METHODS

The sample was obtained from a cross sectional postal survey comparing the health profiles of three United Kingdom military populations, further details of which are described in our original paper.⁹ The three defined cohorts were (a) veterans deployed to the Gulf war; (b) veterans who served in the United Nations Bosnia peacekeeping forces between 1 April 1992 and 6 February 1997;and (c) veterans in active service on 1 January 1991, not deployed to the Gulf war (Era cohort). The subjects were sent a standardised postal questionnaire. After three mailings and telephone reminders, the response rate was 70.4% (n=3531) for the Gulf veterans, 61.9% (n=2050) for the Bosnia cohort, and 62.9% (n=2614) for the Era cohort. To assess potential response bias attempts were made to trace 200 participants who had not responded after two mailings. Characteristics of these non-responders were compared with responders in the main study.⁹

The survey included a shortened measure of symptoms of possible chemical sensitivity derived from the questionnaire of Kipen et al.¹⁰ Subjects were asked whether exposure to any of a list of 11 items brought about symptoms and whether this sensitivity developed subsequent to deployment to the Persian Gulf or Bosnia (or in the case of the Era cohort, after January 1991). A symptom was described as “an awareness of some discomfort or bothersome change—for example, sneezing, runny eyes, pain, swelling, nausea, or trouble concentrating”. Prevalences were calculated for reported sensitivity to each substance and odds ratios (ORs) with 95% confidence intervals (95% CIs) were calculated to enable comparison between cohorts.

RESULTS

Table 1 provides details of self reported chemical sensitivities in the three cohorts. Sensitivity to at least one trigger was reported commonly in all three cohorts but was notably more common among Gulf veterans. Of the Gulf cohort 978 (27.7%) indicated that they were sensitive to at least one trigger compared with 261 (12.7%) of the Bosnia cohort and 371 (14.2%) of the Era cohort (OR: Gulf v Era 2.3, 95% CI 2.0 to 2.6; Bosnia v Era 0.9, 95% CI 0.7 to 1.0). Among the veterans deployed to the Gulf or Bosnia, vehicle exhaust, air pollution, cigarette smoke, cosmetics, and organic chemicals were all often reported as causing symptoms. In the Era cohort, as well as air pollution, newsprint and new office buildings were more likely to trigger symptoms. For the total sample (including all three cohorts), the median number of reported sensitivities was zero, with two sensitivities at the 95th centile. When the cohorts were considered individually the median number of sensitivities remained zero, but the number of sensitivities at the 95th centile differed: Gulf, four; Bosnia, two; Era, one (fig 1). The distribution of reported chemical sensitivities differed significantly between the Gulf and Era cohorts (p<0.001, Mann-Whitney U test), but not the Bosnia and Era cohorts (p=0.34, Mann-Whitney U test).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Distribution of the numbers of chemical sensitivities reported in each military cohort.

DISCUSSION

This paper presents the first prevalence figures for reported chemical sensitivities in a United Kingdom population, albeit a military one. Strengths of the study include the large sample with a relatively high response rate. On the other hand, all responses were questionnaire based and as in previous studies, there was no objective measurement of reported sensitivities.^4,6,7 For our symptom measure we used a shortened version of a validated questionnaire of chemical sensitivity but this was compensated for by its ease of use (the original questionnaire includes 122 items). The higher response rate from Gulf veterans may reflect an increased concern about illness after deployment. Responders to the survey were more likely to be still in service but did not differ from traced non-responders in several health outcomes, including the proportion given medical discharges. Those non-responders who were followed up had worse health perception than responders, although the difference was not significant.⁹

Reported chemical sensitivities were common in all three military cohorts. There have been few previous population based estimates of chemical sensitivities. Bell et al reported a 15% prevalence¹¹ and Meggs et al, 33%¹² but their findings were based solely on reports of feeling ill after smelling various chemical odours. Kreutzer et al conducted the first rigorous large scale survey in which 15.9% reported being “allergic or unusually sensitive to everyday chemicals”, and 11.9% reported sensitivity to more than one type of chemical.⁶ The similarity with our estimate (14.2% in the Era cohort) is surprising given the differences between the populations. We would expect the military sample to be a healthier population than that of Kreutzer et al, and the sample was overwhemingly male (92%). A possible explanation may be that occupational exposures—such as diesel fumes and pesticides, which were commonly reported in the military cohorts,⁹—are likely to be less prevalent in a civilian population. Occupational exposures have been highlighted as a potential cause of chemical sensitivities¹³ although knowledge of occupational exposure goes hand in hand with recall bias, which may also be of relevance in explaining differences between the Gulf and Era/Bosnia cohorts.¹⁴

A striking, secondary finding is that nearly 30% of the veterans of the Gulf conflict reported at least one chemical sensitivity after their deployment. The Iowa study found that Gulf veterans were twice as likely to report illness due to chemical sensitivity when compared with those not deployed (12.8% v 6.1%).⁷ This finding is explored further in another paper,¹⁴ but it reinforces findings reported previously that veterans of the Gulf war report significantly more symptoms of ill health than soldiers not deployed to the Gulf.^9,15 The reasons for this are, as yet, unclear.

The prevalence estimates from this study suggest that a significant proportion of United Kingdom veterans experience problematic sensitivity to everyday chemicals. Previously little priority has been given to researching chemical sensitivities, as is shown by the lack of studies outside of the United States. Objective evidence for a causal link between low level chemical exposures and reported symptoms is lacking, and current publications do not provide definitive answers to these questions. Further studies in the general population are now required, taking account of both health beliefs and chemical exposures, if we are to gain an understanding of the development of chemical sensitivities.