© 2004 Foundation for Promotion of Cancer Research
Cancer Mortality Among Japanese Immigrants and their Descendants in the State of São Paulo, Brazil, 1999–2001
1 Epidemiology and Prevention Division, Research Center for Cancer Prevention and Screening, National Cancer Center, Tokyo, Japan, 2 Fundação Sistema Estadual de Analise de Dados Estatisticos (SEADE) and 3 Nikkei Disease Prevention Center, Santa Cruz Hospital Research Center, São Paulo, Brazil
For reprints and all correspondence: Motoki Iwasaki, Epidemiology and Prevention Division, Research Center for Cancer Prevention and Screening, National Cancer Center, 5-1-1 Tsukiji, Chuo-ku, Tokyo, 104-0045, Japan. E-mail: moiwasak{at}gan2.res.ncc.go.jp
Received July 21, 2004; accepted August 30, 2004
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Background: Only a few studies on the mortality of Japanese immigrants have been conducted in Brazil despite a large population of Japanese immigrants and their different environment and lifestyle from Japanese living in Japan.
Methods: To compare cancer mortality between Japanese in Japan and Japanese immigrants or Brazilians in the state of São Paulo, Brazil, we obtained official death certificates registered during 1999–2001. The standardized mortality ratio (SMR) or the standardized proportional mortality ratio (SPMR) of major cancer sites was calculated for the first generation of Japanese immigrants to Brazil (Japan-born), their Brazil-born Japanese descendants, and native Brazilians using mortality data of Japanese in Japan as a standard.
Results: The SMRs of stomach and colorectal cancer did not differ between the Japan-born residents of Brazil and the native Japanese, but significantly low SMRs were found among the native Brazilians. Compared with the native Japanese, we observed significantly lower SMRs for liver, gallbladder and lung cancer and significantly higher SMRs for prostate, cervical, and brain and nervous system cancer among both the Japan-born residents of Brazil and the Brazilians. Generally, the SPMR results were similar to those of the SMRs. Significantly high SPMRs for breast and uterine cancer were found for both the Japan- and Brazil-born residents of Brazil, although the Japan-born residents had increased SMRs, but not significantly so.
Conclusions: We confirmed the different cancer mortality pattern in the Japanese immigrants from that in Japanese in Japan, thus demonstrating the relative importance of the environment in the development of cancer.
Key Words: cancer • mortality • immigrants • Japan • Brazil
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Studies on migrants, which offer some clues as to the relative importance of genetic and environmental factors in the etiology of cancer, are useful especially when the differences in the mortality rates from or the incidence of cancer and lifestyle are large between the country of origin and the host country. A large number of epidemiological studies of Japanese immigrants and their descendants have been reported in the USA (1–6), especially in Hawaii and California. However, there are few studies on mortality of Japanese immigrants in Brazil despite representing the world's largest population of Japanese outside Japan (7,8). According to a special survey by the Centro de Estudos Nipo-Brasileiros (Center for Japan-Brazil Studies) in 1988, the estimated Japanese population in Brazil was 1 168 000, among whom, 828 000 lived in the state of São Paulo and 290 000 in the city of São Paulo. In addition, the geographical environment and lifestyle in Brazil is quite different from that in Japan or the USA, which suggested the interesting possibility of assessing the mortality rate from or incidence of cancer among Japanese immigrants in Brazil.
We previously showed higher mortality ratios for diabetes mellitus and ischemic heart disease and lower ratios for cerebrovascular disease among Japanese immigrants in the city of São Paulo compared with those among the Japanese in Japan (7). These findings are generally compatible with those among Japanese immigrants in the USA, though the magnitude is relatively small (2). However, the cancer mortality pattern of Japanese immigrants was somewhat different between Brazil and the USA (8). For example, no difference was observed for the mortality ratio of colon cancer between Japanese immigrants in the city of São Paulo and Japanese in Japan, while a significant increase was observed among Japanese immigrants in the USA (1–4). This evidence was based on mortality data from 1979 to 1981 and a relatively small number of deaths, particularly in the second generation. Furthermore, we were not sufficiently aware of the differences in lifestyles between Japanese-Brazilians and native Japanese before conducting cross-sectional studies on Japanese-Brazilians in the city of São Paulo and Japanese living in five prefectures across Japan (9–13). Therefore, to update the differences in the cancer mortality ratios between Japanese in Japan and Japanese immigrants or native Brazilians in the state of São Paulo and to provide new insights into the etiology of cancer occurrence, we analyzed the mortality data from 1999 to 2001 and discuss the observed differences in cancer mortality ratios among these different populations, considering the previously reported results of the cross-sectional studies.
