There are three arms to the study: (1) a serologic survey of representative veterans in three comparison groups; (2) a case-control study of discharges from Veterans Administration hospitals for liver cancer vs. several other major forms of cancer; and (3) a cohort study of mortality based on death certificates and review of available clinical and pathology reports for deaths in the period 19461983. The first phase, published in The New England Journal of Medicine in 1987, involved the comparison of laboratory findings in three groups: (1) men hospital- ized with acute hepatitis in 1942; (2) men vaccinated from contaminated lots of yellow fever vaccine who did not become ill in 1942; and (3) men who entered the Army after contaminated vaccine had been withdrawn. The second phase foundered on small numbers arising from the inadequacy of diagnostic indexing by Veterans Administration hospitals. The third phase makes use of the mortality in three cohorts of about 20,000 each, defined as in the first phase. Only a single carrier was found in the serologic survey, and the mortality survey provides no evidence to believe that the carrier rate in the critical second group was in any way exceptional. The small excess mortality from primary hepatic cell cancer is inconsistent with expectation based on studies in Taiwan and elsewhere. Overall, the findings suggest that the likelihood of liver cancer following infection with the hepatitis B virus depends on age at infection and is low for healthy young males.
