Association of colorectal adenoma with other malignancies in Swedish families

Using the Swedish Family-Cancer Database covering over 11.5 million individuals, estimated relative risks (RRs) for colorectal adenoma were using Poisson's regression. The RR of colorectal adenoma was found to be increased among first-degree relatives of patients with colorectal cancer (2.72; 95% confidence interval=2.46–3.00) and among the offspring and siblings of patients with endometrial and prostate cancers. We also found an increased risk of colorectal adenoma for the offspring of individuals with stomach cancer and leukaemia, and for siblings of those with pancreatic cancer and multiple myeloma. Our results suggest that colorectal adenoma may share a genetic aetiology with cancer even at extracolorectal sites. Increases of colorectal adenoma in families affected by prostate cancer and acute leukaemia cannot be attributed to known cancer syndromes, although the play of chance cannot be excluded.


MATERIALS AND METHODS
The population-based Swedish Family-Cancer Database was created by linking the Multigeneration Register at Statistics Sweden to the Swedish Cancer Registry (Hemminki et al, 2001b). This is based on compulsory reports from pathologists and cytologists, on every cancer diagnosis on surgically removed tissues, biopsies, cytological specimens, bone marrow aspirates and autopsies (Center for Epidemiology, 2004). The Multigeneration Register includes individuals born in Sweden after 1931 and their biological parents and its latest update covers over 11.5 million individuals. Cancer/adenoma data were retrieved from the Swedish Cancer Registry from 1961 to 2004. Database coverage is practically complete but some familial links are missing from offspring born before 1941 and dying in 1960 -1997 reducing the number of fatal cancers among offspring. This is unlikely to bias familial studies because familial and sporadic cases would be reduced proportionately (Hemminki et al, 1998;Hemminki and Vaittinen, 1999). The present study relies on individuals with information about both parents. Family history was restricted to first-degree relatives, parents, and siblings.
From International Classification of Diseases (ICD-7) the following codes were used: for colorectal adenomas 094; CRCs by anatomic site: 1530 (ascending), 1531 (transverse), 1532 (descending), 1533 (sigmoid), 1538 (multiple sites), 1534 and 1539 (other sites), 154 -excluding 1541 (rectum). Follow-up started from the date of birth, immigration or 1 January 1961, whichever occurred last. Follow-up ended on the date of diagnosis of CRC or colorectal adenoma, death, emigration, or the closing date of the study (31 December 2004), whichever came first. Relative risks (RRs) with 95% confidence intervals (CIs) were used to compare adenoma incidence among those whose family was affected by cancer/adenoma with adenoma incidence in the general population. Cases of adenoma and person -years were classified according to gender, family history of cancer, calendar year and age. The distribution of the number of cases in each group was modelled by Poisson's regression. The Genmod procedure of the SAS software was used for the analysis (SAS Version 9.1; SAS Institute, Cary, NC, USA).

RESULTS
The Database included 2943 individuals diagnosed with colorectal adenoma and with information about both parents. The parental generation incorporated 12 458 individuals with colorectal adenoma. The incidence of colorectal adenoma in Sweden has been increasing since the 1960s; the site-specific age-adjusted incidence rates for 1990 -1999 and 2000 -2004 are displayed in Table 1. Table 2 shows some demographic characteristics and RRs. Women were at a slightly lower risk of colorectal adenoma compared to men (RR ¼ 0.89; 95% CI ¼ 0.82 -0.94). As expected, age was associated with risk, which was also increased with a family history of CRC (RR ¼ 2.72; 95% CI ¼ 2.46 -3.00) and particularly of colorectal adenoma (RR ¼ 4.99; 95% CI ¼ 4.12 -6.05). Among the 108 patients with colorectal adenoma and also a family history of this, 55 individuals (in 27 families) had an affected sibling (RR ¼ 9.41; 95% CI ¼ 7.21 -12.3).
Standard Poisson's regression assumes independent observations. Because of possible overdispersion due to clustered family structure, s.e. were adjusted using Pearson's w 2 , divided by the degrees of freedom, resulting in slightly wider CIs. For example, the RR of colorectal adenoma for individuals with a positive family history was 4.99 (95% CI ¼ 3.36 -7.41). However, as this procedure may be sensitive to outlying observations to be expected in our large data set, we show unadjusted CIs and point out the possibly conservative limits due to familial dependence.
The offspring risk of colorectal adenoma was higher when parents had multiple adenomas than when parents presented with single adenomas (RR ¼ 21.6; 95% CI ¼ 5.70 -82.0). Sibling risk was also significantly higher for multiple than for single adenomas in any colorectal site (RR ¼ 20.7; 95% CI ¼ 3.48 -123) (results not shown). Table 4 shows colorectal adenoma RRs with a family history of malignancies other than CRC, but only for sites where at least 10 individuals with a family history of cancer had colorectal adenoma.

DISCUSSION
The present study indicates that individuals with first-degree relatives with colorectal adenoma have an approximately five times higher probability of this neoplasm than those without a family history; the risk being particularly high among relatives of those with multiple adenomas. As in previous studies risk was increased among offspring and siblings of CRC patients (Bonelli et al, 1988;Tung and Wu, 2000;Lindgren et al, 2002).
To our knowledge, there are no studies of colorectal adenoma in relation to a family history of extracolonic cancers. The present data suggest an involvement of colorectal adenoma in HNPCC and associations with the commonest HNPCC cancer sites -endometrium, stomach, pancreas and nervous system -were identified (Lynch and de la Chapelle, 2003).
We also found colorectal adenoma associated with some extracolorectal cancers unrelated to CRC. There was a higher risk of colorectal adenoma for the offspring and siblings of patients with prostate cancer. The siblings of patients with multiple myeloma and the offspring of patients with leukaemia were at an increased risk of colorectal adenoma. Increased risks of CRC are reported among firstdegree relatives of patients with leukaemia (Andrieu et al, 2004;Hemminki and Chen, 2004), myeloma (Hemminki and Li, 2001), and prostate cancer (Slattery and Kerber, 1994). We also found that the offspring of those with adult acute lymphoblastic leukaemia were at a higher risk of colorectal adenoma. Rectal adenoma risk was increased among siblings of women with cervical cancer. An increased risk of rectal cancer before the age of 50 years has been reported among the offspring of women with cervical cancer after the age of 50 years (Hemminki and Chen, 2004), but the risk was not increased among siblings of such women. Moreover, the risk of rectal cancer as a second tumour seems to be increased after a first cervical cancer (Hemminki et al, 2001a). The familial association of rectal adenoma with melanoma should be investigated using independent data.
Our study had the advantage of information on the anatomical location of the colonic tumours, but our analyses were limited by small numbers, although this problem can be addressed when the Database in further updated. During our study, Sweden had no population-based screening programmes for CRC (Hakama et al, 2005). Some individuals in our study may have undergone colonoscopic screening because of a family history of CRC, as the risks reported in Tables 2 and 3 could be overestimated. Another limitation is that only precancerous lesions and gastrointestinal polyps with suspected malignancy are compulsorily reported to the Cancer Registry and some adenomas might be already precancerous. A shortcoming is the lack of data on tobacco smoking, alcohol consumption, or diet. In conclusion, a family history of colorectal adenoma is a risk factor for colorectal adenoma particularly for first-degree relatives of patients with multiple adenomas. Unrelated to known CRC syndromes, we found increases among the offspring and siblings of patients with prostate cancer, and the offspring of patients with leukaemia. However, given the number of tests performed, chance may have operated in some of our findings. Our data may help to understand the adenoma -carcinoma sequence and to develop prevention strategies.