Determinants of cervical cancer rates in developing countries
Article first published online: 16 MAY 2002
Copyright © 2002 Wiley-Liss, Inc.
International Journal of Cancer
Volume 100, Issue 2, pages 199–205, 10 July 2002
How to Cite
Drain, P. K., Holmes, K. K., Hughes, J. P. and Koutsky, L. A. (2002), Determinants of cervical cancer rates in developing countries. Int. J. Cancer, 100: 199–205. doi: 10.1002/ijc.10453
- Issue published online: 11 JUN 2002
- Article first published online: 16 MAY 2002
- Manuscript Accepted: 8 MAR 2002
- Manuscript Revised: 7 MAR 2002
- Manuscript Received: 29 OCT 2001
- National Cancer Institute. Grant Number: CA34493
- cervical cancer;
- human papillomavirus;
- developing countries;
- ecologic analysis;
- sexual behavior
Although cervical cancer (CC) is a leading cause of cancer-related deaths in developing countries, incidence rates vary considerably, ranging from 3 to 61 per 105 females. Identifying determinants of high vs. low rates may suggest population-level prevention strategies. CC rates for 175 countries were obtained from the IARC. Country-specific behavioral, health, economic and demographic measures were obtained from United Nations agencies and other international organizations. Regression analyses performed for 127 low or medium developed countries identified both geography and religion as independently associated with high CC rates. Among behavioral measures, high fertility rates, early age at birth of first child and high teenage birthrates were significantly associated with high CC rates. Countries with high CC rates had fewer doctors per capita, less immunization coverage, more HIV infections and shorter life expectancies. CC rates also tended to be higher in countries with more spending on health and younger, less educated populations. Patterns of CC rates suggest that programmatic approaches, such as promoting delayed childbearing and sexual monogamy, may be appropriate interventions. For countries with high CC rates and some flexibility in their health-care budgets, a once-in-a-lifetime screen of women 30–50 years of age, using Pap smears, direct visual inspection and/or HPV DNA testing, may be cost-effective. Finally, relatively low immunization rates and a shortage of health-care workers in countries with high CC rates suggest potential challenges for introducing prophylactic HPV vaccines. © 2002 Wiley-Liss, Inc.
In developing countries, cervical cancer (CC) ranks second to breast cancer in incidence1 but represents the leading cause of cancer-related deaths among women.2 Of the estimated 470,000 CC cases around the world each year, 380,000 (81%) occur in the developing world.3 Several studies have conclusively determined that HPV plays an etiologic role in the development of invasive CC.4–6 Genital HPV is transmitted through sexual contact,7 and estimates suggest that at least 50% of sexually active adults have had a genital HPV infection.8 Currently, over 120 different HPV types have been identified, of which at least 38 primarily infect the genitalia. Four high-risk types (HPV-16, -18, -31 and -45) account for about 80% of invasive CCs collected from around the world.9
Among developed countries, CC incidence and mortality rates have markedly decreased through active screening and effective treatment.10, 11 However, national CC prevention programs are expensive and laborious and rely on delicate laboratory procedures. Screening programs have rarely been implemented and virtually never sustained in most developing countries.12–14 Among the screening programs that have been introduced, many have had very poor quality and low coverage rates, with little financial support.15, 16 As a result, the less developed countries have CC mortality rates more than twice those of the developed countries with national CC prevention programs.3
Relatively new screening methods, such as HPV DNA testing, now offer a viable alternative or adjunct to conventional Pap smears.17, 18 However, while these methods may prove feasible in a developed country's national screening program, they remain highly laboratory-dependent and are not yet applicable in most developing countries. Direct visual inspection and immediate referral for treatment is another approach that is less laboratory-dependent19 but requires new training of clinical staff.20–22 Thus, technical, financial and personnel constraints provide formidable challenges to implementing effective national CC screening programs in many developing countries in the near future.
Analyses of country-specific CC rates by societal characteristics may help to identify countries with the greatest need and capacity for prevention programs and suggest which population-level interventions may be feasible. With the projected availability of effective HPV vaccines along with technical advances in screening, it is important to define the sociologic, economic and developmental correlates of CC in developing countries and to assess how these correlates may influence the feasibility and design of CC prevention programs.
