Overview
A 2003 longitudinal cohort study by Hotopf et al. examining health outcomes among UK Gulf War veterans four years after initial assessment, comparing them to military personnel deployed to Bosnia and non-deployed "Era" controls.
Study Design
Participants: Military personnel who served in the 1991 Persian Gulf war (Gulf cohort), UN peacekeeping duties in Bosnia (Bosnia cohort), or were serving elsewhere without deployment (Era cohort). All had responded to a previous survey. The study used random stratified sampling with oversampling of women and all male veterans with fatigue scores >8.
Response rate: 71.6% overall, higher in Gulf cohort than other cohorts (P = 0.03).
Key Findings
Health Outcomes at Follow-up (Stage 2)
| Measure | Gulf Stage 1 | Gulf Stage 2 | Change |
|---------|--------------|--------------|--------|
| Fatigue cases | 48.8% | 43.4% | ↓5.4%
| Psychological distress (GHQ) | 40.0% | 37.1% | ↓2.9%
| Physical functioning (SF-36) | 90.3 | 88.7 | ↓1.6 |
Comparative Analysis
Gulf veterans continued to experience poorer health on all outcomes compared to other cohorts, though the gap narrowed slightly over time. The decline in physical functioning was observed across all three cohorts.
Incidence and Persistence Analysis
The study examined whether excess illness at follow-up resulted from higher incidence or greater persistence:
- Incidence: Era controls showed lower fatigue incidence than Gulf veterans (0.5 vs 1.0 odds ratio after adjustment). The difference for general health questionnaire caseness was explained by confounding.
- Persistence: Gulf veterans with fatigue were more likely to remain fatigued at follow-up compared to Bosnia and Era cohorts (corrected odds ratio = 0.7, P = 0.009).
Conclusions
"Gulf war veterans continue to experience symptoms that are considerably worse than would be expected in an equivalent cohort of military personnel. However, Gulf war veterans are not deteriorating and do not have a higher incidence of new illnesses." The authors note the illness may represent "the prodrome of a known disease... [that] has yet to declare itself" or remain poorly understood in terms of conventional biomedical diseases.
Methodological Notes
The study used weighted analyses accounting for sampling strategy. Limitations include potential response bias (lower follow-up rates among those with poorer health at stage 1) and self-report measurement biases.