In 1995, the Agency for Toxic Substances and Disease Registry (ATSDR) began data collection for a study of environmental exposure to volatile organic compounds (VOCs) in drinking water and a variety of adverse pregnancy outcomes at the U.S. Marine Corps Base at Camp Lejeune, Onslow County, North Carolina. This study was undertaken following documentation that environmental exposure to VOCs in drinking water had occurred in the past. At that time, there was no evidence of an increased rate of adverse pregnancy outcomes at Camp Lejeune. However, because fetuses tend to be more sensitive to toxic chemical exposures and many pregnant women had resided in housing areas supplied with contaminated water, it appeared prudent to research the topic. This report describes a study of past exposure to VOC-contaminated drinking water and mean birth weight (MBW), small for gestational age (SGA), and preterm birth in residents of base family housing at Camp Lejeune. The results were based on analysis of live births to women residing in base family housing when they delivered during the period January 1, 1968, through December 31, 1985. Birth certificates were studied from 6,117 tetrachloroethylene (PCE)-exposed women, 141 short-term trichloroethylene (TCE)-exposed women, 31 long-term TCE-exposed women, and 5,681 unexposed women. The following potential confounders and effect modifiers were evaluated: sex of infant, maternal and paternal ages, maternal race, maternal and paternal education, military pay grade, maternal parity, adequacy of prenatal care, marital status, and year of birth. The influence of timing and duration of exposure on potential effects was also explored by linking family base housing records to birth certificate data.

Preterm delivery was not associated with VOC-exposure in any category. For most live births, including all births to women younger than 35 years of age with no prior fetal deaths, there was no association between PCE-contaminated drinking water and MBW or SGA. For the group as a whole, infants whose mothers resided in PCE-exposed areas weighed an average of 24 grams (g) less at birth than infants whose mothers lived in unexposed housing. This difference was too small to be biologically meaningful. After controlling for potential confounders, the overall odds ratio (OR) for PCE and SGA was 1.2 (90% confidence limits [CL]: 1.0, 1.3). These results provide reasonable assurance that PCE-contaminated drinking water did not affect the birth weight of infants of mothers who were younger than 35 years of age and had no medical history of fetal death; this accounted for most base residents exposed to PCE.

Associations between PCE and the study outcomes were observed in two potentially susceptible subgroups: infants of mothers 35 years of age or older and infants whose mothers had histories of fetal deaths. For older mothers, the adjusted difference in MBW between PCE-exposed and unexposed births was -205 g (90% CL: -333, -78), and the adjusted OR was 4.0 (90% CL: 1.6, 10.2) for PCE exposure and SGA. In mothers who had previously had one or more fetal deaths, the adjusted OR for PCE and SGA was 1.6 (90% CL: 1.2, 2.1). In mothers who had previously had two or more fetal deaths, the differences in MBW and SGA between PCE-exposed and unexposed mothers were much larger, but the number of births to women in this group was fairly small. Because associations in these subgroups were not anticipated, these results should be considered exploratory. They are, however, biologically plausible and deserving of followup.

The TCE-exposed groups were both small in number. The difference in adjusted MBW between the long-term TCE-exposed group and the unexposed comparison group was -139 g (90% CL: -277, -1); the OR was 1.5 (90% CL: 0.5, 3.8) for SGA and long-term TCE exposure. This increase was entirely attributable to differences in male infants within the long-term TCE-exposed group. Among males alone, the OR for SGA was 3.9 (90% CL: 1.1, 11.9) and the difference in MBW was -312 g (90% CL: -540, -85). The short-term TCE-exposed group had a lower prevalence of SGA infants, and MBW was slightly higher overall in this group compared with the unexposed group.

The finding and magnitude of reduced birth weight and increased SGA in males within the long-term TCE-exposed group is potentially important. However, the small sample size considerably weakened the evidence for a causal association. Although it is possible to speculate on mechanisms by which such a sex-based difference might arise, this difference was unexpected and could not be explained by known mechanisms of TCE toxicity. These findings warrant followup in a larger TCE-exposed population.

ATSDR had intended to analyze fetal death data, but existing records were too incomplete to be useful. In addition to the main analyses, several substudies were conducted and are presented in Appendices A and B. Important conclusions from these substudies are (1) the housing record data were complete and should have provided reasonable information regarding length of exposure during pregnancy; (2) abstracting medical records is feasible and might enrich the data quality for the subgroups of study participants for which associations between VOC-exposure and MBW and SGA were noted; (3) a limited amount of birth defects data was available from the birth certificate. These data were inadequate for a formal evaluation of associations between VOC exposure and birth defects. Alternative approaches are recommended to study VOC exposure and birth defects if the question remains an issue of public health interest.



The Agency for Toxic Substances and Disease Registry (ATSDR) has a broadly defined legislative mandate to prevent or mitigate adverse human health effects and diminished quality of life resulting from exposure to hazardous substances in the environment. Population-based research conducted to identify links between exposures and specific adverse health effects is a necessary part of this mandate. One exposure-disease relationship that warrants further investigation is the association between volatile organic compounds (VOCs) in drinking water and adverse pregnancy outcomes. Pregnancy outcomes are of particular importance to populations residing on military bases because such populations include a high proportion of reproductive-aged individuals.


The primary objective of this retrospective study was to explore potential associations between previous exposure to VOCs in drinking water and three adverse pregnancy outcomes at the U.S. Marine Corps Base at Camp Lejeune, Onslow County, North Carolina, where tetrachloroethylene (PCE), trichloroethylene (TCE), and 1,2-dichloroethylene (1,2-DCE) were found in drinking water supplies in the 1980s. The three pregnancy outcomes were (1) reduced fetal growth, measured as decreased mean birth weight (MBW) and small for gestational age (SGA); (2) preterm birth; and (3) late fetal deaths. However, because of incomplete data, this third outcome could not be determined.

Three secondary objectives of the study were (1) to validate the quality of housing record information as it was used to assign exposure and duration of exposure (see Appendix A); (2) to evaluate the feasibility and utility of reviewing medical records to enhance the study’s inferences (see Appendix B); and (3) to gather existing information on birth defects (see Appendix B). It was understood from the outset that the third and final objective would be difficult to achieve because of the limited information regarding birth defects available from birth certificates. Nonetheless, the potential relationship between VOCs and birth defects is of such strong public health concern that an attempt was made to evaluate all existing data.

Page last reviewed: November 22, 2000