Occupational exposure to methyl tertiary butyl ether in relation to key health symptom prevalence: the effect of measurement error correction

摘要

In 1995, White et al. reported that methyl tertiary butyl ether (MTBE), an oxygenate added to gasoline, was significantly associated with key health symptoms, including headaches, eye irritation, and burning of the nose and throat, among 44 people occupationally exposed to the compound and for whom serum MTBE measurements were available (odds ratio (OR) = 8.9, 95% CI = [1.2, 75.6]). However, these serum MTBE measurements were available for only 29 per cent of the 150 subjects enrolled. Around the same time, Mannino et al. conducted a similar study among individuals occupationally exposed to low levels of MTBE and did not find a significant association between exposure to MTBE and the presence of one or more key health symptoms among the 264 study participants (OR = 0.60, 95% CI = [0.3, 1.21]). In this article, we evaluate the effect of MTBE on the prevalence of key health symptoms by applying a regression calibration method to White et al.'s and Mannino et al.'s data. Unlike White et al., who classified exposure using actual MTBE levels among a subset of the participants, and Mannino et al., who classified exposure based on job category among all participants, we use all of the available data to obtain an estimate of the effect of MTBE in units of serum concentration, adjusted for measurement error due to using job category instead of measured exposure. After adjusting for age, gender and smoking status, MTBE exposure was found to be significantly associated with a 50 per cent increase in the prevalence of one or more key health symptoms per order of magnitude increase in blood concentration on the log_(10) scale, using data from the 409 study participants with complete information on the covariates (95% CI = [1.00, 2.25]). Simulation results indicated that under conditions similar to those observed in these data, the estimator is unbiased and has a coverage probability close to the nominal value. The methodology illustrated in this article is advantageous because all of the available data were used in the analysis, obtaining a more precise estimate of exposure effect on health outcome, and the estimate is adjusted for measurement error due to using job category instead of measured exposure.