Estimates of the world-wide prevalence of cancer for 25 sites in the adult population

Authors

  • Paola Pisani,

    Corresponding author
    1. Unit of Descriptive Epidemiology, International Agency for Research on Cancer, Lyon, France
    • Unit of Descriptive Epidemiology, International Agency for Research on Cancer, 150 cours Albert Thomas, 69372 Lyon Cedex 08, France
    Search for more papers by this author
    • Fax: +33-472-73-86-50

  • Freddie Bray,

    1. Unit of Descriptive Epidemiology, International Agency for Research on Cancer, Lyon, France
    Search for more papers by this author
  • D. Maxwell Parkin

    1. Unit of Descriptive Epidemiology, International Agency for Research on Cancer, Lyon, France
    Search for more papers by this author

Abstract

In health services planning, in addition to the basic measures of disease occurrence incidence and mortality, other indexes expressing the demand of care are also required to develop strategies for service provision. One of these is prevalence of the disease, which measures the absolute number, and relative proportion in the population, of individuals affected by the disease and that require some form of medical attention. For most cancer sites, cases surviving 5 years from diagnosis experience thereafter the same survival as the general population, so most of the workload is therefore due to medical acts within these first 5 years. This article reports world-wide estimates of 1-, 2–3- and 4–5-year point prevalence in 1990 in the population aged 15 years or over, and hence describes the number of cancer cases diagnosed between 1986 and 1990 who were still alive at the end of 1990. These estimates of prevalence at 1, 2–3 and 4–5 years are applicable to the evaluation of initial treatment, clinical follow-up and point of cure, respectively, for the majority of cancers. We describe the computational procedure and data sources utilised to obtain these figures and compare them with data published by 2 cancer registries. The highest prevalence of cancer is in North America with 1.5% of the population affected and diagnosed in the previous 5 years (about 0.5% of the population in years 4–5 and 2–3 of follow-up and 0.4% within the first year of diagnosis). This corresponds to over 3.2 million individuals. Western Europe and Australia and New Zealand show very similar percentages with 1.2% and 1.1% of the population affected (about 3.9 and 0.2 million cases respectively). Japan and Eastern Europe form the next batch with 1.0% and 0.7%, followed by Latin America and the Caribbean (overall prevalence of 0.4%), and all remaining regions are around 0.2%. Cancer prevalence in developed countries is very similar in men and women, 1.1% of the sex-specific population, while in developing countries the prevalence is some 25% greater in women than men, reflecting a preponderance of cancer sites with poor survival such as liver, oesophagus and stomach in males. The magnitude of disease incidence is the primary determinant of crude prevalence of cases diagnosed within 1 year so that differences by region mainly reflect variation in risk. In the long-term period however different demographic patterns with long-life expectancy in high-income countries determine a higher prevalence in these areas even for relatively uncommon cancer sites such as the cervix. © 2002 Wiley-Liss, Inc.

In previous articles, we presented estimates of the incidence and mortality for 25 common cancers in 23 areas of the world for 1990,1, 2 and this companion article is the first to report corresponding estimates of the prevalence of these cancers world-wide.

Incidence and mortality are the basic measures of disease occurrence adopted in epidemiology to monitor the risk of the disease in populations. They provide evidence for the rational planning of resources and evaluation of cancer control activities. They are also the fundamental data resource for studies on cancer causation. In health services planning, other indexes that express the demand of care are required to develop strategies for service provision. One of these is prevalence of the disease, which measures the absolute number, and relative proportion in the population, of individuals affected by the disease and that require some form of medical attention. Prevalence of a disease in the population is determined by the combination of incidence rate, the proportion of cases dying (fatality) or cured and the general life expectancy in the population. The main difficulty in defining cancer prevalence consists in establishing when a patient is cured, since relapse may occur after many years free of the disease. In most cases however, most of the treatment phase, at least that involving the health services, takes place within the first 5 years from diagnosis. For pragmatic reasons we therefore choose to refer to “prevalent” cases of cancer those individuals still alive who were diagnosed within the last 5 years. Individuals still alive at 5 years are considered “cured.” In fact, the death rates of most cancer patients 5 years after diagnosis are quite close to those “expected” in the general population, although there are exceptions, notably breast cancer, where the risk of death remains higher than normal for many more years.

This article reports world-wide estimates of 1-, 2–3- and 4–5-year point prevalence in 1990 and hence describes the number of cancer cases diagnosed between 1986 and 1990 in subjects aged 15 years or over, who were still alive at the end of 1990. These estimates of prevalence at 1, 2–3 and 4–5 years are applicable to the evaluation of initial treatment, clinical follow-up and point of cure, respectively, for the majority of cancers. In the context of planning health services, they are therefore of more practical use than total prevalence, i.e., the number of persons alive who have ever had cancer at some time in the past.

The method essentially involves estimating partial prevalence for each country of the world based on a function of its respective incidence and observed survival. These are then combined into regional estimates. We identified 2 sources of prevalence data, which can be used to validate our own results: observed prevalence in 1985 in the Nordic countries based on registry data,3 and for Canada, estimates derived according to a similar approach.4

METHOD AND DATA SOURCES

Prevalent cases of age k were estimated from incidence rates and year-specific survival probabilities according to the following formula:

equation image

where ICx is the annual number of new cases at age x, Sx(t) is the proportion of cases diagnosed at age x and alive at time t after diagnosis, and n is the number of years as case or number of years before cured.

