FAMILIAL CARCINOID TUMORS AND SUBSEQUENT CANCERS: A NATION-WIDE EPIDEMIOLOGIC STUDY FROM SWEDEN

Transcription

1 Int. J. Cancer: 94, (2001) 2001 Wiley-Liss, Inc. FAMILIAL CARCINOID TUMORS AND SUBSEQUENT CANCERS: A NATION-WIDE EPIDEMIOLOGIC STUDY FROM SWEDEN Kari HEMMINKI* and Xinjun LI Department of Biosciences at Novum, Karolinska Institute, Huddinge, Sweden Publication of the International Union Against Cancer Carcinoids are rare neuroendocrine tumors, mainly located in the bowel, stomach and lung. Familial risks in carcinoid tumours are not well known apart from multiple endocrine neoplasia 1(MEN1). We used the nation-wide Swedish Family-Cancer Database on 10.1 million individuals for assessment. Carcinoid tumors were retrieved from the Cancer Registry covering the years The offspring generation, aged 0 66 years, accumulated 190 million personyears at risk. The age-adjusted incidence rates were 0.76 for men and 1.29/100,000 for women. Standardized incidence ratios (SIRs) were calculated for offspring when their parents hadacarcinoidoranyothercancer.whenparentspresented with carcinoids, SIRs for offspring were 4.35 (n 8, 95% CI ) for small intestinal and 4.65 (n 4, 95% CI ) for colon carcinoids. If both offspring and parents presented with small intestinal carcinoids, the SIR was (n 4, 95% CI ). Offspring carcinoids were also increased if parents presented with bladder and endocrine gland tumors, the latter association probably partially due to MEN1. Risks for second cancers were increased, particularly at sites where familial risks were found, including carcinoids in the small intestine Wiley-Liss, Inc. Key words: familial risk; colon carcinoids; small intestinal carcinoids; second cancer; incidence rates; multiple endocrine neoplasia Carcinoids are neuroendocrine tumors that secrete vasoactive substances such as serotonin and that give atypical histologic reaction to silver stains. 1 Their annual incidence has been estimated at 1 2/100,000, but because many carcinoids are indolent, their true incidence may be higher (reaching 8.2/10,000 in a Swedishautopsyseries 2 ).Thereportedincidenceofcarcinoidshas shownanincreasingtrendintheunitedstatesandswitzerlandbut not in Denmark. 3 5 The most common sites for carcinoid tumors arethesmallintestine,colon,includingappendix,rectum,stomach and lung, including trachea and bronchus. 1,4 There are differences in incidence between genders, with amoderate overall female excess due to carcinoids in the stomach, appendix and colon. 4,6,7 However, both the increasing incidence and the female excess have been ascribed to improvements in diagnostic medical technology, such as tomography, bronchoscopy, gastroscopy, colonoscopy and ultrasound, which allow detection of indolent tumors. Some of these techniques are more intensely used in women duringthereproductiveyears,probablyexplainingsomeofthesex difference. Theetiologyofcarcinoidtumorsislargelyunknown,apartfrom an association with multiple endocrine neoplasia (MEN) 1, in which foregut (lung, thymus, stomach, duodenum) carcinoids and mutations in the MEN 1gene are manifested. 4,7 Other tumors associated with MEN1 are parathyroid and pituitary adenomas, thymomas and insulomas. There are anecdotal reports on familial carcinoids, but only 1epidemiologic study, which reported 10 carcinoid tumors in relatives of probands, giving arelative risk (RR) of Second tumors are of interest in this regard because they may be informative of genetic predisposition. 9,10 Because benign carcinoid tumors are often diagnosed incidentally at medical examination carried out for other purposes, such as suspicion of malignancy, they can be diagnosed with another synchronous tumor. 4 However, excesses of second tumors have been observed even in follow-up studies of small groups of carcinoid patients. 3,5 Because of the limited population-based data on familial clustering of carcinoid tumors, we examine here familial risks and occurrence of second malignancies after carcinoids using the nation-wide Swedish Family-Cancer Database The Database has been updated in 2001 to include over 10 million individuals and over 1million registered tumors. It offers unique possibilities for reliable estimation of familial risks, because the data on family relationships and cancers were obtained from registered sources of nearly complete coverage. This study is by far the largest study on familial carcinoids published to date. SUBJECTS AND METHODS The Swedish Family-Cancer Database was initially created in the middle of the 1990s by linking an administrative family register on all Swedish families to the Swedish Cancer Registry. 