Assessment of Social Behavior in Non-Human Primate Infants Following Administration of Thimerosal-Containing Vaccines

Background:   In the 1990s, thimerosal (sodium ethylmercurithiosalicylate) was used as a preservative in most pediatric vaccines. While there is currently only one pediatric vaccine formulated with thimerosal that is included in the US pediatric immunization schedule, parental perceptions that vaccines are associated with the onset of neurodevelopmental disorders still continue to impact vaccination rates. 

Objectives:  The objective of this study was to examine whether administration of multiple thimerosal-containing vaccines (TCVs) to non-human primate infants increased the incidence of negative behaviors, such as fear-disturbance, rock-huddle-self-clasp, and stereotypies.

Methods:  We administered vaccines to 6 groups of infant male rhesus macaques (n=12-16/group) using a standardized thimerosal dose. Study groups included the recommended 1990s pediatric vaccine schedule (which included multiple TCVs), an accelerated 1990s primate schedule with or without the measles-mumps-rubella (MMR) vaccine, the MMR vaccine only, and the expanded 2008 vaccine schedule.  We also administered saline injections to age-matched control animals (n=16). Social behavior was evaluated in 40 minute daily playroom sessions for each peer group from 1-12 months of age. Scoring was conducted by a blinded observer in 5 minute focal periods using a coding system of mutually exclusive and exhaustive behaviors. Scored behaviors included passive, explore, withdraw, fear-disturbance, rock-huddle-self-clasp, stereotypy, play, sex and aggression, and could be scored as either a social interaction or a non-social behavior. Data were analyzed using multi-level modeling.

Results:   Overall means and standard deviations for duration and frequency of social and non-social behaviors scored for all infants is shown in Table 1. The duration and frequency of negative behaviors by animals in all groups was very low. In fact, there were no instances of stereotypies recorded across all sessions. Analyses of social interaction data identified a significant Group X Quadratic interaction for negative behaviors (F[5, 752]=2.92, p=0.030). Follow-up contrasts indicated that at 2 months of age, relative to the controls, animals in the 1990s Primate and 2008 groups exhibited significantly fewer negative behaviors (t[752]=-2.47, p=0.034 and t[752]=-2.85, p=0.023), respectively. However, by 12 months of age, there were no significant differences in social behaviors in the experimental groups relative to the control group. Analyses of non-social interaction data also revealed a significant Group X Quadratic interaction for negative behaviors (F[5, 751]=3.68, p=0.021). Follow-up contrasts indicated that at 2 months of age, relative to the control group, the 1990s Primate and MMR groups exhibited significantly fewer overall negative behaviors (t[751]=-4.12, p<0.001) and (t[751]=2.35, p=0.048), respectively. By 12 months of age there were no significant differences in non-social behaviors in the experimental groups relative to the control group. 

Conclusions:   TCVs did not appear to affect the development of social behaviors characteristic of infant macaques of this age. Social and nonsocial behaviors in all study groups, including the MMR only and the expanded 2008 pediatric groups, developed as expected for normal laboratory-reared macaque infants. Of particular relevance under the hypothesis that TCVs may impact behavior, there were very few instances of negative behaviors, such as rocking, self-clasping, and stereotypy, reported across the entire infancy period for all groups.