The method is graphically explained in Figure 1 with an example. The algorithm requires estimation of the distribution of incident cases by single year of age. This was obtained by interpolating the cumulative distribution of cases by age with a logistic function, keeping the total number of cases fixed. Prevalent cases were then summed into the same age-groups as used for incidence.

Figure 1.

Schema illustrating the formula of partial prevalence with the example of calculation of 5-year prevalent cases at age 45. ICx indicates number of incident cases of age x in the year, Sx(t) proportion of cases diagnosed at age x surviving t years.

We assume that (i) incidence rates are constant in the preceding n years; (ii) annual incidence cases occur at mid-year; and (iii) the curve of relative survival with time, in interval (0.5), follows the Weibull distribution. “S” in the above equation refers to actual observed survival amongst cancer patients. Published statistics in survival generally report relative survival, that is, corrected for general mortality. Observed survival must be derived by multiplying relative survival for the probability of surviving until the next birthday, experienced by the general population; this is estimated by:5

equation image

where a is the age-interval width/age group, mx is the all-causes mortality rate, age-, sex- and region-specific.

Such calculations were possible due to the availability of survival estimates for many developed countries, as well as a number of countries in developing areas of the world, all described below, and updated estimates of world-wide incidence for 1990.1

Cancer sites and age groups presented in this report are those presented in our recent publications on incidence and mortality excluding childhood cancer (0–14 years). The geographical areas have been instead regrouped in 11 larger areas for reasons of space (Fig. 2). Detailed and updated (year 2000) estimates for individual countries are available on the Internet (www-dep.iarc.fr). Referring to Figure 2, “Developed or high-income countries” comprise EU and EEA, Eastern Europe, Northern America, Japan and Oceania, while “developing countries” make up the remainder.

Figure 2.

Map showing the 11 world areas and abbreviations used in the text.

Since the main scope of this article is to describe the methods used to produce the estimates, we chose to present 1990 figures rather than the more recent set available on line because the former could be validated against published data on actual counts of prevalent cancer cases.3, 4

Estimates are reported for all cancer sites combined, cancer of the oral cavity (ICD-9 140–145), nasopharynx (147), oropharynx (146, 148–149), oesophagus (150), stomach (151), colon and rectum (153 and 154), liver (155), pancreas (157), larynx (161), lung (162), melanoma of skin (172), female breast (174), cervix uteri (180), corpus uteri (182), ovary (183), prostate (185), testis (186), bladder (188), kidney and urinary tract (189), brain and central nervous system (191–192), thyroid (193), non-Hodgkin lymphoma (200,202), multiple myeloma (203), Hodgkin's disease (202) and leukaemia (204–208). Calculations were based on 4 age-groups: 15–44, 45–54, 55–64 and 65+, and are reported separately by sex.

Sources of survival data

Three sources of site-specific survival data were available:

(i) The IARC study on cancer survival in developing countries estimated relative survival of cases diagnosed between 1982 and 1993, in several communities served by a population-based cancer registry, in China, Cuba, India, the Philippines and Thailand. An overall age and site-specific survival rate for developing countries was calculated as the arithmetic mean of registry-specific survival rates. Similarly, rates for China and India were obtained by taking the unweighted mean of the survival rates from Chinese and Indian registries, respectively.

(ii) The Surveillance, Epidemiology, and Ends Results (SEER) programme in the United States produced estimates of relative survival based on registry data from 9 areas covering approximately 10% of the U.S. population, of cases diagnosed between 1973 and 1994.6 We used all-races rates for the 2 sexes combined.

(iii) The EUROCARE I study recently estimated survival rates of incident cases between 1985 and 1989 in 43 cancer registries in 17 countries in Europe.7 Observed age-, sex- and site-specific rates were derived directly from the EUROCARE I database. Where possible, country-specific survival estimates were used, based on contributing regional registries. For countries where no local survival data were available, an appropriate pooled dataset comprising registry data from several countries was used.

A more detailed description of how these 3 datasets were compiled is given in an article describing cancer mortality world-wide in 1990.2 In addition, local sources of survival were used for stomach cancer in Japan8 and melanoma of the skin in Australia,9 where prognosis was likely to differ from that in SEER and EUROCARE I. For composite sites such as oral cavity and mouth, we used the weighted average of sub-site survival rates.

We used the same set of relative survival rates for men and women, given that their relative survival rates are approximately equal for most cancers if other risk factors and competing causes of death are adequately controlled for.10 Relative rates were converted to observed rates using region-specific probabilities of death (all causes) around 1990.11 Estimates were available for the following 8 areas of the world: Established Market Economies (EME), Former Socialist economies of Europe (FSE), India (IND), China (CHN), Other Asia and Islands (OAI), Sub-Saharan Africa (SSA), Latin American and the Caribbean (LAC) and the Middle Eastern Crescent (MEC). Thus mid-year observed survival rates for 1–5 years were calculated for every country by age and sex from the corresponding relative survival rates and appropriate death probabilities. The appendix shows the sources of survival for each area of the world and, where required, the corresponding source of the probability of death from which observed survival was calculated. Hence the observed survival rates calculated for developing countries were used as proxies of survival for all of the countries of Africa, Melanesia, Micronesia and Polynesia and for the majority of the Asian countries. Separate survival rates were calculated for the populations of China and India, respectively. Various EUROCARE I datasets estimated survival in Europe, New Zealand and Australia (excepting local survival data used for melanoma of the skin), while SEER all-race survival rates estimated cancer survival in Northern America and Japan (excepting local rates used for stomach cancer).

