Is Pancreatic Cancer a Preventable Disease?

Pancreatic cancer is one of the most formidable types of cancer a patient and his/her physician must face. These cancers are difficult to treat due to their inaccessible location, proximity to other vital organs, and inherently aggressive pattern of growth. Although advances in surgical techniques, radiation, and chemotherapy have provided incremental improvements in the length and quality of life, less than 5% of patients with pancreatic cancer will live beyond 5 years.¹ Therefore, it is both surprising and gratifying that pancreatic cancer should be emerging as a form of cancer that might be preventable, at least in part through modification of lifestyle habits such as diet, exercise, and smoking.

Cigarette smoking has long been recognized as an important determinant of pancreatic cancer risk² ; however, there are no other established, modifiable risk factors. The article by Michaud and colleagues³ in this issue of THE JOURNAL provides new information regarding the associations of height, obesity, and physical activity with risk of pancreatic cancer. The investigators used data from 2 well-designed, large cohort studies, the Nurses' Health Study and the Health Professionals Follow-up Study. Among 46 648 men aged 40 to 75 years and 117 041 women aged 30 to 55 years at baseline, 350 incident pancreatic cancer cases were identified during 20 years of follow-up. Analyses showed that tall height and greater body mass index (BMI) were independently and positively associated with pancreatic cancer risk, whereas physical activity of moderate intensity was inversely associated with risk. It is particularly interesting that among individuals with a BMI of less than 25 kg/m², total physical activity (both moderate and vigorous) was not related to risk of pancreatic cancer, whereas among overweight and obese individuals (BMI ≥25 kg/m²), low total physical activity was associated with a higher risk.

A higher risk of pancreatic cancer among persons with higher BMI and lower physical activity supports the hypothesis that hyperinsulinemia could play an important role in pancreatic carcinogenesis. Indeed, an association between abnormal glucose metabolism and pancreatic cancer has long been suspected. While in some cases, impaired glucose tolerance, clinical diabetes, or both appear to be a result of the tumor,^{4 - 7} a considerable amount of data now exist suggesting that in some cases these could be predisposing factors in pancreatic carcinogenesis. For example, a meta-analysis of epidemiologic studies estimated a 2-fold greater risk of pancreatic cancer associated with diabetes diagnosed at least 5 years prior to either diagnosis of pancreatic cancer or pancreatic cancer death.⁸ More recently, we reported a 2.2-fold higher risk of pancreatic cancer mortality for men and women whose postload plasma glucose level was 200 mg/dL (≥11.1 mmol/L) or more at baseline compared with those patients whose level was 119 mg/dL (≤6.6 mmol/L) or less.⁹

Two aspects of the findings of Michaud et al³ lend further indirect support for the insulin hypothesis. First, the form of activity most clearly associated with reduced risk was sustained exercise at a moderate intensity level, such as walking or hiking, rather than vigorous activity for shorter durations. This is consistent with the literature on the effects of various forms of exercise on the prevention or treatment of impaired glucose tolerance.¹⁰ Second, the group that appeared to benefit most from physical activity included those participants who were obese, followed by the group who were overweight, precisely those participants whose glucose tolerance and insulin levels could be improved, even in the absence of weight loss, by moderate exercise. Moreover, in a large case-control study of risk factors for pancreatic cancer, individuals who ate only 1 major meal a day (and presumably ate smaller amounts of food or little food throughout the day otherwise) had a 50% lower risk than those who ate 3 or more major meals per day.¹¹ It is possible that avoidance of large meals might be one way to reduce overall insulin secretion. The pathogenic relationship of pancreatic cancer to insulin could, in part, be explained by anatomy: the exocrine cells of the pancreas, which give rise to fatal pancreatic cancers,¹² are exposed to extremely high concentrations of insulin because their blood supply passes through the islet cell region.¹³

A growing body of intriguing evidence indicates that many pancreatic cancers could be prevented through lifestyle modifications. Silverman et al¹⁴ estimated that the proportion of pancreatic cancer cases attributable to cigarette smoking is approximately 25%. In the United States, the prevalence of current cigarette smoking decreased from 41.9% to 24.0% between 1965 and 1998.¹⁵ It is plausible that the decrease in smoking might have led to a decline in the incidence of pancreatic cancer allowing for a latency period. Data from the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute indicate that between 1973 and 1998 the average annual change in pancreatic cancer incidence was −0.04% for all men and women with the greatest decrease occurring in white men (−0.09%).¹ During the same period, the prevalence of obesity, including childhood obesity, increased quite dramatically.¹⁵ However, it is not yet clear how the secular trends in pancreatic cancer incidence have been affected by the increase in obesity.

Regardless, the establishment of obesity and perhaps physical inactivity as causal factors for pancreatic cancer could have a profound impact on the current and future estimates of the preventable fraction of this disease. Using conservative assumptions about relative risks and the prevalence of obesity and inactivity, such as those reported by Michaud et al,³ these factors could account for as much as 15% of pancreatic cancer cases beyond those attributable to smoking. It is important to note that if the actual relative risks have been underestimated, for example, due to random error in measuring physical activity or the use of indirect rather than direct measures of insulin exposure, the true preventable fraction could be even larger.

The results of the study by Michaud et al³ should be confirmed in other epidemiologic studies, but meanwhile, several other issues remain to be addressed. Specifically, it is important to determine whether pancreatic cancer risk is affected by the composition of the diet, in particular glycemic index, as well as the frequency and size of meals. In addition, if the hypothesis that exposure to high concentrations of insulin is correct, additional studies should determine if risk is greatest among those with abdominal obesity. Other investigations might also evaluate whether medications intended to treat glucose intolerance by promoting the secretion of insulin could play a role in the development of pancreatic cancer in certain patients. While these issues require additional exploration, it is increasingly clear that pancreatic cancer, challenging as it may be, could be prevented through behavioral and lifestyle changes.