tions. Controlled epidemiologic studies provide the best evidence for the examination of associations of vaccination and subsequent adverse events. In evaluations of the studies presented to the committee, additional weight was given to those that (1) used active surveillance rather than self-reports of postimmunization events; (2) included sufficiently large numbers of subjects; (3) had clearly specified, objective criteria for the definition of adverse events; and (4) had sufficiently long postimmunization follow-up intervals to allow identification of later-onset events. Those studies that included a suitable unimmunized comparison group or in which evaluators were blinded to the subjects’ vaccination status were especially useful to the committee.

Among the concerns that have been raised about the anthrax vaccine is that contaminants in the vaccine product are producing adverse health outcomes. Laboratory analyses have been conducted to test for the presence of two suggested contaminants, mycoplasma and squalene.

Mycoplasma contamination of the anthrax vaccine has been suggested as a possible cause of illness among Gulf War veterans (Nicolson et al., 2000). Mycoplasmas are a distinctive type of bacteria that lack cell walls (Baseman and Tully, 1997). Many mycoplasma species appear to be harmless constituents of the normal human microbial flora, but others are associated with various diseases, including atypical pneumonia, genitourinary infections, joint diseases, and opportunistic infections in persons with compromised immune systems. Mycoplasma contamination is considered possible in vaccines produced in cell cultures, and FDA requires testing to demonstrate the absence of such contamination (Hart et al., 2002). For vaccines like AVA that are not derived from cell cultures, contamination is considered unlikely. However, to respond to the concerns about AVA, DoD commissioned two nonmilitary laboratories to conduct studies with samples from five AVA lots. The samples were obtained from eight different DoD vaccination clinics.

Hart and colleagues (2002) have reported on the results of those tests. Tests for the presence of live organisms were conducted at the National Cancer Institute’s Mycoplasma Laboratory, which is located in Frederick, Maryland, and operated by Science Applications International Corporation. No mycoplasma colonies could be cultivated from the vaccine samples tested. In addition, tests of a vaccine sample deliberately inoculated with mycoplasma showed no presence of viable organisms after 24 hours. Sepa-