Several studies have shown cervical cancer to have a heterogeneous distribution in the United States, with substantial differences in mortality rates and incidence by socioeconomic status, both overall and within specific racial or ethnic subgroups.1 Rural areas such as Appalachia, part of the South, the Texas–Mexico border, and the central valley of California, which have high concentrations of poor people, African Americans, or Hispanics, have been specifically identified as areas with high mortality from cervical cancer.2,3 In contrast, geographic incidence (as opposed to mortality) has not been well characterized, because for many years cancer registry data were not available for areas other than the 12 sites included in the National Cancer Institute's (NCI's) Surveillance, Epidemiology, and End Results Program (SEER), which represented less than 14% of the U.S. population.4
Clearly, our approach to preventing cervical cancer in the United States should include more focused attention on pockets of high risk. High efficacy regarding two vaccines against human papillomavirus (HPV) have been published; one has already been approved by the Food and Drug Administration and recommended by the Advisory Committee on Immunization Practices (a quadrivalent formulation against HPV 16, 18, 6, and 11), and another, a bivalent formulation against HPV 16 and 18, will be licensed soon. When available and administered to the eligible populations these vaccines can prevent about 70% of all cervical cancers, those caused by HPV 16 and HPV 18, in this country and worldwide.5 Accordingly, a comprehensive surveillance system is needed to monitor the burden and characteristics of existing cases and to ensure that the populations with a disproportional burden of disease are effectively reached with vaccination and screening. Using data from U.S. population-based cancer registries, we characterize the burden of invasive cervical cancer and examine geographic and histologic variation in its incidence by race or ethnicity.
This analysis included incidence data from population-based registries participating in the Centers for Disease Control and Prevention's (CDC's) National Program of Cancer Registries (NPCR) and the SEER program.6 The study was approved by the CDC's Institutional Review Board. National Program of Cancer Registries data were reported to CDC as of January 31, 2005, and SEER data were reported to NCI as of November 1, 2004. Both NPCR and SEER data are collected and reported using standard data items and uniform codes and procedures as documented by the North American Association of Central Cancer Registries.7 We included 39 states and the District of Columbia (DC), all of which continuously met the standards for high-quality data for the 5-year period 1998–2002. Combined, the NPCR and SEER data included 87% of the U.S. population for the period.
The primary site and histology of the cancers were coded using the International Classification of Diseases for Oncology, Third Edition (ICD-O-3).8 For comparability with earlier studies based on SEER data, we used the SEER site recode for defining cervical cancer,9 using site codes of cervix (C53). Squamous cell carcinoma was defined as ICD-03 histology codes 805–813; adenocarcinoma, ICD-03 814, 816, 818–822, 825–850, 852–855, 857, 894, and 9110; and adenosquamous carcinoma, ICD-03 856 and 8015. Other carcinomas were defined as ICD-03 801–802, 803–804, 815, 817, 823–824, 851, and 858–867.8,10 All other histology codes were identified as “cancer not otherwise specified (NOS)” or “non-carcinoma malignancy” and were combined in “other malignancies.” We did not examine in situ cases, because the registries were not required to report these data after 1996. Cases were staged using the SEER summary stage system (1997 version for cases diagnosed in 1998–2000,11 2000 version for 2001–200212); the differences between versions are minor for cervical cancer. The categories are localized, regional, distant, and unstaged. Localized stage was defined as confined to the cervix; this classification corresponds to International Federation of Gynecology and Obstetrics (FIGO) stages IA1, IA2, IB, and not further specified. In “regional stage” the disease had spread beyond the cervix by direct extension to adjacent organs or tissues and/or to regional lymph nodes; this corresponds to FIGO stages IIA, IIB IIIA IIIB, and III(NOS). Distant stage was defined as disease with distant site(s)/lymph node(s), corresponding to FIGO stages IV, IVA, and IVB.13 Cases with insufficient data were categorized as unstaged.
The SEER*Stat 6.1.4 statistical software package (National Cancer Institute, http://www.seer.cancer.gov/publicdata/access.html) was used for all analyses. Rates were calculated by age (in years), race (white, African American, and Asian or Pacific Islander), ethnicity (Hispanic, non-Hispanic), and geography, as defined regionally by the U.S. Census (Northeast, South, West, or East) or by individual state. Coverage by region for this study was 98% for the Northeast and Midwest, 96% for the West, and 68% for the South. Hispanic ethnicity was tabulated independently of race. Because the work on improving the identification in the registries of American Indian and Alaska Native cases is still in progress, we have not presented data for this group. Their data are included in the overall summary analyses, however. All analyses involving histology were limited to histology-confirmed cancers (97% of all cases). Average annual incidence rates were computed as the sum of the cases reported during 1998–2002 divided by the sum of the annual population denominators, which were obtained from the 2000 U.S. Census.6 Incidence rates were directly age adjusted by 5-year age groups to the 2000 U.S. standard population and are expressed per 100,000 females of all ages; 95% confidence intervals (CIs) were obtained using the gamma method.14 Rate ratios (RRs) and their corresponding 95% CIs were also calculated.15 Statistical significance was assessed with a two-tailed P value of .05.
