Mass screening for colorectal cancer is not justified in most developing countries

Authors


Abstract

Most developed countries insist on the prevention of colorectal cancer (CRC) and offer screening to the population. Mass screening is proposed to both sexes in the population aged 50 years or more. Colonoscopy is then offered to persons having a positive faecal occult blood test. This recommendation should not be diffused over the world without considering the distinct dimensions of the risk and resources in developed and developing countries. A national screening policy is legitimate in developed countries like Japan, South Korea, and in North America and Europe. On the other hand, a mass screening policy for CRC is not recommended in most developing countries. The limited amount of resources attributed to health care for cancer should concern other indications in the control of common cancers, particularly in the cervix or liver. Indeed the risk of CRC is very low in most regions of Africa, and in some countries of South America and Asia. © 2009 UICC

Screening for colorectal cancer

The worldwide burden of colorectal cancer (CRC) was estimated at 1,227,000 incident cases in 2002 and 528,500 deaths.1 The figures justify, in complement to primary prevention,2 a policy of secondary prevention based on early detection of cancer and its premalignant precursors, assumed to be completely curable by a localized treatment, in countries where CRC accounts for a significant part of morbidity and mortality. Wilson and Jungner3 established 50 years ago a list of 10 criteria justifying the application of a generalized screening policy by the health authorities. The first criterion is that the condition should be an important health problem. In developed countries, CRC plays a significant role in the general morbidity and mortality of the population and the first criterion is fulfilled. In most developing countries the incidence of CRC is much lower and the application of a generalized screening policy by the health authorities may be questioned. The objective of this study is to review in distinct geographic areas the dimension of the risk and the available facilities for early diagnosis and prevention of CRC.

Abbreviations:

ASR, age-standardized rate; CRC, colorectal cancer; FOBT, fecal occult blood test.

The role of colonoscopy in early detection

Colonoscopy, taking into account the role of nonpolypoid precursors, is the gold standard procedure for early detection, and mini-invasive treatment, of those lesions.4–6 There is indirect evidence from nonrandomized trials that endoscopic polypectomy reduces the incidence of CRC: a 76% reduction has been shown, in comparison with various controls, in the prospective cohort of the National Polyp study in the United states7 and guidelines for surveillance colonoscopy have been established.8, 9 Drawbacks of endoscopy include a limited compliance, a small risk of severe complications, the high cost of the material and trained specialists. The prevalence of adenomatous polyps in adult populations of various developed countries is in the range of 30%; this means that the majority of precursors never progress to cancer, and points to a potential risk of overtreatment. Flexible sigmoidoscopy10, 11 only explores the distal colon and the rectum, but has better acceptance than colonoscopy and can be performed at a lesser cost.

Mass screening protocols

Mass screening of CRC is proposed by Health Authorities to a targeted population (all persons, both sexes, aged 50–74 years) and is based on a relatively simple filter test, the faecal occult blood test (FOBT) which can be offered by the National Health Authorities in campaigns repeated at biennial intervals; colonoscopy being only proposed to persons positive to the test. Despite limitations, the FOBT using Guaiac or Immunochemical method is estimated to reduce mortality from CRC by 15%.12, 13 During the Minnesota trial, a large proportion of the screenees for CRC14 were submitted to colonoscopy and a 20% reduction in the cumulative incidence of CRC was observed.

In complement to mass screening, individual nonorganized screening for CRC will occur at the initiative of person's conscious of his health, through primary colonoscopy. Nonorganized screening and screening organized by the Health Authorities address to different targets in the population and may have a complementary impact on incidence and mortality curves. The impact of organized screening can be evaluated during the follow-up of the intervention in terms of benefit, drawbacks and cost. In the United States, there is no organized screening, but a large diffusion is given, through the federal institution Medicare, to recommendations on the prevention of CRC and indications (and reimbursement) of screening procedures (FOBT and colonoscopy). The decrease of the mortality rate (32.5%) of CRC shown in the SEER Registries15 during the period 1975–2003 is attributed to progress in early detection and in treatment of advanced disease; during the same period the decrease of the incidence rate (19.4%) results from endoscopic treatment of precursors, rather than from changes in the risk profiles. Overall the declining figures of mortality and incidence confirm the benefit of this strategy.

