Title

A Spatial Epidemiological Analysis of Oral Clefts and Volatile Organic Compounds in Texas

Document Type

Dissertation

Comments


Presented to the Graduate Council of Texas State University-San Marcos in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy, December 2007.
Committee Members Approved:
Dr. F. Benjamin Zhan, Chair
Dr. Denise Blanchard-Boehm
Dr. Yongmei Lu
Dr. Jean Brender
Approved:
Dr. J. Michael Willoughby, Dean of the Graduate College

Abstract

Existing research comparing models of residential exposure to volatile organic compounds (VOCs) is insufficient. A comparison of these exposure models is necessary for determining whether changes in exposure measurements may influence the results. The risk of oral cleft birth defects was found to be associated with exposures to organic solvents with most findings suggesting a significant association. Using a case-control approach, this dissertation investigated how two exposure measurements possibly identified the association between residential exposures to air concentrations of VOCs released through the use of solvents and the occurrence of oral cleft defects in Texas. Model 1 was based on an air dispersion model and two interpolation methods - inverse distance weight (IDW) and ordinary kriging (OK) - and model 2 was based on distance proximity. This research also considered the risk of oral clefts according to maternal age and additional risk factors associated with this congenital malformation.

The results of the risk estimate analysis did not confirm a statistically significant increased odds ratio for the analysis of VOC concentrations for any of the models when median exposure values were considered. However, when exposure values were classified into quintiles, the model results were significant. Model 2 estimates identified absence of risk according to the furthest distance from the site for mothers of all ages (OR 0.69 95% CI 0.52-0.93).

Also, results from both models revealed significant associations according to maternal age groups. These estimates were more consistent for mothers at the oldest age group as associations for mothers at the youngest age group (younger than 20) and the third age group (25-29) were only identified by results of model 1. Model 1 also identified risk for mothers at 35 years of age or older living in areas with VOC concentrations higher than the median concentration value for both the IDW (adjusted OR 1.9, 95% CI 1.06-3.6) and the OK interpolation (OR 2.06, 95% CI 1.12-3.79). However findings of model 2 do not indicate associations for the analysis based on median distance values.

In the analysis regarding concentration quintiles, model 1 identified risk for mothers at the oldest age group if they lived in areas with higher VOC concentrations as the odds ratios based on the OK interpolation were significant for the 4th (OR 2.97, 95% CI 1.05-8.41) and 5th quintiles (OR 3.57, 95% CI 1.2-10.85). The results of the IDW interpolation also identified significant risk according to the largest exposure level (OR 3.8, 95% CI 1.3-10.1). In the case of model 2, the results of the analysis concerning older mothers indicated that there was a decrease in risk according to the second and third quintile distances as they presented significant decreased odds ratios (0.27, 95% 0.09- 0.85 and 0.29, 95% 0.1-0.86, respectively).

In addition, regarding the logistic regression analysis examining possible risk factors, results indicated that male gender of the child (OR 1.40, 95% CI 1.17-1.7), low birthweight (OR 1.7, 95% CI 1.23-2.36), and 12 years of maternal education (OR 1.27, 95% CI 1.02-1.8) were positively associated with the risk of oral clefts. On the other hand, African-American ethnicity proved to be a protective factor against these birth defects (OR 0.68, 95% CI 0.48-0.96).

These findings indicate that, among other factors, exposure to solvent-VOCs may contribute to the risk of having a child with oral clefts and also suggest that the comparison of exposure assessment in studies examining exposure to air pollution needs to be considered. The exposure models appear to provide distinct outcomes as model 1 identified the presence of risk according to different age groups, while model 2 identified the absence of risk and for only one age group. Furthermore, the results of the statistical analyses considering diverse maternal age groups and VOC levels of exposure may contribute to the awareness of the potential susceptibility of older mothers to differing levels of exposure to air pollution.