|
METHODS |
|---|
|
TOPAbstractINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
POPULATION DATA
The denominator data for Japanese-Brazilians living in the state of São Paulo (JBs) were provided by the Brazilian census conducted by the Fundação Instituto Brasileiro de Geografia e Estatistica (Brazilian Geographic and Statistic Institute Foundation). The population of Japanese living in Japan (JJs) was acquired from the 2000 population census (14). The population of Brazilians living in the state of São Paulo, Brazil (BBs) was obtained by the Brazilian census in 2000 through the Fundação Instituto Brasileiro de Geografia e Estatistica.
MORTALITY DATA
We obtained official death certificates from the Fundação Sistema Estadual de Analise de Dados Estatisticos [State System Foundation for Statistical Data Analysis (SEADE)]. We chose all death certificates registered during 1999–2001, when the place of birth was specified as Japan or the fathers' and/or mothers' names were of Japanese origin, and their addresses were in the state of São Paulo. In this study, we defined the former as the first-generation of Japanese-Brazilian (JBs-I) and the latter as the Brazil-born Japanese descendants (JBs-II). The number of JJ deaths was obtained from national vital statistics in 2000 published by the Ministry of Health, Labour and Welfare in Japan (15). The number of BB deaths in 2000 was provided by the SEADE. According to the International Classification of Diseases 10th Revision (ICD10), deaths from cancer in all sites were coded as C01–C97 and codes for each cancer site are described in Tables 3 and 4.
|
|
STATISTICAL METHODS
In this study, the estimated JB population by sex and age (5-year interval) was available for only JBs-I (Table 1). We limited death cases to subjects >40 years of age, because of the low number of deaths seen in those JBs under 40 years of age, particularly in JBs-I (Table 2).
|
|
Standardized mortality ratios (SMRs) and 95% confidence intervals (CIs) of major cancer sites were determined to compare the cancer mortality in JJs with that of JBs or BBs. The expected number of deaths was calculated using sex-, age- (5 year interval) and site-specific mortality rates in 2000 for JJs and sex- and age-specific populations for JBs or BBs in the same year. A 95% CI was computed using Byar's approximation (16). Because the population of JBs-II was not available, we calculated SMRs only for JBs-I in this study. Standardized proportional mortality ratios (SPMRs) and 95% CIs of major cancer sites for 2000 were calculated for JBs-I and JBs-II using the sex-, age- (5-year interval) and site-specific proportions as a proportion of deaths from all cancer sites among JJs.
|
RESULTS |
|---|
|
TOPAbstractINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
The population of JBs-I in 2000 is shown in Table 1. The majority of the population was elderly, with subjects of 70 years of age and older accounting for almost half of the population.
During the 3 year interval from 1999 to 2001 in the state of São Paulo, out of 5826 deaths from all causes (3258 men and 2568 women), 1095 deaths from all cancer sites (715 men and 380 women) were found among JBs-I aged over 40 years, and out of 6242 deaths from all causes (3730 men and 2512 women), 1557 deaths from all cancer sites (837 men and 720 women) were identified among JBs-II aged over 40 years (Tables 2 and 3).
The SMRs and 95% CIs of the major cancer sites for JBs-I and BBs are shown in Table 3. Relative to JJs, the all-cause mortality ratios were significantly increased but the mortality ratios from all cancer sites were significantly decreased among JBs-I for both men and women. For cancer of the lip, oral cavity and pharynx, esophagus, larynx, and corpus uteri and leukemia, BBs had significantly higher mortality ratios than JJs except for leukemia among men, but no such significant increase was seen in JBs-I. In terms of esophageal cancer, male JBs-I had a significantly lower mortality ratio than male JJs but that for female JBs-I was not significantly lower. For stomach and colorectal cancer, BB mortality ratios were significantly lower than those for JJs, but those for JBs-I were similar to those for JJs. The same pattern was observed for pancreatic cancer and malignant lymphoma among men. Compared with JJs, both JBs-I and BBs had significantly decreased mortality ratios for liver, gallbladder and lung cancer regardless of sex. Significantly increased mortality ratios were observed for prostate, cervical, and brain and nervous system cancer among both JBs-I and BBs. For men, we found non-significant increased mortality ratios for skin and bladder cancer among JBs-I but a significant increase among BBs. For women, we found a non-significant increase in mortality ratios for skin, breast, uterine and bladder cancer among JBs-I but a significant increase among BBs. The mortality ratios for ovarian cancer and malignant lymphoma among women, and leukemia among men did not differ between JJs, JBs-I and BBs.
The SPMRs and 95% CIs of major cancer sites for JBs-I, JBs-II and BBs are shown in Table 4. Generally, results for the SPMRs were similar to the SMRs, although somewhat inconsistent results were seen in stomach and colorectal cancer. Among men, the SPMRs of lip, oral cavity and pharynx and larynx cancer were significantly elevated for JBs-II and BBs but not for JBs-I. Compared with JJs, the SPMRs of stomach and colorectal cancer were significantly lower for BBs. However, the stomach cancer SPMR was significantly increased for male JBs-I and significantly decreased for female JBs-II. For colorectal cancer, a significantly high SPMR was observed for male JBs-II, but neither JBs-I nor female JBs-II demonstrated any significant difference in the SPMRs as compared with JJs. For liver, gallbladder and lung cancer, lower SPMRs were observed for JBs-I, JBs-II and BBs than for JJs. In particular, significantly decreased liver cancer SPMRs among men were found for JBs-I (SPMR = 58), JBs-II (SPMR = 43) and BBs (SPMR = 29). The SPMRs of skin, prostate, and brain and nervous system cancer were significantly elevated among male JBs-I, JBs-II and BB subjects. The SPMRs of breast, uterine, cervical, and brain and nervous system cancer were significantly increased among female JBs-I, JBs-II and BB subjects.
|
DISCUSSION |
|---|
|
TOPAbstractINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
We confirmed the different cancer mortality pattern in JBs from that in JJs, thus demonstrating the relative importance of the environment in the development of cancer. In this study, we observed significantly different mortality ratios of liver, gallbladder, lung, prostate, and brain and nervous system cancer among JBs-I from those among JJs regardless of gender. Significant differences between two groups were found in mortality ratios of esophagus for men and cervix uteri for women. Meanwhile, JBs-I had relatively similar mortality ratios of stomach and colorectal cancer to JJs. Basically, the present findings showed much the same overall pattern as the previously published study based on mortality data from 1979 to 1981 (Figs 1 and 2) (8). However, the present study newly revealed that mortality ratios among JBs-I differed significantly from those among JJs in the following cancer sites: lung for men; gallbladder and cervix uteri for women; and brain and nervous system cancer for both men and women, while these ratios were statistically non-significant differences in the previous report. The SMR of breast cancer changed from 70 in 1980 to 139 in 2000, although this was not statistically significant (Figs 1 and 2). Our updated results strongly support the previous findings and consolidate the evidence that several sites of cancer in the Japanese could be prevented by modifications in environmental factors.
|
|
Differences in lifestyle and dietary habits constitute a major component in the environmental changes experienced by migrant populations, which have been associated with the development of cancer. According to the cross-sectional studies of randomly selected Japanese-Brazilians in the city of São Paulo and Japanese living in five prefectures across Japan (9–11), the Japanese-Brazilians consumed fewer cigarettes and less alcohol and had a higher body mass index than the native Japanese. For dietary habits, the frequency of consuming bread, green vegetables, beef, chicken, dairy products, cheese, other beans and coffee was much higher among Japanese-Brazilians than among Japanese, while pickled vegetables, pork, fish, miso soup, soybean and its products, and green tea were consumed less. In short, the dietary habits of Japanese-Brazilians shifted toward the pattern seen in Western countries (9,10). These differences might explain the observed cancer mortality pattern, such as the low mortality ratios of esophageal and lung cancer and high mortality ratios of breast and prostate cancer.
