Inbreeding and cancer incidence in human isolates
Rudan, Igor
Abstract This study investigates the incidence of cancer in isolate populations. Thorough anthropological research over the past 3 decades has established islimd populations in Middle Dalmatia, Croatia, as outstanding examples of genetic isolates. The number of cancer cases on 5 islands (Brac, Hvar, Korcula, Vis, and Lastovo) over a 20-year period (1971-1990) has been extracted from the data of the Croatian Cancer Registry. The population of coastal Dalmatia, characterized by similar environmental factors but a different population genetic structure, was used as a control population. The leading hypothesis was that, if there were genes or gene complexes (especially with recessive inheritance) responsible for genetic susceptibility to certain types of cancer, then the incidence of those cancer types should be greater in reproductively isolated island populations than in a control population because of increased manifestation of such genes or gene complexes caused by inbreeding. Furthermore, the cancer incidence should increase along with greater reproductive isolation (i.e., greater geographic distance of the islands from the mainland). After adjusting the data for sex and age, I confirmed the hypothesis: Island populations have greater total cancer incidence than the control population for both sexes. The excess incidence on the islands shows an almost linear correlation with geographic distance from the mainland. The cancer sites primarily responsible for the excess incidence are bladder cancer in males, and breast, ovarian, brain, and large bowel cancer in females, predominantly in the younger age groups.
KEY WORDS: CANCER INCIDENCE, GENETIC EPIDEMIOLOGY, DEMOGRAPHY, ISOLATE POPULATIONS, CROATIA
A number of extrinsic and intrinsic factors contribute to the development of cancer, a disease still far from satisfactorily controlled in most communities worldwide. Current geographic and ethnic differences in cancer incidence are considered a result of primarily environmental influences (e.g., nutrition, pollution, smoking, chronic irritation, chemical carcinogens, irradiation). However, for some specific cancer sites inherited (genetic) susceptibility cannot be neglected, although it is probably less important than extrinsic factors. Genetic susceptibility sometimes characterizes whole nations; some examples are nasopharyngeal carcinoma (common in Chinese populations), bladder cancer (common in Native Americans), and testicular cancer and Ewing’s sarcoma (extremely uncommon among blacks in Africa and the United States) (Weiss et al. 1984, 1986; Fraumeni et al. 1989). In other types of cancer there is striking evidence of familial aggregation. Among these types the most frequently cited are colorectal cancer (Weiss and Chakraborty 1984; Bishop and Thomas 1990; Scapoli et al. 1994; Houlston et al. 1995; Hall et al. 1996), breast and ovarian cancer (Andrieu et al. 1991; Houlston et al. 1991; Iselius et al. 1991; Bishop 1992, 1993, 1994; Thompson 1994; Kerber and Slattery 1995; Struewing et al. 1995), medullary carcinoma of the thyroid gland (Bergholm 1989; Bellis et al. 1995), and melanoma (Goldstein, Dracopoli et al. 1994; Goldstein, Fraser et al. 1994; Goldstein and Tucker 1995).
Studies of cancer-prone families show that inheritance of cancer susceptibility does not have a unique pattern. There is evidence of autosomal dominant inheritance (e.g., retinoblastoma, familial colorectal and breast cancer, nevoid baseocellular carcinoma), autosomal recessive inheritance (e.g., ataxia-telangiectasia, susceptibility to leukemias and lymphomas, gastric cancer), and X-chromosome-linked recessive inheritance (e.g., Wiskott-Aldrich syndrome) (Anderson 1982; German 1985; Mulvihill 1985; Fraumeni et al. 1989).
Here, I test the hypothesis that, if there are genes or gene complexes (especially with recessive inheritance) responsible for genetic susceptibility to certain types of cancer, then the incidence of those cancer types should be greater in reproductively (genetically) isolated populations because of increased manifestation of such genes or gene complexes caused by inbreeding.
Materials and Methods
Island Populations. Thorough anthropological research over the past 3 decades has established island populations in Middle Dalmatia, Croatia, as outstanding examples of genetic isolates (Rudan et al. 1987; Rudan, Bennett et al. 1990; Rudan, Finka et al. 1990). As some of the last persisting isolates among contemporary European human groups, the 5 islands chosen for this study (Brac, Hvar, Korcula, Vis, and Lastovo) have features that make them suitable for studies of genetic epidemiology. Some of those characteristics are specific ethnohistory, known migrations that have occurred in the last millennium, their continuing mutual reproductive isolation, and welldocumented effects of various extrinsic events that through generations directly influenced the populations’ biological and genetic formation.