11,12 For each child there are data on both parents at time of birth. Each person has been assigned aunique technical identification number (which is different from the national identification number, personal number ), allowing construction of families, for example, through the mother. The present publication is the first on the fourth version of the Family-Cancer Database. It includes all persons born in Sweden after 1931 with their biologic parents, totalling over 10.2 million individuals. 14 Itwas updated at the beginningof2001toincludecancersfromthenationwideswedish Cancer Registry from the years The Database is organized by 3.2 million families, with parents and offspring. Because of the longitudinal structure of the Database, the older generations of offspring have become parents. Thus the numbers of individuals in the parental (6.2 million) and offspring (7.0 million)categoriesexceedthenumberofindividuals,10.2million. However, this does not imply duplication of cases. The completeness of cancer registration in the 1970s has been estimated to be over 95% and is now considered to be close to 100%. The percentage of cytologically or histologically verified cases of colorectal cancers has been 98 99%, but no specific data on carcinoid tumors are available. 15 The Family-Cancer Database has an incomplete linkage from deceased offspring to parents, particularly among those offspring born between 1932 and Of atotal of 7.0 million offspring, 216,000 had died by the end of follow-up, December 31, Parental information was missing from 15,000 offspring who had adiagnosis of cancer (8.9% of all offspring cancers). This deficit is unlikely to cause any effect on familial risk estimates in this study. The Swedish Cancer Registry is based on compulsory notification of cases. 15 Carcinoid tumors are to be reported irrespective of their malignancy. Even tumors found at autopsy are included, but Grant sponsor: the Swedish Cancer Society. *Correspondence to: CNT Novum, Karolinska Institute, Huddinge, Sweden. Fax: kari.hemminki@cnt.ki.se Received 23 February 2001; Revised 23 April 2001; Accepted 8May 2001 Published online 3August 2001; DOI /ijc.1473

2 CARCINOID TUMORS IN SWEDEN 445 the numbers are small compared with clinically diagnosed tumours, e.g., autopsied cases accounted for 2% of male and female colonic and female small intestinal tumors and 12% of male small intestinal tumours in However, we lack individual data on the autopsied cases. A 4-digit diagnostic code according to the seventh revision of the International Classification of Diseases (ICD-7) was used. The following ICD-7 codes were pooled: upper aerodigestive tract cancer, codes 161 (larynx) and (lip, mouth, pharynx), except for code 142 (salivary glands) and leukemia, codes (leukemias), 208 (polycythemia vera) and 209 (myelofibrosis). The histologic classification was used to define carcinoid tumor, pathologic anatomic diagnosis (PAD) code 086. Family history information was collected on all first-degree relatives (parents, siblings and children), but the parent-offspring relationship was mainly used in the present study. In 4 families 2 siblings presented with a carcinoid; each of these were counted as individual events. In no family were 2 parents affected. All tumor incidence rates were based on the data in the Family-Cancer Database. The risk of carcinoids was calculated for offspring whose parents presented with carcinoid or any other specified tumor and their risk was compared with the rate of carcinoid tumous among all offspring. This is the basic type of analysis in genetic epidemiology, testing the hypothesis that the familial ties, either inherited or environmental, influence the risk of disease. Follow-up was started at birth or January 1958, whichever came latest. Follow-up was terminated on death, emigration or the closing data of the study, December 31, Standardized incidence ratios (SIRs) were calculated as the ratio of observed (O) to expected (E) number of cases. The expected numbers were calculated from 5-year-age-, sex- and tumor type-specific standard incidence rates. 16 Confidence intervals (95% CI) were calculated assuming a Poisson distribution. 