California Department of Public Health logo: three likenesses of people colored blue, green, and orange  
Last Edited: 8/5/2013

An investigation of spontaneous abortions following a metam sodium spill into the Sacramento River
Written: 1993

    On July 14, 1991, shortly before 10 p.m., a tank car derailed and released 19,000 gallons of the herbicide metam sodium into the Sacramento River six miles north of the town of Dunsmuir in Northern California. Local residents were exposed to the toxic by-products, mainly methylisothiocyanate. A community-based health survey focusing on chronic effects including pregnancy outcomes was conducted in April 1992 in the Dunsmuir area. A total of 969 households (81%) participated.

   Since metam sodium has been associated with increased incidence of fetal loss in animals and there had been anecdotal repons of spontaneous abortions following the spill, the relationship between spontaneous abortion (fetal loss before 20 weeks gestation) and possibility of exposure was examined. Seventy-four pregnancies were reponed during the study period (7/90-2/92). A follow-up interview was conducted (October 1992) to confirm pregnancy information and obtain permission to review the medical records of women who reported a spontaneous abortion .
   Five spontaneous abortions were reported during the April 1992 survey among fetuses exposed in the first trimester. One spontaneous abortion has been confirmed to date giving a confirmed spontaneous abortion rate of 14% (1/7). If all five were confirmed the rate would be 45% (5/11). The rates of reported and confirmed spontaneous abortion in the unexposed group (26%, 21%), in the exposed post-spill conception group (26%, 23%) and in the unexposed post-spill conception group (20%, 20%) were similar.
   Using only confirmed spontaneous abortions, this study does not show a higher rate of spontaneous abortion among fetuses exposed in the first trimester. Including the four unconfirmed spontaneous abortions exposed in the first trimester increases the rate above the unexposed, but the statistical confidence interval around this rate is very wide. The higher than usual rates of confirmed spontaneous abortion even in unexposed groups and a review of birth certificates suggests that women whose pregnancies ended in live births did not all participate in our study.

Full report (Size:  2 MB)