Fluoride Exposure and Its Potential Reproductive Risk Among Semiconductor Fabrication Workers
Chris Woods - Univ of California at Berkeley (SSA Journal Volume 11 Number 1 - Spring 1997 pp. 47 - 58 )

In response to preliminary studies indicating an increased risk of spontaneous abortion among women working in the semiconductor industry, the Semiconductor Health Study (SHS), a large, multi-disciplinary investigation, was launched. A complex, tiered exposure assessment was developed with the objective of clarifying a potential causal relationship between specific agents and increased risk of spontaneous abortion. The results of the SHS indicated an increased risk associated with working in a fabrication room, specifically, working in masking, dopants, and thin film, and with exposure to photoresist chemicals and fluoride. However, the results of the analysis did not indicate the expected dose-response relationship between the calculated relative risk and the exposure score derived for workers using the exposure assessment algorithm. As a follow-up to the SHS, this project used the SHS exposure algorithm to retrospectively evaluate the exposure potential associated with workers for whom fluoride samples were collected comparing the exposure scores to the results of the samples to look for a correlation. This comparison was sued to identify possible explanations regarding the lack of dose response seen in the historical SHS cohort study and to provide lessons learned for future studies and applications of this type of algorithm for exposure assessment. The results of this study suggest that the algorithm used in the SHS was misclassifying some workers with low to moderate exposure to the moderate or high exposure groups. If the air samples are representative of the worker's "true exposure", this misclassification provides an explanation of the non-linear dose response found by the SHS investigators. The statistical analysis supported the relative risks reported by the SHS to be associated with fluoride exposure; the mena sample values were 10.04 Чg/m^3, 2.37 Чg/m^3 and 16.83 Чg/m^3 (low, medium and high exposure groups, respectively) with a predicted relative risk for the SHS of 1.91, 1.53, and 2.15. The highest air sample results were found in the group with the highest relative risk, with a linear relationship down to the group with the lowest air sample results. This study provides the semiconductor industry with preliminary quantitative results of fluoride exposure which may be associated with increased spontaneous abortion risk. In addition, insight was gained regarding potential improvements to the algorithm design and exposure assessment to better classify subjects in future investigations.