The objectives of this research project were as follows: (i) to describe the geographic distribution of CC rates; (ii) to define associations between country-specific CC rates and predominant religions, sexual behaviors, public health measures, economic indicators and development and demographic characteristics; and (iii) to suggest potential interventions suitable for resource-limited CC prevention programs.
This analysis is of potential interest to epidemiologists interested in understanding the global distribution and ecologic correlates of CC in developing countries, to investigators actively pursuing new screening tests and prophylactic HPV vaccines and to health officials interested in identifying potential population-level interventions to reduce CC rates.
MATERIAL AND METHODS
In this ecologic study, we used country-specific data compiled from several existing sources. The IARC's GLOBOCAN-2000 database provided CC incidence and mortality rates for 175 countries.3 These data estimate country-specific age-standardized rates (per 100,000 female population) for the year 2000 based on the “most recent data available”, which the authors cite as from “generally 3–5 years earlier”.23 We excluded Haiti from all analyses because both its incidence and mortality rates were extreme outlying values (4.6 and 5.1 SDs higher than the mean, respectively),24 not comparable to other countries in the Caribbean region and based on estimated preliminary data from a small sample in the early 1980s.25 We excluded Guam and Puerto Rico from the analyses due to minimal statistical reporting for territories. The Federated States of Micronesia and French Polynesia, listed as regions in the IARC database, had country-level data sufficient to warrant inclusion. Country-specific data were collected for a total of 172 countries listed in the IARC database.
The UNDP Human Development Report 2000 provided the HDI data, which measure a country's development status on the basis of life expectancy, educational attainment and adjusted real income.26 After an initial comparison of rates, the 45 countries classified as high human development (“developed countries”) were excluded from the analyses under the assumption that they have had the capacity to sustain a national CC prevention program. Therefore, subsequent analyses were restricted to the 127 low and medium human development (“developing”) countries. None of these countries, to our knowledge, has conducted a sustainable and effective national CC prevention program over the last decade.
Country-level statistics were compiled for 14 sexual and reproductive behavior, 17 public health, 10 economic and 13 development and demographic indicators from various published sources. The major data sources included the UNDP's Human Development Report 2000, the WHO's World Health Report 2000,27 UNICEF's The State of the World's Children 2000,28 UNSD data in 2000,29 UNAIDS's Report on the Global HIV/AIDS Epidemic,30 the World Bank's World Development Indicators 200031 and Macro International's Demographic and Health Surveys.32 All population statistics, estimated for the year 2000, were obtained from the UNSD. The United States Central Intelligence Agency's The World Factbook 2000 provided each country's predominant religion.33
Statistical analyses concentrated on the 127 developing countries and were weighted by each country's adult female population to adjust for the precision of sampling methods. All regression analyses were performed using a robust variance to account for ecologic and population differences. Subanalyses were conducted without India and China to determine their impact on the results. Stata Version 634 was used for conducting all statistical analyses.
Mean CC incidence rates and SDs, summarized by geographic region and predominant religion, were analyzed for statistical significance (p < 0.05) using a 1-way analysis of variance test and compared using Bonferroni multiple-comparison tests. Univariate linear regression statistics examined each independent variable, with CC incidence and mortality rates as dependent variables. Independent variables with fewer than 30 observations and no significant association with CC rates (male age at first sexual intercourse, females using condoms with nonregular sex partners and divorces) were excluded from the analyses. Variables measured on a continuous scale were explored for linear fit, abnormal patterns and residual distribution.
Countries were categorized into approximate tertiles as having low (<20), medium (20–30) or high (>30) CC incidence rates per 105 females. The mean and SD for each independent variable were separately summarized among the low and high CC incidence groups. The low vs. high means for each indicator were analyzed for statistical significance (p < 0.05) using 2-sample t-tests for independent samples with unequal variances. The same statistical analyses were conducted among countries separated into low (<10/105) and high (>16/105) mortality groups.
