|Home | About | Journals | Submit | Contact Us | Français|
D. Coggon. MRC Epidemiology Resource Centre, University of Southampton, UK
ObjectivesTo develop and deliver a voluntary testing program for UK veterans of the 1990/91 Gulf War and the Balkans campaign, who wished to know whether they had been significantly exposed to depleted uranium (DU).
MethodsUnder the supervision of the Depleted Uranium Oversight Board, a test was developed based on the measurement of uranium isotopes in 24 h samples of urine by highly sensitive mass spectrometry methods. Initial pilot work showed that the test was capable of detecting an increase in the U238:U235 ratio as low as 144 in samples containing less than 5 ng per litre uranium (the U238:U235 ratio of natural uranium being 137.9). Biokinetic calculations indicated that this level of sensitivity would be more than adequate to detect DU excretion in veterans that (according to mainstream medical opinion) would be of practical relevance to health, either through radiation effects or through chemical toxicity. The main testing program was widely advertised, and was available on request to eligible veterans during September 2004 to January 2006. The continuing performance of the two analysing laboratories was monitored by blinded analysis of duplicate samples, and of spiked samples prepared by a third laboratory.
ResultsFindings from the duplicate analyses and from blinded analysis of spiked samples confirmed that the expected performance of the two analysing laboratories was maintained over the course of the testing program. None of the 464 veterans who were tested in the main program had detectable exposure to DU. Total 24 h excretion of DU exceeded 30 ng in eight samples, the highest value being 497 ng.
ConclusionBecause participants were self‐selected, their exposures may not have been representative of all who were eligible for testing. Nevertheless, the findings suggest that clinically significant exposure of British military personnel to DU in the Gulf War and Balkans, if it occurred at all, was relatively uncommon. The occasional high excretion of natural uranium presumably reflected higher than average dietary or other environmental exposures.
Key wordsbiomarker; depleted uranium; military
B. G. Miller1, A. P. Colvin1, P. A. Hutchison1, H. Tait1, S. Dempsey1, D. Lewis2, C. A. Soutar1. 1Institute of Occupational Medicine; 2Institute of Naval Medicine
ObjectivesThe UK Ministry of Defence controls a biological monitoring program that includes testing for uranium in personnel who served in the conflict in Iraq in 2003. To help interpret the results, the MoD commissioned this study, to quantify the distribution of urinary uranium concentrations and isotope compositions, in personnel who had not served in that conflict.
MethodsThe study used a cluster sampling approach, to select and visit bases where we might recruit a representative mix of ranks, genders and occupational groups (combat, support and auxiliary). A standardised protocol and recruitment questionnaire were used. The urine samples collected were analysed for uranium and creatinine concentrations and (where possible) for uranium isotope 238U/235U ratio.
ResultsIn all, samples from 732 eligible subjects were analysed. Uranium concentrations ranged up to >400 ng/l, somewhat higher than reference values quoted for the USA, but much lower than recorded in granite areas (eg, Finland). Isotope ratio measurements were available for samples with the highest concentrations; these all had a natural isotope signature, and no evidence of depleted uranium. On average, urinary uranium concentrations were lower in officers than in other ranks; they differed also across the services, the Navy being lowest and the Army highest.
ConclusionThe levels give no concern for health risks in the personnel studied. Since even the highest values were from natural sources, we assume the observed differences represent differences in ingestion of natural uranium. The data could be used to construct reference distributions, but these would need to be tailored to the distribution by service and rank of the target population.
Key wordsuranium; urinary excretion; depleted uranium
M. R. Sim1, H. Kelsall1, D. McKenzie1, K. Leder1, J. Ross2, J. Ikin1, A. Forbes1. 1Department of Epidemiology and Preventive Medicine, Monash University; 2Australian Defence Force
ObjectivesRecall bias can be a serious limitation in studies of war veterans, especially in cross‐sectional studies where exposure is often measured by self‐report, due to the absence of reliable records. In our study of Australian Gulf War veterans we have shown some small, but significant, associations between vaccinations, based on self‐report, and several health outcomes. In the current study we aimed to validate the reported vaccination data in those veterans who had recorded data and to repeat the analyses with health outcomes, using the recorded data.
MethodsOf the 1456 Australian Gulf War veterans who took part in the original study, 654 provided their WHO vaccinations booklet for copying. The reported (collected by postal questionnaire) and recorded vaccination data for these subjects were compared. Some analyses undertaken in the original study using reported vaccination data were repeated using the recorded vaccination data.
