OBJECTIVE: To evaluate the screening efficacy and importance of atypical squamous cells and atypical glandular cells in predicting subsequent cervical cancer risk.
METHODS: This national cohort study in Taiwan analyzed associations between Pap test screening frequency and findings in 1995–2000 and subsequent risk of squamous cell carcinoma and adenocarcinoma after 2002. Women aged 30 years or older in 1995 without a cervical cancer history were included. Multivariate-adjusted hazard ratios and their 95% confidence intervals (CIs) were assessed using Cox regression analysis.
RESULTS: During a total follow-up of 31,693,980 person-years in 2002–2008, 9,471 squamous cell carcinoma and 1,455 adenocarcinoma cases were newly diagnosed, resulting in 2,067 deaths. The risk of developing and dying from squamous cell carcinoma decreased significantly with increasing attendance frequency between 1995 and 2000 (all P values for trend <.001). Women who attended more than three screenings in 1995–2000 had 0.69-fold and 0.35-fold decrease in incidence and mortality of adenocarcinoma, respectively, compared with women who never attended any screenings. Abnormal cytologic findings were significant predictors of the incidence and mortality of cervical cancers. The adjusted hazard ratio (95% CI) of developing squamous cell carcinoma was 29.94 (22.83–39.25) for atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesions, and the adjusted hazard ratio (95% CI) of developing adenocarcinoma was 49.43 (36.49–66.97) for atypical glandular cells.
CONCLUSION: Significant reductions in cervical adenocarcinoma occurred in women who attend three or more annual screenings in 6 years. High-grade atypical squamous cells and atypical glandular cells are important predictors of subsequent adenocarcinoma and squamous cell carcinoma.
LEVEL OF EVIDENCE: II.
Screening frequency and atypia are important predictors of cervical cancer risk, and adenocarcinoma is significantly reduced in women who attended three or more screenings in 6 years.
Department of Obstetrics and Gynecology, College of Medicine and the Graduate Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, the Genomics Research Center, Academia Sinica, the School of Public Health, National Defense Medical Center, and the Health Promotion Administration, Ministry of Health and Welfare, Executive Yuan, Taipei, Taiwan.
Corresponding author: Chien-Jen Chen, ScD, Genomic Research Center, Academia Sinica, 128 Academia Road, Section 2, Nankang, Taipei 115, Taiwan; e-mail: email@example.com.
Supported by the Health Promotion Administration, Ministry of Health and Welfare, Executive Yuan, Taipei, Taiwan.
The authors thank Health Promotion Administration, Ministry of Health and Welfare, Executive Yuan, Taipei, Taiwan, for financial support of the project (Project number A1011113) and the members of the Taiwan Cervical Cancer Screening Task Force for management of data included in this analysis: Huang-Cheng Hsu, Ching-Hsun Ho, Hsiang-Yun Cheng, Yi-Wen Wang, Jian-Jhih Lin, I-Ling Chen, Ya-Chuan Chen, and Chiao-Ju Wu.
Financial Disclosure The authors did not report any potential conflicts of interest.
Cervical cancer is one of the leading cancers in Taiwan.1 A screening program dramatically decreases the chance of squamous cell carcinoma and reveals the importance of further preventing adenocarcinoma. It is well known that screening programs have decreased the overall cervical cancer incidence. However, they have disproportionately affected squamous cell carcinomas resulting in an increased percentage of adenocarcinomas. Previous studies showed an increasing time trend of percentage rates for adenocarcinoma throughout European countries.2 Recent case–control studies found a significant reduction (37–43%) in adenocarcinoma among those screened.3,4 However, low adenocarcinoma incidence results in difficulties in evaluating its screening efficacy.
Women with abnormal and atypical cytology have a higher chance of developing cervical cancer.5–7 The Bethesda System further categorized atypical squamous cells (ASC) and atypical glandular cells (AGC) from atypia, and differences in long-term cervical cancer risk between atypia and squamous intraepithelial lesions have rarely been compared through cohort studies.
Pap test screening in Taiwan has been reimbursed annually for women aged 30 years or older since 1995.8 There were 26 million screenings performed between 1995 and 2008 with an average triennial screening participation of 53% after 2000.9 This program has successfully reduced incidence of all cervical cancers by 54% and mortality by 52%.9,10 The proportion of adenocarcinoma among all cervical cancers increased from 6.8% in 1995 to 14.8% in 2008.10
Because there are at least 260 adenocarcinoma cases in Taiwan per year, it is feasible to evaluate the effect of screening on the adenocarcinoma burden. Based on the advantage of having several national registries in Taiwan, this study aimed to evaluate the efficacy of screening in reducing incidence and mortality of both cell types.
