SYNOPSES OF PAPERS
In the paper by Seeff, we learn about a study whose purpose was to examine whether hepatitis B infection in an area of low endemicity conveys the same risk of subsequent liver cancer as it does in endemic areas. The opportunity for this study was an outbreak of hepatitis among World War II (WW II) Army servicemen involving some 50,000 cases; it is now known, because of this study, that this outbreak was the largest point-source outbreak of hepatitis B ever recorded. Seeff reports limited case-control mortality data as well as serologic data on three groups: those servicemen who received hepatitis-contaminated yellow fever vaccine and were hospitalized, those who received the vaccine and who were asymptomatic, and a third group who did not receive the contaminated vaccine. In the serologic study, a large proportion of subjects in the first two groups had markers for the hepatitis B virus--97.7 percent and 77 percent, respectively, versus 13 percent for the control group. Of greater interest was the fact that hepatitis B surface antigen was found for only one person in the first two groups, yielding an overall carrier rate of less than 0.5 percent, unexpectedly low given the prevailing view that 5 to 10 percent of acutely hepatitis B virus-infected persons are supposed to become carriers. It is not surprising that this striking result has begun to overturn conventional thinking in hepatitis research.
In the Carmelli et al. paper, heritability estimates for tobacco, alcohol, and coffee use are produced using data from the Medical Followup Agency's (MFUA) large panel of WW II twins. The heritability estimates produced are very much in line with similar, previous estimates, yet they have the additional feature that they are derived using multivariate statistical models as well as the usual univariate ones. It is the size of the twin panel as well as the availability of longitudinal covariate data (which are used to adjust the heritability estimates) that permits these more powerful analyses, illustrating the fact that even when veteran status, per se, is relatively unimportant, veteran cohorts can nevertheless provide significant opportunities for research. But it should be noted that in at least one regard the veteran twins in the MFUA panel are not so typical:
their reported history of tobacco use (“ever smoked”) was a very striking 82 percent. This high rate is attributed to the distribution of free cigarettes to WW II soldiers.
In the paper by LeDuc et al., we learn of a “very interesting set of sera” collected by the Hemorrhagic Fever Commission, formed by the Army during the Korean conflict to deal with the “new” disease encountered by U.S. troops in Korea. In the subsequent decades, the cause of this new disease was identified and named (“hantaan virus,” after the Hantaan River in the endemic area of Korea), and a serologic test for the agent was developed. Thus did these sera become very interesting: it was now possible to test them with the new assays and determine whether the Korean conflict epidemic was actually due to the newly identified hantaan virus. LeDuc et al. report that it was--some 94 percent of the time the original clinical diagnoses were accurate. Having verified the clinical diagnosis, it is now possible to follow up the men who provided these sera and study the long-term sequelae of hantaan virus infection.
Moreover, not only is all of this fascinating from a medical point of view, but there are also potential public health ramifications. Specifically, studies of Korean patients residing in urban centers and diagnosed as having hemorrhagic fever with renal syndrome led to the identification of a new virus, now named the Seoul virus, related to but distinct from the hantaan virus. Of special interest was the fact that this new virus had been found in domestic rats rather than in its usual vector, the striped fieldmouse. Studies of the rat population in the inner city of Baltimore have shown that these Seoul-like viruses are also common there, and studies of inner-city Baltimore residents done at Johns Hopkins Hospital found that they had a five-fold higher prevalence rate of seropositivity for this virus; the most common diagnosis among seropositives in the group was hypertensive renal disease. If past hantaan viral infection is associated with subsequent development of chronic renal disease, as the evidence in this paper suggests, then such illnesses represent a multi-million-dollar public health problem.
The paper of Engdahl and Page reports data from one of the unique cohorts discussed earlier. Data are presented on depressive symptoms showing that former prisoners of war (POWs) still have notable psychiatric sequelae nearly forty years after their release from captivity, a three- to five-times-higher prevalence of depressive symptoms than expected. More important, this higher rate of depressive symptomatology is statistically linked with severity of treatment during captivity and with demographic factors such as years of education and age at capture. Such results not only have relevance to the current medical treatment of POWs but also
provide valuable insights into post-traumatic adaptation in the more general sense.