16 The small number of familial carcinoids did not allow a further stratified analysis by age of onset or calendar period. SIRs for second cancers were calculated in a similar way, starting follow-up from the diagnosis of the carcinoid. All individuals, parents and offspring were included in the analysis of second events. Even synchronous second cancers were included, but the follow-up time was divided in 3 periods, allowing the assessment of the effect of follow-up time. Sibling risks were calculated by the cohort method described earlier. 13 RESULTS The Family-Cancer Database covered the years from the Swedish Cancer Registry and included 1,933 offspring and 4,713 parents with carcinoid tumors (Table I). Because of different age distributions of the offspring and parental generations, the proportions at affected sites differed between the generations. The colon was the most common site for sons, daughters and mothers, whereas for fathers it was the small intestine. Colonic, mainly appendiceal carcinoids are diagnosed at a relatively young age, compared with small intestinal tumors, which are common at older ages. Apart from the bowel, the lung was the only other common site, with around 10% of the cases. The age-specific incidence of all offspring carcinoids is shown in Figure 1 by gender. Up to age 40, the incidence was about twice as high in females compared with males, but the difference became smaller at higher ages. The age-adjusted incidence rates were 0.76 for men and 1.29/100,000 for women aged 0 66 years. Familial risks for offspring carcinoid tumors were calculated by parental cancer (Table II). Only those parental sites were listed that were associated with at least 5 offspring carcinoids; thus the number in row any cancer is higher than the sum of the listed cases. The offspring generation in the Database accumulated a total of 190 million person-years at risk and 55.8 million for those whose parents presented with cancer. No parental cancer was associated with a risk of carcinoids in sons. Carcinoids in daughters were associated with parental small intestinal and bladder tumors. Among all offspring, the SIRs were 2.59 (95% CI ), 1.39 (95% CI ) and 1.66 (95% CI ) by parental small intestinal, bladder and endocrine tumors, respectively. Among the 8 parental small intestinal tumors, all were carcinoids. Among the 20 endocrine gland tumors, 14 were parathyroid adenomas, 3 were pituitary adenomas, 1 was a pituitary squamous cell carcinoma, 1 was a thymoma and 1 was an adrenal adenoma. We also calculated cancer risks to siblings from probands who presented with carcinoid tumors (data not shown). The only site that associated with carcinoids was endocrine glands (n 10, SIR 2.59, 95% CI ; 130,572 person-years). Of these, 5 were parathyroid adenomas and in 1 family a parent was also affected with the same tumor. Three were pituitary adenomas and in 1 family a parent was also affected by a parathyroid adenoma; the remaining 2 were solitary cases of thymoma and insuloma. Offspring carcinoids were assessed by parental carcinoids (shown in Table III). In all, 17 concordant pairs were found, giving an SIR of 4.31 (95% CI ). Of these parental sites, 8 were small intestine and 4 were colon. Parental carcinoids in the small intestine associated with offspring small intestinal carcinoids in 4 pairs, resulting in a high SIR of (95% CI ). Among the 17 pairs of offspring and parents, 7 presented with carcinoids in the same organ. Among the total of 16 small intestinal carcinoids, diagnosed in parent or offspring, none was in the duodenum, 6 were at other sites, 1 was a multifocal tumor and in 9 cases the site was not specified. However, among the offspring small intestinal carcinoids, 5 of 6 were nonduodenal. In 1 offspring-parent pair, both members presented with pulmonary carcinoids and in another with pulmonary-small intestinal carcinoids. SIRs for second cancers following all male (father and son) carcinoids are shown in Table IV, with a total of 16,701 personyears. The SIRs were very high, 15 for all sites, during the first year of follow-up; synchronous tumors were included. Tumors in TABLE I NUMBER OF CASES OF CARCINOID IN OFFSPRING AND PARENTS BY CANCER SITES Offspring Parents Cancer site Son Daughter Offspring Fathers Mothers Parents No. % No. % No. % No. % No. % No. % Stomach Small intestine Colon Rectum Liver Pancreas Lung Ovary Endocrine glands Others All