A multivariate ANCOVA model for CC incidence was developed using the major geographic and religious categories and continuous variables with more than 100 observations and significant univariate associations (p < 0.10) with incidence rates. From each group of highly correlated indicator variables, only 1, which appeared to measure the common factor, was included. For example, the 4 measures of immunization rates were highly correlated and, therefore, only 1 indicator (measles immunization) was included in the model. Nonsignificant variables (p > 0.10) were singularly removed from the model in a backward fashion and reexamined in the final model. Highly correlated variables were analyzed to include the one indicator with the best fit. Countries included in the final model were explored for any abnormal patterns and selection biases. A multivariate regression model was developed with the same observations and indicators to obtain individual coefficients and p values.
Country-specific CC rates ranged from 3.0 to 61.4 cases per 105 females for all 172 countries. The mean incidence rate among the 127 developing countries (19.2 cases/105) was nearly twice that of the 45 developed countries (10.6/105). Likewise, the mean CC mortality rate of developing countries (9.9 deaths/105) was more than 2.5 times that of the developed countries (3.9/105). In addition, the estimated case fatality ratio (CC mortality/CC incidence) among developing countries (53%) was significantly higher than that for developed countries (37%) (p < 0.001).
The global distribution of the adult female population and CC deaths by human development category (least, middle and most developed) are shown in Figure 1. The 37 least developed countries, with 10% of the world's population, had 14% of the total mortality. The 88 middle developed countries, excluding India and China, with 32% of the population, had 34% of the deaths. India, with 16% of the population, had the largest proportional burden of mortality (30%). Conversely, China, with 22% of the total population, had only 8% of CC mortality. The 45 most developed countries had 20% of the adult female population and 14% of total CC deaths.
Geography and religion.
Among developing countries, both geography and religion were independently associated with CC rates. All 7 countries with the lowest incidence are predominantly Muslim; 6 are located in the Middle East and 1 in west Asia (Table I). Conversely, all 7 countries with the highest incidence are predominantly Christian; 2 are located in Latin America and 5 in eastern or southern Africa.
|Lowest CC rates||Highest CC rates|
|Syrian Arab Republic||3.0||Tanzania||61.4|
The geographic distribution of CC indicates that the Latin American (35.8 cases/105) and African (27.9/105) regions had the highest incidence rates (Table II). Asia and southern and eastern Europe had rates of 16.3 and 16.1 cases/105, respectively. The Middle Eastern region had the lowest rate (5.6/105). The difference in incidence rates between the 5 major geographic regions was statistically significant (p < 0.001, F = 13.3).
|Geographic region||Number of countries||Mean CC rates2||Min, max|
|Latin America||21||35.8||23.9, 61.1|
|Central America||8||40.4||25.0, 61.1|
|South America||9||34.2||31.3, 58.1|
|East Africa||13||43.2||26.5, 61.4|
|Southern Africa||8||39.0||24.7, 61.1|
|Central Africa||8||25.2||16.6, 31.7|
|West Africa||16||20.3||16.3, 51.8|
|North Africa||6||17.0||6.8, 23.4|
|South-central Asia||5||30.2||26.5, 30.7|
|Without India||4||27.6||26.5, 28.9|
|Southeast Asia||8||18.4||12.0, 24.6|
|West Asia||10||8.1||4.2, 24.6|
|East Asia||3||5.4||5.2, 18.0|
|Without China||2||15.5||15.3, 18.0|
|Southern Europe||5||17.4||13.3, 19.6|
|Eastern Europe||8||16.0||9.3, 31.5|
|Middle East||9||5.6||3.0, 9.4|
The geographic distribution of mortality rates paralleled the distribution of incidence. However, Africa (15.3 deaths/105) had a slightly higher rate than Latin America (14.3/105). Asia and southern/eastern Europe had rates of 9.1 and 6.0/105, respectively. The Middle Eastern region had the lowest rate (3.0/105). The difference in mortality rates was also statistically significant (p < 0.001, F = 7.9).