MATERIAL AND METHODS
The cervical cancer screening program in Taiwan has provided annual Pap test screening through the National Health Insurance since 1995. Participation is voluntary. The study population consists of 4,701,796 women who were older than age 30 years in 1995, screened between 1995 and 2000, and did not have a known history of cervical cancer before January 1, 2002. The unique national identification number, age, and vital status of each woman in the study were derived from the Taiwan household registry database, which was found to be very complete (99.3%) when compared with the census data.11
Pap test results before mid-1998 were reviewed and coded according to The Bethesda System and after that according to The 2001 Bethesda System nomenclature.12 It is mandatory for all Pap test results to be reported to the cervical cancer screening registry, which covers more than 99% of Pap test examinations reported in the National Health Insurance claim data. Cytologic screening data are collected from approximately 100 certified laboratories, and the quality of diagnosis is evaluated regularly by the Taiwan Society of Pathology. Women with high-grade or worse lesions were referred to the follow-up system of local government health departments.9
In this study, cervical cancer screening history from 1995 to 2000 in each woman was derived from the cervical cancer screening registry. The screening attendance frequency of each woman was calculated as the number of years she received a Pap test screening from 1995 to 2000 and classified into five groups: none, 1 year, 2 years, 3 years, and 4 or more years. The most severe cytologic findings during the same period were also obtained from the same registry.
To ascertain the incident and fatal cases of cervical cancer in the study, the list of participants was electronically linked to the National Cancer Registry and National Death Certification Registry. The Taiwan National Cancer Registry was launched in 1979 to collect information on all incident cancer cases.13 The registry is considered to be complete with the percentage of cervical cancer identified by death certificate at less than 1% of all incident cervical cancer cases after 2000 and is accurate with a percentage of cervical cancer morphologic verification at greater than 99% after 2002.10 It is mandatory for all death certificates in Taiwan to be reported through the Internet within 7 days after deaths and greater than 99% of death certificates from cervical cancer are issued by medical doctors. The accuracy of cause-of-death coding is very high, showing excellent agreement between death certificate reviewers and original coders for malignant neoplasms (κ=0.94).14
Patients with cervical cancer were identified from the National Cancer Registry using the International Classification of Diseases for Oncology, Third Revision T-code C53 from January 1, 2002, through December 31, 2008. Histologic types were identified from the morphology codes in the same database. Death from cervical cancer was defined by the underlying causes of death code 180 in death certificates using the International Classification of Diseases, 9th Revision, Clinical Modification.
The person-years at risk of dying from cervical cancer were calculated from January 1, 2002, to the date at death or December 31, 2008, whichever came first; whereas the person-years at risk of developing cervical cancer were calculated from January 1, 2002, to the date of the diagnosis of newly developed cervical cancer, the date at death, or December 31, 2008, whichever came first. Incidence rates were calculated by dividing the number of cervical cancer cases by the person-years at risk of developing cervical cancer. Mortality rates were calculated by dividing the number of cervical cancer deaths by the person-years at risk of dying from cervical cancer. The approximate 95% confidence intervals of incidence and mortality rates under the Poisson distribution were calculated. The associations between screening attendance frequency and findings for cervical cancer incidence and mortality were assessed with Cox' proportional hazards models adjusting for age and education to obtain multivariate-adjusted hazard ratios (HRs) and its 95% confidence intervals (CIs) for each variable. All statistical analyses performed in the study were done with SAS 9.1.
This study is approved by the Research Ethics Committee at the National Taiwan University Hospital and is registered in the ClinicalTrials.gov (Identifier NCT01015079). The computerized linkage of all national profiles was conducted by the Taiwan Cervical Cancer Screening Task Force according to the standardized protocols. Permission to use the national profiles in this study was obtained from the Health Promotion Administration in Taiwan.
Among 4,701,796 eligible women, 41.7%, 24.5%, 14.8%, 9.4%, and 9.6%, respectively, attended none, one, two, three, and four or more screenings during 1995–2000 as shown in Figure 1. During a follow-up of 31,693,980 person-years in 2002–2008, 9,471 cervical squamous cell carcinoma cases and 1,455 adenocarcinoma cases were newly diagnosed. There were 2,067 cervical cancer deaths, including 1,775 from squamous cell carcinoma and 292 from adenocarcinoma (Fig. 1). The case-fatality rates were 18.7% and 20.1%, respectively.
In 2002–2008, there were 31,693,980 and 31,792,735 follow-up person-years for the estimation of cervical cancer incidence and mortality as shown in Table 1, respectively. The incidence of squamous cell carcinoma increased with increasing age (P for trend <.001), whereas the incidence of adenocarcinoma decreased significantly with age (0.56% for each 5-year increment, P for trend <.001). The mortality from both squamous cell carcinoma and adenocarcinoma also increased with increasing age (P for trend <.001). The risk of developing and dying from adenocarcinoma and squamous cell carcinoma decreased significantly in a gradient of increasing attendance frequency from 1995 to 2000 (P for trend <.001). Among 2,740,106 women who attended one or more screenings, abnormal cytologic findings in 1995–2000 were significantly associated with increased incidence of and mortality from squamous cell carcinoma and adenocarcinoma.
Figure 2 shows the age-specific incidence and mortality of squamous cell carcinoma and adenocarcinoma according to screening attendance frequency. The incidence and mortality of squamous cell carcinoma decreased significantly in all age groups for women who attended one or more screenings when compared with those who never attended, whereas the incidence and mortality of adenocarcinoma decreased significantly in young age groups for women who attended three or more screenings when compared with those who never attended.
Table 2 shows the multivariate-adjusted HRs of developing and dying from squamous cell carcinoma and adenocarcinoma during 2002–2008 for variables included in the Cox proportional hazards models. Both age and years of schooling were significantly associated with the incidence and mortality of cervical cancer. After adjustment for age and years of schooling, attending one screening during 1995–2000 conferred significantly lower HRs of 0.53 and 0.37 for developing and dying from squamous cell carcinoma compared with nonattendees, respectively, and also had a significantly lower HR of 0.65 for developing adenocarcinoma during 2002–2008. The trends for HRs for incidence of and mortality from squamous cell carcinoma and incidence of adenocarcinoma decreased with increased screening attendance frequency (all P for trend <.001). A stepwise increase in screening numbers was significantly associated with a decrease in squamous cell carcinoma incidence. Women who attended two or more screenings had significantly lower HRs of dying from squamous cell carcinoma compared with nonattendees, but there was no further significant decrease with screening numbers greater than twice. Significantly decreased HRs of adenocarcinoma incidence were only observed in women who attended screening more than three times compared with nonattendees and women who attended screenings more than four times compared with women who attended only once or twice (all P value <.005).
Table 3 shows the multivariate-adjusted HRs of developing and dying from squamous cell carcinoma and adenocarcinoma in 2002–2008 according to the cytologic findings from women who attended screening during 1995–2000. Abnormal findings were significant predictors of the incidence and mortality of adenocarcinoma and squamous cell carcinoma during 2002–2008 after adjustment for age, years of schooling, and screening attendance frequency. Women with high-grade ASC or AGC generally had higher risk of developing squamous cell carcinoma and adenocarcinoma compared with those of normal and low-grade squamous intraepithelial lesions. Women with high-grade squamous intraepithelial lesions or more severe lesions had the highest risk of developing and dying from squamous cell carcinoma, whereas women with AGC had the highest risk of adenocarcinoma. The adjusted HR (95% CI) of developing squamous cell carcinoma was 29.94 (22.83–39.25) for high-grade ASC, and the adjusted HR (95% CI) of developing adenocarcinoma was 49.43 (36.49–66.97) for AGC. The adjusted HR (95% CI) of dying from squamous cell carcinoma was 35.12 (18.01–68.47) for high-grade ASC, and the adjusted HR (95% CI) of dying from adenocarcinoma was 46.67 (21.62–100.71) for AGC.
This study included the entire population in Taiwan to evaluate risks of major cervical cancer cell types according to different screening frequencies. A previous simulation study in Taiwan that used screenings in the early years of the program expected that it would reduce incidence by 29% and mortality by 50% by 2001,8 but numbers were actually reduced 14.9% and 25.7%, respectively.9 In a previous time trend study,9 a significant decrease in adenocarcinoma incidence was found after 1995. These studies demonstrated that the screening program could reduce disease burden but provided limited information on screening efficacy in specific cell types and among different screening frequencies.
Although a significant decrease in adenocarcinoma incidence was observed in frequently screened women, the Pap test did not reduce the risk of adenocarcinoma as much as it reduced risk of squamous cell carcinoma. For women who attended screenings nearly annually, HRs of squamous cell carcinoma incidence and mortality were 0.18 and 0.09, respectively, but those in adenocarcinoma were much higher at 0.66 and 0.56, respectively. Significantly lower fatality from adenocarcinoma in women who attended at least one screening was also observed (data not shown). The cytobrush may facilitate the detection of adenocarcinoma in situ and the excision procedure is considered to be effective,15,16 but there is doubt that treatment would reduce adenocarcinoma.17–19 These may be the reasons why Pap test screening limitedly reduces adenocarcinoma only in women who attend frequent screenings.