In all, 59,848 cases of invasive cervical cancer were identified during 1998–2002. Annual cases declined from 12,720 in 1998 to 11,071 in 2002; the incidence rate declined from 10.2 cases per 100,000 (95% CI 10.1–10.4) in 1998 to 8.5 per 100,000 (95% CI 8.4–8.7) in 2002 (data not shown). For the period as a whole, the average annual incidence rate was highest among Hispanics (14.8, 95% CI 14.5–15.1), followed by African-American women (13.5, 95% CI 13.2–13.7). Rates among Asian or Pacific Islander (8.9, 95% CI 8.5–9.3) and white women (8.9, 95% CI 8.8–9.0) were similar (Table 1).
Rates rose with age in all racial or ethnic subgroups except whites, and thus the highest differences by race or ethnicity were among older women. For example, the overall RR (of African-American women compared with white women) was 1.5, but among women aged 65 and older, this ratio was 2.5 (Table 2). Among Asian or Pacific Islander, overall rates were lower or the same as those for white women, but at age 50 years and older Asian or Pacific Islander had significantly higher rates than whites. Hispanic women had an overall rate 66% greater than that for non-Hispanics (RR 1.66; 95% CI, 1.62–1.70, P<.05).
The histologic classification of cases was 70% squamous cell carcinoma (n=41,628), 18% adenocarcinoma (n=10,849), 4% adenosquamous carcinoma, 5% other carcinomas, and 1.5% other malignancies (data not shown). Rates for each histologic subtype by race or ethnicity are provided (Table 3). Over the study period, squamous cell carcinoma rates were significantly higher among African-American (RR 1.67, P<.05) and Hispanic women (RR 1.75, P<.05) than among their white or non-Hispanic counterparts. In contrast, adenocarcinoma rates were significantly lower among African-American women (RR 0.88 using white women as the referent, P<.05). Both adenocarcinoma rates (RR 1.39, P<.05) and adenosquamous carcinoma rates (RR 1.71, P<.05) were significantly higher among Hispanic women than among non-Hispanics (Table 4).
Over one half of the cases were staged as local; 30%, regional; 9%, distant; and 9%, unstaged. Across all stages, African-American women had rates 1.2 to 2 times those of white women (Tables 3 and 4). Similarly, by stage Hispanic women had an RR of 1.5 to 1.9 when the comparison was with non-Hispanic women. Minimal differences were seen between white and Asian or Pacific Islander women.
The 40 participating jurisdictions (39 states and DC) are shown in Figure 1 ranked by their incidence rates; these ranged from 6.6 to 12.3 per 100,000 females. The highest rates were in DC, followed by Kentucky, West Virginia, Florida, and Louisiana. The South had the highest rate (10.6, 95% CI 10.5–10.8) of any region. Rates among white women were significantly higher in the South than elsewhere (Tables 5–6). Rates among African-American women were lowest in the West (9.3, 95% CI 8.6–10.1).
Histologic differences by geography were noted. Squamous cell carcinoma rates were highest in the South (7.6, 95% CI 7.5–7.7); in other regions the rate was 6.1–6.3 per 100,000 (Table 7). The squamous cell carcinoma rates were highest among African-American and Hispanic women; in the West, squamous cell carcinoma rates were much higher among Hispanics than African Americans (Fig. 2). Adenocarcinoma and adenosquamous carcinoma rates differed very little by region (Table 7), but differences were seen by race or ethnicity, because adenocarcinoma rates were highest among Hispanics in all regions (Fig. 2).
Higher rates of cancer for all stages were noted in the South than elsewhere (Table 8). The burden of unstaged cancers was highest in the South and lowest in the West. The rates of distant and unstaged cases exhibited some variation by region and race. In the Northeast, the rate of distant cancer was clearly higher for African-American women than for other groups, and in all four regions, African-American women and Hispanic women had higher rates of distant stage (Fig. 2). Rates of unstaged cancers were similar across race or ethnicity in the West; in the Midwest, rates of these cancers were highest among African-American women and then Hispanics. In the Northeast, high, similar rates of unstaged cancer were found among African Americans and Hispanics; and in the South, rates of unstaged cancer were highest among Hispanics, followed by African-American women (Fig. 2).
Our study reports on rates of cervical cancer in the United States using combined data from NPCR and SEER. Previous analyses were restricted to SEER sites and thus had limited coverage, but our analysis covered 39 states and Washington, DC, allowing us to examine geographic patterns for most of the U.S. population. We also obtained greater coverage of Hispanic and African-American women as well as women from rural areas. This report of 59,848 incident cases of invasive cervical cancer is large and geographically comprehensive, including data from registries representing seven-eighths of the U.S. population.
We confirmed that in the United States there is a 50% higher incidence of cervical cancer among African-American (compared with white) and 66% higher incidence among Hispanic women (compared with non-Hispanic).16,17 These groups and Asian or Pacific Islander women were sufficiently well represented in our data sources to develop robust estimates. We also demonstrated geographic differences by histologic type and by stage. Differences in incidence patterns in the United States have often been attributed to differences in the rates of Pap testing and patterns of diagnosis and treatment of preinvasive disease,2 but recent studies report similar rates of Pap tests for African-American and white women,18 while showing lower rates of Pap testing among Hispanic women.19 The difference in incidence among these groups seen in our data may be due also, in substantial measure, to differences, by race, in the follow-up of abnormal Pap tests,
In the present study, Asian or Pacific Islander women, as a whole, had rates of cervical cancer similar to those of white women. This may be seen as surprising, because many studies have shown lower screening rates and higher mortality rates for immigrant women, many of whom are Asian or Hispanic.20 Because most registries do not collect information about specific Asian subpopulations, we were unable to verify earlier reports that certain groups (Vietnamese, Korean, and Hmong women) have higher incidence rates of cervical cancer.21,22
We found that incidence rates increased with age among all racial or ethnic groups except white women. The rate ratio (RR) comparing African-American with white women started to increase in the 40–49 age group, whereas the RR comparing Hispanics with non-Hispanics began to increase a decade earlier and then continued to rise. A younger age at diagnosing cervical cancer among Hispanic women, which is consistent with our findings, has been reported previously.23,24 The differences in age-specific prevalence of HPV infection in white, African-American, Hispanic, and non-Hispanic women, if combined with lower screening rates and poorer rates of follow-up among nonwhites, may account for the higher age-specific incidence rates found among nonwhite women. Although it is reported that the prevalence of HPV infection declines sharply with age in most areas of the world, reports from Latin America indicate that age-specific rates for the infection rise again for women in their mid 40s to 50s.25 Our findings confirm the need to continue screening older women as recommended by guidelines and to find better strategies for access to screening of women of color.
Our analysis shows that squamous cell carcinoma rates are two-thirds higher among African-American women than in whites. Adenocarcinoma and adenosquamous carcinomas represented over one fifth of all invasive cervical cancers in our analysis, and there is concern about an increase in adenocarcinoma in younger white women.26 Our finding of higher rates of adenocarcinoma for white than for African-American women is consistent with a trend analysis of SEER data from 1976 to 2000.16 The reproducibility of histologic classification of these cervical cancers has not been measured, however, and variations in the use of histologic criteria may have affected the observed case distribution.27 The risk of adenocarcinoma may be increased by a variety of hormonal-based risk factors, such as obesity, parity, and use of oral contraceptives,28 and these may vary by race or ethnicity.29 In our analysis of adenocarcinoma, we found a higher RR in Hispanic women (compared with non-Hispanics) and lower RR among African Americans (compared with whites). A recent SEER analysis of 1992–1996 data showed a higher proportion of adenosquamous carcinoma but not adenocarcinoma among Hispanic women (compared with non-Hispanic groups).24 Two recent studies of invasive cervical cancers worldwide showed that although HPV 16 is the most common type detected in squamous cell carcinoma, HPV 18 (followed by HPV 16) is the most common HPV type detected in adenocarcinoma.5,28 We have no evidence, however, indicating that the prevalence of these types is higher among Hispanic women than among non-Hispanic women. Indeed, we have no clear explanations for the finding of a higher apparent risk of adenocarcinoma among Hispanic women and suggest that this issue be studied in more detail. Regardless of the finding, the effect that the HPV vaccine would have on adenocarcinomas is significant, given that the lesions may arise in the endocervix and are less likely to be detected by screening with a Pap test.
We found that the South had higher rates of cervical cancer than other regions, but some states in the Northeast and Midwest also had high rates. The geographical disparities are mostly due to the higher rates of squamous cell carcinoma in the South; there was very little regional variation in adenocarcinoma or adenosquamous cancer. Several factors, such as type-specific HPV prevalence and smoking prevalence, could explain the observed geographic differences; however, we have limited regional U.S. data on the subject of type-specific HPV distribution in the United States. Although minor differences have been reported in screening practices using data from the Behavioral Risk Factor Surveillance System, Yabroff et al2 did not find any conclusive differences in geographic patterns in the follow-up of abnormal screening tests. The lower incidence rates seen for unstaged cancers in the West (compared with other regions) may be due to differences in reporting or differences in the characteristics of patients or tumors or possibly an increased rate of referral in the West to an out-of-state tertiary care center, resulting in incomplete data.
Several limitations exist in this geographical analysis. First, the proportion of the covered population for the Southern region was only about 70%. Still, it must be noted that this analysis was based on the states that had 5 continuous years of high-quality data, and registries from the South that met this standard for just 1 or 2 years had to be excluded. We suspect that the difference between the South and other regions would be even higher had every Southern state been included. Finally, because we used population-based registries, we did not have patient data on country of birth, tobacco use, oral contraceptive use, hysterectomy, HPV status, or socioeconomic status. Recent studies and our own subanalysis of 1998–2001 data showed that women at higher risk of cervical cancer (both in the overall population and among specific racial or ethnic groups) are more likely to reside in areas with high poverty and low educational levels (Saraiya M, Ahmed F, King J, Krishnan S. Disparities in cervical cancer incidence and mortality, National Program of Cancer Registries, 1998–2001. International Papillomavirus Conference, Vancouver, Canada, April 30, 2005.)1
In conclusion, rates of cervical cancer in the United States have decreased over time, although significant racial, ethnic, and geographic differences continue to be seen. The current surveillance system used by cancer registries will provide a consistent means of measuring trends in the incidence of invasive cervical cancer after the introduction of a prophylactic HPV vaccine in mid 2006, although the effect on invasive cervical cancer might not be seen for 20 years. In the future, having information on vaccine status in the future or linking with an immunization registry will allow a more measured effect of vaccination on cancer outcomes. More specifically, a comprehensive surveillance system is needed to monitor characteristics of cases (who, where, and what type) and to ensure that we focus on populations in need of vaccination services, cervical cytology, or both.
Introduction of the new HPV vaccine gives gynecologists an additional tool to reduce cervical cancer morbidity and mortality. The populations not reached by current screening methods are at greatest risk of being missed by vaccine initiatives. Cervical cancer screening at the recommended intervals will need to continue for both vaccinated and nonvaccinated females. Gynecologists can play an important role in advocating for cervical cancer screening of older women and women of color. They can also advocate for and provide cervical cancer screening and vaccination to underserved women and ensure that there are no missed opportunities for screening women who might already be receiving care for other reasons.30 Before and during the implementation of HPV vaccine, we must document where differences exist by age, race or ethnicity, histology, geographic region and other characteristics so that we can monitor the success of both interventions—two approaches that will ultimately reduce both morbidity and mortality from cervical cancer.
1. Singh GK, Miller BA, Hankey BF, Edwards BK. Persistent area socioeconomic disparities in U.S. incidence of cervical cancer, mortality, stage, and survival, 1975–2000. Cancer 2004;101:1051–7.
2. Yabroff KR, Lawrence WF, King JC, Mangan P, Washington KS, Yi B, et al. Geographic disparities in cervical cancer mortality: what are the roles of risk factor prevalence, screening, and use of recommended treatment? J Rural Health 2005;21:149–57.
3. Freeman H, Wingrove B. Excess cervical cancer mortality: a marker for low access to health care in poor communities. NIH Report No. 05-5282. Rockville (MD): National Cancer Institute, Center to Reduce Cancer Health Disparities; 2005.
4. Ries LAG, Harkins D, Krapcho M, Mariotta A, Miller BA, Feuer EJ, et al. National Cancer Institute. SEER cancer statistics review, 1975–2003. Based on November 2005 SEER data submission, posted to the SEER Web site, 2006. Available at: http://seer.cancer.gov/csr/1975_2003/
. Retrieved December 3, 2006.
5. Clifford GM, Smith JS, Plummer M, Munoz N, Franceschi S. Human papillomavirus types in invasive cervical cancer worldwide: a meta-analysis. Br J Cancer 2003;88:63–73.
6. U.S. Cancer Statistics Working Group. United States cancer statistics: 1999–2002 incidence and mortality Web-based report version. Available at: http://www.cdc.gov/cancer/npcr/publications/
. Retrieved December 13, 2006. Atlanta (GA): Department of Health and Human Services, Centers for Disease Control and Prevention, and National Cancer Institute; 2005.
7. Wingo PA, Jamison PM, Hiatt RA, Weir HK, Gargiullo PM, Hutton M, et al. Building the infrastructure for nationwide cancer surveillance and control—a comparison between the National Program of Cancer Registries (NPCR) and the Surveillance, Epidemiology, and End Results (SEER) Program (United States). Cancer Causes Control 2003;14:175–93.
8. World Health Organization. International classification of disease for oncology. 3rd ed. Geneva (Switzerland): World Health Organization; 2003.
10. Berg JW. Morphological classification of human cancer. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer epidemiology and prevention. 2nd ed. New York (NY): Oxford University Press; 1996. p. 28–44.
11. Surveillance, Epidemiology and End Results Program. Summary staging guide for the cancer Surveillance, Epidemiology, and End Results (SEER) Program. Bethesda (MD): National Institutes of Health, National Cancer Institute; 1977.
12. Surveillance, Epidemiology and End Results Program. Summary staging manual 2000. Bethesda (MD): National Institutes of Health, National Cancer Institute; 2005.
13. Kosary CL. FIGO stage, histology, histologic grade, age and race as prognostic factors in determining survival for cancers of the female gynecological system: an analysis of 1973–87 SEER cases of cancers of the endometrium, cervix, ovary, vulva, and vagina. Semin Surg Oncol 1994 Jan;10:31–46.
14. Fay MP, Feuer EJ. Confidence intervals for directly standardized rates: a method based on the gamma distribution. Stat Med 1997;16:791–801.
15. Fay MP. Approximate confidence intervals for rate ratios from directly standardized rates with sparse data. Commun Stat Theory Methods 1999;28:2141–60.
16. Wang SS, Sherman ME, Hildesheim A, Lacey JV Jr, Devesa S. Cervical adenocarcinoma and squamous cell carcinoma incidence trends among white women and black women in the United States for 1976–2000. Cancer 2004;100:1035–44.
17. 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;102:765–73.
18. Ahluwalia IB, Mack KA, Murphy W, Mokdad AH, Bales VS. State-specific prevalence of selected chronic disease-related characteristics — Behavioral Risk Factor Surveillance System, 2001. MMWR Surveill Summ 2003;52:1–80.
19. Swan J, Breen N, Coates RJ, Rimer BK, Lee NC. Progress in cancer screening practices in the United States: results from the 2000 National Health Interview Survey. Cancer 2003;97:1528–40.
20. Seeff LC, McKenna MT. Cervical cancer mortality among foreign-born women living in the United States, 1985 to 1996. Cancer Detect Prev 2003;27:203–8.
21. Kwong SL, Chen MS Jr, Snipes KP, Bal DG, Wright WE. Asian subgroups and cancer incidence and mortality rates in California. Cancer 2005;104 suppl: 2975–81.
22. Mills PK, Yang RC, Riordan D. Cancer incidence in the Hmong in California, 1988–2000. Cancer 2005;104 suppl:2969–74.
23. Schorge JO, Lea JS, Garner EO, Duska LR, Miller DS, Coleman RL. Cervical adenocarcinoma survival among Hispanic and white women: a multicenter cohort study. Am J Obstet Gynecol 2003;188:640–4.
24. Patel DA, Barnholtz-Sloan JS, Patel MK, Malone JM Jr, Chuba PJ, Schwartz K. A population-based study of racial and ethnic differences in survival among women with invasive cervical cancer: analysis of Surveillance, Epidemiology, and End Results data. Gynecol Oncol 2005;97:550–8.
25. Franceschi S, Herrero R, Clifford GM, Snijders PJ, Arslan A, Anh PT, et al. Variations in the age-specific curves of human papillomavirus prevalence in women worldwide. Int J Cancer 2006;119:2677–84.
26. Hildesheim A, Berrington de Gonzalez A. Etiology and prevention of cervical adenocarcinomas. J Natl Cancer Inst 2006;98:292–3.
27. Kurman R, Norris HJ, Wilkinson EJ. Tumors of the cervix, vagina, and vulva. atlas of tumor pathology. Washington (DC): Armed Forces Institute of Pathology; 1992.
28. Castellsague X, Diaz M, de Sanjose S, Munoz N, Herrero R, Franceschi S, et al. Worldwide human papillomavirus etiology of cervical adenocarcinoma and its cofactors: implications for screening and prevention. J Natl Cancer Inst 2006;98:303–15.
29. Chandra A, Martinez GM, Mosher WB, Abma JC, Jones J. Fertility, family planning, and reproductive health among US women: data from the 2002 National Survey of Family Growth. Vital Health Stat 23 (25). 2005.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
30. 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;88:2283–9.