No difference was observed in stomach cancer mortality ratios between JBs-I and JJs, while the mortality ratio of this cancer among Japanese immigrants in the USA was significantly lower than that among native Japanese (1–3). This discrepancy could be explained by the degree of Westernization in dietary habits. Although the dietary habits of Japanese-Brazilians shifted toward the pattern seen in Western countries (9,10), they still consumed more traditional and salted Japanese foods as compared with Japanese-Americans in Hawaii. For example, 15% of male Japanese-Brazilians aged 40–49 years consumed miso soup almost every day, and 4% of them consumed pickled vegetables (10), while only 2% of male Japanese-Americans aged 45–69 years in Hawaii consumed miso soup almost every day in the 1960s (17). Furthermore, salt excretion levels in 24 h urine among male Japan-born residents aged 40–59 years in São Paulo were 14 g/day for 21 volunteers originally from Iwate, Akita and Nagasaki Prefecture, and 8.7 g/day for 12 volunteers from Okinawa Prefecture, respectively (unpublished data). These levels were almost comparable with the Japanese in Japan.
Meanwhile, diets high in vegetables possibly reduce the risk of stomach cancer (18). Although Japanese-Brazilians frequently consumed vegetables more than Japanese in Japan (10,11), the mortality ratio of stomach cancer among JBs-I was not lower than that among JJs. That is partly because even a small amount of vegetables might be enough to protect against stomach cancer and any additional vegetable intake might not offer any greater degree of reduction in the risk of stomach cancer, as one of the prospective studies in Japan showed that subjects who consumed yellow or white vegetables on >1 day per week had a significantly decreased risk of stomach cancer as compared with those who consumed them <1 day per week, but further reduction in risk was not observed with the increased frequency of their consumption (19).
As other evidence in support of the lack of any difference in the mortality ratios of stomach cancer between JBs and JJs, we previously observed a slightly higher prevalence of atrophic gastritis among Japanese residents in the city of São Paulo than in Japanese located in the area in Japan where atrophic gastritis was most prevalent (12,13).
It is a challenge to interpret the mortality pattern of colorectal cancer for JBs, because two conflicting factors might be associated with the mortality pattern of colorectal cancer as follows. First, an elevated mortality ratio of colorectal cancer in JBs-I was expected, as an increase in the colorectal cancer mortality and incidence rates was observed among Japanese immigrants in the USA (1–4). Compared with Japanese in Japan, Japanese-Brazilians consumed more red meat and fat and had a higher body mass index (10,11), which could have been possible risk factors for colorectal cancer (20). Secondly, a low mortality ratio of colorectal cancer in JBs-I was expected, because the colorectal cancer mortality ratio among BBs was significantly lower than for JJs. Thus, the Brazilian environment might well offer potentially protective factors because it is a low-risk country for colorectal cancer. In addition, frequent intake of vegetables among the Japanese-Brazilians might be related to lowering the mortality ratio of colorectal cancer regardless of high red meat and fat consumption (10,11). In the present study, no difference in colorectal cancer mortality ratios was found between JBs-I and JJs, which might be due to the effect of these conflicting factors.
Persistent infection with hepatitis B virus or hepatitis C virus is one of the major etiological factors for liver cancer (21). For hepatitis B surface antigen, carrier rates were similar between Japanese immigrants in the city of São Paulo (8) and Japanese in Japan (22). However, we do not know to what extent hepatitis C infection accounts for the difference in mortality ratios of liver cancer between JBs and JJs because of no available data.
Timing of immigration is of great etiological importance. If the effect of factors in early life is large on the development of cancer, the mortality ratios for residents who emigrated in later life must be similar to those in the country of origin. On the other hand, if the effect of factors in later life is large, the mortality ratios for residents who emigrated in later life must approximate to those in the host country. In this study, the SMR of prostate cancer was significantly higher in JBs-I than in JJs and this cancer was significantly more frequent for both JBs-I and JBs-II. Although JBs-I include those who emigrated to Brazil either early or later in life, we were not able to examine the SMR of prostate cancer for only those who emigrated in later life because of the lack of any availability of data for the age of the subjects at immigration. However, a previous study of Japanese immigrants in Los Angeles County showed that the incidence rates of prostate cancer among Japanese immigrants were similar to those of US-born residents, regardless of the age at immigration (6). Furthermore, Yatani et al. reported no significant difference in the rate of latent prostate cancer between Japanese immigrants in Hawaii and Japanese in Japan (23). These previous studies might support the interpretation of our result that environmental factors in later life are more important for the development of prostate cancer than those in early life.
For breast cancer, there was an increasing tendency seen in the SMR for JBs-I, but it was not statistically significant and an excess risk of breast cancer was not striking for JBs-I relative to that of prostate cancer. One of the possible reasons for the equivocal result is that those who emigrated in early life among JBs-I might account for the elevated mortality ratio among JBs-I, because breast cancer was significantly more frequent for both JBs-I and JBs-II than for JJs. However, a previous study of Japanese immigrants in Los Angeles County clearly showed that the incidence rates of breast cancer among US-born residents and residents who immigrated in early life were almost identical, but the rate among residents who immigrated in later life was intermediate between the rates of US-born residents and Japanese in Japan (6). Considering the previous study, our result suggests that factors in early life make a more substantial contribution to the development of breast cancer.
Significantly raised mortality ratios of brain and nervous system cancer in JBs indicate the important role of environmental factors in the development of this cancer, but what kinds of factors contributed to bring about the change in its occurrence remains unclear. Otherwise, the present findings might be explained by an artifact, such as a difference in the coding of death certificates, because one study showed that both Japanese immigrants in the USA and Japanese in Japan had similar incidence rates (5).
Comparisons of mortality ratios between Japan and Brazil must be interpreted cautiously because of several methodological considerations as follows. Firstly, although one study reported that information on cause of death was quite accurate, with a low proportion of deaths certified as being due to senility or other ill-defined conditions (24), differences in the quality of death certificates due to differences in medical practice and coding of death certificates between Japan and Brazil might exist and lead to misclassification of cancer sites. Secondly, SPMRs must obviously be interpreted with more caution than SMRs, because a proportionate excess can reflect either an excess in the absolute rate for that disease, or a deficit in the absolute rates for some of the other causes. Thirdly, we must keep in mind that Japanese immigrants to Brazil were self-selected and may not be representative of the population of Japan. Fourthly, mortality ratios have been derived for many cancer sites and, as a consequence, some of the findings deemed to be statistically significant may have occurred merely by chance. Even if the findings might not be due to chance, some of them seemed to be unreliable, because of a small number of deaths.
Allowing for these methodological issues, we confirmed the different cancer mortality pattern in JBs from that in JJs and highlighted the relative importance of the environment in the development of cancer.
|
Acknowledgments |
|---|
This work was supported by Grants-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology, for Cancer Research, for the second term comprehensive 10-year strategy for cancer control, and for the Risk Analysis Research on Food and Pharmaceuticals from the Ministry of Health, Labour, and Welfare of Japan.
|
References |
|---|
|
TOPAbstractINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
1 Smith RL. Recorded and expected mortality among the Japanese of the United States and Hawaii, with special reference to cancer. J Natl Cancer Inst 1956;17:459–73.[Web of Science][Medline]
2 Haenszel W, Kurihara M. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 1968;40:43–68.[Web of Science][Medline]
3 Locke FB, King H. Cancer mortality risk among Japanese in the United States. J Natl Cancer Inst 1980;65:1149–56.[Web of Science][Medline]
4 Shimizu H, Mack TM, Ross RK, Henderson BE. Cancer of the gastrointestinal tract among Japanese and white immigrants in Los Angeles County. J Natl Cancer Inst 1987;78:223–8.[Web of Science][Medline]
5 Tominaga S. Cancer incidence in Japanese in Japan, Hawaii, and western United States. Natl Cancer Inst Monogr 1985;69:83–92.
6 Shimizu H, Ross RK, Bernstein L, Yatani R, Henderson BE, Mack TM. Cancers of the prostate and breast among Japanese and white immigrants in Los Angeles County. Br J Cancer 1991;63:963–6.[Web of Science][Medline]
7 Tsugane S, Gotlieb SL, Laurenti R, Souza JM, Watanabe S. Mortality and cause of death among first-generation Japanese in São Paulo, Brazil. Int J Epidemiol 1989;18:647–51.
8 Tsugane S, Gotlieb SL, Laurenti R, de Souza JM, Watanabe S. Cancer mortality among Japanese residents of the city of São Paulo, Brazil. Int J Cancer 1990;45:436–9.[Web of Science][Medline]
9 Tsugane S, Hamada GS, Souza JM, Gotlieb SLD, Takashima Y, Todoriki H, et al. Lifestyle and health related factors among randomly selected Japanese residents in the city of São Paulo, Brazil, and their comparisons with Japanese in Japan. J Epidemiol 1994;4:37–46.
10 Tsugane S, Hamada GS, Karita K, Tsubono Y, Laurenti R. Cancer patterns and lifestyle among Japanese immigrants and their descendants in the city of São Paulo, Brazil. Gann Monogr Cancer Res 1996;44:43–50.
11 Tsugane S. Cancer patterns and lifestyle among Japanese Brazilian in São Paulo. J Epidemiol 1996;6:S169–73.[Medline]
12 Tsugane S, Kabuto M, Imai H, Gey F, Tei Y, Hanaoka T, et al. Helicobacter pylori, dietary factors, and atrophic gastritis in five Japanese populations with different gastric cancer mortality. Cancer Causes Control 1993;4:297–305.[CrossRef][Web of Science][Medline]
13 Tsugane S, Fahey MT, Hamada GS, Kabuto M, Miyakawa VY. Helicobacter pylori infection and atrophic gastritis in middle-aged Japanese residents of São Paulo and Lima. Int J Epidemiol 1999;28:577–82.
14 Statistics Bureau, Ministry of Public Management, Home Affairs, Posts and Telecommunications. 2000 Population Census of Japan, Volume 2–1. Tokyo: Statistics Bureau, Ministry of Public Management, Home Affairs, Posts and Telecommunications 2001 (in Japanese).
15 Statistics and Information Department Minister's Secretariat Ministry of Health, Labor and Welfare. Vital Statistics of Japan. Tokyo: Health and Welfare Statistics Association 2000 (in Japanese).
16 Breslow N, Day NE. Statistical methods in cancer research, vol. II. The design and analysis of cohort studies. Lyon: IARC Scientific Publication 82 1987;69–71.
17 Severson RK, Nomura AM, Grove JS, Stemmermann GN. A prospective study of demographics, diet, and prostate cancer among men of Japanese ancestry in Hawaii. Cancer Res 1989;49:1857–60.
18 International Agency for Research on Cancer. IARC Handbooks of Cancer Prevention, Volume 8 Fruit and Vegetables. Lyon: IARC Press 2003.
19 Kobayashi M, Tsubono Y, Sasazuki S, Sasaki S, Tsugane S. Vegetables, fruit and risk of gastric cancer in Japan: a 10-year follow-up of the JPHC Study Cohort I. Int J Cancer 2002;102:39–44.[CrossRef][Web of Science][Medline]
20 World Cancer Research Fund and American Institute for Cancer Research. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington DC: American Institute for Cancer Research 1997;216–251.
21 IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 59: Hepatitis Viruses. Lyon: IARC 1994.
22 Yuasa T, Tagaya I, Sekine T, Nishioka K. Hepatitis B antigen and antibody prevalence of Japanese sera selected from the 1972 year's collection at National Serum Bank, National Institute of Health of Japan. Jpn J Med Sci Biol 1981;34:181–90.[Medline]
23 Yatani R, Chigusa I, Akazaki K, Stemmermann GN, Welsh RA, Correa P. Geographic pathology of latent prostatic carcinoma. Int J Cancer 1982;29:611–6.[Web of Science][Medline]
24 Bouchardy C, Mirra AP. Cancer in Italian migrant populations. São Paulo, Brazil. IARC Sci Publ 1993;123:103–16.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||