Giovannucci and colleagues [1] do not present biologically plausible mechanisms in their effort to define a protective effect of aspirin against colorectal cancer. Evaluation of possible mechanisms is essential for the design of future studies and preventive strategies and for understanding the pathophysiology of colorectal cancer. An increasing biologically plausible possibility [2, 3] is that prolonged use of aspirin protects by causing progressive depletion of stored iron. In an iron-depleted state, less iron is available for carcinogenesis through free-radical-mediated mechanisms and for direct promotion of tumor growth.

Even small doses of aspirin are associated with increased blood loss from the stomach [4]. With repeated daily use of aspirin over a period of years, substantial losses of stored iron are likely. A decrease in the stored iron level is possible even in patients with bleeding that is detectable only by very sensitive methods. Amounts of blood lost because of regular aspirin use [4] can cumulatively exceed menstrual blood losses. Averaged over the month, typical menstrual blood losses amount to 1 to 2 mL per day. A single daily aspirin tablet causes approximately 1 to 10 mL of blood loss per day, with much individual variation [4].

Evidence for improving effectiveness with duration of use [1] suggests that more than the immediate antiplatelet effects of aspirin may be involved. In cases of prolonged use, aspirin may have an important second mode of action, that is, induction of iron depletion. Progressive iron loss from mucosal microbleeding, generally considered an undesirable side effect, could be a major protective mechanism of prolonged aspirin use. Aspirin-induced iron loss may be protective against both cancer [2, 3] and heart disease [5]. Establishing iron loss as a protective effect of prolonged aspirin use is important because iron loss without an increased risk for hemorrhage can be achieved by other methods, such as regular phlebotomy.