Validation

Two sources of prevalence of selected cancers were used to validate the method adopted. The Association of the Nordic Cancer Registries (Denmark, Finland, Iceland, Norway and Sweden) have published 1- and 2–5-year point prevalence at 1/1/87 based on observed cancer registrations.3 The National Cancer Institute of Canada has published 5-year prevalence estimates for Canada in 1990.4 Our estimates are compared with the other 2 sources in Table I by means of the ratios of prevalence to incidence from each source.

Table I. Ratios of 5-Year Prevalence to Incidence by Sex and Site in A) Our Data Set, B) 1/1/87/ in the Nordic Countries and C) 1990 in Canada
Site (ICD 9)Nordic CountriesCanada
A) Our estimatesB) Cancer registries (3)A) Our estimatesC) Local estimates (4)
Males
 Mouth (140–145)3.43.4
 Oropharynx, hypopharynx &  NOS (146, 148–149)2.11.6
 Nasopharynx (147)3.02.8
 Oesophagus (150)0.90.9
 Stomach (151)1.21.11.11.3
 Colon/rectum (153–154)2.82.52.32.7
 Liver (155)0.30.3
 Pancreas (157)0.50.40.40.5
 Larynx (161)3.63.4
 Lung (162)1.00.90.91.1
 Melanoma (172)3.83.7
 Prostate (185)3.53.02.53.0
 Testicular (186)4.44.94.64.2
 Bladder (188)3.43.33.13.7
 Kidney (189)2.52.12.22.6
 Brain and central nervous system  (191–192)1.61.6
 Non-Hodgkin's lymphoma (200,  202)2.72.8
 Leukaemia (204–208)1.42.2
 All sites but 1732.22.22.02.3
Females
 Mouth (140–145)3.53.4
 Oropharynx, hypopharynx and  NOS (146, 148–149)2.51.8
 Nasopharynx (147)2.22.4
 Oesophagus (150)1.51.0
 Stomach (151)1.11.21.31.5
 Colon/rectum (153–154)3.02.62.72.8
 Liver (155)0.40.4
 Pancreas (157)0.40.40.40.6
 Larynx (161)3.63.5
 Lung (162)1.11.01.11.2
 Melanoma (172)4.24.4
 Breast (174)3.94.13.63.8
 Cervix (180)3.33.43.73.7
 Corpus (182)4.14.03.44.1
 Ovary (183)2.92.42.62.3
 Bladder (188)3.43.13.63.5
 Kidney (189)2.22.22.22.7
 Brain and central nervous system  (191–192)1.51.6
 Non-Hodgkin's lymphoma (200,  202)3.02.8
 Leukaemia (204–208)1.42.1
 All sites but 1733.12.82.72.9

Overall, the results suggested that, for the majority of cancers, our estimates of prevalence were reasonable. As might be expected, our figures were generally closer to those published for Canada than for the Nordic countries because the former are also estimates based on a similar approach but with the advantage that they could use the appropriate survival experienced by their case population. For Canada, the largest difference was observed for leukaemia in both sexes (1.4 estimated vs. 2.2 for men and 2.1 for women), indicating a pessimistic value for survival and prostate cancer in men (2.5 vs. 3.0), possibly reflecting more optimistic prognosis in the era of screening. In women, the prevalence of cancers of the pharynx (ICD-9 146,148–149) appears substantially overestimated relative to incidence (2.5 vs. 1.8) as well as cancer of the ovary to a lesser extent. Otherwise, the differences in the prevalence:incidence ratios are not substantial.

Differences, comparing our estimates with the actual prevalence in the Nordic countries around the same period, were of the same magnitude as those observed in the comparison with Canada. The largest errors were observed for the pharynx in both men and women, pointing to an overestimate of survival in our data, and for the ovary. All other ratios differ by less than 0.5.

RESULTS

The results are given in Tables II–VI as numbers of affected individuals per 100,000 population, by geographical region. Table II shows 5-year prevalence of all cancers and population totals (thousands). Tables III–V give proportions of cases surviving 1, 2–3 and 4–5 years, respectively, by cancer site, sex and region. The 3 measures in Tables III–V estimate the cases undergoing initial standard treatment (1 year), those under clinical follow-up (2–3 years) and those in remission and entering long-term follow-up (4–5 years). These statistics can be summed to give the prevalence of all cases diagnosed within 5 years (Tables II,VI).

