effective against nonsquamous cancers.
Prior knowledge The effectiveness of cervical cancer screening programs differs widely in different populations, but the reasons are unclear. Routine and comprehensive audits have been proposed as an ethically required component of screening. Study design A comprehensive nationwide audit of the effectiveness of the organized cervical cancer screening program in Sweden. Contribution Not having had a Pap smear taken within the recommended screening interval was the most important risk factor for cervical cancer in the presence of a screening program. The screening program was equally effective for women of all ages, including women younger than 30, and was also effective against nonsquamous cancers. Implications Compliance with screening recommendations and high population coverage of screening are vital for success. Limitations Attendance to the screening program (ie, the relationship between invitations issued and tests performed) was not studied because only some counties kept identifiable information on when and to whom invitations had been issued.
|
Cervical cancer is the second most commonly diagnosed cancer in women worldwide; in 2002, half a million incident cases of cervical cancer were reported and there were more than a quarter of a million deaths from this disease (1). In many industrialized countries, the incidence of cervical cancer has decreased by the use of Pap smear screening. In Sweden, for example, the overall incidence of cervical cancer declined by 67% over a 40-year period, from 20 cases per 100 000 women (world standard rate) in 1965 to 6.6 cases per 100 000 women in 2005 (2).
The major cause of cervical cancer is infection with oncogenic types of human papillomavirus (HPV) (3). Although vaccines to prevent infection with specific oncogenic HPVs are now available, it will take many years for their effects on cervical cancer morbidity to be known (4). Until then, cervical cancer screening programs will remain the primary preventive strategy for this cancer (5).
The rationale for cervical cancer prevention by screening is primarily to reduce the incidence of the disease through the detection and removal of precancerous lesions and secondarily to reduce disease progression through the detection of invasive cancers in the early stages, thereby improving chances for a cure and reducing mortality. Cytological screening for cervical cancer (ie, via a Pap smear) is highly effective in reducing the incidence of squamous cell cancer, although its effect on the incidence of adenocarcinoma is less well established (6,7). However, the effect of screening programs on cervical cancer morbidity at the population level has varied substantially among different countries, primarily because of differences in the implementation of organized screening and coverage of the target population (8–12). Cervical cancer still occurs even in countries that have an organized program for cytological screening of cervical cancer, but the reasons need to be clarified in more detail to enable targeted improvement. Although the need for routine and comprehensive audits of screening programs that include the entire population that is being targeted by a program has been emphasized repeatedly (13–15), such audits have so far not been implemented. Audits may greatly assist in improving the effectiveness of screening programs that are intended to prevent cancer by evaluating the extent of the morbidity that can be affected by changes in program organization, screening methods, or treatment and follow-up strategies for treating abnormalities detected by screening (8,11,16).
We developed a protocol for a comprehensive nationwide audit of the Swedish cervical cancer screening program. The aims of the audit were to set the standards for routine monitoring of the reasons why cervical cancers still occur despite organized screening and to assist in prioritizing efforts to improve the effectiveness of screening programs.
|
Subjects and Methods |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The Swedish Cervical Cancer Screening Program
Organized screening for cervical cancer was introduced in Sweden between 1967 and 1977 (17). The national guidelines for organized cervical cancer screening recommend Pap smear screening for all women between the age of 23 and 60 years, with 3-year screening intervals for women aged 23–50 years and 5-year intervals for women aged 51–60 years (18). Women older than 60 years are not invited to the cervical cancer screening program in Sweden because women who have been screened up to age 60 and have had no previous abnormal smears are considered to be at low risk for cervical cancer. Because health care in Sweden is organized at the county level, there are minor differences in how these national recommendations are implemented, particularly in terms of age limits and screening intervals (18). For example, in 5 of 21 counties, women who are eligible for invitation to the screening program are identified by birth cohort, whereas in the other 16 counties, eligible women are identified by the time that has elapsed since their last Pap smear. Information on all Pap smears taken within or outside the organized screening program in a county is stored in a common database; therefore, women who have Pap smears taken outside the organized program are invited for screening but not until a time period equal to a screening interval has elapsed.
