Annals of Oncology Advance Access originally published online on September 5, 2007
Annals of Oncology 2008 19(1):163-167; doi:10.1093/annonc/mdm414
epidemiology |
How common is familial cancer?
1 Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany
2 Center for Family and Community Medicine, Karolinska Institute, Huddinge, Sweden
* Correspondence to: Dr K. Hemminki, Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 580, D-69120 Heidelberg, Germany. Tel: +49-62-21421800; Fax: +49-62-21421810; E-mail: k.hemminki{at}dkfz.de
 |
Abstract |
|---|
 Top Abstract introductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Background: Family history of a disease may point to its heritable or environmental etiology. It can be described by the proportion of the familial disease, i.e. same disease in two or more family members. A family history always needs to be specified as to the number of generations covered and their ages.
Patients and methods: Proportions of site-specific familial cancers (familial proportions) were calculated using the Swedish Family-Cancer Database, the largest dataset of its kind in the world, with cancers from the Swedish Cancer Registry. Familial proportions refer to the offspring population up to age 72 years when their parents or siblings were diagnosed with a concordant (same) cancer.
Results: A total of 34 cancer sites and 205 638 cases were covered. Prostate cancer showed the highest familial proportion of 20.15%, followed by breast (13.58%) and colorectal (12.80%) cancers. Salivary gland cancers showed the lowest familial proportion of 0.15%, but bone, laryngeal, anal, connective tissue and other genital cancers also remained <1%. The familial proportion depended on the prevalence of the particular cancer and on its familial risk.
Conclusions: The derived familial proportions can justifiably be used in statements ‘X% of the patients had a family history of the cancer’.
Key words: clinical genetic counselling, familial risk, referral of familial cancer, family history, cancer prevalence, familial proportion
|
introduction |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Medical texts on cancer etiology discuss increasingly heritable causes and related genetic mechanisms, ‘X% of the patients with cancer Y report a family history’ [1, 2]. Such data are indeed relevant because family history and twin concordance are the only population-level pieces of evidence informative of a possible heritable etiology, with an important caveat that environmental causes of cancer may also cause familial aggregation [2, 3]. Known cancer syndromes with identified gene defects explain only a small proportion of the known familial aggregation of any common cancers [4–8]. Whenever family history is discussed, it needs to be specified as to the number of generations covered (first degree, second degree, etc.) and to the ages of at least the youngest generation. Both these variables have a drastic impact on the implications of a family history. Nevertheless, an unqualified reference to a family history is commonplace. In expert reviews on specific cancers, unreferenced statements are made ‘X% of the patients have a family history of the cancer’, and this figure is then cited as a proof of heritable etiology in numerous subsequent publications by other authors. The multiauthor treatise, The Genetic Basis of Human Cancer carries 18 chapters on site-specific cancers, of which nine make a reference to a family history, but only two specify what is meant [9]. Furthermore, in most literature on familial cancer, patients have been asked about cancers in their relatives without confirming the reports. The accuracy of reporting a family history depends on many factors, including the type of cancers; the reporting is notoriously inaccurate for many internal cancers [10, 11].
An unambiguous definition of the proportion of cases with an affected relative (familial proportion) is possible through a nationwide familial dataset, such as the Swedish Family-Cancer Database, with registered families and medically verified cancers from the Swedish Cancer Registry [12, 13]. Using this resource, we define the proportion of concordant (same) familial cancers for all cancers in a population that has reached age 72 years. We divide the families into three distinct types through the relationship of the affected individuals: parent–offspring pairs, two or more siblings or parent and two or more siblings. These distinct types have implications to the causes of familial aggregation. Common cancers occur in families more frequently than rare cancers but we demonstrate that the prevalence does not alone explain the familial aggregation.
|
subjects and methods |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
The Family-Cancer Database has been created by linking information from the Multi-Generation Register, national censuses, Swedish Cancer Registry and death notifications [12, 13]. Data on family relationships were obtained from the Multi-Generation Register, where children (the offspring generation) born in Sweden in 1932 and later are registered with their biological parents (parental generation) as families. Register linkages were carried out using the individually unique national registration number (different from the national identification number). The Swedish Cancer Registry is based on compulsory reports of diagnosed cases provided by physicians (clinical report) and by pathologists or cytologists (pathology report); the coverage of the cancer registration is currently considered to be close to 100%. The database (MigMed2) was updated in 2006 to include the cancer cases from years 1961 to 2004 [13].
Calculations were based on individuals with identified parents (7.45 million offspring aged 72 years or below), who were followed from birth, immigration date or 1961, whichever came latest, until diagnosis of any cancer, death, emigration or 31 December 2004, whichever came first (median length of follow-up 33 years). Individuals affected by cancer during the follow-up period were further classified according to a family history of cancer at the concordant site. Three different types of family history were considered: individuals with an affected parent (parent as proband) and with no affected sibling, individuals with at least one affected sibling and without an affected parent and individuals with one affected parent and at least one affected sibling. The cumulative risk of site-specific cancer before age 73 years was also estimated for the general offspring population and for individuals with at least one affected sibling. Cumulative rates were calculated as the sum of age-specific incidence rates multiplied by the width of the age groups (<20, 20–24, 25–29, ..., 65–69 and 70–72 years). Then, the cumulative risk was computed as 100 x (1 – ecumulative rate/100), where e is the base of the natural logarithm [14].
|
results |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
We analyzed the proportion of concordant familial cancers in a 72-year-old population of offspring whose parents' ages were not limited, covering 34 cancer sites (and additionally colorectum) and 205 638 cases (Table 1). Prostate cancer showed the highest familial proportion of 20.15%, followed by breast (13.58%) and colorectal (12.80%) cancers. Salivary gland cancers showed the lowest familial proportion of 0.15%, but bone, laryngeal, anal, connective tissue and other genital cancers also remained <1%.
|
The most common familial clusters were the affected parent–offspring pairs which accounted for 68.7% of all familial cases, for example, 2478/3587 = 69.1% for prostate, 63.5% for breast and 80.1% for colorectal cancers (Table 1). Affected siblings in families where parents were not diagnosed with a concordant cancer accounted for 23.9% for prostate, 32.1% for breast and 15.9% for colorectal cancers. Small intestinal, testicular and bone tumors and Hodgkin's disease were notable exceptions as for these cancers affected siblings were more numerous that affected offspring of affected parents. Families with affected parents and multiple affected offspring were rare, accounting for 4.5% of all familial cases and for some sites no such families were found; thyroid (18.1%) and small intestinal (14.3%; however only one family) tumors were exceptions with high proportions. Families with multiple affected members were also common for prostate cancer (7.0%), while for breast cancer (4.4%) the percentage was close to the average.
Figure 1 shows the dependence of the cumulative risk of cancer in siblings of affected individuals on the cumulative risk of that cancer in the whole population. The results include cancers for which at least 10 affected siblings were recorded. The diagonal line in the log–log scale shows the expected cumulative risks for siblings of affected individuals if these individuals were at the same risk as anyone in the population; this risk increases linearly as a function of the population prevalence of the particular cancer. All observed cumulative risks lay above the diagonal, with a distance to the diagonal proportionate to the familial risk of cancer for this site.
|
|
discussion |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
The present familial proportions show the prevalence of familial cancers in a 72-year-old Swedish population. About 85% (13 707 of 16 075) of the familial cancers in this population were diagnosed from years 1990 to 2004. The Swedish incidence patterns of cancer are similar to the Northern and Western European and North American white populations [15]. During the 1990s, large increases in incidence have taken place for breast cancer because of the implementation of the nationwide mammography screening program [16–19] and for prostate cancer because of opportunistic prostate-specific antigen screening [20]. Cervical cancer screening was implemented in Sweden in the 1960s, while for colorectal cancer no national programs have been launched. Practically, all cancers in the Swedish Cancer Registry are histologically verified and thus diagnostic accuracy is uncompromised [21]. We have noted earlier that family members may be diagnosed close in time for some screening-detected cancers, but such lead-time bias has hardly any effect on the overall familial risks [22].
The familial proportion depends on the age-specific prevalence of the particular cancer and its familial risk. The familial proportions were calculated for mutually exclusive family types of affected parent–offspring and sibling pairs and of affected parents and multiple siblings. Thus, the familial proportion is strictly defined within two generations of first-degree relatives. The proportions were high for the common cancers of the prostate (20.15%), breast (13.58%) and colorectum (12.80%). They exceeded 5% even for lung, stomach and urinary bladder cancers. These percentages can justifiably be used in statements ‘X% of the patients have a family history of the cancer’, as long as the population age and family structures are explained. When the population ages, the cumulative risks of cancer increase. However, as an opposite trend, familial risks tend to be lower for older individuals [23], and thus the net effect of aging on the familial proportion depends on the balance between these two trends, which will be specific to the cancer type. A word of caution is in order about the comparability of the data between cancers sites. The offspring generation was truncated at age 72 years, which is barely over the median diagnostic age of cancer in the Swedish Cancer Registry (70 years). Thus, the present study scores larger proportions of early onset than late onset familial cancers.
Many publications give familial risks (for which familial proportion needs to be calculated but which often remain unreported) for ‘first-degree relatives’, whom clinicians never see. Here, we define familial proportions to a specific family relationship to correspond the clinical situation. Moreover, the defined family relationship may have mechanistic implications: parent–offspring concordance, particularly when several offspring are affected, may arise through dominant heritable effects while sibling concordance may be due to recessive effects, when parents remain unaffected [24]. Environmental factors may cause similar aggregation and siblings may be affected if exposures, e.g. infections, take place during childhood. Familial proportions may even witness therapeutic improvements in early onset cancers: survival for Hodgkin's disease and testicular cancer has drastically improved in the 1980s with affected siblings surviving but affected parents dying before the end of their reproductive period. The familial proportion would then be relatively higher for siblings than for parent–offspring pairs.
How generalizable are the present results? Limited data are available on familial proportions from outside Sweden. The proportions can be calculated from the year 1994 publication on The Utah Population Database, with the distinction that this source counted familial cancers for all first-degree relatives of cases, e.g. considering three generations, instead of the present two generations, although probably the young offspring generation contributed a few familial cancers [25]. The results on familial proportions were curiously similar to the present ones, for example, for prostate 21.7% (present 20.15%), breast 14.0% (13.58%) and colorectal 10.8% (12.80%) cancers. The agreement in these common cancers is excellent even though the covered periods were different, yet overlapping. For less common cancer sites the agreement was not as good, but the present number of familial cases, about five times larger than that in the Utah study, should guarantee precision even down to rare familial cancers. A more indirect way of examining the generalizability of the result would be to compare familial risk estimates for common cancers. The familial risks of common cancers from the Swedish, Utah and Icelandic population-based family datasets and the results from pooled familial risk analysis agree within a few decimal units [24–30]: familial risks (parent–offspring) are 
1.8 for breast, 2.0 for colorectal and 2.3 for prostate cancers, in spite of different study designs, population definitions and observed heterogeneity in the pooled studies. As a final point, although founder mutations for heritable cancers differ both qualitatively and quantitatively in populations, only some of the most common mutants of cancer susceptibility genes, such as the 999del15 variant for BRCA2 in Iceland [31], influence familial risks of common cancers. Thus, in the absence of such strong founder effects, the present familial proportions should be generalizable into populations with approximately similar cancer incidence and pattern.
|
funding |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
Deutsche Krebshilfe; Swedish Cancer Society; Swedish Council for Working Life and Social Research; European Union (LSHC-CT-2004-503465).
|
Acknowledgements |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
The Family-Cancer Database was created by linking registers maintained at Statistics Sweden and the Swedish Cancer Registry. The authors declare no conflict of interests.
Received for publication July 12, 2007. Accepted for publication July 16, 2007.
|
References |
|---|
|
TopAbstractintroductionsubjects and methodsresultsdiscussionfundingAcknowledgementsReferences |
|---|
1. Guttmacher AE, Collins FS, Carmona RH. The family history—more important than ever. N Engl J Med (2004) 351:2333–2336.
2. Hemminki K, Lorenzo Bermejo J, Försti A. The balance between heritable and environmental aetiology of human disease. Nat Rev Genet (2006) 7:958–965.[CrossRef][Web of Science][Medline]
3. Hopper JL, Bishop DT, Easton DF. Population-based family studies in genetic epidemiology. Lancet (2005) 366:1397–1406.[CrossRef][Web of Science][Medline]
4. Lynch HT, de la Chapelle A. Hereditary colorectal cancer. N Engl J Med (2003) 348:919–932.
5. Nagy R, Sweet K, Eng C. Highly penetrant hereditary cancer syndromes. Oncogene (2004) 23:6445–6470.[CrossRef][Web of Science][Medline]
6. Peto J, Houlston R. Genetics and the common cancers. Eur J Cancer (2001) 37:S88–S96.[Web of Science][Medline]
7. Pharoah PD, Antoniou A, Bobrow M, et al. Polygenic susceptibility to breast cancer and implications for prevention. Nat Genet (2002) 31:33–36.[CrossRef][Web of Science][Medline]
8. Syrjäkoski K, Vahteristo P, Eerola H, et al. Population-based study of BRCA1 and BRCA2 mutations in 1035 unselected Finnish breast cancer patients. J Natl Cancer Inst (2000) 92:1529–1531.
9. Vogelstein B, Kinzler K. The Genetic Basis of Human Cancer (2002) 2nd edition. New York, NY: McGraw-Hill.
10. Chang E, Ekström Smedby K, Hjalgrim H, et al. Reliability of self-reported family history of cancer in a large case-control study of lymphoma. J Natl Cancer Inst (2006) 98:61–68.