The paper by Kang et al. (which appeared in the January 1991 issue of the Journal of the American Public Health Association) makes innovative use of general population data to study the health effects of dioxin on Vietnam veterans. In contrast to the kinds of studies discussed earlier, wherein data from military or veteran populations are important because of their applicability to the general population, in this study the general population data are important because they pertain to a certain subset of military veterans. Notwithstanding the irony of this particular situation, the use of routinely gathered samples from the general population to study the effects of environmental exposures in a special subpopulation is commendable.
The data themselves come from a bank of adipose tissue samples assembled under the auspices of the U.S. Environmental Protection Agency. Although the target population for the samples was meant to be all non-institutionalized persons in the United States, the invasive nature of adipose tissue collecting limited the actual sampled population to decedents who died from external causes (90 percent) and surgical patients (10 percent). The samples were then analyzed to determine the presence of 2,3,7,8-TCDD (tetrachlorodibenzo-p-dioxin), one of the toxic contaminants of Agent Orange, a defoliant sprayed on parts of Vietnam during the Vietnam conflict. The records of the Agent Orange spraying missions (the so-called HERBS tape) were used to derive an Agent Orange exposure index. The study found that mean dioxin levels did not differ among Vietnam veterans, non-Vietnam veterans, and civilian controls, with or without adjustment for confounding factors. In addition, none of the surrogate measures of Agent Orange exposure was associated with adipose tissue dioxin levels.
The last paper presented at the conference was Jablon's paper on radiation risk studies in military populations. In this insightful paper, Jablon divides studies of radiation risk in humans into two classes: “scientific studies” and, for lack of a better term, “population studies.” This classification is central to the paper, for the two kinds of studies yield very different kinds of information--in the first case, quantitative information concerning the risk of radiation carcinogenesis, and in the second, evidence (or lack of it) that some given radiation exposure was likely to have caused excess cancer. Studies of veterans, Jablon asserts, generally fall into the second class, due to the fact that accurate exposure data are seldom available. Notwithstanding the difficulty, if not impossibility, of deriving quantitative radiation risk data from studies of military veterans, Jablon goes on to identify some of the advantages of studying military veterans, pointing out the now familiar themes of (1) the
availability of large, well-identified cohorts and (2) the potential for large-scale, relatively quick-and-easy mortality follow-up. He further concludes that studies of radiation carcinogenesis in veterans are of strategic importance because very large populations must be studied over very long time intervals, and because the federal government has undertaken responsibility to compensate individuals for injuries from military service.
Finally, although a panel discussion was held on opportunities for research in military and veteran populations, formal written presentations by the panel members were not submitted. In lieu of a transcript of the panel's remarks, these proceedings include a paper kindly provided by Dr. Kang on the data sources he uses in his studies of veterans. This paper describes three automated information systems, the Beneficiary Identification and Records Locator Subsystem (BIRLS), the Patient Treatment File (PTF), and the Agent Orange Registry, and discusses not only the information these data bases provide but also their strengths and weaknesses.
The BIRLS file is a huge file of VA beneficiaries that may be used to ascertain the vital status of war veterans and to obtain copies of their death certificates; the completeness of its mortality ascertainment, however, is still under study. The PTF file is likewise a large file, but it contains only computerized VA hospital discharge information (VA hospitalization is not a “benefit,” so BIRLS and PTF are independent files). PTF has served as the source of a number of clinical studies, but the quality of its diagnostic information requires that studies include independent diagnostic verification: only about one-half of putative cancer diagnoses were verified during hard-copy record review in the examples Dr. Kang cites. The Agent Orange Registry is a different kind of file altogether, being a record of a special health examination of some 200,000 Vietnam veterans who presented themselves to a VA medical center. Each of these resources can be used either as a sampling frame for studies or a source of morbidity or mortality follow-up information.
wfp