The results of anthropological investigations carried out on these 5 islands and the demographic and ethnohistorical characteristics of the islands have been amply published in the world literature (Rudan et al. 1987, 1992, 1994; Rudan, Bennett et al. 1990; Rudan, Finka et al. 1990; Sujoldzic et al. 1989; Simic and Rudan 1990; Roberts et al. 1992; Janicijevic et al. 1994; Roguljic et al. 1997). To summarize a few important characteristics relevant to the basic genetic structure of these populations, I note the following facts. The mentioned studies have indicated that less than 10% of the parents of current inhabitants were born outside the islands. The level of endogamy (product of the percentage of mothers and fathers whose offspring still inhabit their birthplace) amounts to more than 75% (Rudan et al. 1987; Rudan, Bennett et al. 1990; Rudan, Finka et al. 1990). The coefficient of inbreeding determined from isonymy data is as high as 40% in some villages, and the coefficient of kinship reaches 10%. Finally, because of a positive attitude toward isonymous marriages, which results from some historical and cultural traditions, such marriages often account for more than 35% of all marriages (Roguljic et al. 1997). All these findings indicate a high level of inbreeding in populations of island isolates from Middle Dalmatia, Croatia.
Control Population. To avoid as many potential biases as possible and to provide some firm and indisputable cancer incidence rates for all cancer sites, the control population had to be large. Also, the control population should have been exposed to similar environmental influences as those on the islands, especially regarding nutrition habits (intake of fish, wine, and olive oil) and pollution (extremely low). Finally, the genetic structure of the control population had to be different from that in the island isolates. In other words, instead of reproductive isolation, the control population had to be outbred and characterized by excessive immigration and emigration over the past several decades.
The population of coastal Dalmatia meets these criteria and was selected as the control population. Coastal Dalmatia is composed of 19 districts with 800,000 inhabitants, which is a number large enough to allow accurate cancer incidence rates. The ethnohistorical and demographic events in the area resemble those on the islands, and nutritional habits are also similar. Although pollution may be somewhat greater than on the islands because of traffic and industry, this should not present a major obstacle in comparison of cancer rates with those on the islands. Finally, there is no evidence of inbreeding in the control population of coastal Dalmatia. On the contrary, this region has experienced both considerable emigration to northern Croatia and to Western countries during the 20th century and immigration as a result of summer tourism that developed in the region after World War II. Figure 1 shows the geographic location of the island and control populations.
Source of Data and Period of Study. Incidence data on cancer cases in the island and control populations were obtained from the Cancer Registry of Croatia, which is part of the Croatian Institute of Public Health. This registry was founded in 1959, and it began registering all cancer cases in Croatia in 1962. The electronic database was introduced in 1968, and from that point the data are easily accessed. Since 1994 the Registry has been a member of the International Association of Registries of Cancer, the European Network of Cancer Registries (ENCR), and the database on Cancer Incidence and Mortality (EUROCIM), and it is therefore considered a viable cancer registry. Indicators of the quality of data in the Registry are 55% of registered new cases are histologically confirmed, 22% of registered cases are notified by death certificate only, and the mortality/incidence ratio amounts to 0.77 (Croatian Institute for Public Health 1997).
The cancer sites specifically analyzed in this study included cancers of the stomach, bowel, rectum, pancreas, larynx, lungs, prostate, bladder, kidney, and brain in males; these sites represent the 10 most frequent sites of occurrence in Croatia (not in exact order). In females the cancer sites analyzed were those of the stomach, bowel, rectum, pancreas, lungs, breast, cervix utery, corpus utery, ovary, and brain.
Defining the study period represented a considerable problem. Because of the relatively small number of inhabitants on the islands, it was desirable for the study period to be as long as possible to record more cases and to obtain reliable incidence estimates for all major cancer sites. Another preference was to have a census of the population in the middle of the study period for purposes of later standardization according to sex and age. Finally, the study period had to end before 1991, when the war in this area began to bring waves of refugees to the region, including the islands. For these reasons the period 1971-1990 was chosen, and all cancer incidences were based on the census of 1981.