Identification of Case and Control Subjects
All cases of invasive cervical epithelial carcinoma and unspecified uterine cancer (19) that were reported to the nationwide Swedish Cancer Registry for the 3-year period from January 1, 1999, through December 31, 2001, were retrieved. Swedish physicians and pathologists are required to report all new cancers to this registry (2,20). To verify that all cervical cancer cases included in the cancer register were actually cervical cancers, we obtained the archival histological specimens of all registered cervical cancers and subjected them to histopathologic rereview by an expert pathologist (WR). No archival histological specimens were available for 40 cases. However, we included these 40 cases in the analysis because the clinical and pathology records for these cases indicated that they had been reviewed and that the reviews had verified the original diagnosis of primary invasive cervical carcinoma.
A total of 1335 cervical cancer cases were reported to the Cancer Registry from January 1, 1999, through December 31, 2001. Upon histopathologic rereview, we excluded 45 cases of cancers of noncervical origin, 1 case of nonepithelial cervical cancer, 47 cases of noninvasive cancers, and 23 cases that were recurrences of cervical cancer according to the cancer register. Among the 89 unspecified uterine cancers that were reported to the Cancer Registry during this time period, 11 were found to be cervical cancers upon histopathologic re-review and were included in the analysis. Altogether, 1230 case subjects were included. Of 6150 age-matched control subjects, 26 were excluded because they had been previously diagnosed with cervical cancer; thus, 6124 control subjects were included.
All case subjects were classified by their age at diagnosis and by clinical tumor stage (microinvasive [stage IA], localized [stage IB], or advanced [stage II or higher]) according to International Federation of Gynecology and Obstetrics (FIGO) guidelines (21) (Table 1). Information about the tumor stage at diagnosis for each case subject was collected from regional databases or from clinical records that were kept by the clinical departments that reported the cases to the cancer registry. All case subjects were also classified by histological type (squamous cell carcinoma, adenocarcinoma, adenosquamous carcinoma, or "other epithelial tumor," which included neuroendocrine, poorly differentiated, and small cell carcinomas). The overall distribution of case subjects by age and stage is shown in Figure 1. For the analysis, we then divided all subjects into three age groups: those younger than age 30, for whom the benefit of screening has been questioned; those aged 30–65, for whom Pap smears should be taken within the recommended screening interval; and those older than 65, who are expected to be at low risk of cervical cancer provided that they have had only normal Pap smears up to the age of 60 (Table 1).
|
|
Each case subject was matched by year of birth to five control subjects who were randomly selected from the National Population Register at Statistics Sweden (Örebro, Sweden). Each control subject had to be alive on the date of diagnosis for the case subject and not have been diagnosed with cervical cancer up to the date of diagnosis for the respective case subject. All residents of Sweden are assigned a unique personal identification number that allows the linkage of data, for example, between health care registers and population registers. We used these numbers to identify all case and control subjects and to link them to all registers and clinical records.
This study was approved by the Ethical Review Boards in Uppsala, Stockholm, Umeå, Gothenburg, and Lund and the joint board of Örebro and Linköping, Sweden.
Screening History
The National Cervical Cancer Screening Register (Stockholm, Sweden) collects all cervical cytology and histopathology reports for women who underwent cervical cancer screening in Sweden beginning in 1993; these reports are provided by the diagnostic databases at all 30 pathology and cytology laboratories in Sweden. Each report notes the personal identification number of the woman who was tested, the date of the test, the specimen number, a code that identifies the laboratory that analyzed the test, and one or more diagnostic codes. All diagnostic codes for cytology and all combinations of codes used in these reports were converted to the 14 Systematized Nomenclature of Medicine codes that are currently in use in Sweden (22). Cytology data for women who were tested in 1995 or later are nearly 98% complete; the missing data are from a single laboratory that provides services to 1.3% of the population of Sweden. Cytology data for women who tested in 1993 and 1994 are 93% and 97% complete, respectively.
An evaluable screening history was defined as the availability of all smears and cervical histopathology specimens during the 6-year period that started 6.5 years before and ended 6 months before the date of cervical cancer diagnosis for case subjects or the corresponding date for the control subjects. We chose this period because 1) 6 years corresponds to two screening rounds in the 23- to 50-year age group and 2) 6.5 years was the longest time for which we had access to nationwide data for all case and control subjects in this study. In addition, the process of evaluation and treatment of an abnormal smear normally takes up to half a year. Pap smears or biopsies that were taken during the 6 months before the date of cancer diagnosis were therefore not considered as part of the evaluable screening history for case or control subjects but rather as tests that led to the detection of invasive cancer (16).
