TABLE 3

Association Between Thimerosal Exposure and Autism Outcomes

Exposure MeasureUnadjusted Model Results (No Covariates)Covariate Adjusted Model Results
1-U Difference in Exposure, OR (95% CLs)a2-SD Difference in Exposure, OR (95% CLs)b1-U Difference in Exposure, OR (95% CLs)a2-SD Difference in Exposure, OR (95% CLs)b
Case-children with ASD (n = 256) vs controls (n = 752)
    Prenatal1.007 (0.990, 1.025)1.125 (0.846, 1.495)1.007 (0.988, 1.026)1.119 (0.827, 1.513)
    Birth to 1 mo (28 d)0.973 (0.898, 1.054)0.894 (0.644, 1.240)0.970 (0.889, 1.059)0.883 (0.617, 1.264)
    Birth to 7 mo (214 d)0.992 (0.966, 1.020)0.887 (0.582, 1.351)0.967 (0.937, 0.999)c0.597 (0.360, 0.990)c
    Birth to 20 mo (609 d)0.991 (0.965, 1.016)0.862 (0.533, 1.336)0.968 (0.938, 0.998)c0.598 (0.317, 0.971)c
Case-children with AD (n = 187) vs controls (n = 724)
    Prenatal1.010 (0.991, 1.030)1.179 (0.862, 1.614)1.011 (0.990, 1.032)1.196 (0.855, 1.674)
    Birth to 1 mo (28 d)1.010 (0.927, 1.100)1.040 (0.732, 1.478)1.029 (0.935, 1.132)1.123 (0.759, 1.661)
    Birth to 7 mo (214 d)0.991 (0.962, 1.022)0.875 (0.545, 1.404)0.958 (0.924, 0.994)c0.516 (0.290, 0.916)c
    Birth to 20 mo (609 d)0.992 (0.964, 1.021)0.884 (0.520, 1.449)0.962 (0.928, 0.996)c0.544 (0.265, 0.938)c
Case-children with ASD with regression (n = 49) vs controls (n = 652)
    Prenatal1.031 (0.993, 1.072)1.656 (0.885, 3.095)1.039 (0.997, 1.083)1.860 (0.945, 3.660)
    Birth to 1 mo (28 d)0.938 (0.794, 1.108)0.769 (0.390, 1.519)0.901 (0.761, 1.067)0.653 (0.327, 1.303)
    Birth to 7 mo (214 d)0.936 (0.880, 0.994)c0.355 (0.138, 0.915)c0.906 (0.848, 0.968)c0.214 (0.076, 0.600)c
    Birth to 20 mo (609 d)0.953 (0.900, 1.009)0.473 (0.154, 1.170)0.925 (0.869, 0.985)c0.297 (0.081, 0.764)c
  • Covariates for ASD models: birth weight, maternal age, birth order, breastfeeding duration, family income, maternal health care–seeking behavior (Kotelchuck inadequacy of prenatal care, use of cholesterol screening, use of Papanicolaou test screening), maternal exposures during pregnancy with study child (alcohol use, folic acid use, viral infection, lead exposure), and early childhood health conditions (anemia between 6 and 30 months of age; pica before 3 years of age). CLs indicates confidence limits. Covariates for AD models: birth weight, maternal age, birth order, breastfeeding duration, family income, maternal health care–seeking behavior (Kotelchuck inadequacy of prenatal care, use of cholesterol screening, use of Papanicolaou test screening), maternal exposures during pregnancy with study child (folic acid use), and early childhood health conditions (anemia between 6 and 30 months of age; pica before 3 years of age). Covariates for ASD with regression models: birth weight, maternal age, family income, maternal education level, maternal exposures during pregnancy with study child (alcohol use).

  • a OR for autism associated with a 1 U increase in exposure. For prenatal exposure, 1 U = 1 μg of ethylmercury. For postnatal exposure, 1 U = 1 μg of ethylmercury per 1 kg of body weight at time of vaccine or immunoglobulin receipt.

  • b OR for autism associated with an increase in exposure equal to 2 SD units of the exposure measure. For the measure of prenatal exposure, 2 SDs = 16.34 μg. Two SDs of the birth-to-1-month measure is 4.08 μg/kg U. Similarly, 2 SDs of birth-to-7-month and birth-to-20-month exposures are 15.56 μg/kg and 17.82 μg/kg.

  • c 95% CLs for the OR does not include 1.000.