Design of the Study, Statistics, and Hypotheses. The design of this study is relatively simple. The idea was to determine cancer incidences in a large control population of coastal Dalmatia using the data of the Croatian Cancer Registry. All incident cancer cases in this region from 1971 to 1990 were used to calculate the average annual cancer incidence according to age groups and specific cancer site. The same procedure was also performed for each of the 5 island populations. After that, the obtained incidences were adjusted for sex and age through direct standardization to the Standard World Population, as suggested by Parkin et al. (1997). After standardization, the following four hypotheses were tested.
The first hypothesis was that the total cancer incidence should be greater in island populations than in the control population as a result of the combined effect of recessive inheritance of susceptibility to cancer and high inbreeding in island populations. Second, the cancer incidence rates in island populations should especially differ from control rates in younger age groups, because “inherited” cancer tends to develop earlier in life than “acquired” cancer. These two hypotheses were tested using Fisher’s chi-square test.
The third hypothesis was that the total cancer incidence should vary among the islands, increasing along with reproductive isolation, that is, along with greater sea distance from the mainland. The sea distance from the mainland was determined as the shortest likely distance between the main harbor center on the mainland (which is Split; see Figure 1) and the main harbor on the particular island. The deviation of the total cancer incidence on each island from the control population was then correlated with geographic distance from the mainland using a simple linear regression. The last hypothesis was that the cancer sites mostly responsible for any cancer incidence excess in the island populations should include cancers with a well-established genetic predisposition, such as colorectal, breast, and ovarian cancer. This hypothesis was also tested using Fisher’s chi-square test.
Results
Table 1 shows the average annual cancer incidence rates in the control population and in the 5 island populations, standardized per 100,000 Standard World Population.
The incidence in the control males was 189.6 and was greater on all islands except Lastovo, ranging from 192.5 in Brac to 210.3 in Vis. On the most distant island of Lastovo the incidence of only 99.5 was significantly lower than the control population value (p
In females the incidence for the control population, 134.4, was exceeded in all 5 island populations, ranging from 151.0 on Brac to 196.0 on Lastovo. The excess reached statistical significance on Hvar (p
The analysis of age-specific incidence for island population males showed that cancer incidence in younger age groups (
Figure 2 shows the deviations from expected (control) incidence in males for each island population. These deviations were then compared with the shortest likely sea distance from the mainland harbor. The negative deviation from expected incidence recorded on Lastovo was also considered positive for purposes of linear regression. The regression analysis revealed a very significant correlation (r = 0.791) between deviation in incidence and geographic distance from the mainland, which is an indirect measure of reproductive isolation. In addition, Figure 3 shows the deviations from expected (control) incidence in female island populations. The correlation coefficient between incidence excess and geographic distance from mainland in females is extremely high (r = 0.991).
Site-specific male cancer incidence showed no clear patterns (Table 2). The only significant differences from the control population occurred in stomach cancer incidence on Vis (25.5 vs. 13.3, p
Site-specific female cancer incidence showed some differences. The incidence of breast cancer in the control population (26.5) was lower than in all island populations, namely, on Brac (30.4), Hvar (35.7, p
Discussion
The existence of reproductively isolated populations with known ethnohistory, abundance of data regarding population structure, and coverage by a viable cancer registry for over 30 years is unusual and prompted this study of cancer patterns using the postulates of genetic epidemiology and genetic demography.
The first hypothesis stated that the total cancer incidence should be greater in island populations than in the control population as a result of the combined effect of recessive inheritance of susceptibility to cancer and high inbreeding in island populations (Hauck and Martin 1984; Seminara and Obrams 1994; Malkin and Knoppers 1996). At this point, there is a need to clarify why inbreeding necessarily raises the incidence of cancer. Most cancer-susceptibility genes are anti-oncogenes, which are recessive at the cell level. Therefore it is generally thought that it is impossible to survive fetal development with two inactive copies. Theoretically, for such genes a recessive effect on adult-onset cancers could not be expected in an inbred population because the affected offspring would abort or never even be detected. However, one should allow the possibility of other kinds of genes, tumorpromoting ones, which are also well established. Furthermore, many genetic disorders that are characteristic of inbred isolates often are connected with increased risk of various cancers, suggesting another possible mechanism. Apart from that, there is enough experimental evidence in animals to suggest that inbreeding indeed raises the risk of various cancers (Crawford et al. 1987; Kawaguchi et al. 1997; Hennings et al. 1997). However, there were no studies investigating cancer patterns in inbred human populations before this study.
The confirmation of the first hypothesis still raises several questions and biases. Some concerns include the fact that inbreeding raises recessive homozygosity of fortuitous genes and that one cannot predict that a particular trait will have increased frequency in inbred populations. Such a finding might reflect some other influence, such as more complete cancer surveillance in smaller populations or a lesser degree of cultural variation. In addition, the inbreeding effect is largely due to chance founder effects that bring some recessive alleles to high frequency in the isolated, inbred populations. Therefore, just by virtue of having more variation, the outbred population may have lower cancer rates than the isolate, even if that has nothing to do with inbreeding. However, if the effect of inbreeding is real, it would suggest that a kind of general, probably large number of genes contribute to the variation in cancer and, when homozygous, can raise the risk of cancer somewhat.
In any case, the first hypothesis was indeed confirmed, the effect being more pronounced in females (Table 1; Figures 2 and 3). The only exception was a low total cancer incidence in males on Lastovo (Table 1; Figure 1). This low incidence can be explained in 2 ways: (1) It may be a statistical artifact resulting from the small population size or (2) it may be a result of a founder effect, if the founding population of Lastovo by chance did not carry the same frequency of genes for susceptibility to cancer as did the other founding populations. With the smaller population size the increased effect of random mating may have led to loss of cancer genes, or inbreeding may have caused increased homozygosity of the noncancer alleles, leading to a decrease in incidence instead of an increase (Thurnon and Robertson 1979). Therefore in the male population of Lastovo the decreased total cancer incidence probably represents a result of inbreeding in the same way as do the increased incidences in populations of the other islands. That is the reason that the deviation from expected incidence in Lastovo males was considered positive in the linear regression analysis (Figure 2).
The second hypothesis claimed that cancer incidence rates in island populations should especially differ from control rates in younger age groups because inherited cancer tends to develop earlier in life than acquired cancer (Schneider et al. 1986; Hansen and Cavenee 1987; Fraumeni et al. 1989; Claus 1995). This hypothesis was not supported by the male population results (Table 1), but it was confirmed in female populations (Table 1). The results in females clearly indicate that the differences in cancer incidence between the control and the island populations resulted almost entirely from differences in age-specific rates for the age groups 0-24 years and 25-44 years.
According to the third hypothesis, the total cancer incidence should vary among the islands, increasing with reproductive isolation, that is, with greater sea distance from the mainland. Although it was optimistic to expect this hypothesis to be clearly confirmed, the linear regression results yielded a coefficient of correlation between 2 variables that was very high in males (r = 0.791; Figure 2) and incredibly high in females (r = 0.991; Figure 3). This was probably the strongest confirmation of the effect of the specific genetic structure of the islands on cancer incidence in their populations.
The fourth hypothesis stated that the cancer sites mostly responsible for total cancer incidence excess in island populations should include cancers with a well-established genetic predisposition, such as colorectal, breast, and ovarian cancer. However, the analysis of site-specific cancer incidence in males did not reveal any clear pattern. The cancer sites that primarily contributed to the differences in total cancer incidence across the island populations (Table 2) were lungs for Brac, lungs and prostate for Hvar, bladder for Korcula, and stomach, lung, and bladder for Vis, whereas the total incidence was decreased on Lastovo. Bowel and rectum cancer did not contribute as expected.
In females the incidences of breast and ovarian cancer were clearly increased in island populations, as expected, knowing the patterns of inheritance for these 2 cancers. Furthermore, the incidences of bowel, brain, and stomach cancer were also increased on the islands compared with the control population, indicating some role of genetic susceptibility in development of these cancers (Table 3). The increased incidence of cervical cancer on 2 islands (Hvar and Vis) probably represents an artifact, because some doctors started reporting cervical carcinoma in situ to the Croatian Cancer Registry as early as 1975, whereas others (particularly, on Lastovo, Korcula, Brac, and the mainland) did so only later, which could cause the discrepancies in total incidence of cervical cancer (invasive + in situ).
Two interesting findings emerge from the analysis of site-specific cancer incidence in female populations. The first was an exceptionally low incidence of rectal cancer on Vis in both males and females (Tables 2 and 3). It is possible that this low incidence is another manifestation of a founder effect; that is, the founding population of Vis did not bring to the island genetic susceptibility to rectal cancer.
The second interesting finding was the negative correlation between breast and ovarian cancer rates in the same population. Today, it is well known that the gene BRCA1 carries the susceptibility to both breast and ovarian cancer (Bishop 1994; Cornelisse et al. 1996; Marx 1997). A lifetime penetrance of this gene is somewhat greater for breast cancer than for ovarian cancer (Fraumeni et al. 1989; Bishop 1994; Kuska 1997). In my research breast cancer incidence was significantly increased in 3 island populations (Hvar, Korcula, and Vis), whereas ovarian cancer incidence was significantly increased in the other 2 island populations (Brac and Lastovo). It is probable that such findings represent the result of accumulation of the BRCA1 gene, which more likely is expressed as breast cancer in some populations and as ovarian cancer in others (Kuska 1997). That also explains the previously mentioned negative correlation between breast cancer incidence and ovarian cancer incidence on each island. Finally, taking into consideration that more female cancers have a pronounced genetic component than male cancers, it is clear that all the presumed patterns were much more obvious in female than in male populations.
This study demonstrates how reproductively isolated populations can be used as a model for research in genetic epidemiology and demography. The accuracy of such investigations results from the large number of subjects under investigation and a long follow-up period. This study confirmed the presence of a degree of direct genetic influence on the development of excess cancer incidence in reproductively isolated populations, which is attributable to the specific genetic structure of such populations.
Acknowledgments This research was supported by the Ministry of Science and Technology of the Republic of Croatia through grant 01960101 as part of the project “Anthropological Investigations of the Population Structure of Croatia: Biomedical Approach.” Further support was furnished by the United States-Croatian Joint Commission for Scientific Cooperation through the Smithsonian Institution, Washington, D.C. (grant JF259), and by the Croatian League Against Cancer (CLAC). I would like to thank Pavao Rudan, Principal Investigator of the project; Linda A. Bennett, the Project Co-Principal Investigator; Francine C. Berkowitz, Director of International Relations of Smithsonian Institutions; Damir Eljuga, General Secretary of CLAC; and Marija Strnad, Head of Croatian Cancer Registry, for their continuous friendship and support.
Received 9 March 1998; revision received 4 September 1998.
Literature Cited
Anderson, D.E. 1982. Familial predisposition. In Cancer Epidemiology and Prevention, D. Schottenfeld and J.F. Fraumeni Jr., eds. Philadelphia, PA: W.B. Saunders, 483485. Andrieu, N., F. Clavel, A. Auquier et al. 1991. Association between breast cancer and family
malignancies. Eur. J. Cancer 27:244-248.
Bellis, G., M. De Braekeleer, C. Calmettes et al. 1995. Genetic demography of medullary thyroid carcinoma in France. Colleg. Antropol. 19:327-336.
Bergholm, U.I. 1989. Medullary thyroid carcinoma in Sweden. Diss. Abstr. Int. [C], 50:686. Bishop, D.T. 1992. Family history of breast cancer: How important is it? J. Med. Genet. 29:152-153.
Bishop, D.T. 1993. The involvement of genetic factors in susceptibility to breast cancer. Eur. J. Cancer Prev. 2(suppl. 3):125-130.
Bishop, D.T. 1994. BRCAI, BRCA2, BRCA3 . . . a myriad of breast cancer genes. Eur. J. Cancer
30A:1738-1739.
Bishop, D.T., and H.J. Thomas.1990. The genetics of colorectal cancer. Cancer Surv. 9:585604. Claus, E.B. 1995. The genetic epidemiology of cancer. Cancer Surv. 25:13-26. Cornelisse, C.J., R.S. Cornelis, and P. Devilee. 1996. Genes responsible for familial breast
cancer. Pathol. Res. Pract. 192:6844693.
Crawford, M.H., E.J. Devor, D.H. O’Rourke et al. 1987. Genetics of primate lymphomas in a baboon (Papio hamadryas) colony of Sukhumi, USSR. Genetica 73:145-157. Croatian Institute for Public Health. 1980-1997. Cancer Incidence in Croatia. Bulletins 1-18.
Zagreb, Croatia: Croatian Institute for Public Health.
Fraumeni, J.F., Jr., R.N. Hoover, S.S. Devesa et al. 1989. Epidemiology of cancer. In Cancer: Principles and Practice of Oncology, 3d ed., V.T. De Vita Jr., S. Hellman, and S.A. Rosenberg, eds. Philadelphia, PA: J.B. Lippincott, 198-227. German, J. 1985. Recessive inheritance in human cancer. In Genetics in Clinical Oncology, R.S.K. Chaganti, and J.L. German, eds. New York: Oxford University Press, 90-102. Goldstein, A.M., and M.A. Tucker. 1995. Genetic epidemiology of familial melanoma. Der
matol. Clin. 13:605-612.
Goldstein, A.M., N.C. Dracopoli, M. Engelstein et al. 1994. Linkage of cutaneous malignant melanoma/dysplastic nevi to chromosome 9p, and evidence for genetic heterogeneity. Am. J. Hum. Genet. 54:489496.
Goldstein, A.M., M.C. Fraser, W.H. Clark Jr. et al. 1994. Age at diagnosis and transmission of invasive melanoma in 23 families with cutaneous malignant melanoma/dysplastic nevi. J. Natl. Cancer Inst. 86:1385-1390.
Hall, N.R., D.T. Bishop, B.M. Stephenson et al. 1996. Hereditary susceptibility to colorectal cancer: Relatives of early onset cases are particularly at risk. Dis. Colon Rectum 39:739-743.
Hansen, M.F., and W.K. Cavenee. 1987. Genetics of cancer predisposition. Cancer Res. 47:5518-5520.
Hauck, W.W., and O. Martin. 1984. A statistical test for detection of ancestral genetic contri
butions to disease occurrence in finite populations. Genet. Epidemiol 1:383-400. Hennings, H., D.T. Lowry, S.H. Yuspa et al. 1997. New strains of inbred SENCAR mice with increased susceptibility to induction of papillomas and squamous cell carcinomas in skin. Mol. Carc*nog. 20:143-150.
Houlston, R.S., A. Collins, F. Kee et al. 1995. Segregation analysis of colorectal cancer in Northern Ireland. Hum. Hered. 45:4148.
Houlston, R.S., A. Collins, J. Slack et al. 1991. Genetic epidemiology of ovarian cancer: Seg
regation analysis. Ann. Hum. Genet. 55:291-299.
Iselius, L., J. Slack, M. Littler et al. 1991. Genetic epidemiology of breast cancer in Britain. Ann. Hum. Genet. 55:151-159.
Janicijevic, B., M. Bakran, S.S. Papiha et al. 1994. Serogenetic analysis in the study of the
population structure of the eastern Adriatic (Croatia). Hum. Biol. 66:991-1003. Kawaguchi, T., H. Nakajima, and T. Hongyo. 1997. Consecutive maintenance of human solitary and hereditary colorectal polyps in SCID mice. Cancer Detect. Prev. 21:148-157. Kerber, R.A., and M.L. Slattery. 1995. The impact of family history on ovarian cancer risk:
The Utah population database. Arch. Intern. Med. 155:905-912. Kuska, B. 1997. Researchers expand genetic epidemiology of BRCA genes. J. Natl. Cancer Inst. 89:844-845.
Malkin, D., and B.M. Knoppers. 1996. Genetic predisposition to cancer: Issues to consider.
Sem. Cancer Biol. 7:49-53.
Marx, G. 1997. Possible function found for breast cancer genes. Science 276:531-532.
Mulvihill, J.J. 1985. Clinical ecogenetics: Cancer in families. New Engl. J. Med. 312:15691578.
Parkin, D.M., S.L. Whelan, J. Ferlay et al. 1997. Cancer Incidence in Five Continents, v. VII. IARC Scientific Publication 143. Lyon, France: International Agency for Research on Cancer.
Roberts, D.F., Z.M. Noor, S.S. Papiha et al.1992. Genetic variation in Brac, Croatia. Ann. Hum.
Biol. 19:539-557.
Roguljic, D., I. Rudan, and P. Rudan. 1997. Estimations of inbreeding, kinship, genetic distance, and population structure from surnames: Example from the island of Hvar, Croatia. Am. J. Hum. Biol. 9:595-407.
Rudan, P., J.L. Angel, B. Bennett et al. 1987. Anthropological Investigations of Eastern Adriatic, v. 1, Biological and Cultural Microdifferentiation of the Rural Populations on the Island of Korcula and Peljesac Peninsula. Zagreb, Croatia: Croatian Anthropological Society.
Rudan, P., L.A. Bennett, B. Finka et al. 1990. Anthropological Investigations of Eastern Adriatic, v. 3, Biological and Cultural Microdifferentiation of the Rural Populations on the Island of Brac. Zagreb, Croatia: Croatian Anthropological Society. Rudan, P., B. Finka, B. JaniCijevic et al. 1990. Anthropological Investigations of Eastern Adriatic, v. 2, Biological and Cultural Microdifferentiation of the Rural Populations on the Island of Hvar. Zagreb, Croatia: Croatian Anthropological Society. Rudan, P., B. Janicijevic, D. Simic et al. 1994. Migration in the eastern Adriatic and their consequences on population structure: Analyses of census data. In Mediterranean Societies: Tradition and Change, A. Lopagic, ed. Zagreb, Croatia: Croatian Anthropological Society.
Rudan, P., A. Sujoldzic, D. mimic et al. 1992. Population structure in the eastern Adriatic: The influence of historical processes, migration patterns, isolation, and ecological pressures, and their interaction. In Isolation, Migration, and Health, D.F. Roberts, N. Fujiki, and K. Torizuka, eds. Cambridge, England: Cambridge University Press, 204-218. Scapoli, C., M. Ponz de Leon, R. Sassatelli et al. 1994. Genetic epidemiology of hereditary non-polyposis colorectal cancer syndromes in Modena, Italy: Results of complex segregation analysis. Ann. Hum. Genet. 58:275-279.
Schneider, N.R., W.R. Williams, and R.S. Chaganti. 1986. Genetic epidemiology of familial
aggregation of cancer. Adv. Cancer Res. 47:1-36.
Seminara, D., and G.I. Obrams. 1994. Genetic epidemiology of cancer: A multidisciplinary approach. Genet. Epidemiol 11:235-254.
Simic, D., and P. Rudan. 1990. Isolation by distance and correlation analysis of distance measures in the study of population structure: Example from the Island of Hvar. Hum. Biol. 62:113-130.
Struewing, J.P., L.C. Brody, M.R. Erdos et al. 1995. Detection of eight BRCA1 mutations in 10 breast/ovarian cancer families, including one family with male breast cancer. Am. J. Hum. Genet. 57:1-7.
Sujoldzic, A., V. Jovanovic, J.L. Angel et al. 1989. Migration within the Island of Korcula.
Ann. Hum. Biol. 16:483-489.
Thompson, W.D. 1994. Genetic epidemiology of breast cancer. Cancer 74:279-287. Thumon, T.F., and K.P. Robertson. 1979. Genetic considerations in human cancer incidence.
Public Health Rep. 94:471-476.
Weiss, K.M., and R. Chakraborty. 1984. Multistage risk models and the age pattern in familial polyposis coli. Cancer Invest. 2:443-48.
Weiss, K.M., R. Chakraborty, P.E. Smouse et al. 1986. Familial aggregation of cancer in Laredo, Texas: A generally low risk Mexican-American population. Genet. Epidemiol. 3:121143.
Weiss, K.M., R.E. Ferrell, C’.L. Hanis et al. 1984. Genetics and epidemiology of gall bladder disease in New World native peoples. Am. J. Hum. Genet. 36:1259-1278.
IGOR RUDAN1
1Department of Medical Statistics, Epidemiology, and Medical Informatics, Faculty of Medici University of Zagreb, School of Public Health “A. Stampar,” Rockefellerova 4, 10000 Zagreb, Croatia.
Human Biology, April 1999, v. 71, no. 2, pp. 173-187.
Copyright 1999 Wayne State University Press, Detroit, Michigan 48201-1309
Copyright Wayne State University Press Apr 1999
Provided by ProQuest Information and Learning Company. All rights Reserved
