specific exposures associated with this syndrome (see Chapter 2). Because these studies used an incomplete study design (i.e., lacked an unexposed comparison group) and demonstrated other weaknesses, they have been strongly criticized on methodological grounds (Cowan et al., 1996; Gordon et al., 1997; Landrigan, 1997; Gray et al., 1998; Wolfe et al., 1998). This section reviews the Haley studies and other epidemiologic investigations of the association between Gulf War veterans’ symptoms and PB exposure.

In an initial survey designed to search for syndromes characteristic of Gulf War veterans, Haley et al. (1997b) studied the questionnaire responses of 249 men (41 percent of 606 males from a reserve naval mobile construction battalion [i.e., Seabees]) living in five southeastern states. All respondents had been called to active duty during the Gulf War, and there were no survey responses from nondeployed personnel. Characteristics of the participants (n = 249) and non-participants11 (n = 357) of the 24th Reserve Naval Mobile Construction Battalion (RNMCB-24) indicate that members of the battalion were, on average, older than most deployed forces, with a mean age of 41 years for participants and 37 years for nonparticipants. Participants and nonparticipants were similar in race or ethnicity, education, active reserve status, wartime military rank, and wartime job ranking (Haley et al., 1997b). However, large differences between participants and nonparticipants, respectively, were noted for percentage reporting serious health problems since the war (70 percent versus 43 percent) and percentage unemployed at the time of the survey (11 percent versus 3 percent). Of 249 individuals who responded to the survey several years after deployment, 145 (58 percent) had retired from the military, and the rest were still serving in the battalion. Symptoms included in the survey were those commonly associated with post–Gulf War illness in clinical examinations performed by teams of DoD and Department of Veterans Affairs physicians. From survey responses, the authors used factor analysis12 to identify six clinical syndromes: (1) impaired cognition, (2) confusion–ataxia, (3) arthromyoneuropathy, (4) phobia–apraxia, (5) fever–adenopathy, and (6) weakness–incontinence.

Psychological testing indicated that veterans with any of the six syndromes had the same psychological profile, which differed only in clinical severity but did not represent posttraumatic stress disorder. Those with syndromes 2 (confusion– ataxia) and 4 (phobia–apraxia) had increased self-reported occupational disability compared to the others. The low participation rate (41 percent) of veterans in the battalion suggests that results may have been affected by selection bias, in that


This term refers to members of the battalion who were contacted by investigators but chose not to participate. Information about nonparticipants was obtained from the wartime roster and from a telephone survey of randomly selected nonparticipants.


Factor analysis is a mathematical–statistical technique used to define the primary dimensions from batteries of test items. Factor analysis aggregates survey responses into statistical groupings of factors that may or may not have biological plausibility or clinical relevance. It is recognized that factor analysis has the potential to generate syndromes that may not be reproduced when a new population is examined.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement