Acute effects of nitrogen dioxide after accidental release.

OBJECTIVES: Following an accidental release of nitrogen dioxide from a railroad tank car containing nitrous tetroxide, the authors undertook a study of the health effects of the release, measuring the association between acute low level exposure and pulmonary symptoms.
METHODS: The authors reviewed the records of three emergency departments, surveyed 80 emergency department patients, 552 community residents, 21 chemical plant workers, and 29 emergency workers, and conducted a case-control study. Pulmonary case status was defined as having an objective pulmonary finding noted on the emergency department record, reporting that the onset of symptoms was subsequent to the release, and being within the city limits at the time of the release. Self-reported case status was defined as reporting one or more symptoms consistent with exposure to nitrogen dioxide in the week after the release and having been within the city limits at the time of the release. Control subjects were survey respondents who reported no symptoms in the week after the release and had been within the city limits at the time of the release. Chemical exposure was characterized by proximity to, direction from, and being outdoors within one hour after the release. Duration of potential exposure was not measured. Logistic regression was used to estimate odds ratios and 95% confidence interval for symptoms by exposure level, adjusted for age, sex, smoking, and preexisting pulmonary conditions.
RESULTS: Local emergency department visits increased fivefold in the week after the release. The most common complaints recorded in a systematic sample of 528 visits in the first 30 hours after the release were headache (31%), burning eyes (30%), and sore throat (24%). Objective pulmonary findings were recorded for 41 (5%) patients in the week before and 165 (4%) in the week after the release. The odds of being a pulmonary case increased by 40% for each quarter-mile increment in proximity to the release (odds ratio [OR] 1.4; 95% confidence interval [CI] 1.1, 1.7), while the odds of being a self-reported case increased by 20% for each quarter-mile increment in proximity (OR 1.2, 95% CI 1.1, 1.4). People who met the pulmonary case definition were 2.5 times (CI 1.3, 4.8) more likely than control subjects to have been outdoors and 6.4 times (CI 3.2, 12.6) more likely to report a preexisting pulmonary condition. Self-reported cases were 2.6 times (95% CI 1.8, 3.8) more likely than control subjects to have been outdoors and 1.9 times (95% CI 1.1, 3.1) more likely to report a preexisting pulmonary condition.
CONCLUSIONS: Emergency department visits increased five-fold, but serious acute health effects were uncommon. People who met the pulmonary case definition were six times more likely to report pulmonary symptoms than those without preexisting conditions. This study was not designed to determine any potential long-term effects of exposure.