Worldwide burden of CRC in developed and developing countries

The legitimacy of organized screening is based on a realistic evaluation of the burden of CRC estimated from the mortality attributable to this tumour, its proportion among the other causes of cancer deaths, and the importance of noncancer causes of death. The annual number of deaths from CRC in a population is expressed in a rate per 100,000 persons; this crude rate is relevant to the actual burden in the country. Comparisons of the risk between countries differing in the distribution of age classes, requires adjustment to a world population standard: the adjusted or age-standardized rate (ASR) of mortality for 100,000 persons. The annual number of incident cases, another parameter of the burden of the disease is expressed in crude and ASR of incidence for 100,000 persons.

The IARC Globocan database1 provides for each country, for the year 2002, an estimation of the crude and age-adjusted rates of incidence and mortality for all sites of cancer. Global estimations are also available for the comparison between developed countries of North America and Europe and developing countries of Asia, Africa and South America. As shown in Table I, the global incidence of CRC is higher in developed than in developing countries and the difference is more pronounced for the crude rate than for the ASR which takes into account the difference in ages classes.

Table I. Worldwide Estimation in 2005 of the Crude and ASR of incidence for CRC in Developed and Developing Countries for Men and Women
CRC incidence/100,000MenWomen
Crude rate/100,000ASR/100,000Crude rate/100,000ASR/100,000
  1. Source: Globocan 2002, IARC database.1

  2. ASR, age-standardized rate; CRC, colorectal cancer.

Developed countries60.840.050.926.6
Developing countries7.710.26.510.7

Cancer registries, based on the population of a region or a country provide precise data and confirm the higher crude and age-adjusted incidence of CRC in developed countries. As an example, for the period 1998–2002,16 when the crude incidence in the Miyagi registry (Japan) is compared with that observed in the Mumbai registry (India), the ratio is as high as 28:1 for men and 22:1 for women (Table II).

Table II. Crude and ASR of Incidence for CRC, for Men and Women, in Cancer Registries from Developed and Developing Countries (Period 1998–2002)
CountriesMenWomen
Crude rate/100,000ASR rate/100,000Crude rate/100,000ASR rate/100,000
  1. The countries are listed by decreasing figures for the crude rate. The age-standardized rate allows to compare the risk between countries without taking into account the differences in the age structure of the respective populations.

  2. Source: Cancer Incidence in V Continents, IARC Scientific publication 160.14

  3. ASR, age-standardized rate; CRC, colorectal cancer.

Japan (Miyagi)106.258.471.432.4
France69.548.149.825.0
Slovakia62.651.943.227.3
USA, SEER, 9 registries50.337.648.327.3
China (Shangai)40.727.040.522.9
Korea (Seoul)29.336.122.221.2
Brazil (Goiania)25.817.613.817.6
Thailand (Chiang Mai)10.29.69.78.7
Egypt (Gharbiah)4.36.33.44.4
India (Mumbai)3.85.63.24.2
Zimbabwe (Harare)2.87.22.16.2

Persons aged 55 years or more are the target of organized screening policies. Table III, based on the Globocan database1 shows, for the year 2002, the number of persons in this age range in some developed and developing countries. The table also gives the estimation of incident cases of CRC occurring during the same year in the corresponding population. The effectiveness of a screening protocol is linked to the number of persons aged 55 years or more to be explored to find one incident case of CRC. In most developed countries this number is not over 500. In developing countries the ratio is in the range of 1 to 2,000 to 6,000 persons. Therefore, despite the small proportion of persons deserving screening in developing countries, the effectiveness of organized screening is poor.

Table III. Estimations of Incident Cases of CRC in Persons Aged 55 Years or More (Both Sexes) in 2002 and in 2025 in Various Countries and Number of Persons in the Population in the Same Age Classes
Countries20022025
No. persons aged 55+ yr (in million)No. incident CRC in persons Aged 55+ yrRatio incident CRC to no. persons Aged 55+ yrNo. persons aged 55+ yr (in million)No. incident CRC in persons Aged 55+ yrRatio incident CRC to no. persons Aged 55+ yr
  1. Source: Globocan 2002, IARC database.1

  2. CRC, colorectal cancer.

USA61.0141,0001/430101.0245,0001/410
Slovakia1.12,4501/4501.63,5001/460
Japan39.083,0001/47052.0117,0001/440
France15.229,7001/51022.344,0001/510
China185.099,0001/1,870399.0206,0001/1,940
South Africa3.91,9001/2,0506.03,6001/1,670
Kenya1.84001/4,5003.37101/4,650
India114.020,0001/5,700227.042,0001/5,400
Nigeria8.31,3501/6,15016.01,5501/10,300

Perspectives in the worldwide evolution of the burden of CRC

Temporal variations in the incidence of cancer are linked to environmental factors and also to changing age distribution in the population, because the annual number of cases increases in relation to the increased proportion of older age classes.

The trend to increasing incidence of CRC is confirmed in the aging population of some European countries or in Japan where the proportion of persons aged 65 years or more, both sexes, is expected to increase from 18% in 2005 to 29% in 2025.1 An increasing incidence has also been confirmed in some metropolitan areas in Asia, such as Singapore or Hong Kong, justifying an analysis of the cost/effectiveness of developing screening in those areas.17

On the other hand, the increasing population of most developing countries shows no trend to aging for the next two decades: in Nigeria the proportion of persons aged 65 years, both sexes, which was 3.8% in 2005 is expected to be 4% in 2025.1 In Table III, the estimated ratio of incident CRC to the number of persons aged 55 years or more in the population for the year 2025 is almost the same as in 2002. In two Asiatic countries, i.e. China and India, the burden of CRC in 2025 increases with a doubling of incident cases; however, the number of persons to be screened for the detection of one incident CRC case is similar.

Mass screening for cancer in low-resource countries

The WHO-Global InfoBase-Impact of cancer database18 shows previsions on the proportion of cancer among the five major causes in mortality in each country for 2005 and 2030. The temporal trend in cancer mortality (all causes) varies between developed and developing countries, as shown in Table IV which includes the respective proportions of death by cancer or by communicable diseases in some countries: cancer represents 31.7% of death causes in Japan and only 4.4% in Nigeria; the respective proportions of deaths from communicable diseases in the two countries are 11.7 and 68.9%.

Table IV. Respective Impact of Mortality by Cancer All Causes (Both Sexes) and by Communicable Diseases in the General Mortality in Persons Under Age 70, in Various Countries, with Previsions for 2005 and 2030 (Based on 2002 Who Burden of Disease Estimates)
Countries% Global mortality in 2005% Global mortality in 2030
CancerCommunicable diseasesCancerCommunicable diseases
  1. Source: World Health Organization-Global InfoBase—Impact of cancer.15

Japan31.711.729.910.2
France28.36.127.94.3
Slovakia23.43.425.22.7
USA23.25.922.44.7
China20.29.323.65.8
India8.036.211.921.9
South Africa5.367.75.966.4
Kenya4.768.28.055.1
Nigeria4.468.96.857.8

In developing countries the question arises, however, on the eventual positive impact of a mass screening policy for localizations of frequent cancers, i.e. breast cancer, cancer of the cervix, hepatocarcinoma19–24 noteworthy: around 80% of the worldwide burden of hepatocarcinoma and cancer of the cervix occurs in developing countries. In the same countries CRC has a very small impact on the general mortality and accounts for only 30% of the worldwide burden despite their large population. The estimated number of incident cases of CRC occurring in 20051 in developed and developing countries is summarized in Table V, which also includes data for breast, liver and cervix sites.

Table V. Estimation of the Number of Incident Cases of Cancer Occurring in 2005 in Developed and in Developing Countries
Site of cancerWorld no. of casesDeveloped countries no. of casesDeveloping countries no. of cases and % Total
  1. Source: Globocan 2002, IARC database.1

Colorectal (both sexes)972,350665,700306,850 (31%)
Liver (both sexes)623,400110,400513,000 (82%)
Breast (women)1,150,200636,100514,100 (44%)
Cervix (women)492,80083,400409,400 (83%)

Liver cancer (mainly hepatocarcinoma) is the sixth most common cancer and the third commonest cause of death from cancer in the world. Most cases (82%) occur in developing countries, with the highest rates in sub-Saharan Africa, Eastern and South-Eastern Asia. According to the World Health Organization, approximately 350 million people are chronically infected with hepatitis virus B (HBV) and 170 million with hepatitis virus C (HCV) worldwide. The fraction of liver cancer attributable to an infectious factor (HBV and HCV) in developing countries was estimated in 2002 at 58.8% for HBV and 33.4% for HCV;23 other causes include Aflatoxin, whereas alcohol consumption is a significant causal factor in developed countries. Screening for liver cancer with Alpha-foetoprotein and ultrasound has been attempted in high risk populations, but the early diagnosis has not proved to be effective in reducing mortality.24 HBV vaccination is a more justified alternative option, but economic constraints limit its efficacy. As yet there is no vaccine against HCV.

Breast cancer incidence and mortality rates are increasing worldwide, including countries in Africa and Asia.19 Early detection linked to appropriate treatment is currently the most effective strategy to reduce breast cancer mortality. However, mammography screening is expensive and resource demanding, thus not feasible in developing countries. Increasing the awareness of the women on breast cancer symptoms and signs, and providing readily accessible and effective diagnostic and treatment services seem the best response for populations in low-income settings.

Cancer of the cervix is frequent in low-resources countries which bear 80% of the global burden of cervical cancer.20 In developed countries, screening for cervical cancer has shown to be effective in decreasing the incidence and the mortality from cervical cancer. Implementation of population-based screening programmes is also recommended in developing countries, using highly sensitive screening tests. The “screen-diagnose-treat” approach, to avoid recalls, is performed to target women in one setting on site and has also proven to be effective.21, 22 Besides cervical cancer screening, HPV vaccination takes into account the role of the infectious agent and offers an emerging option for cervical prevention.

In regions of developing countries where specific localizations of cancer show a very high incidence, corresponding programmes of screening deserve to be developed and associated to measures of primary prevention which play the major role. Such programmes have shown to be effective in reducing the incidence and mortality from cancer. This applies to screening for cervical cancer in India,25 for oral cancer in the Trivandrum area in India,26 and to screening for oesophageal cancer in Linxian, China.27, 28

Low-resource countries suffer, in addition to their limited resources, from multiple problems which impair the organization of a national policy of screening. Those factors are connected to the structure of the health care system, the identification of the population, the follow-up of persons selected as positive, the legal grounds for recalling of persons and for the distribution, analysis and reimbursement of tests distributed by the government. Concerning the specific development of screening for CRC, the intervention is based on the availability of colonoscopy procedures, which entail an expensive cost for the material and its maintenance and an appropriate number of well-trained specialists. This results in a frequent lack of infrastructure in endoscopic detection, treatment of curable neoplastic lesions and follow-up. The establishment of a national screening policy is, therefore, not justified in most low-resources countries. In this situation, early detection of CRC will occur only in the settings of individual screening in urbanized areas.

Conclusions

In most developed countries, CRC shares a significant proportion of the principal causes of deaths; the level of the risk and the aging population justifies the legitimacy of a screening policy; Japan and Korea, North America and most European countries enter this group. In most developing countries the risk is very low; communicable diseases have a considerable weight, and the age structure of the population should not change in the short term. The resources of the health authorities, often low in those countries, should be attributed to other priorities than screening for CRC. This applies to African countries, some countries in South America and Asia, and particularly to India. Finally, organized screening deserves to be developed from now, in complement to individual screening, in limited regions of large emerging countries where there is a shift to Western lifestyle and an aging population. This applies to the most industrialized sectors of China (e.g., Hong Kong), India (e.g., Mumbai) or Brazil (e.g., region of Rio Grande do Sul and the urban area of Sao Paulo).

Ancillary