In this study, the adenocarcinoma incidence peaked at ages 35–44 years old, whereas squamous cell carcinoma incidence peaked at older ages. The adenocarcinoma incidence in England increases rapidly after 30 years of age and plateaus at 50–69 years,18 whereas incidence in The Netherlands peaks at 30–44 years.20 The age-specific adenocarcinoma incidence was similar in women with various screening attendance frequencies in this study. The different age patterns of these two cell types may be the result of the difference in the natural history of type-specific human papillomavirus (HPV) infections. HPV types 16 and 18 were detected in 70–94% of adenocarcinomas and 59–74% of squamous cell carcinomas in Taiwan.21,22 In a recent study in Taiwan, among women affected with high-grade squamous intraepithelial lesion cytology, HPV types 16 and 58 were prevalent in women aged 30–49 years,23 and HPV type 52 was prevalent in older women (21%).24 During 16 years of follow-up, the persistent infection rate was 15.5% for HPV types 16 and 18 and 38.7% for types 52 and 58.23 This age difference in prevalence of various HPV types may explain why the incidence of squamous cell carcinoma increased with age and why adenocarcinoma peaked at young and middle ages.
Despite the rarity of high-grade ASC (less than 1% in 6 years in this study) and AGC (0.14% in 6 years), they were both equally or even more clinically important than high-grade squamous intraepithelial lesions. Previous study in patients with atypical squamous cells of undetermined significance in Taiwan found different cervical cancer risks between patients who had previous Pap tests and those who were new for screening,5 but it did not evaluate the risks of developing the two cell types of cervical cancer among atypical squamous cells of undetermined significance, high-grade ASC, and other squamous intraepithelial lesions. AGC is heterogeneously composed of a wide variety of squamous, endocervical, and endometrial lesions. Although it is limitedly reproducible,25 it has been suspected to increase the risk of cervical neoplasia, endometrial hyperplasia, and adenocarcinomas of the endometrium, ovary, and rectum.26 The sample size in most clinical studies was too small to accurately estimate the risk by specific histology. In this study, high-grade ASC and AGC were both associated with adenocarcinoma and squamous cell carcinoma. It is essential to follow these lesions more intensively.
There are several limitations that should be noted. In Taiwan, women with a high risk of cervical cancer usually have earlier sexual debut, more sexual partners, more deliveries of children, and less attendance of Pap test. However, these important factors associated with cervical cancer are not available in this study. We attempted to eliminate some of these confounders by adjusting our HRs by age and schooling years. Moreover, women with abnormal cytology may have been advised to receive frequent Pap tests, and these frequent screenings may have resulted in early detection of cervical cancer after 2002, thus underestimating the effect of screening on preventing cervical cancer death.
Through this study, we found a significant reduction in cervical adenocarcinoma in women who attended three or more annual screenings in 6 years, which was more frequent than that recommended by most organized screening programs. Women affected with high-grade ASC or AGC have significantly higher risks of subsequent adenocarcinoma and squamous cell carcinoma.
1. Chen CJ, You SL, Lin LH, Hsu WL, Yang YW. Cancer epidemiology and control in Taiwan: a brief review. Jpn J Clin Oncol 2002;32(suppl):S66–81.
2. Bray F, Carstensen B, Møller H, Zappa M, Zakelj MP, Lawrence G, et al.. Incidence trends of adenocarcinoma of the cervix in 13 European countries. Cancer Epidemiol Biomarkers Prev 2005;14:2191–9.
3. Andrae B, Kemetli L, Sparén P, Silfverdal L, Strander B, Ryd W, et al.. Screening-preventable cervical cancer risks: evidence from a nationwide audit in Sweden. J Natl Cancer Inst 2008;100:622–9.
4. Sasieni P, Castanon A, Cuzick J. Screening and adenocarcinoma of the cervix. Int J Cancer 2009;125:525–9.
5. Cheng WF, Huang CY, You SL, Chen CJ, Hu CH, Chen CA. Clinical significance of cytologic atypical squamous cells of undetermined significance. Obstet Gynecol 2009;113:888–94.
6. Chhieng DC, Gallaspy S, Yang H, Roberson J, Eltoum I. Women with atypical glandular cells: a long-term follow-up study in a high-risk population. Am J Clin Pathol 2004;122:575–9.