3 446 HEMMINKI AND LI FIGURE 1 Age-specific incidence rate for carcinoid tumors in offspring in the Family-Cancer Database. TABLE II SIR FOR CARCINOID IN OFFSPRING BY PARENTAL CANCER 1 Parental cancer Son Daughter All offspring No cancer 1, , , Upper aerodigestive tract Stomach Small intestine Colon Rectum Liver Pancreas Lung Breast Cervix Endometrium Ovary Prostate Kidney Urinary bladder Melanoma Skin Nervous system Endocrine glands Non-Hodgkin s lymphoma Hodgkin s disease Myeloma Leukemia Any cancer the small intestine, colon and endocrine glands had SIRs of over 60. In the subsequent period, 1 10 years of follow-up, the overall SIR had decreased to However, the excess in small intestinal tumors remained high, with an SIR of All but 1 of the small intestinal tumors were carcinoids, for which the SIRs were about twice as high as for all small intestinal tumors. Among colon tumors, only a few were carcinoids but the SIRs were very high. The other sites that were in excess included colon, prostate, squamous cell carcinoma of the skin, endocrine glands and lymphatic tissue (non-hodgkin s lymphoma). In the longest follow-up period, only small intestinal and prostatic cancers remained elevated and the pancreas emerged as a new site. Among the 12 second small intestinal carcinoids noted in Table IV, 7 had a primary site in the small intestine, 2 in the colon and 3 in other locations. All 6 of the second colonic carcinoids had the small intestine as the primary site. None of the first colon tumors were located in appendix. Second cancers in all female carcinoids, with an accumulated 23,189 person-years, were fewer, but the overall pattern was similar to that found in the men (Table V). Small intestinal, colon and endocrine tumors were markedly increased initially, but only small intestinal tumors, particularly carcinoids, remained at a high level throughout the follow-up period. At the longest follow-up,

4 CARCINOID TUMORS IN SWEDEN 447 TABLE III SIR FOR CARCINOID IN OFFSPRING BY PARENTAL CARCINOID 1 Parents Cancer site Son Daughter All offspring Fathers All carcinoid Small intestine Colon Mothers All carcinoid Small intestine Colon Parents All carcinoid Small intestine Colon TABLE IV SIR FOR SECOND CANCER IN MEN FOLLOWING CARCINOID 1 Follow-up interval (yr) Second cancer site Upper aerodigestive tract Stomach Small intestine Carcinoid Colon Carcinoid Rectum Liver Pancreas Lung Prostate Kidney Urinary bladder Melanoma Skin Nervous system Endocrine glands Non-Hodgkin s lymphoma Myeloma Leukemia Any sites increased risks were observed for tumors of the upper aerodigestive tract, breast, endometrium, ovary, kidney, skin (melanoma) and bone marrow (myeloma), in addition to small intestine. Even the overall SIR of 3.31 at 10 years of follow-up was higher than that of 1.75 for 1 10 years of follow-up. Of the 8 second small intestinal carcinoids, 3 had the colon as the first site and the remaining had other locations. Of the 7 colonic carcinoids, 3 had the colon and 3 the small intestine as the first site. DISCUSSION All common cancers have many genetic and environmental causes, i.e., they are heterogeneous. Most carcinoids are assumed to be sporadic and the only convincing evidence for a heritable cause is the presence of carcinoids in MEN 1 syndrome. 17 This features the occurrence of mainly foregut carcinoids with adenomas in parathyroid, pituitary and pancreatic islet glands. 18,19 It is likely, but difficult to prove only based on epidemiologic data, that part of the clustering we observed here of carcinoid and endocrine gland tumors was due to MEN1. However, our data provide unique population-based evidence that there is a consistent link between carcinoid and endocrine gland tumors, shown between offspring and parents, between siblings and between first carcinoids and subsequent endocrine tumors. Carcinoids are endocrine tumors, related physiologically to other endocrine gland tumors even though they are grouped separately in the ICD classification; thus the shared familial risks may not be surprising. The present study provided convincing evidence for familial risks in carcinoid tumors, corroborating earlier anecdotal evidence and a small epidemiologic study. 8 The familial risk for any carcinoid tumor in offspring by affected parent was 4.31; the risk was essentially independent of the site of parental carcinoid (small intestine or colon). However, if both generations presented with small intestinal carcinoids, the SIR was (n 4, 95% CI ). Two arguments suggest that the observed familial clustering was independent of MEN 1. First, in only 1 of 17 pairs of the affected offspring-parent pairs were the carcinoids located in the foregut (lungs), which is the preferential location of carcinoids in MEN 1. Second, none of the offspring-parent or sib-sib pairs concordant for carcinoid had other family members with endocrine tumors, typically present in MEN 1 syndrome. The only other site that was associated with carcinoids in offspring was urinary bladder, with an SIR of 1.39 ( ). Because of the indolent nature of some carcinoids, they are often found incidentally, perhaps in search of other malignancies. This is the reason for the large number of synchronous neoplasms detected at the diagnosis of carcinoids. 