The distribution of CC incidence by predominant religion is presented in Table III. The predominantly Christian (35.2 cases/105) and predominantly Hindu (30.6/105) countries had the highest rates. Countries whose predominant religion was classified as “Indigenous/Local,” all located in sub-Saharan Africa, had a rate of 30.3 cases/105. The predominantly Orthodox and Muslim countries had rates of 16.0 and 15.6/105, respectively. The predominantly Buddhist countries had the lowest rate (7.7/105), but the rate increased to 20.7/105 after removing China. The difference in incidence rates between the predominant religions was highly significant (p < 0.0001, F = 54.1).
|Predominant religion||Number of countries||Mean CC rates2||Min, max|
|Christian Protestant||7||40.9||24.7, 56.2|
|Christian Roman Catholic||25||33.4||13.3, 61.1|
|Muslim Sunni||15||10.6||3.0, 44.9|
|Muslim Shi'a||2||7.6||3.3, 9.0|
|Without China||9||20.7||15.3, 28.9|
Mortality rates were similar to incidence rates, except that predominantly Hindu (17.4 deaths/105) and Indigenous (16.5/105) countries had slightly higher rates than predominantly Christian countries (16.0/105).
Sexual/reproductive behavior, public health, economic and development/demographic measures.
Relationships between CC incidence rates and sexual/reproductive behavior, public health, economic and development/demographic indicators are presented in Table IV, with comparisons between developing countries with low (<20 cases/105) vs. high (>30/105) incidence rates. Many sexual and reproductive behaviors, including higher total fertility and early female age at birth of first child, were significantly associated with increased CC rates. Percent of males using condoms with nonregular sex partners, teenage birth rate and females with nonregular partners were positively associated with higher incidence. Although any contraception usage was negatively associated, this finding was in part attributable to China's high contraception usage. An analysis without China revealed that both contraception measures were positively associated with CC but neither was statistically significant. Countries with higher CC rates had higher percentages of males with nonregular sex partners and a younger female age at first marriage.
|Indicator||All developing countries||Developing countries with low incidence rates||Developing countries with high incidence rates|
|Females using any contraception (%)||106||−0.25||0.23||38||62||42||45|
|Males using condom with nonregular sex partner (%)||31||0.28||0.21||8||20||17||37|
|Total fertility rate (births/woman)||127||4.2||0.17||46||2.4||49||3.2|
|Total births to women under 20 (%)||57||0.86||0.14||25||12||21||17|
|Female age at birth of first child (years)||58||−2.4||0.10||16||21.1||28||20.0|
|Females with nonregular sex partners (%)||35||0.90||0.19||6||3.4||17||7.9|
|Female age at first sex intercourse (years)||50||−1.4||0.055||8||18.8||28||18.6|
|Polygamy (% of marriages)||33||0.21||0.052||8||20||14||24|
|Males with nonregular sex partners (%)||38||0.14||0.040||8||17||18||32|
|Females using oral contraception (%)||97||0.29||0.028||33||5.1||38||5.2|
|Female age at first marriage (years)||61||−0.67||0.015||16||18.8||29||18.0|
|Cigarette consumption (number/year)||84||−0.013||0.49||28||1478||35||441|
|Male disability-adjusted life expectancy (years)||126||−0.92||0.30||46||58.4||49||51.9|
|Male life expectancy (years)||127||−0.94||0.28||46||65.4||49||59.3|
|Female disability-adjusted life expectancy (years)||126||−0.73||0.24||46||60.9||49||54.0|
|Female life expectancy (years)||127||−0.74||0.23||46||69.1||49||62.2|
|Infants with low birthweight (%)||103||0.52||0.21||34||10||46||24|
|HIV seroprevalence (/105 adults)||126||0.0016||0.15||46||343||49||1,940|
|Child mortality rate (deaths/1,000)||126||0.10||0.15||46||59||49||94|
|Number of doctors (/105)||104||−0.040||0.090||37||135||40||64|
|Children immunized for measles (%)||125||−0.27||0.18||46||86||48||72|
|Children immunized for polio (%)||126||−0.28||0.16||46||89||49||76|
|Children immunized for DPT (%)||126||−0.26||0.15||46||87||49||76|
|Infant mortality rate (deaths/1,000)||127||0.16||0.13||46||44||49||63|
|Children immunized for TB (%)||122||−0.19||0.051||44||89||47||83|
|Number of nurses (/105)||99||−0.0054||0.011||36||203||38||57|
|Percent of gross domestic product spent on health||126||5.3||0.38||46||3.3||49||5.3|
|Private spending on health ($/capita)||126||0.075||0.072||46||52||49||89|
|Female unemployment rate (%)||54||0.57||0.086||24||6.3||21||9.7|
|Male unemployment rate (%)||54||0.80||0.083||24||5.0||21||6.4|
|Total (private and public) spending on health ($/capita)||126||0.028||0.056||46||100||49||157|
|Public spending on health ($/capita)||126||0.028||0.022||46||46||49||59|
|Total public expenditures on health (%)||120||0.33||0.016||41||7.0||48||6.5|
|Gross domestic product (US $/capita)||127||0.00045||0.0022||46||1118||49||1310|
|Total expenditures on health (%)||126||0.020||0.0010||46||37||49||28|
|Total schooling, male (years)||49||−3.4||0.26||16||11.7||22||10.3|
|Population under age 15 (%)||127||0.85||0.23||46||28||49||34|
|Male adult illiteracy rate (%)||114||0.37||0.17||40||12||46||26|
|Male population over age 60 (%)||127||−1.9||0.15||46||8.5||49||6.6|
|Female population over age 60 (%)||127||−1.0||0.12||46||10.9||49||7.7|
|Total schooling, female (years)||49||−1.9||0.096||16||10.8||22||10.5|
|Female adult illiteracy rate (%)||114||0.16||0.072||40||26||46||42|
|Urban population with access to adequate sanitation (%)||94||0.14||0.026||31||74||38||78|
|Urban population with access to safe water (%)||94||−0.12||0.019||27||85||38||81|
|Percent of population urban||127||0.032||0.0021||46||42||49||41|
|Rural population with access to safe water (%)||96||0.020||0.0019||27||53||41||66|
|Rural population with access to adequate sanitation (%)||92||0.0017||0.000||30||28||38||17|
|Indicator2||All developing countries|
|Number of doctors (/105)||−0.061||<0.0001|
|Children immunized for measles (%)||0.30||0.0001|
|Female disability-adjusted life expectancy (years)||−0.77||0.0002|
|Female adult illiteracy rate (%)||−0.25||0.0003|
|Infants with low birthweight (%)||0.39||0.002|
Among public health measures, life expectancy and number of doctors were negatively associated with CC rates. The negative association of cigarette consumption with CC was primarily attributable to a high prevalence of smoking in China, which had a low CC rate, and India's low smoking prevalence and high CC rate. After analyzing this relationship without India and China, a small negative association was not statistically significant. Tuberculosis, percent of low-birthweight infants, HIV and child mortality were positively associated with higher CC incidence. Countries with high CC rates had fewer immunized children, fewer nurses and higher infant mortality. All public health measures examined, except smoking, were more strongly associated with mortality rates than incidence.
Although most economic indicators were not significantly associated with CC rates, more health spending was associated with higher CC rates. The percent of gross domestic product spent on health, private spending on health and Gini index, a measure of income inequality, were positively associated with higher CC incidence. Countries with higher CC rates had more total spending on health but a smaller percent of total expenditures spent on health. However, some of these findings were partly due to China's relatively minimal health spending and low CC rates. Furthermore, the gross domestic product was not significantly higher among countries with low CC rates.
Among development and demographic characteristics, percent of population under age 15 was associated with higher CC rates, while percent of male population over 60 and percent of female population over 60 were associated with lower CC rates. Increased schooling of males and females was associated with lower CC rates. However, education results were partially influenced by relatively good schooling in China and high illiteracy in India, and associations were no longer significant after removing those 2 countries. The HDI was not associated with CC incidence but was positively associated with CC mortality (p = 0.02). Urban and rural development measures and percent urbanization had little relation to CC rates. Most development and demographic measures had stronger associations with mortality rates than incidence.
In the multivariate model (Table V), number of doctors had a strong negative association with CC incidence. Relationships between CC rates and percentages of immunized children and female illiteracy measures changed sign in the multivariate model. This was primarily because of poor immunization and high illiteracy rates in the Middle East and Asia (countries with low CC rates) and low illiteracy rates in Latin America and Europe (countries with high CC rates). Female life expectancy continued to be negatively associated and percentage of infants with low birthweight continued to be positively associated with CC rates. Predominant religion and geographic region remained independently associated with incidence rates. Compared with the 45 countries excluded from the model due to missing information, the 82 countries in the multivariate model did not appear to be different in terms of predominant religion, geographic region, female disability-adjusted life expectancy and child mortality rates.
Among the 127 developing countries without effective screening and treatment programs, striking differences in CC rates were observed. Both geography and religion were independently associated with CC rates. Middle Eastern and predominantly Muslim and Buddhist countries tended to have the lowest rates, while Latin American and predominantly Christian and Hindu countries generally had the highest rates. Countries with high incidence rates tended to have more total fertility, younger ages at first childbirth and a higher percentage of males with nonregular sex partners. Countries with high CC rates also had fewer years of life expectancy and fewer doctors, more infants with low birthweight and more adults with tuberculosis and HIV. Unexpectedly, the few economic spending measures that were associated with incidence rates showed positive relationships, suggesting increasing rates with improving economies. Finally, countries with high CC rates tended to have younger, less educated and more illiterate populations.
An ecologic study of this type has limitations. First, while ecologic analyses are useful for describing patterns of disease occurrence in relation to regional, social and demographic patterns, they cannot be used for estimates of individual risk because variables are not measured at the individual level and the temporal sequence of events is undetermined.35 Second, current incidence rates undoubtedly reflect exposures that occurred decades ago, and it is not known whether and how behavioral, economic and demographic measures may have changed during this time. Available data do suggest, however, that CC rates for developing countries have not changed substantially over the last 30 years. Third, the validity of country-level data is unknown, and some data, such as those describing sexual behaviors, were not recorded for many countries, particularly countries in the Middle East. Fourth, and perhaps most importantly, information on the geographic distribution of HPV types is not complete. The International Biological Study on Cervical Cancer described the regional variations of several high-risk HPV types in cancers, but HPV prevalence in the general population was not determined.9 Despite these limitations, findings from this ecologic analysis can help to inform decisions concerning the need for additional research and the allocation of health-care resources.
Differences in CC rates between developed and developing countries are undoubtedly due in part to levels of human development, economic resources and the capacity to sustain a national CC program. However, after excluding the most developed countries, levels of development, economic income and amount of spending on health did not appear to have a clear association with CC rates. Similar findings within India led others to suggest that “environmental differences” must account for patterns of widely varying incidence rates.36 The results of this study led us to suggest that differences in social factors, health programs and sexual behaviors may partially explain global and regional patterns of CC rates. For example, India and China, although geographic neighbors with similarly large populations and poor economies, have quite different rates of CC (30.7/105vs. 5.2/105, respectively). Since the early 1970s, China's incidence rate has steadily declined, most likely from the government's social changes and improved health programs.37 After markedly reducing prostitution, improving case finding and therapy, and destigmatizing sexually transmitted diseases, China nearly eliminated sexually transmitted infections in the 1960s.38 Meanwhile, India, like most developing countries, has not fully addressed many sexual and reproductive health issues and has seen increasing rates of some sexually transmitted infections. This example not only illustrates the need to develop appropriate intervention strategies for individual countries but also suggests that population-level programs that attempt to reduce exposures to sexually transmitted infections could have an impact on decreasing HPV infections and lowering CC rates.
Social class, education, economic opportunities and demographic changes largely influence regional and cultural patterns of sexual and reproductive behaviors.39 More modern societies, with higher levels of schooling and more urbanization, tend to abandon traditional patterns of behavior and to favor delayed childbearing, casual partnerships and commercial sex.40 Blanc and Way41 determined that African societies have a larger proportion of adolescents who have ever had sex or been married than Latin American and Middle Eastern societies, which may account for some of the geographic patterns observed with CC rates. In addition to early onset of sexual activity, levels of casual partnerships and sexual networking may be important contributing factors in countries with elevated CC rates. Carael et al.42 suggested that the prevalence of nonmarital sex is low among conservative countries characterized by “restrictive sexual values”, such as Sri Lanka and Ethiopia (predominantly Buddhist and Muslim countries), but higher in less restrictive countries, such as Lesotho and Tanzania (predominantly Christian countries).
In addition to societal pressures, several studies have described the important role of male sexual behaviors, such as number of prostitutes and extramarital partners, in transmitting HPV to their regular female partners.43–46 The finding of an almost 2-fold increase in the percentage of men reporting sex with nonregular partners in high- vs. low-incidence countries is consistent with results from de Sanjose et al.,47 who found that husbands who visited prostitutes significantly increased their wives' risk of developing invasive CC. White et al.48 determined that a male's younger age at sexual debut and number of sex partners before marriage are significantly associated with the probability of men having extramarital sexual relationships. In developing countries where female monogamy is more commonly practiced than male monogamy, CC rates may be determined more by male than female sexual behaviors.49
Differences in the oncogenicity of variants of high-risk HPV types may explain some CC patterns. For example, the HPV-16 Asian-American and African variants, which are more commonly found in Native-American and African populations, respectively, appear to be more highly oncogenic than the prototype European HPV-16 variant.50, 51 Prevalence information that details the geographic distribution of HPV types and variants is needed to better understand their relationship with global CC patterns.
While ongoing as well as new and potentially less expensive screening technologies require health personnel and a public health infrastructure, which are generally lacking in countries with high rates, the distinct patterns of CC rates and economic spending levels suggest that certain programmatic approaches may be feasible for some developing countries. First, countries with high rates might decrease CC by promoting delayed onset of sexual activity and fertility. Second, efforts to encourage sexual monogamy and discourage casual and commercial sex networks might also reduce CC rates. Since many countries with increased CC rates appear to have funds to spend on health care, performing once-in-a-lifetime screening of women 30–50 years of age may be cost-effective in some countries.52 Finally, the relatively low immunization rates and shortage of medical personnel in countries with high rates suggest potential challenges to the introduction of prophylactic HPV vaccines.
We thank Ms. C. Nelson for assistance with the manuscript.
- 3GLOBOCAN 2000: Cancer incidence, mortality, and prevalence worldwide. International Agency for Research on Cancer CancerBase 5, Version 1.0. Lyon: IARC, 2001., , , .
- 12Feasibility of screening for premalignant lesions in an African environment. In: WilliamsAO, O'ConorGT, de ThéGB, JohnsonCA, eds. Virus-associated cancers in Africa. IARC Sci Publ 63. Lyon: IARC, 1984. 451–63., , .
- 13Organization of screening programs in developing countries with reference to screening for cancer of the uterine cervix in India. In: HakamaM, MillerAB, DayNE, eds. Screening for cancer of the uterine cervix. IARC Sci Publ 76. Lyon: IARC, 1986. 273–88., .
- 14WHO. Control of cancer of the cervix uteri. A WHO meeting. Bull World Health Organ 1986;64: 607–18.
- 23GLOBOCAN 2000: Database guide. Lyon: IARC, 2001., , , .
- 24Fundamentals of Biostatistics, 4th ed. Harrisonburg, VA: Duxbury Press, 1995..
- 26United Nations Development Programme. Human development report 2000. New York: Oxford University Press, 2000.
- 27WHO. The world health report 2000, health systems: improving performance. Geneva: WHO, 2000.
- 28United Nations Children's Fund. The state of the world's children 2000. New York: UNICEF, 2000.
- 29United Nations Statistics Division. Demographic and Social Indicators. Accessed at www.un.org/Depts/unsd/social/ on 25–26 Sept 2000.
- 30Joint United Nations Programme on HIV/AIDS. Report on the global HIV/AIDS epidemic. Geneva: UNAIDS, 2000.
- 31World Bank. World development indicators 2000. Herndon, VA: World Bank, 2000.
- 32Macro International. Demographic and health surveys, Measure DHS+. Washington, DC. Macro International, 2000.
- 33United States Central Intelligence Agency. The world factbook 2000. Washington DC: US Government Printing Office, 2000.
- 34StataCorp. Stata statistical software: release 6.0. College Station, TX: Stata Corporation, 1999.
- 47Social differences in sexual behaviour and cervical cancer. In: KogevinasM, PearceN, SusserM, BoffettaP, eds. Social inequalities and cancer. IARC Sci Publ 138. Lyon: IARC, 1997. 309–17., , , .