To avoid differential exclusion of screening tests for case subjects, we assessed the screening histories of all subjects during a window of screening exposure that was exactly one screening interval long (as defined by the subject’s age) and that ended 6 months before the cervical cancer diagnosis (23,24). A woman was considered to have been tested within the recommended screening interval if she was 53 years or younger and had a smear taken from 6 to 42 months (0.5–3.5 years) before a cervical cancer diagnosis; for women aged 54–65 years, for whom a 5-year screening interval applies, the smear had to be taken from 6 to 66 months (0.5–5.5 years) before the cervical cancer diagnosis. We also assessed whether women aged 66 years or older had had a smear taken within 0.5–6.5 years before a cancer diagnosis. Screening was defined as normal if at least one smear had been taken within the recommended interval of the screening program and no cytological abnormality was reported in the period 0.5–6.5 years, that is, more than one screening interval, before the cancer diagnosis of the case subject. A screened woman was considered to have had an abnormal smear if a test was taken within the recommended interval and at least one smear had atypical squamous cells of uncertain significance, a low-grade squamous intraepithelial lesion, or a high-grade squamous intraepithelial lesion in the period 0.5–6.5 years before cancer diagnosis. A smear that was labeled "not satisfactory for evaluation" was also classified as abnormal unless it was followed by a normal smear. An abnormal smear that was followed by a report of a histopathologic specimen was considered to indicate an assessment with a biopsy by a gynecologist. Pap smears that were taken during the 6 months before a cervical cancer diagnosis were analyzed separately to see if there was a shift toward lower FIGO stages among case subjects whose cancers were presumed to be screen detected. For this analysis, symptomatic cases were defined as those for which a Pap smear was taken less than 1 month before diagnosis or for which no Pap smear was taken during the 6 months before diagnosis and screen-detected cases were defined as those for which a Pap smear was taken 1–6 months before diagnosis.
Statistical Methods
The Pearson 
2 test was used to test for independency in FIGO stage distribution between case subjects with and without a Pap smear within the recommended screening interval, between case subjects with normal and abnormal Pap smears, and between screen-detected and symptomatic cases. Odds ratios (ORs) for cervical cancer associated with screening history were estimated with 95% confidence intervals (CIs) in conditional logistic regression models that were stratified by birth year. In these analyses, histological type was classified as squamous (squamous cell carcinoma) or nonsquamous (adenocarcinoma, adenosquamous carcinoma, small cell carcinoma, poorly differentiated carcinoma, and neuroendocrine carcinoma). When analyzing subgroups, we sometimes had to use unconditional logistic regression that controlled for age in 5-year age classes because case subjects and their matched control subjects ended up in different risk sets. We performed interaction tests using a conditional logistic regression model to estimate P values for homogeneity in odds ratios for women who were tested vs women who were not tested over categories of age at diagnosis and FIGO stage. SAS statistical software (version 9.1; SAS Institute, Cary, NC) was used for all analyses. This study was powered to detect an odds ratio of 1.34 or higher, given a 10% prevalence of a risk factor in the study population. All statistical tests were two-sided.
|
Results |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
A total of 1230 case subjects—all of whom were classified by age at diagnosis, FIGO stage, and histopathologic type—were included in this study (Table 1). Approximately two-thirds of the case subjects (n = 840) were younger than 66 years at diagnosis and thus had fully evaluable screening histories. The distribution of case subjects by age at diagnosis and FIGO stage is presented in Figure 1.
Sixty-three (5%) of the 1230 case subjects were younger than age 30 years at diagnosis (Table 1); only one case subject was younger than 23 years at diagnosis (data not shown). Squamous cell carcinoma was the predominant histological type in all age groups, and adenocarcinomas were more common among women who were 30–65 years old at diagnosis than among women who were younger than 30 or older than 65 at diagnosis (Table 1).
Of the 39 women who were diagnosed with a small cell, poorly differentiated, or neuroendocrine cervical carcinoma, 29 did not have a Pap smear taken during the recommended screening interval and 28 had advanced-stage disease (data not shown). Twenty-six of these 39 women were older than 60 at diagnosis (data not shown).
A total of 789 (64%) of all case subjects and 394 (83%) of the 477 case subjects who were diagnosed with advanced cervical cancer (FIGO stage II or higher) did not have a Pap smear taken during the recommended screening interval (Table 2). Approximately 50% of the case subjects who did not have a Pap smear during the recommended screening interval were diagnosed with advanced disease, compared with less than 19% of those who did have a Pap smear during the recommended interval (PFIGO stage < .001) (Table 2). The proportion of cancers at higher FIGO stages increased steadily with increasing age at diagnosis (Figure 1), and 55% of the advanced-stage cancers (264/477) were diagnosed in women who were older than 65 and did not have a Pap smear within 0.5–6.5 years before diagnosis (data not shown).
|
Women who had not had a Pap smear taken within the recommended screening interval had a higher risk of cervical cancer than women who had been tested (OR = 2.52, 95% CI = 2.19 to 2.91) (Table 3). Risks of both squamous cell carcinoma (OR = 2.97, 95% CI = 2.51 to 3.50) and nonsquamous (primarily adenocarcinoma) cervical carcinoma (OR = 1.59, 95% CI = 1.20 to 2.11) were increased.
|
Among screened women, those whose Pap smears were abnormal had a higher risk of cervical cancer than those whose smears were normal (OR = 7.55, 95% CI = 5.88 to 9.69) (Table 3). Of the 1230 case subjects, 141 (11.5%) had an abnormal Pap smear. Of the 6124 control subjects, 191 (3.1%) had an abnormal Pap smear. However, only 4.2% (20/477) of the case subjects with advanced cancer had an abnormal Pap smear. Among women who had an abnormal Pap smear, those who did not have a follow-up biopsy had a higher risk of cervical cancer than those who did (OR = 1.89, 95% CI = 1.19 to 3.02). The risk of any cervical cancer for women without a Pap smear vs women with a Pap smear was statistically significantly increased with increasing FIGO stage (ORmicroinvasive cancer = 1.70, 95% CI = 1.28 to 2.26; ORlocalized cancer = 2.10, 95% CI = 1.71 to 2.59; ORadvanced cancer = 4.82, 95% CI = 3.61 to 6.44; Phomogeneity < .001) (Table 3). We observed a similar pattern of increasing risk with increasing FIGO stage for nonsquamous cervical carcinomas, but it was not statistically significant (Phomogeneity = .29). The risk of cervical cancer for women without vs with a Pap smear was similarly increased for all age groups (Phomogeneity = .96; Table 3).
The FIGO stage distribution for all case subjects (n = 1230) was 20.4% stage IA disease, 40.8% stage IB, and 38.8% stage II or higher. For the presumably screen-detected cases (ie, those with a Pap smear taken 1–6 months before diagnosis; n = 305), the stage distribution was further shifted toward earlier stages; 46.5% had stage IA, 45.6% had stage IB, and 8.9% had stage II or higher compared with symptomatic cases (PFIGO stage < .001; Table 2). We detected no increase in screening intensity among either the case subjects or the matched control subjects during the year before the window of screening exposure (data not shown).
|
Discussion |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Here, we report results of a comprehensive nationwide audit of an organized cervical cancer screening program. The key features of the audit were 1) use of nationwide cancer registry to identify cervical cancer cases; 2) histopathologic rereview of diagnostic slides; 3) inclusion of unspecified uterine cancers in the rereview, which identified additional verified cervical cancers; 4) nationwide and comprehensive collection of all cytological and histopathologic data, including smears taken outside the organized screening program; 5) control subjects who were identified from the nationwide population register; and 6) permission from ethical review boards to evaluate the effect of the program on the entire population without having to obtain informed consent, using data on individual subjects from national registers. Because we used the same source of screening history data for the case and control subjects and a window of exposure to screening that was equal to one screening interval, we were able to avoid selection, recall, and testing biases (23–25). As expected, we found that not having had a Pap smear taken within the recommended screening interval was the most important risk factor for cervical cancer in the presence of a screening program: approximately 64% of all cervical cancers and 83% of the advanced cases were diagnosed in women who were not tested. Screening was also associated with decreased risks of cervical cancer among women who were younger than 30 or who were older than 65 years. Together, these results indicate that compliance with screening recommendations and high population coverage of screening are vital for success and that attention should be paid to underscreened older women.
Only a few attempts have been made to comprehensively audit an organized screening program (11,26). The largest audit published to date is from the United Kingdom (16). In that study, the odds ratios of acquiring invasive cervical cancer of stage IB or higher after 3 years since the last negative smear were 0.28 and 0.12 at ages 20–39 and 40–54, respectively, compared with women with no previous negative smear. After 5 years, the corresponding odds ratios were 1.03 and 0.39, respectively. Although the UK audit used a different approach to estimate the relative protection of cervical cancer screening (time since last negative smear) and only included invasive cervical cancers of stage IB or higher, their results are in accordance with what we found, with one exception. The UK audit concluded that screening afforded less protection against cervical cancer among younger women, whereas we found no evidence that younger women without a Pap smear within the recommended screening interval should be at a relatively lower risk than older women. The reason for this discrepancy could be the different methodological approaches used, but it is also possible that there are real differences in the natural history of the disease or in the effectiveness of cervical cancer screening programs among countries.
Epidemiological evaluation of the effectiveness of screening commonly relies on the identification of trends in incidence or mortality rates over time. However, because the background risks may also change over time, this approach can be problematic, particularly when analyzing subgroups that have small numbers of cases because they tend to have limited statistical power in analyses that evaluate short time spans. Two such subgroups that have been difficult to study with respect to the effect of screening are women younger than 30 and women diagnosed with adenocarcinoma (26,27). In both of these groups, the incidence of cervical cancer has not declined since screening was introduced, and in some countries incidence has even increased (6,9,26–30). Our access to complete databases, combined with the fact that both the case and control subjects were identified from the nationwide population register, made it possible to study the effectiveness of the Swedish cervical cancer screening program in these subgroups in a reliable manner and with adequate statistical power. There was clear evidence that the screening program was effective in reducing the incidence of cervical cancer in women younger than 30 and also that it reduced the incidence of adenocarcinoma (Table 3). One interpretation of the stable, or even increasing, incidences of cervical cancer in women younger than 30 years of age and of adenocarcinoma of the cervix is that they are the net result of increases in background risks and the counteracting effect of the screening program (31–33).
Screening of women younger than 30 years of age has been described as having little benefit because of the faster progression from precursors to cancer and a higher proportion of nonsquamous cell cancers compared with older age groups (16,27,34). In this audit, the lowest proportion of nonsquamous cell cervical cancer cases was in the under-30 age group. Although our data support screening of women who are younger than 30 years of age, the fact that only one case of invasive cervical cancer occurred in a women who was younger than the lower age limit of the screening program supports the view that screening of women younger than 23 is unnecessary.
We found that women older than 65 who had a Pap smear taken during the previous 6.5 years had a statistically significant lower risk of cervical cancer compared with those who did not have a Pap smear taken during that time (Table 3). However, women older than 65 with a Pap smear in the last 6.5 years may also have participated to a greater extent in screening before the age of 60 than women of the same age who did not have a recent Pap smear. Because we did not have data to study this hypothesis, the finding should not be taken as evidence that screening should be extended after the age of 60 in women who have previously had normal Pap smears. Our finding that screening is associated with reduced risks for advanced-stage cervical cancers is in accordance with results of previous studies (9,35,36).
The strongly elevated risk of cervical cancer (OR = 7.55, 95% CI = 5.88 to 9.69) associated with having an abnormal smear can be seen as a failure in the screening system (Table 3). Unless an abnormal smear signals an already invasive cancer, having an abnormal Pap smear per se should not necessarily be associated with an increased risk of cervical cancer because the subsequent assessment and possible treatment should protect a woman from invasive cancer; indeed this is the very foundation of screening. The finding that 11.5% of the cervical cancer case subjects in our study had an abnormal Pap smear as opposed to only 3.1% of control subjects implies that the risk of cervical cancer could also be reduced by improving the management of abnormal Pap smears. Furthermore, women with abnormal Pap smears who did not have a follow-up biopsy had a higher risk of cervical cancer than those with a follow-up biopsy (OR = 1.89, 95% CI = 1.19 to 3.02). There are several reasons why an abnormal smear may not be followed by a biopsy: 1) the smear may be overlooked or ignored by the medical professionals; 2) the woman may refrain from follow-up; 3) the follow-up may have been only a repeated Pap smear, which has been an accepted and common procedure in Sweden for low-grade and borderline abnormalities; and/or 4) a follow-up examination with colposcopy may not have found any lesion on which to perform a biopsy. This study cannot differentiate between these possibilities. However, the fact that 7% of the cancer cases in our study were diagnosed in women who had an abnormal smear that was not followed up with histopathology indicates that the assessment of abnormal smears is an area that needs improvement, as has been pointed out in several studies (10,11,37,38).