11. Murff HJ, Spigel DR, Syngal S. Does this patient have a family history of cancer? An evidence-based analysis of the accuracy of family cancer history. JAMA (2004) 292:1480–1489.
12. Hemminki K, Li X, Plna K, et al. The nation-wide Swedish Family-Cancer Database: updated structure and familial rates. Acta Oncol (2001) 40:772–777.[CrossRef][Web of Science][Medline]
13. Hemminki K, Granström C, Sundquist J, Lorenzo Bermejo J. The updated Swedish family-cancer database used to assess familial risks of prostate cancer during rapidly increasing incidence. Heredit Cancer Clin Pract (2006) 4:186–192.
14. dos Santos Silva I. Cancer Epidemiology: Principles and Methods (1999) Lyon, France: IARC.
15. IARC (ed.). GLOBOCAN 2000. Cancer Incidence, Mortality and Prevalence Worldwide (2001) Lyon, France: IARC Press.
16. Hemminki K, Rawal R, Lorenzo Bermejo J. Mammographic screening is dramatically changing age-incidence data for breast cancer. J Clin Oncol (2004) 22:4652–4653.
17. Hemminki K, Lorenzo Bermejo J. Effects of screening for breast cancer on its age-incidence relationships and familial risk. Int J Cancer (2005) 117:145–149.[CrossRef][Web of Science][Medline]
18. Moller B, Weedon-Fekjaer H, Hakulinen T, et al. The influence of mammographic screening on national trends in breast cancer incidence. Eur J Cancer Prev (2005) 14:117–128.[CrossRef][Web of Science][Medline]
19. Olsson S, Andersson I, Karlberg I, et al. Implementation of service screening with mammography in Sweden: from pilot study to nationwide programme. J Med Screen (2000) 7:14–18.
20. Hemminki K, Rawal R, Bermejo JL. Prostate cancer screening, changing age-specific incidence trends and implications on familial risk. Int J Cancer (2005) 113:312–315.[CrossRef][Web of Science][Medline]
21. Center for Epidemiology. Cancer Incidence in Sweden 2002 (2004) Stockholm, Sweden: The National Board of Health and Welfare.
22. Lorenzo Bermejo J, Hemminki K. Familial risk of cancer shortly after diagnosis of the first familial tumor. J Natl Cancer Inst (2005) 97:1575–1579.
23. Hemminki K, Li X, Czene K. Familial risk of cancer: data for clinical counseling and cancer genetics. Int J Cancer (2004) 108:109–114.[CrossRef][Web of Science][Medline]
24. Hemminki K, Li X. Familial risks of cancer as a guide to gene identification and mode of inheritance. Int J Cancer (2004) 110:291–294.[CrossRef][Web of Science][Medline]
25. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst (1994) 86:1600–1607.
26. Amundadottir LT, Thorvaldsson S, Gudbjartsson DF, et al. Cancer as a complex phenotype: pattern of cancer distribution within and beyond the nuclear family. PLoS Med (2004) 1:e65.[CrossRef][Medline]
27. Collaborative Group of Hormonal Factors in Breast Cancer. Familial breast cancer: collaborative reanalysis of individual data from 52 epidemiological studies including 58 209 women with breast cancer and 101 986 women without the disease. Lancet (2001) 358:1389–1399.[CrossRef][Web of Science][Medline]
28. Johns LE, Houlston RS. A systematic review and meta-analysis of familial colorectal cancer risk. Am J Gastroenterol (2001) 96:2992–3003.[CrossRef][Web of Science][Medline]
29. Johns LE, Houlston RS. A systematic review and meta-analysis of familial prostate cancer risk. BJU Int (2003) 91:789–794.[CrossRef][Web of Science][Medline]
30. Butterworth AS, Higgins JP, Pharoah P. Relative and absolute risk of colorectal cancer for individuals with a family history: a meta-analysis. Eur J Cancer (2006) 42:216–227.[CrossRef][Web of Science][Medline]
31. Thorlacius S, Sigurdsson S, Bjarnadottir H, et al. Study of a single BRCA2 mutation with high carrier frequency in a small population. Am J Hum Genet (1997) 60(5):1079–1084.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
Related articles in Ann Oncol:
- in this issue
Ann Oncol 2008 19: 1.[Extract] [FREE Full Text]
This article has been cited by other articles:
|
|
 |
|
  M. Leu, K. Czene, and M. Reilly Bias correction of estimates of familial risk from population-based cohort studies Int. J. Epidemiol., October 13, 2009; (2009) dyp304v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Leu, M. Reilly, and K. Czene Evaluation of Bias in Familial Risk Estimates: A Study of Common Cancers Using Swedish Population-based Registers J Natl Cancer Inst, September 17, 2008; 100(18): 1318 - 1325. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||