Table II. Population (Thousands) and All Site 5-Year Prevalence of Cancer Cases (Thousands), by Sex and Region
 Population (1,000)Thousands of 5-year prevalent cases5-year prevalence
MenWomenMen %Women %Men and women %
Sub-Saharan Africa266,681325.8422.70.250.310.28
Middle East and Northern Africa175,677183.3236.70.200.280.24
Latin America and Caribbean281,686469.5757.80.340.530.44
North America217,7111,572.01,639.31.501.451.48
China and OEA891,0531,134.41,037.40.250.240.24
Japan100,805529.6469.91.080.910.99
South Eastern Asia279,650264.2453.60.190.320.26
South Central Asia769,154796.11,156.20.200.310.25
Eastern Europe239,249773.81,031.40.700.800.75
EU and EEA334,0301,837.12,098.31.141.211.18
Oceania20,036106.1116.51.061.161.11
Developed907,5584,811.95,344.61.111.131.12
Developing2,668,1743,180.14,075.30.240.310.27
All3,575,7327,991.99,419.90.450.530.49
Table III. Prevalent Cases Diagnosed Within a Year, as Proportion of the Population per 100,000, By Site (ICD9), Sex and Geographic Area
 Mouth (140–145)Other pharynx (146, 148, 149)Nasopharynx (147)Oesophagus (150)Stomach (151)Colon/rectum (153, 154)Liver (155)Pancreas (157)Larynx (161)Lung (162)Melanoma (172)Breast (174)Cervix (180)
MFMFMFMFMFMFMFMFMFMFMFFF
SSA4.22.30.70.21.00.33.11.34.02.82.82.33.41.30.60.51.60.21.80.61.41.614.421.8
MENA2.11.40.60.21.30.71.00.82.92.13.83.70.70.40.50.44.40.66.71.50.70.823.06.6
LAC4.92.02.30.50.50.12.80.910.26.08.69.10.70.51.51.34.40.88.82.82.21.938.230.5
North America8.95.33.21.10.70.33.41.26.43.947.545.11.30.73.33.28.02.147.127.913.510.5125.911.2
China and OEA1.00.50.20.12.81.48.53.923.610.69.27.86.02.00.50.31.40.411.24.30.20.213.15.4
Japan3.92.01.50.30.60.27.01.4103.951.848.335.711.73.54.32.84.30.227.39.80.50.442.213.5
South-Eastern Asia2.41.81.40.54.62.01.10.63.42.26.35.63.01.10.50.42.10.58.43.10.30.321.816.2
South-Central Asia7.75.25.31.40.50.24.13.63.02.02.62.30.50.30.50.44.40.85.11.30.30.320.019.3
Eastern Europe10.22.43.10.50.60.33.00.724.516.525.826.61.11.32.72.315.41.050.69.14.04.756.518.8
EU and EEA13.33.28.20.91.00.35.01.417.010.950.046.22.91.43.22.713.61.046.69.87.010.6106.215.7
Oceania18.98.03.50.71.00.33.01.87.64.546.641.21.50.72.21.96.20.828.110.428.527.891.018.0
Developed10.53.44.90.80.80.34.31.226.115.143.139.63.01.43.22.711.51.245.314.07.68.390.515.2
Developing3.92.42.10.61.80.84.82.610.85.45.95.32.91.10.60.52.90.57.72.60.60.619.515.0
World5.52.72.80.61.60.74.72.314.57.915.014.43.01.21.21.15.00.716.85.62.32.638.315.1
 Corpus uteri (182)Ovary (183)Prostate (185)Testis (186)Bladder (188)Kidney (189)Brain, CNS (191, 192)Thyroid (193)NHL (200)Hodgkin's (201)Multiple Myeloma (203)Leukaemia (204–208)Other sitesAll sites but skin (140–208 but 173)
FFMMMFMFMFMFMFMFMFMFMFMF
SSA2.14.09.20.33.11.30.70.50.30.30.81.64.43.01.70.80.50.41.41.125.618.372.883.2
MENA3.43.43.20.910.32.51.00.71.41.00.82.72.92.41.30.80.20.32.11.310.812.659.673.2
LAC7.15.316.12.15.31.62.41.52.51.71.24.23.72.61.61.00.60.62.82.416.519.4101.4141.9
North America22.714.7117.24.529.38.711.67.35.34.12.86.515.812.43.22.54.54.09.46.826.723.8373.8351.8
China and OEA1.82.60.70.22.20.71.00.60.40.80.21.40.90.70.30.30.10.11.71.36.77.879.167.9
Japan6.28.710.41.311.23.26.43.12.11.61.87.97.55.00.50.21.81.95.43.716.115.0277.8220.3
South-Eastern Asia3.14.92.80.72.20.60.80.50.80.51.14.52.72.00.60.30.40.32.01.710.49.758.184.2
South-Central Asia1.84.02.40.72.40.70.70.51.20.80.71.72.11.31.10.50.30.32.01.413.312.960.783.0
Eastern Europe16.615.016.33.514.83.29.36.34.23.11.33.94.73.33.72.31.71.96.65.216.817.8223.9219.3
EU and EEA19.814.653.16.431.57.311.56.55.03.61.84.012.09.43.52.13.93.69.06.827.125.2332.4313.4
