Clustering of concordant and discordant cancer types in Swedish couples is rare

Background: Spouses are exposed to common environmental cancer risk factors during adulthood. Investigating the aggregation of cancer in couples might provide valuable insights into cancer development. Methods: The 2008 update of the Swedish Family-Cancer Database includes over 2 million couples with at least one child in common with one single partner. We quantified the contribution of shared adulthood environment by standardised incidence ratios (SIRs) and population attributable fractions (PAFs). Estimated SIRs were used to build an etiological map reflecting the similarity of cancers by adult environmental exposures. Results: Increased risks of concordant types amongst spouses were found for lung, upper aerodigestive tract and skin cancers (SIRs from 1.24 to 1.97), which are probably related to shared exposure to smoking and UV radiation. PAFs were low with the highest value of 1.46% for uterus cancer in wives of men affected by prostate cancer. Further analysis, based on all non-sex-specific concordant and discordant types, revealed a clustering of lung, stomach, pancreas and bladder cancers sharing smoking as a risk factor. This aggregation was used as a cut-point to identify further "novel" clusters. Conclusion: Shared lifestyles including smoking and drinking habits as well as human papilloma virus infection (HPV) might be associated with an excess of cancer incidence amongst spouses. We observed significantly an increased risk for smoking-related cancers such as lung, upper aerodigestive tract and oesophageal cancers. The present population-based study confirms that the lifestyle shared by spouses plays a minor role in cancer causation. Only strong environmental risk factors such as smoking seem to influence cancer development in adulthood. The proposed etiological map based on 24 cancer types identifies novel clusters - for example, non-Hodgkin lymphoma and leukaemia, bone cancer and myeloma - that are not completely explained by established risk factors. Some of the identified clusters relied on reproduced associations between cancer risks amongst husband and wives; however, the role of chance cannot be excluded. (C) 2010 Elsevier Ltd. All rights reserved.