The main focus of our audit was to evaluate the effectiveness of an organized screening program in preventing invasive cervical cancer through the detection and further assessment of precursors, procedures that can take up to 6 months. Therefore, we omitted from the analysis any smears that were taken within 6 months of a diagnosis from both the case subjects and the matched control subjects. This approach does not exclude smears that potentially could have prevented disease that had already advanced to the occult invasive phase (24). For all case and control subjects, we used the same "window of exposure" to screening, which corresponded to exactly one age-related screening interval that ended 6 months before the case subject's diagnosis, so we can conclude that there was no differential exclusion of screening tests for the case subjects (23,24).
Smears taken within 6 months of diagnosis of the cancer could not have prevented the invasive cancer because it was already there before the assessment and treatment process was finished. These smears are, however, interesting in another respect. Screen detection of invasive cancer may be beneficial to the prognosis of a cancer patient due to detection at earlier stages. Such a benefit seems to be present given that the stage distribution among presumably screening-detected invasive cancers in our study was shifted toward lower stages.
A weakness of this audit was that we were not able to study attendance to the screening program—that is, the relationship between invitations issued and tests performed—because only some counties kept identifiable information on when and to whom invitations had been issued. Future audits should examine whether a lack of screening is due to the failure to issue invitations or to nonresponse to a screening invitation. Quality assurance of the "fail-safe" systems in use is important to ensure that all eligible women are indeed invited and that the screening tests are taken without delay (13,39–41).
In summary, this study highlights the benefit of reliable population-based registers in the analysis of reasons for remaining incidence of cervical cancer in spite of existence of an organized screening program. As expected, imperfect population coverage of screening was by far the most important risk factor for incident cervical cancer. Abnormal smears, particularly if not followed up by a biopsy, were also an important risk factor, which suggests that improved follow-up programs are warranted. On the other hand, it was possible to demonstrate a beneficial effect of screening in younger women and against adenocarcinomas of the cervix.
|
Funding |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The Swedish Cancer Society (02-6988); Research and Development Center of Gävleborg County (2002-03-27).
|
NOTES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The National Cervical Cancer Screening Registry was supported by the Swedish National Board of Health and Welfare.
The funding bodies had no role in data extraction and analyses, in the writing of the manuscript, or in the decision to submit the manuscript for publication.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
1. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin (2005) 55(2):74–108.
2. Socialstyrelsen. Cancer Incidence in Sweden 2005 (2007) Stockholm, Sweden: National Board of Health and Welfare.
3. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol (2002) 55(4):244–265.
4. Goldie SJ, Kohli M, Grima D, et al. Projected clinical benefits and cost-effectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst (2004) 96(8):604–615.
5. Franco EL, Harper DM. Vaccination against human papillomavirus infection: a new paradigm in cervical cancer control. Vaccine (2005) 23(17–18):2388–2394.[CrossRef][Web of Science][Medline]
6. Bray F, Carstensen B, Moller H, et al. Incidence trends of adenocarcinoma of the cervix in 13 European countries. Cancer Epidemiol Biomarkers Prev (2005) 14(9):2191–2199.
7. Bergstrom R, Sparen P, Adami HO. Trends in cancer of the cervix uteri in Sweden following cytological screening. Br J Cancer (1999) 81(1):159–166.[CrossRef][Web of Science][Medline]
8. Madlensky L, Goel V, Polzer J, Ashbury FD. Assessing the evidence for organised cancer screening programmes. Eur J Cancer (2003) 39(12):1648–1653.[CrossRef][Web of Science][Medline]
9. Bray F, Loos AH, McCarron P, et al. Trends in cervical squamous cell carcinoma incidence in 13 European countries: changing risk and the effects of screening. Cancer Epidemiol Biomarkers Prev (2005) 14(3):677–686.
10. Andrae B, Smith P. Clinical impact of quality assurance in an organized cervical screening program. Acta Obstet Gynecol Scand (1999) 78(5):429–435.[CrossRef][Web of Science][Medline]
11. Sung HY, Kearney KA, Miller M, Kinney W, Sawaya GF, Hiatt RA. Papanicolaou smear history and diagnosis of invasive cervical carcinoma among members of a large prepaid health plan. Cancer (2000) 88(10):2283–2289.[CrossRef][Web of Science][Medline]
12. Leyden WA, Manos MM, Geiger AM, et al. Cervical cancer in women with comprehensive health care access: attributable factors in the screening process. J Natl Cancer Inst (2005) 97(9):675–683.
13. Sasieni P, Cuzick J. Routine audit is an ethical requirement of screening. BMJ (2001) 322(7295):1179.
14. Slater DN, Milner PC, Radley H. Audit of deaths from cervical cancer: proposal for an essential component of the National Screening Programme. J Clin Pathol (1994) 47(1):27–28.
15. McKenzie CA, Duncan ID. The value of cervical screening in women over 50 years of age–time for a multicentre audit. Scott Med J (1998) 43(1):19–20.[Web of Science][Medline]
16. Sasieni P, Adams J, Cuzick J. Benefit of cervical screening at different ages: evidence from the UK audit of screening histories. Br J Cancer (2003) 89(1):88–93.[CrossRef][Web of Science][Medline]
17. Dillner J. Cervical cancer screening in Sweden. Eur J Cancer (2000) 36(17):2255–2259.[CrossRef][Web of Science][Medline]
18. Socialstyrelsen. Gynekologisk cellprovskontroll. (1998) Stockholm, Sweden: National Board of Health and Welfare.
19. Loos AH, Bray F, McCarron P, Weiderpass E, Hakama M, Parkin DM. Sheep and goats: separating cervix and corpus uteri from imprecisely coded uterine cancer deaths, for studies of geographical and temporal variations in mortality. Eur J Cancer (2004) 40(18):2794–2803.[CrossRef][Web of Science][Medline]
20. Mattsson B, Wallgren A. Completeness of the Swedish Cancer Register. Non-notified cancer cases recorded on death certificates in 1978. Acta Radiol Oncol (1984) 23(5):305–313.[Web of Science][Medline]
21. Quinn MA, Benedet JL, Odicino F, et al. Carcinoma of the cervix uteri. FIGO 6th Annual Report on the Results of Treatment in Gynecological Cancer. Int J Gynaecol Obstet (2006) 95(suppl 1):S43–S103.[CrossRef][Medline]
22. KVAST Swedish Society of Pathology. Exfoliativ cytologi—Vaginalcytologi. http://www3.svls.se/sektioner/pa/Vaginalcytologiv2.pdf. (in Swedish) 2006. Accessed February 3, 2008.
23. Weiss NS. Application of the case-control method in the evaluation of screening. Epidemiol Rev. (1994) 16(1):102–108.
24. Weiss NS. Case-control studies of the efficacy of screening tests designed to prevent the incidence of cancer. Am J Epidemiol (1999) 149(1):1–4.
25. Cohen MM. Using administrative data for case-control studies: the case of the Papanicolaou smear. Ann Epidemiol (1993) 3(1):93–98.[Medline]
26. Sigurdsson K, Sigvaldason H. Effectiveness of cervical cancer screening in Iceland, 1964–2002: a study on trends in incidence and mortality and the effect of risk factors. Acta Obstet Gynecol Scand (2006) 85(3):343–349.[Web of Science][Medline]
27. Chan PG, Sung HY, Sawaya GF. Changes in cervical cancer incidence after three decades of screening US women less than 30 years old. Obstet Gynecol (2003) 102(4):765–773.[CrossRef][Web of Science][Medline]
28. Anttila A, Pukkala E, Soderman B, Kallio M, Nieminen P, Hakama M. Effect of organised screening on cervical cancer incidence and mortality in Finland, 1963-1995: recent increase in cervical cancer incidence. Int J Cancer (1999) 83(1):59–65.[CrossRef][Web of Science][Medline]
29. Bulk S, Visser O, Rozendaal L, Verheijen RH, Meijer CJ. Cervical cancer in the Netherlands 1989-1998: Decrease of squamous cell carcinoma in older women, increase of adenocarcinoma in younger women. Int J Cancer (2005) 113(6):1005–1009.[CrossRef][Web of Science][Medline]
30. Herbert A, Singh N, Smith JA. Adenocarcinoma of the uterine cervix compared with squamous cell carcinoma: a 12-year study in Southampton and South-west Hampshire. Cytopathology (2001) 12(1):26–36.[CrossRef][Web of Science][Medline]
31. Zheng T, Holford TR, Ma Z, et al. The continuing increase in adenocarcinoma of the uterine cervix: a birth cohort phenomenon. Int J Epidemiol (1996) 25(2):252–258.
32. Vizcaino AP, Moreno V, Bosch FX, Munoz N, Barros-Dios XM, Parkin DM. International trends in the incidence of cervical cancer: I. Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer (1998) 75(4):536–545.[CrossRef][Web of Science][Medline]
33. Mitchell H, Hocking J, Saville M. Improvement in protection against adenocarcinoma of the cervix resulting from participation in cervical screening. Cancer (2003) 99(6):336–341.[CrossRef][Web of Science][Medline]
34. Rieck GC, Tristram A, Hauke A, Fielder H, Fiander AN. Cervical screening in 20-24-year olds. J Med Screen (2006) 13(2):64–71. discussion 62–63.
35. Nygard JF, Nygard M, Skare GB, Thoresen SO. Screening histories of women with CIN 2/3 compared with women diagnosed with invasive cervical cancer: a retrospective analysis of the Norwegian Coordinated Cervical Cancer Screening Program. Cancer Causes Control (2005) 16(4):463–474.[CrossRef][Web of Science][Medline]
36. Kinney W, Sawaya GF, Sung HY, Kearney KA, Miller M, Hiatt RA. Stage at diagnosis and mortality in patients with adenocarcinoma and adenosquamous carcinoma of the uterine cervix diagnosed as a consequence of cytologic screening. Acta Cytol (2003) 47(2):167–171.[Web of Science][Medline]
37. Andersson-Ellstrom A, Seidal T, Grannas M, Hagmar B. The pap-smear history of women with invasive cervical squamous carcinoma. A case-control study from Sweden. Acta Obstet Gynecol Scand (2000) 79(3):221–226.[CrossRef][Web of Science][Medline]
38. Kenter GG, Schoonderwald EM, Koelma IA, Arentz N, Hermans J, Fleuren GJ. The cytological screening history of 469 patients with squamous cell carcinoma of the cervix uteri; does interval carcinoma exist? Acta Obstet Gynecol Scand (1996) 75(4):400–403.[Web of Science][Medline]
39. Department of Health Publications. NHSCSP: Guidelines on Failsafe Actions. Publication No. 21, 2004. Sheffield, England.
40. Quinn M, Babb P, Jones J, Allen E. Effect of screening on incidence of and mortality from cancer of cervix in England: evaluation based on routinely collected statistics. BMJ (1999) 318(7188):904–908.
41. Howe A, Owen-Smith V, Richardson J. Health authority cervical screening recall policies and time since last smear: a retrospective cohort analysis in the north west England. J Med Screen (2003) 10(4):184–188.
Manuscript received August 20, 2007; revised February 15, 2008; accepted March 4, 2008.
Editorial about this Article
Related Article in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2008 100: 605-606.
J Natl Cancer Inst 2008 100: 603.
This article has been cited by other articles:
|
|
 |
|
  N. Brewer, N. Pearce, M. Jeffreys, B. Borman, and L. Ellison-Loschmann Does screening history explain the ethnic differences in stage at diagnosis of cervical cancer in New Zealand? Int. J. Epidemiol., October 6, 2009; (2009) dyp303v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Sasieni, A. Castanon, and J. Cuzick Effectiveness of cervical screening with age: population based case-control study of prospectively recorded data BMJ, July 28, 2009; 339(jul28_2): b2968 - b2968. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Strander At what age should cervical screening stop? BMJ, April 24, 2009; 338(apr24_1): b809 - b809. [Full Text]
|
|
|
|
 |
|
 J. Cuzick Routine Audit of Large-Scale Cervical Cancer Screening Programs J Natl Cancer Inst, May 7, 2008; 100(9): 605 - 606. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||