7. Silfverdal L, Kemetli L, Sparén P, Andrae B, Strander B, Ryd W, et al.. Risk of invasive cervical cancer in relation to clinical investigation and treatment after abnormal cytology: a population-based case-control study. Int J Cancer 2011;129:1450–8.
8. Koong SL, Yen AM, Chen TH. Efficacy and cost-effectiveness of nationwide cervical cancer screening in Taiwan. J Med Screen 2006;13(suppl 1):S44–7.
9. Chen YY, You SL, Chen CA, Shih LY, Koong SL, Chao KY, et al.. Effectiveness of national cervical cancer screening programme in Taiwan: 12-year experiences. Br J Cancer 2009;101:174–7.
10. Taiwan Cancer Registry Task Force. Taiwan Cancer Registry annual report. Taipei (Taiwan): Department of Health, Executive Yuan; 1995–2008.
11. Chen CN. An evaluation of Taiwan Census 2000: a comparison of census and registered populations. J Popul Stud 2002;25:1–56.
12. Solomon D, Davey D, Kurman R, Moriarty A, O'Connor D, Prey M, et al.; Forum Group Members; Bethesda 2001 Workshop. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 2002;287:2114–9.
13. Chiang CJ, Chen YC, Chen CJ, You SL, Lai MS. Cancer trends in Taiwan. Jpn J Clin Oncol 2010;40:897–904.
14. Lu TH, Lee MC, Chou MC. Accuracy of cause-of-death coding in Taiwan: types of miscoding and effects on mortality statistics. Int J Epidemiol 2000;29:336–43.
15. Andersen ES, Nielsen K. Adenocarcinoma in situ of the cervix: a prospective study of conization as definitive treatment. Gynecol Oncol 2002;86:365–9.
16. Salani R, Puri I, Bristow RE. Adenocarcinoma in situ of the uterine cervix: a metaanalysis of 1278 patients evaluating the predictive value of conization margin status. Am J Obstet Gynecol 2009;200:182.e1–5.
17. 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;113:1005–9.
18. Sasieni P, Adams J. Changing rates of adenocarcinoma and adenosquamous carcinoma of the cervix in England. Lancet 2001;357:1490–3.
19. Smith HO, Tiffany MF, Qualls CR, Key CR. The rising incidence of adenocarcinoma relative to squamous cell carcinoma of the uterine cervix in the United States—a 24-year population-based study. Gynecol Oncol 2000;78:97–105.
20. van de Nieuwenhof HP, Massuger LF, de Hullu JA, van Ham MA, van Dijck JA, Siebers AG, et al.. Significant decrease of adenocarcinoma in situ not reflected in cervical adenocarcinoma incidence in the Netherlands 1989-2003. Br J Cancer 2008;98:165–7.
21. Chen CA, Liu CY, Chou HH, Ho CM, Twu NF, Kan YY, et al.. The distribution and differential risks of human papillomavirus genotypes in cervical preinvasive lesions: a Taiwan Cooperative Oncologic Group Study. Int J Gynecol Cancer 2006;16:1801–8.
22. Ding DC, Hsu HC, Huang RL, Lai HC, Lin CY, Yu MH, et al.. Type-specific distribution of HPV along the full spectrum of cervical carcinogenesis in Taiwan: an indication of viral oncogenic potential. Eur J Obstet Gynecol Reprod Biol 2008;140:245–51.
23. Chen HC, Schiffman M, Lin CY, Pan MH, You SL, Chuang LC, et al.. Persistence of type-specific human papillomavirus infection and increased long-term risk of cervical cancer. J Natl Cancer Inst 2011;103:1387–96.
24. Chen HC, You SL, Hsieh CY, Schiffman M, Lin CY, Pan MH, et al.. Prevalence of genotype-specific human papillomavirus infection and cervical neoplasia in Taiwan: a community-based survey of 10,602 women. Int J Cancer 2011;128:1192–203.
25. Simsir A, Hwang S, Cangiarella J, Elgert P, Levine P, Sheffield MV, et al.. Glandular cell atypia on Papanicolaou smears: interobserver variability in the diagnosis and prediction of cell of origin. Cancer 2003;99:323–30.
© 2014 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
26. Cheng WF, Chen YL, You SL, Chen CJ, Chen YC, Hsieh CY, et al.. Risk of gynaecological malignancies in cytologically atypical glandular cells: follow-up study of a nationwide screening population. BJOG 2010;118:34–41.