ResultsThere was considerable variability in agreement for individual vaccinations, with substantial agreement for plague and polio, moderate agreement for hepatitis B, hepatitis A, typhoid and pertussis, and fair to slight agreement for anthrax cholera, MMR and tuberculosis vaccines. There was fair agreement between reported and recorded vaccinations for different deployment‐related time periods and for vaccination clustering (>5 vaccinations within a week or less). Our previously found associations of increasing number of reported vaccinations in dose–response relationships with total number of symptoms, functional impairment and poorer physical health measured by the SF‐12, were not observed when recorded vaccination data were used.
ConclusionOur findings demonstrate that recall bias was an important factor in our previous analyses of health outcomes and reported vaccination data. Future veteran health studies should not rely on reported vaccination data and analyses should be based on independently recorded vaccination data.
Key wordsveteran health; vaccinations; validation study
V. Baste1, T. Riise2, B. E. Moen2. 1Section for Occupational Medicine, Unifob; 2Section for Occupational Medicine, University of Bergen
ObjectivesConcern is growing about exposure to electromagnetic fields and male reproductive health. Many Royal Norwegian Navy employees are exposed to non‐ionising radiation while working near equipment emitting electromagnetic radiation.
MethodsIn a cross‐sectional study among military men employed in the Royal Norwegian Navy (1950–2002), information about work close to equipment emitting radio‐frequent electromagnetic radiation (high frequent aerials, communication equipments, radar), 1‐year infertility and children was obtained from a paper questionnaire in a large survey of work and health.
ResultsAmong 10497 respondents (response rate 63%), 22% had worked close to high‐frequent aerials to a “high” or “very high” degree. Reported infertility increased significantly along with increasing exposure to radiofrequency electromagnetic fields. In a logistic regression, odds ratio (OR) for infertility among those who had worked closer than 10 m to high‐frequent aerials at a “very high” degree, relative to those who reported no work near high‐frequent aerials was 1.86 (95% CI 1.46 to 2.37), adjusted for age, smoking habits, alcohol consumption and exposure to organic solvents, welding and lead. Similar adjusted OR for those who had worked closer than 10 m to high‐frequent aerials to a “high”, “some” and “low” degree were 1.93 (95% CI 1.55 to 2.40), 1.52 (95% CI 1.25 to 1.84) and 1.39 (95% CI 1.15 to 1.68), respectively. When analysing the subgroup of those reporting infertility who not have biological children the ORs were even higher. However, the degree of exposure to radiofrequency radiation and the number of children were not associated.
ConclusionAn increasing degree of self‐reported work close to equipment emitting radio‐frequent electromagnetic radiation among navy personnel was significantly linearly associated with increasing infertility. For those who could be considered involuntarily childless, the association was even stronger.
Key wordsmale infertility; radiofrequency electromagnetic fields; occupational exposure
K. M. Venables, L. M. Carpenter. University of Oxford
ObjectivesThe UK government has had a program of research at Porton Down since WW1 on military capability under chemical warfare conditions, including the use of countermeasures. 30000 people took part in a “service volunteer programme” and it is unclear if this has led to any long‐term health effects. In 2002 we started the epidemiological investigation; research into the history and ethics is also underway at the University of Kent.
MethodsIn 2002–03, we carried out a review of the available exposure records, a pilot exposure assessment on a sample of 150, a pilot of data linkage with national mortality records for a second sample of 500, and a postal survey of 436 self‐selected support group members using the SF‐36 and symptoms questionnaires used previously for Gulf War veterans. In 2003–07, we carried out a cohort study of 19247 veterans (1941 to 1989), taking follow‐up to 2004 and incorporating an exposure assessment which focused on vesicants (eg, mustard gas) and nerve agents (eg, sarin). This included a comparison group of similar veterans who did not go to Porton Down.
ResultsThe exposure records were substantially complete and an exposure assessment was carried out for 17304 veterans. Over the period, 690 different chemicals were used. Over 14000 veterans were exposed to at least one in IARC groups 1, 2a or 2b. However, at an individual level, exposures were, in general, short, few, and included respiratory or other protection. The support group members reported higher scores in the SF‐36 than comparable members of the public but showed no definitive cluster of specific morbidity. Data linkage proved feasible, utilising identifiers abstracted from military personnel files, and, in the cohort study, 7305 had died up to 2004.
ConclusionIt proved feasible to carry out a cohort study of a large and representative group of these veterans and to select comparison veterans. The findings will help inform UK policy on veterans' health. The findings may also have some applicability in other settings, for example in understanding the long‐term effects of nitrogen mustard cancer chemotherapy, or in civil defence planning for deliberate releases of chemical warfare agents.
Key wordshuman experimental data; cohort studies; chemical warfare agents