Following significant advances in imaging to aid in patient selection for definitive treatment and improvement in surgical technique and perioperative care, prognosis for resectable pancreatic cancer has improved appreciably. Chemotherapy has a significant role to play in selected cases [18
]. However, it does appear that further significant improvement in outcome from the illness will be directly related to the ability to detect the disease early and institute prompt management. This will require identification of high-risk groups in whom targeted screening can be employed and early or precursor lesions recognized [19
] and this has been demonstrated successfully in familial forms of the disease [20
] and has been found to be cost-effective [21
In our prospective hospital-based case-control study, we have seen that pancreatic cancer patients smoked more than our control group and an ever-smoker individual had a 3-times higher risk for the development of pancreatic cancer than a non-smoker. These are well-recognised findings. In addition, however, there were other significant results; smokers on average developed the cancer about 6-7 years earlier than nonsmokers which was independent of a family history of malignancy and has been previously described on the basis of WHO cancer mortality data and SEER cancer incidence data [22
]. More importantly a family history of malignancy in first-degree relatives appeared to decrease the amount of tobacco exposure (as measured by pack-years) required for the development of pancreatic cancer. The earlier onset of the disease was however not related to FDR status.
It is accepted that familial pancreatic cancer appears to develop at an earlier age as compared to its sporadic counterpart, and tobacco exposure is the most important factor influencing the penetrance of the FPC gene [14
]. Smokers in FPC [23
] and in hereditary pancreatic cancer syndromes, specifically hereditary pancreatitis patients [25
], develop the disease about 10 years earlier, demonstrating the interaction between an inherited susceptibility to cancer and an environmental carcinogen. A recent report has described gene-environment interaction in a study of cases only, although the sample size was large [26
]. We have now shown for the first time that smokers who also have a family history of cancer develop the disease at a lower level of exposure. In smokers, the disease also appears to develop earlier. This might be due to continued or faster accumulation of genotoxic mutations secondary to a variety of factors, one of which might be an inefficient DNA repair mechanism. Other genetic and environmental factors might play a role and this will need further elucidating. For example, a recent report has shown an earlier age of onset of pancreatic cancer in those who had a high BMI during their teen and younger years [27
The groups of index cases and controls with and without a family history of cancer were comparable given their similar age distribution and gender distribution. We have obtained history of cancer in FDR from index cases and controls and it is known that such information is reliable and accurate especially with regard to FDRs [28
]. The reliability of information obtained, however, decreases with regard to other relatives [29
], and we have therefore restricted our study to data on first-degree relatives. It has been suggested that, if anything there is under reporting of family history of cancer especially with regard to individuals with colorectal neoplasms [31
]. Other details of the illness in the FDR such as age of onset (of the cancer in the relative) are unreliable especially in older probands and we have therefore not utilised such data in our study [29
]. We have not performed genetic analysis in this group of patients to confirm that they are not familial cancers as most familial pancreatic cancers are not due to known mutations. It is likely that our patients represent sporadic malignancies due to the fact that the age distribution of the group of patients is normal and there was no difference in the mean (SD) of the age at diagnosis of the index cases in the FDR+ and FDR− groups (65.93 (10.67) and 64.57 (12.38) years).
We have also demonstrated in this prospective group of patients, that those with a family history of cancer as evidenced by the occurrence of a malignancy in an FDR are not only at twice the risk of developing pancreatic cancer (OR 1.98 (95% CI: 1.15–3.38)) but more importantly require less of a genotoxic exposure as compared to those who do not have such a genetic vulnerability (). Just under 2/3rds of FDR+ index cases (n
= 36; 59%) had just a single first-degree relative with malignancy. In the FDR+ group, there was a decreased tobacco exposure required for the development of adenocarcinoma of pancreas but this did not depend upon the number of relatives with malignancy, as the FDR > 1 group did not demonstrate a significantly decreased cumulative tobacco exposure. It is well accepted that a family history of cancer is a risk factor for most cancer types. With respect to adenocarcinoma of the pancreas, a recent meta-analysis of seven case-control and two cohort studies involving 6,568 pancreatic adenocarcinoma cases concluded that a family history of adenocarcinoma of the pancreas conferred double the risk (1.80 (95% CI: 1.48–2.12)) for the disease in individuals with such a history compared to those without [32
]. A recent cohort study from the PanScan consortium [33
] and prospective followup of participants of Cancer Prevention Study-II [34
] have suggested an association between family history of various cancers especially prostate cancer and pancreatic cancer. An important additional finding from our study is confirmation that the presence of any malignancy in FDR, apart from dermatological and primary brain malignancies, appears to confer an increased risk for pancreatic adenocarcinoma. We have not performed specific FDR malignancy associated risk analyses in view of the small size of our study population. This is, however, intended for the future when a sufficiently large number of cases have been accrued.
In the presence of a family history of malignancy (i.e., increased susceptibility), a decreased dose of an environmental carcinogen is sufficient to cause cancer (cumulative tobacco exposure in FDR+ (30 (24.77) versus FDR− (44.69 (28.47) (P
= .00)). It is possible that the decreased tobacco dose demonstrated in the caFDR+ group is due to a genetic or other environmental factor which potentiates the genotoxic effect of tobacco-derived carcinogen by either impairing the processing of tobacco-derived carcinogen into inactive metabolites or causing the inefficient or incomplete repair of genetic damage induced by it. Genetic factors such as poor DNA repair, impaired carcinogen metabolism and environmental factors may interact in the development of tobacco-related cancers, including that of the lung, bladder and head and neck [35
]. There is some evidence for this in pancreatic carcinogenesis too from molecular epidemiological studies: the presence of XRCC2 Arg188His polymorphism modulates risk for pancreatic cancer amongst smokers [39
]; XPD gene polymorphisms—exon 10 Asp(312)Asn and exon 23 Lys(751)Gln polymorphisms—influence risk for smoking associated adenocarcinoma of the pancreas [40
]; XRCC1 399Gln allele determines susceptibility to smoking induced pancreatic cancer [41
]; deletion polymorphism in GSTT1 is associated with an increased risk of adenocarcinoma of the pancreas amongst Caucasians [42
]. None of these studies, however, has ascertained the risk for smokers carrying these genotypes in the presence of a family history of malignancy. Our findings point to the presence of a high-risk group for adenocarcinoma of the pancreas. This cohort needs further characterisation and replication in larger population based and molecular epidemiological studies.
Identifying risk might help stratify individuals for pancreatic cancer screening but screening is not well established, the pickup rate is low and the false positive rate is relatively high. Surgery usually means a total pancreatectomy with all its potential complications. If we are able, however, to better quantify the risk, the benefits might be greater and identifying genetic and environmental factors is important. With the completion of the human genome project and advances in molecular epidemiological techniques, these low penetrant/polymorphic genes should become more frequently identified and their function understood; for example, genome-wide association studies have identified smokers with a non-O blood group as a significant high risk group for pancreas cancer as compared to nonsmokers of non-O blood group (OR 2.68 (95% CI: 2.03–3.54)) [43
]. Similarly identification of high-risk groups such as smokers with a positive family history of cancer could have implications for the earlier diagnosis by making screening for the disease possible leading to the prospect of long-term survival if not cure for more patients.