Oceania13.010.859.74.216.85.87.65.74.23.31.44.212.79.92.11.63.63.28.85.827.723.9295.9294.3
Developed18.014.054.84.624.26.110.36.34.63.41.95.010.68.03.12.13.33.18.16.023.221.7308.9287.2
Developing2.73.74.00.63.21.01.00.71.00.80.62.22.21.60.90.50.30.31.91.512.112.172.084.0
World6.76.416.31.68.32.33.32.21.81.50.93.04.33.31.40.91.01.03.42.714.814.7129.5137.8
Table IV. Prevalent Cases Diagnosed 2–3 Years Earlier, as Proportion of the Population per 100,000, By Site (ICD9), Sex and Geographic Area
 Mouth (140–145)Other pharynx (146, 148, 149)Nasopharynx (147)Oesophagus (150)Stomach (151)Colon/rectum (153, 154)Liver (155)Pancreas (157)Larynx (161)Lung (162)Melanoma (172)Breast (174)Cervix (180)
MFMFMFMFMFMFMFMFMFMFMFFF
SSA6.43.60.90.31.60.63.11.34.73.34.23.33.71.40.50.42.40.41.80.62.12.523.735.1
MENA3.22.20.80.32.01.20.90.83.32.45.55.40.70.50.40.36.60.96.51.51.11.137.910.6
LAC7.43.12.90.60.70.22.81.011.56.812.513.20.70.51.21.06.41.28.42.73.32.962.649.3
North America15.18.94.31.51.20.53.41.27.74.779.375.81.40.72.72.413.83.650.930.225.720.0234.520.0
China and OEA1.60.80.30.14.32.08.54.125.911.713.211.26.82.20.50.32.10.616.56.40.40.322.08.5
Japan6.63.52.00.31.00.37.01.4188.593.081.060.111.43.33.42.27.40.429.310.60.90.779.223.6
South-Eastern Asia3.72.71.80.77.13.21.10.63.92.69.28.23.21.10.40.33.10.78.13.00.40.436.126.2
South-Central Asia11.88.26.81.90.70.34.03.53.42.24.03.50.50.30.40.36.51.14.71.20.40.532.931.7
Eastern Europe14.23.93.40.60.90.42.20.628.419.435.137.10.70.92.41.823.91.747.79.86.38.195.731.2
EU and EEA20.75.110.81.31.40.44.41.420.913.873.068.62.31.02.31.822.71.745.410.312.319.6188.026.8
Oceania30.513.94.51.01.50.42.71.99.25.768.061.11.50.81.81.610.01.426.910.853.853.1160.630.4
Developed16.35.66.21.11.20.43.91.138.421.766.061.02.71.22.52.018.92.045.215.013.715.3162.226.0
Developing6.03.82.70.82.71.34.82.712.06.08.67.73.21.20.50.44.20.89.23.20.90.932.324.3
World8.54.23.60.92.31.04.62.318.410.122.521.83.11.21.00.87.81.118.06.44.04.766.724.8
 Corpus uteri (182)Ovary (183)Prostate (185)Testis (186)Bladder (188)Kidney (189)Brain, CNS (191, 192)Thyroid (193)NHL (200)Hodgkin's (201)Multiple Myeloma (203)Leukaemia (204–208)Other sitesAll sites but skin (140–208 but 173)
FFMMMFMFMFMFMFMFMFMFMFMF
SSA3.96.512.80.54.92.11.10.80.50.41.63.16.44.42.71.30.70.51.71.437.827.1101.9127.8
MENA6.15.54.31.516.23.91.61.12.01.41.65.04.23.42.11.30.30.52.71.615.818.483.3113.3
LAC12.78.422.63.78.12.53.52.33.52.52.27.85.13.52.61.60.70.73.52.924.128.3137.4218.3
North America42.223.0216.39.055.116.519.512.37.45.65.612.826.220.16.04.76.25.414.410.440.936.4612.1593.5
China, & OEA3.24.00.90.43.31.11.40.90.61.20.42.61.31.00.50.50.10.12.31.89.811.5100.998.1
Japan11.714.519.22.721.16.110.85.13.02.33.615.412.58.20.90.42.42.68.65.924.522.9447.9373.5
South-Eastern Asia5.78.14.01.23.40.91.30.81.20.72.08.63.82.91.00.50.40.32.52.115.314.478.1130.8
South-Central Asia3.36.43.31.13.71.01.10.71.71.11.23.23.01.81.90.70.40.42.41.719.519.082.4127.1
Eastern Europe29.820.922.36.422.24.813.710.05.23.92.37.27.04.96.34.12.22.68.76.922.524.1283.8330.5
EU and EEA35.620.982.112.352.211.717.810.56.24.83.27.318.714.76.53.95.24.812.39.738.335.5470.6499.2
Oceania23.115.890.88.127.59.311.99.15.34.52.78.019.415.33.83.04.84.212.08.339.934.2436.6477.4
Developed32.820.792.88.941.410.416.310.25.94.53.69.416.912.75.63.84.34.111.58.733.431.2455.5462.9
Developing4.86.05.51.14.91.51.51.01.41.21.24.23.12.21.50.80.40.32.41.917.817.995.9127.1
World12.29.926.73.013.83.85.13.42.52.11.85.66.55.02.51.61.31.34.63.721.621.4183.1216.1
