A case of self-experimentation.

In this issue of Cancer Epidemiology, Biomarkers & Prevention, Cheng et a!. (1) report that two of the investigators in their study (M. Root and Z. Cheng) each ingested 1.0 g of AFB1,2 a group 1 hepatocarcinogen (2). The purpose of their experiment was to monitor the urinary excretion, over a 10-day period, of a metabolite of AFB1, AFM1 . This is the first reported instance in which humans were intentionally dosed with this toxin. Here, we examine the propriety of this experiment in the context of self-experimentation, in general. Self-experimentation has a lengthy, though largely unpublicized, tradition in medical research. Over the years, selfexperimenters have made many valuable contributions to modicine. For example, in 1929, Werner Forsmann made an incision in his own arm and inserted a catheter through his vein, directly into his heart, confirming his belief that the heart could be accessed directly without requiring invasive surgery (3). Although such a procedure could have resulted in severe complications or death, Forsmann came out of the procedure virtually unscathed. He ultimately won the Nobel Prize in Medicine and was praised by the Nobel Committee for his courage in using himself as his first subject. More recently, another important discovery through self-experimentation was made by Dr. Barry Marshall. Marshall was convinced that peptic ulcer disease had an infectious etiology, but his medical and scientific colleagues were skeptical. To test his hypothesis, Marshall ingested isolated Helicobacter pylon (4). Soon after, he developed gastritis, thus providing critical support for his theory. Fortunately for Marshall, he treated himselfwith antibiotics and apparently eradicated the infection. A less successful seif-expenment was performed by John Hunter, who infected himself with gonorrhea to better understand the venereal disease, only to find that he had also inadvertenfly infected himself with syphilis (5). This self-administered infection led to a life-long illness with syphilis, as well as decades of confusion within the scientific community regarding the etiology and disease spectram of gonorrhea. As these examples illustrate, the ultimate success or failure of self-experiments often has more to do with the particular luck of the researcher than the validity of the hypothesis; a self-experiment gone awry may derail the progression toward scientific discovery for years, as well as cause irreversible harm