Quality of data
All of the data used for this analysis had appeared in the series Cancer incidence in five continents (Waterhouse et al.,1982; Muir et al.,1987; Parkin et al.,1992, 1997). This constitutes the most reliable source of international cancer incidence statistics available, and peer-reviewed procedures are applied to decide on the inclusion or otherwise of the results in each volume. In this analysis, close to 175,000 cases were considered, of which 133,000 were squamous-cell carcinomas. For the study period, histological verification of diagnosis was high (>90%) and the fraction of cases included based on death certificate information was low (<5%). Squamous-cell carcinomas represented approximately 80% of all tumors of the cervix, while cases without histological diagnosis and those with ill-defined histology (grouped as “other and unspecified cancers”) generally comprised <10%. There were no significant trends in incidence of this latter group (e.g., reflecting improved diagnostic methods) that would have influenced the observed trends in squamous-cell carcinomas. The statistical methods used in the analysis of the secular trends are standards in cancer epidemiology and were developed to separate the effects of age from the effects of calendar period and the effects linked to generation as reflected by the year of birth.
Geographic variation in incidence rates of squamous-cell carcinoma of the cervix
The average annual incidence of cervical cancer in each country varies greatly, from the high rates in Latin America and Hispanic populations in the United States to the low rates in Israel. In Europe, low rates are observed in the Nordic countries, where organized screening was effectively introduced as early as 1960 and coverage has been consistently high. Denmark, however, has retained relatively high rates of squamous-cell cervical carcinoma. Screening programs have evolved differently in the individual Nordic countries. In Finland and Sweden, there are nationwide cervical cancer screening programs, and the overall attendance rate in Finland is 70% to 80% of the female population, whereas only a few counties in Denmark, including the most populous ones, have organized screening programs and the attendance rate is low (Hakama, 1997; Sigurdsson, 1999).
The geographical variation in the incidence of cervical cancer observed between countries can also be observed between populations within them. For example, in France, there is an almost 2-fold variation between the rates in the Herault region (14.1) and the Tarn (7.7). Similar variations are observed in Spain and Italy. The populations within countries should theoretically benefit from similar quality of screening and medical-care programs, but inequalities or cultural differences could modify the impact of the programs in some areas. Alternatively, the observed differences could be related to differences in exposure to the key risk factors (namely, sexual behavior patterns and prevalence of human papillomavirus infection).
Singularities in trends
In the majority of countries, there has been a progressive decline in the incidence of invasive squamous-cell carcinoma, while in 8 countries studied, rates were essentially stable. The United Kingdom recorded an upward trend in the incidence and mortality due to cervical cancer among women below the age of 50, but this trend was not paralleled among older women. In fact, there are marked birth cohort–specific differences in the risk of cervical cancer in the United Kingdom, which, in its early years, the screening program was unable to hide (Parkin et al.,1985). However, after 1988, the program was substantially improved, mostly by reinforcement of the centralized (and computerized) services of call and recall of women, resulting in a substantial increase in coverage (from 42% in 1988 to 85% in 1995) and a new reversal of the secular trend, with a consistent decrease in the incidence and mortality rates in the young age groups from 1994 onward (Patnick, 1997; Gibson et al.,1997; Quinn et al.,1999). Accelerated declines in cervical cancer incidence and mortality have been also observed in Scotland (Walker et al.,1998).
An increase in incidence and mortality of cervical carcinomas among young women was noted in previous studies in other European countries (Beral et al.,1994). Slovakia and Slovenia, for example, show increases in the incidence rates of squamous-cell cervical cancer among young women. It is likely that deficiencies of screening programs in these countries can explain such changes (Vlasák et al.,1991; Pompe Kirn et al.,1992).
Our study included mostly developed countries with either established screening programs or high levels of opportunistic screening. In an earlier report (Vizcaino et al.,1998), we found that in many countries [e.g., the white population of the United States, Australia, New Zealand (non-Maori), Japan (Osaka), the Chinese population of Singapore, Denmark and Sweden] there has been an increase in the incidence of adenocarcinoma among younger women, despite a concomitant decrease in squamous-cell carcinoma. This is coherent with an underlying increase in risk of both histological types of cervical cancer but less efficient prevention of adenocarcinoma by conventional screening. In an evaluation of the screening program in Iceland, it was calculated that the sensitivity of the standard cytology for identifying pre-invasive squamous-cell carcinoma following a 3-year interval was 81% but only 42% for pre-invasive adenocarcinoma (Sigurdsson, 1995). In a case-control study in Australia, Mitchell et al. (1995) observed that the risk of adenocarcinoma was not reduced by screening.
We conclude that the likely explanation of the pattern reported here is that squamous carcinoma of the cervix is being largely prevented by screening programs in developed countries. As a result, incidence is actually declining, or relatively stable, when perhaps an increase in risk might have been anticipated due to a change in sexual life-style favoring transmission of HPV. In general, it is accepted that organized programs, with regular screening of a defined population (such as in the Nordic countries and the United Kingdom), are more efficient than spontaneous or opportunistic screening (Hakama et al.,1985; Sigurdsson, 1999). However, even “opportunistic” screening, although wasteful of resources, can clearly achieve a beneficial result if it is sufficiently intensive, as the results for several countries, including the United States, demonstrate.
The situation in developing countries is, however, far from satisfactory. Although some cytological screening is done, there are no organized programs and the testing is often distributed inefficiently among the population and of poor quality (Lazcano Ponce et al.,1998). As a result, in several Latin American countries, for example, there has been a very limited impact on the incidence of cervical cancer, despite relatively large numbers of cytological examinations, as in Cuba (Fernandez et al.,1996) and Costa Rica (Herrero et al.,1997).
In developing countries with no resources for screening, there are few long time series available to monitor cancer incidence, but data from Uganda (Wabinga et al.,1999) indicate that, at least in some areas, substantial increases in the incidence of cervical cancer may have occurred.
In conclusion, developed countries should continuously monitor the coverage and the quality of cytology-based screening programs. Particular attention should be given to the evaluation of alternatives to reduce false-negative results in detecting cervical adenocarcinoma. In most developing countries, an effort is required to validate and introduce simple and more effective screening techniques. In the long run, primary prevention with an HPV vaccine might be a more effective solution to the control of this disease.