Table V. Prevalent Cases Diagnosed 4–5 Years Earlier, as Proportion of the Population per 100,000, By Site, Sex and Geographic Area
 Mouth (140–145)Other pharynx (146, 148, 149)Nasopharynx (147)Oesophagus (150)Stomach (151)Colon/rectum (153, 154)Liver (155)Pancreas (157)Larynx (161)Lung (162)Melanoma (172)Breast (174)Cervix (180)
MFMFMFMFMFMFMFMFMFMFMFFF
SSA5.12.90.60.21.20.41.70.73.02.13.22.62.40.90.20.21.80.31.00.41.62.018.828.5
MENA2.51.70.50.21.50.90.50.42.11.54.24.20.50.30.20.24.90.73.50.80.80.830.08.5
LAC5.82.51.80.40.50.21.70.57.14.29.610.10.50.30.60.54.80.94.51.52.52.249.440.3
North America12.87.52.91.00.90.31.80.64.82.966.764.40.80.41.21.111.73.029.117.424.218.8212.417.6
China and OEA1.30.60.20.13.11.54.82.414.56.79.78.34.61.50.30.21.40.412.55.10.30.217.26.8
Japan5.62.91.30.20.80.23.70.7150.475.168.151.06.31.71.51.06.20.316.66.10.90.672.320.1
South-Eastern Asia2.92.11.20.45.42.60.60.32.41.67.16.42.10.70.20.12.30.54.41.70.30.328.721.4
South-Central Asia9.26.44.31.30.60.32.01.82.11.43.22.80.30.20.20.14.80.82.10.60.30.326.326.5
Eastern Europe10.13.22.00.40.70.31.00.217.912.424.426.50.30.41.20.917.91.523.95.84.76.776.125.6
EU and EEA15.84.16.60.90.90.32.00.813.79.353.051.11.20.51.10.818.11.423.55.910.417.6157.622.7
Oceania24.412.22.80.71.00.21.31.16.03.849.545.40.90.40.90.88.01.113.76.150.450.4135.325.9
Developed12.64.73.90.70.80.31.90.628.015.750.947.71.40.61.20.915.01.724.08.712.213.9139.322.0
Developing4.72.91.70.52.11.02.71.56.93.66.55.92.20.80.30.23.10.66.12.30.70.725.520.0
World6.63.42.20.61.80.82.51.312.06.817.317.02.00.70.50.46.00.910.44.03.54.255.720.5
 Corpus uteri (182)Ovary (183)Prostate (185)Testis (186)Bladder (188)Kidney (189)Brain, CNS (191, 192)Thyroid (193)NHL (200)Hodgkin's (201)Multiple Myeloma (203)Leukaemia (204–208)Other sitesAll sites but skin (140–208 but 173)
FFMMMFMFMFMFMFMFMFMFMFMF
SSA3.65.48.60.43.91.70.90.60.40.31.52.94.93.42.21.00.50.31.10.927.920.174.1100.4
MENA5.64.62.91.313.13.11.30.91.51.11.44.73.22.61.81.10.20.31.91.111.613.561.588.9
LAC11.56.915.83.36.32.02.71.82.81.92.07.33.82.62.11.30.50.52.42.017.520.698.5171.5
North America39.017.3192.29.151.515.516.710.55.44.05.512.622.216.85.64.43.73.211.38.232.528.8512.6508.0
China and OEA2.83.20.60.42.50.80.90.60.51.10.32.51.00.80.40.40.10.11.71.47.18.468.273.0
Japan11.111.617.12.619.85.79.24.42.31.73.514.810.66.90.80.41.51.66.94.819.418.2355.0313.4
South-Eastern Asia5.36.92.71.12.70.71.00.60.90.61.88.22.92.20.80.40.30.21.81.511.210.656.0104.0
South-Central Asia3.05.32.21.02.90.80.90.61.30.91.23.02.21.31.60.60.30.21.41.014.214.058.199.6
Eastern Europe26.414.414.65.916.43.610.58.13.42.62.16.75.33.75.33.71.31.66.04.916.017.5190.8253.2
EU and EEA31.614.858.211.942.19.514.08.54.03.42.86.814.711.65.83.63.22.98.77.227.825.9339.5398.6
Oceania20.311.563.57.922.27.69.57.43.53.32.47.515.212.13.32.73.02.58.56.130.125.3328.0389.8
Developed29.614.975.28.635.08.913.28.44.03.23.49.013.710.25.03.42.72.58.56.525.123.5346.2376.7
Developing4.34.93.81.03.91.21.10.81.11.01.14.02.41.71.20.70.20.21.71.313.013.167.398.6
World11.07.621.12.811.43.24.12.81.81.61.65.35.13.92.11.40.80.83.32.715.915.8134.9172.3
Table VI. Cumulative 1–5 Year Prevalence per 100,000 of the 5 Most Prevalent Cancer Sites in Men and Women by Region
MenWomen
Developed countriesDeveloping countriesDeveloped countriesDeveloping countries
Prostate222.8Stomach29.8Breast392.0Breast77.3
Colon and rectum160.0Lung23.0Colon and rectum148.2Cervix59.4
Lung114.6Colon and rectum21.1Endometrium80.4Colon and rectum18.9
Bladder100.6Mouth and pharynx14.6Cervix63.2Stomach14.9
Stomach92.4Prostate13.3Stomach52.5Ovary14.6

It may be worth reminding the reader that prevalence and incidence are not directly comparable because their units are not commensurable: incidence rates measure number of events per unit of population during a defined time period, which is normally a year. Prevalence measures the proportion of population affected at the same point in time, e.g., a day. Therefore, for example, an annual crude incidence rate of 381/100,000 in men, all sites, in developed countries means on average 1 new case diagnosed in a population of 100,000 population every day, while the corresponding 1-year prevalence estimate suggests that 309 of those 100,000 are being treated at the same time.

Since prevalence is intended to measure the absolute load on health services, standardisation by age would not be appropriate.

The highest prevalence of cancer is in North America (Table II) with 1.5% of the population above age 15 affected by the disease, diagnosed in the previous 5 years (0.5% in years 4–5 of follow-up and 0.6% in years 2–3, Tables IV,V). This corresponds to over 3.2 million individuals affected. Western Europe and Oceania show very similar percentages with 1.2% and 1.1% of the population affected, respectively (about 3.9 and 0.2 million cases, respectively). Japan and Eastern Europe form the next batch with 1.0% and 0.7%, followed by Latin America and the Caribbean (overall prevalence of 0.4%), and all remaining regions are around 0.2%. Cancer prevalence in developed countries is very similar in men and women, 1.1% of the sex-specific population, while in developing countries, the prevalence is some 25% greater in women than men, reflecting a preponderance of cancer sites with poor survival such as liver, oesophagus and stomach in males.

On average, in developed countries 0.3% (309 and 287 every 100,000 men and women, respectively) of the adult population is undergoing treatment for cancer (Table III), 0.5% is under clinical follow-up (Table IV) and 0.4% is in remission (Table V). The corresponding figures for developing countries once rounded are all equal to 0.1% actually all slightly below this value, in both males and females, a prevalence about one quarter that in high-income countries. This reflects both the lower overall cancer risk and poorer survival. The poor survival from cancer in developing countries is in part due to the high incidence of cancers characterised by high fatality anywhere in the world such as cancers of the liver, oesophagus and stomach, and in part due to poorer outcome, particularly for cancers requiring sophisticated or costly treatment.12

For most cancer sites, 2–3-year prevalence (Table IV) is greater than the prevalence in the first year after diagnosis (Table III), and that of cases at 4–5 years after diagnosis (Table V). The exceptions are some cancer sites characterised by extremely poor survival such as oesophagus, pancreas and liver, and the prevalence of cured cases is lower than that of cases receiving treatment (Tables III and V).

The cancer site by far most prevalent in both high- and low-income countries is cancer of the breast in females with 392 and 77 cumulative cases in 100,000 women diagnosed within 5 years and alive at the same time in developed and developing countries, respectively (Table VI). In developed countries they can be distinguished in 0.1% under treatment (Table III), a similar rate in remission (Table V) and 0.2% in the intermediate phase (Table IV). These represent 16%, 18% and 19% of the respective all site totals. In the high-risk region the second most prevalent site is cancer of the prostate with overall 220 cases per 100,000 population (age 15+ years) diagnosed within 5 years (1–5 year prevalence) and representing about 10% of all prevalent cases in men. In both sexes of developed countries colorectal cancer is second (about 150 cases every 100,000 population). It is followed in men by lung (115/100,000), bladder (101/100,000) and stomach (92/100,000). The highest values of 1- to 5-year prevalence in developing countries are observed in women for the breast (77/100,000) and cervix (59/100,000) followed by the stomach in men (30/100,000) (Table VI). Breast and gynaecological cancers account for 52% of prevalent cases in women in both high- and low-income countries (Tables VI,II). In developed countries they account for 35% of the cases followed by 7% for endometrium, 6% cervix and 4% ovary. The corresponding distribution in developing countries is breast 25%, cervix 19%, ovary 5% and endometrium 4%. Twenty percent of 5-year prevalent cases among men in developed countries are prostate cancers followed by cancers of the colon and rectum (14%) and lung (10%), overall 45% of all prevalent cases (Table VI). In men in developing countries, the stomach is the single most common site (13%) followed by lung (10%) and colorectal cancer (9%) (Table VI).

Crude cancer prevalence appears very low in developing countries compared with developed countries, even for cancer sites rare in the latter regions such as the cervix. Most of the difference is explained by different demographic patterns with high-income countries having long life expectancy in age groups when the risk of the disease is highest. Age standardisation reduced differences between developed and developing countries at every site, except hemopoietic malignancies and brain tumours, which are common in children and are not included in this set of estimates (data not shown).

The magnitude of disease incidence is the primary determinant of crude prevalence so that differences by region mainly reflect variation in risk. The maximum variation of cancer prevalence is observed in the presence of active mass screening, which inflates incidence and, at the same time, improves survival. This explains the over 100-fold difference in prevalence rates of melanoma between Australia and New Zealand (Oceania) and China, as well as the 70-fold difference in stomach cancer prevalence rate between Japan, where it is the most common cancer site and screening has been in place for decades, and Arabic countries (MENA) a very low-risk region. The largest difference is recorded for prostate cancer between North America (overall 5-year prevalence is 526 cases/100,000 men) and China (2 prevalent cases every 100,000 population).

DISCUSSION

Prevalence is a relevant measure in the context of health services planning and care delivery. It reflects the number of individuals in the community requiring at the same time a defined care procedure. In this context, it is informative if it reflects the number of subjects being in a given status rather than that of subjects ever diagnosed with the disease. For chronic diseases that require continuous treatment, such as diabetes and hypertension, the 2 measures in fact coincide; but this is not the case for cancer at several sites which, when treatment is successful, can be considered cured. Moreover, any index in the domain of planning ought to be up-to-date in order to be helpful; in fact projections in the near future would be even more informative in this context. Our simple approach allows useful figures to be obtained quite rapidly.

In different circumstances, prevalence of all ever-affected individuals is a more relevant index, as when estimating the proportion of the population that carry any kind of physical or psychological disabilities left by the disease,11 or to predict excess disease burden induced by cancer treatment among cancer survivors.

Cancer prevalence can be estimated directly from population-based cancer registry data by simply counting the number of registered cases still alive and present at a specified point in time.13–15 This approach requires not only the long-term registration of new cases but also the complete follow-up of the vital status of the population over many years. In fact, registries provide measures of the prevalence of survivors from cancer diagnosed within a limited time period. This depends on the time span covered by the registry; when this is very long, prevalence is likely to be very close to the total;13 otherwise it is referred to the concerned period.3, 16 Population surveys have been a practical alternative,13, 17 although this approach leads to the underestimation of the true values.19

Several estimation-based methods have been proposed. A mathematical model for chronic disease morbidity20 has been used to estimate breast cancer prevalence in the Lombardy cancer region of Italy21 and in the USA, combining modelling with incomplete cancer prevalence figures from cancer registries.22 Coldman et al.23 have reproduced cancer prevalence in British Columbia by a model they developed that incorporates a rate of cure. Colonna et al.24 estimated the prevalence of common cancer sites in France by modelling incidence and mortality.

In systematically estimating the prevalence of cancer for each country of the world, however, such advanced statistical modelling seemed inappropriately complex, given that the resulting prevalence estimates by country would often be based on incidence, mortality and survival figures, which were themselves derived from estimates. Bearing in mind the uses of this measure described above, and given the satisfactory results of the validation, we feel that this straightforward approach and our estimates are accurate enough for use in the planning of cancer health services. The prevalence of all survivors would drastically overestimate the cost of assistance since the large majority of long-term survivors do not require medical attention. In a recent survey conducted in the USA,17 the prevalence of adult cancer survivors was estimated by retrospective interviews. Despite the documented overall underestimation obtained by this method that increases with time since diagnosis,17, 19, 25 5-year prevalent cases represented only 35% of all the cases surveyed. The measure of partial prevalence would be improved by incorporating cases that develop recurrence irrespective of time since diagnosis. This would require measures of relapse rates by cancer site among all cases, which are not generally available. A step in this direction has been recently made by Colonna et al.26 who assessed the health status of a representative sample of 5-year prevalent cases of colorectal cancer, recorded by 5 cancer registries. They estimated that 19% of the 5-year prevalent cases experienced metastatic or local relapse after initial treatment and incorporated this measure in their estimate of the partial prevalence.

The aim of this article was to describe the method, its implied assumptions and data sources, and to validate the figures obtained. Estimates have been updated to prevalence in year 2000 and made available on the Internet (www-dep.iarc.fr/globocan/globocan.html).

APPENDIX

Source of Survival and Probability of Death by Area
Area and countriesSource of survivalAll-causes mortality (from Murrey and Lopez)11
1. Sub-Saharan Africa (SSA)Developing countriesSSA
2. Middle East and Northern Africa (MENA)Developing countriesMEC
3. Latin America and CaribbeansDeveloping countriesLAC
4. Northern AmericaSEEREME
5. ChinaChinaCHN
Mongolia, Democratic People's Republic of Korea, Republic of Korea, Hong KongOAI
6. JapanSEER except stomach cancer, which is from (8)EME
7. South Eastern AsiaDeveloping countriesOAI
8. South-Central AsiaDeveloping countries
Kazakhstan, KyrgyzstanFSE
Afghanistan, Islamic Republic of Iran, Pakistan, Tajikistan Turkmenistan and UzbekistanMEC
Bangladesh, Bhutan, Nepal, Sri LankaOAI
IndiaIND
9. Eastern EuropeObserved survival from EUROCARE 1 (7)  pooled estimate of Estonia, Poland, Slovakia  and Slovenia.FSE
10. EU and EEA, European Union and European Economic AreaObserved survival from EUROCARE 1 (7)
Denmark, Finland, Sweden, Italy, Spain, Austria, France, Germany, SwitzerlandCountry-specific
Norway, IcelandDenmark, Finland, Sweden and Iceland pooled
United Kingdom, IrelandEngland and Scotland
Albania, Bosnia and Herzegovina, Croatia, Republic of Macedonia, Malta, Slovenia, YugoslaviaEstonia, Poland, Slovakia and Slovenia pooled
Italy, SpainCountry-specific
Greece, PortugalSpain and Italy pooled
Austria, France, Germany, Netherlands, SwitzerlandCountry-specific
Belgium, LuxembourgNetherlands and Germany (observed survival,  age and sex-specific)
11. OceaniaEUROCARE 1 observed survival. Pool of all European registries except:
AustraliaMelanoma of the skin, relative survivalEME
MelanesiaDeveloping countriesOAI
Micronesia and PolynesiaDeveloping countriesOAI

Ancillary