2,4 Our data agree on the large excess risk of noncarcinoid tumors in the first year of follow-up. However, there was an overall excess of second cancers even during subsequent follow-up periods, indicating that the reason is not only an incidental finding of other neoplasms upon diagnostic procedures. Almost all tumor notifications registered at the Swedish Cancer Registry bear a histologic or cytologic verification,

5 448 HEMMINKI AND LI TABLE V SIR FOR SECOND CANCER IN WOMEN FOLLOWING CARCINOID 1 Follow-up interval (yr) Second cancer site Upper aerodigestive tract Stomach Small intestine Carcinoid Colon Carcinoid Pancreas Lung Breast Endometrium Ovary Kidney Melanoma Skin Nervous system Thyroid gland Endocrine glands Connective tissue Non-Hodgkin s lymphoma Myeloma Leukemia Any sites excluding false reporting as the cause for the increase in second cancers. On the other hand, any treatment-related effects would be expected to show up only about a decade after diagnosis of the first cancer; some evidence to this effect was observed among women, who had an increase in the overall SIR after 10 years of follow-up. However, it is likely that the increases in small intestinal, colon and endocrine tumors in men and women, observed particularly during 1 10 years of follow-up, represent true susceptibility of the carcinoid patients. Interestingly, these were the sites also showing the highest familial risk, suggesting that familial, perhaps polygenic, factors are important causes of second malignancies. REFERENCES 1. Kulke M, Mayer R. Carcinoid tumors. N Engl J Med 1999;340: Berge T, Linell F. Carcinoid tumours: frequency in a defined population during a 12-year period. Acta Pathol Microbiol Scand (A) 1976;84: Westergaard T, Frisch M, Melbye M. Carcinoid tumors in Denmark and the risk of subsequent cancers. A population-based study. Cancer 1995;76: Modlin I, Sandor A. An analysis of 8305 cases of carcinoid tumors. Cancer 1997;79: Levi F, Te V, Randimbison L, et al. Epidemiology of carcinoid neoplasms in Vaud, Switzerland, Br J Cancer 2000;83: Newton J, Swerdlow A, dos Santos Silva I, et al. The epidemiology of carcinoid tumours in England and Scotland. Br J Cancer 1994;70: Hamilton S, Aaltonen L, editors. Tumours of the digestive system. Lyon: IARC, Babovic-Vuksanovic D, Constantinou C, Rubin J, et al. Familial occurence of carcinoid tumors and association with other malignant neoplasms. Cancer Epidemiol Biomarkers Prev 1999;8: Dong C, Hemminki K. Multiple primary cancers at colon, breast and skin (melanoma) as models for polygenic cancers. Int J Cancer 2001; 92: Dong C, Hemminki K. Second primary neoplasms in 633,964 cancer patients in Sweden, Int J Cancer 2001;93: Hemminki K, Vaittinen P, Kyyrönen P. Age-specific familial risks in common cancers of the offspring. Int J Cancer 1998;78: Hemminki K, Vaittinen P. Familial cancers in a nation-wide familycancer database: age distribution and prevalence. Eur J Cancer 1999; 35: Hemminki K, Vaittinen P, Dong C, et al. Sibling risks in cancer: clues to recessive or X-linked genes? Br J Cancer 2001;84: Hemminki K, Vaittinen P, Dong C. Endometrial cancer in the Family- Cancer Database. Cancer Epidemiol Biomarkers Prev 1999;8: Centre for Epidemiology. Cancer incidence in Sweden Stockholm: Centre for Epidemiology, Esteve J, Benhamou E, Raymond L. Statistical methods in cancer research. Lyon: IARC, Kumar V, Cotran R, Robbins S. Basic pathology. Philadelphia: WBSaunders, Lindor N, Greene M. A concise handbook of family cancer syndromes. J Natl Cancer Inst 1998;90: Monson J. The epidemiology of endocrine tumours. Endocr Rel Cancer 2000;7: Bergfeldt K, Einhorn S, Rosendahl I, et al. Increased risk of second primary malignancies in patients with gynecological cancer. Acta Oncol 1995;34: Swerdlow A, Barber J, Hudson G, et al. Risk of second malignancy after Hodgkin s disease in a collaborative British cohort: the relation to age at treatment. J Clin Oncol 2000;18: Metayer C, Lynch C, Clarke E, et al. Second cancers among long-term survivors of Hodgkin s disease diagnosed in childhood and adolescence. J Clin Oncol 2000;18: