Skip Navigation Links
Centers for Disease Control and Prevention
 CDC Home Search Health Topics A-Z
NIP header
Health Care Professionals
bullet NIP HOME
bullet First time visitor?
bullet About NIP
bullet Data and Statistics
bullet International Efforts
bullet Links to other web sites 
bullet Glossary/ Acronyms 

NIP sub-sites:
bullet ACIP
bullet Flu Vaccine
bullet Immunization Registries
bullet Vaccines for Children Program
bullet CASA (Clinic Assessment Program)
bullet AFIX (Grantee Assessment)
bullet VACMAN

NIP Site Search
For Immunization Information, call the
CDC-INFO Contact Center:
English and Spanish

Get Acrobat Reader
Get Adobe Reader

Vaccine Safety > VSD
CDC Research on the Safety of Thimerosal Containing Vaccines

  At a glance: CDC has conducted a study using computerized data from 3 HMOS to see whether there were associations between vaccines that used thimerosal as a preservative and a wide range of neurodevelopmental disorders. Results found no consistent significant associations across the HMOs. Other studies of the safety of thimerosal containing vaccines are underway.  

Contents of this page:
Related pages and information:

What research is CDC doing to assess the safety of thimerosal-containing vaccines?
  • Vaccine Safety Datalink (VSD) screening study. CDC has conducted a study using computerized data from 3 HMOS to see whether there were associations between vaccines that used thimerosal as a preservative and a wide range of neurodevelopmental disorders. Results (published in 2003) found no consistent significant associations across the HMOs. Details on this study follow on this page.
  • The thimerosal follow-up study compares performance on a number of neurodevelopmental measures (such as speech and language skills, attention, coordination, tics, and hearing) among children who were exposed to different amounts of thimerosal through vaccination and other exposures. This study was designed to look more closely at inconsistent findings from the earlier, 2003 VSD study listed above. It does not assess autism or Autism Spectrum Disorders (ASD) as possible outcomes of thimerosal exposure through vaccination. The study includes over 1,000 children 7 to 10 years of age. The researchers collected information from the childrenís medical records, gave them standardized tests, and interviewed their parents about things that could influence the results of the study (such as a motherís diet during pregnancy and parental health seeking behavior). Although CDC funded this study, 13 non-CDC expert consultants gave input on the study design, conduct, and analysis. All of the information (data) needed for this study has been collected. Initial preliminary results are being reviewed by the expert consultants and the CDC study investigators. Researchers will continue to analyze the data and final results are expected to be published in 2006.
  • The thimerosal and autism case-control study. The possibility that exposure to mercury may cause autism has probably caused the greatest concern about exposure to thimerosal as a preservative in childhood vaccines. Although the weight of scientific evidence, including results from the VSD study, do not support this association, autism can be difficult to diagnose and study data based on autism diagnosis from medical computerized databases can be difficult to verify. This study is designed to address these issues.
Vaccine Safety Datalink (VSD)Study

November 2003
  • CDC takes vaccine safety seriously, and this study, a part of the Vaccine Safety Datalink project, is an example of CDC's ongoing efforts to examine issues of concern to parents and healthcare providers.
  • The results of this study are consistent with the weight of scientific evidence to date, which does not support that neurodevelopmental disabilities such as autism are caused by vaccines containing thimerosal. Additional studies are being conducted to help answer additional questions and to ensure that a comprehensive assessment of thimerosal-related health issues is done.
  • Today, none of the vaccines used routinely in the U.S. to protect preschool children against 12 infectious diseases actually contain thimerosal as a preservative. In addition, the measles-mumps-rubella (MMR), polio, pneumococcal conjugate (PCV) and varicella (chickenpox) vaccines never contained thimerosal.

Questions and Answers on the Vaccine Safety Datalink Study
  1. What were the major findings and conclusions from the CDC’s Vaccine Safety Datalink study?

The final results of this study found no consistent statistically significant associations between exposure to vaccines that contained thimerosal as a preservative and a wide range of neurodevelopmental problems, including autism, attention deficit disorder (ADD), language delays, sleep disorders, emotional disorders, and tics. None of the results found any associations with autism or ADD. Since it is necessary to first establish there is an association between two things (often called “variables”), the results from this study suggest there is not a “cause and effect” relationship between thimerosal and autism or ADD.

In the first phase of the research, there were some statistically significant associations between exposure to thimerosal-containing vaccines and two categories of neurodevelopmental problems, “tics” and language delay. However, these results were not consistent—that is, the relationships were only found with one of the health maintenance organization (HMO) databases, rather than in all three that were used in the study. Such a pattern suggests that an association does not exist, but that further research should be done.

As a result of the inconsistent finding involving “tics” and language delay, one of the major conclusions was that additional research be done to further examine the issue of exposure to thimerosal in vaccines. The CDC is currently involved in this additional research.


  1. What do the study’s results mean for parents?

The study did not find consistent evidence to suggest that thimerosal might cause any of the conditions studied. In particular, the study found that thimerosal exposure was not associated with autism, which adds to the evidence from other studies that thimerosal does not increase the risk of autism.

Since 1999 thimerosal has been removed from the majority of recommended childhood vaccines for preschool children. In fact today, with the exception of some flu vaccine, none of the routinely recommended childhood vaccines used in the U.S. to protect preschool children against 12 infectious diseases contain thimerosal as a preservative. These include vaccines for measles, mumps, rubella, chickenpox, hepatitis B, diphtheria, tetanus, pertussis (whooping cough), Haemophilus influenzae type b (Hib), polio, influenza (flu) and pneumococcal disease. In some of the recommended childhood vaccines minute amounts of thimerosal can be found; this is an unavoidable byproduct of the production process and has nothing to do with thimerosal being used as a preservative in the vaccines. No harmful effects have ever been reported or found with these amounts of thimerosal in vaccine.


  1. This study looked at thimerosal in vaccines. Why is there a concern about this vaccine preservative?

There was a concern because thimerosal contains a type of mercury called ethylmercury. Since exposure to small amounts of mercury can be harmful, concerns were raised about the very tiny amount of ethylmercury in thimerosal. However, the safety concerns related to thimerosal were not based on any evidence that connects thimerosal to harmful health outcomes. Rather, the concerns were related to federal government guidelines and theoretical estimates about how much mercury people can be safely exposed to without possible risk of serious health effects.

In 1999, the Food and Drug Administration (FDA) conducted a review of the mercury content in drugs and medicines, including vaccines. As part of this review, the FDA concluded that the use of thimerosal as a preservative in some recommended childhood vaccines could result in a six-month-old infant being exposed to an amount of mercury that exceeded the Environmental Protection Agency (EPA) recommendations. However, this same amount would not exceed the FDA, the Agency for Toxic Substances and Disease Registry (ATSDR), or the World Health Organization (WHO) guidelines for methylmercury intake (Ball et al., 2001). Thimerosal contains ethylmercury. Methylmercury is a related compound and has been more thoroughly researched than ethylmercury. Thus federal safety standards are based on information we have about methylmercury.

The FDA’s review found no evidence of harm caused by the small amounts of thimerosal in vaccines, except for minor local reactions (Ball et al., 2001). In July 1999, as a precautionary measure, the Public Health Service (including the CDC and FDA), the American Academy of Pediatrics, and vaccine manufacturers agreed that thimerosal levels in vaccines should be reduced or eliminated, and the FDA committed to expediting the review of new vaccines that do not contain thimerosal.

Although there is no evidence that any vaccine or vaccine additive increases the risk of neurodevelopment disorders, such as autism, CDC recognizes that this issue is of concern to parents and others. As a result, CDC is committed to investigating this issue to the fullest extent possible, using the best scientific methods available.

For more information on thimerosal in vaccines, visit:


  1. This study was based on computerized data from the Vaccine Safety Datalink (VSD) project. What is the VSD?

The Vaccine Safety Datalink (VSD) Project is a collaborative project involving the Centers for Disease Control and Prevention (CDC) and several large health maintenance organizations (HMOs). The VSD was established primarily to assess vaccine safety issues in the United States through analyses of large-linked databases collected at the HMOs as part of their routine administration of health services. The databases contain the vaccination and medical records of millions of children and adults.

The VSD is a cost-effective and valuable tool for monitoring vaccine safety and evaluating research questions. The computerized data sets that make up the VSD data base, however, are not collected for specific research purposes and as a result have certain limitations to how they can be used. Definitive vaccine safety studies in the VSD usually require additional data collection, such as detailed review of hardcopy medical records or standardized interviews of parents. Neither of these data sets are available in the computerized data bases.

For more information on the VSD, see Vaccine Safety Data Sharing Process.


  1. What was the purpose of the Vaccine Safety Datalink (VSD) database study?

The purpose of this Vaccine Safety Datalink database study was to determine whether there were associations between vaccines that used thimerosal as a preservative (i.e., an ingredient to prevent the growth of harmful bacteria and fungi) and a wide range of neurodevelopmental problems, including autism, attention deficit disorder, language delays, sleep disorders, emotional disorders, and tics. Researchers from CDC and participating Health Maintenance Organizations (HMOs) analyzed the computerized vaccination and medical records of over 140,000 children from birth up to 9 years of age.


  1. There have been reports that one of the CDC scientists who worked on this study now works for a vaccine manufacturer and that this was never mentioned in the publication of the study. Is this true?

Yes. From the time the study started to the time it was published took about 4 years and during that time one of four CDC scientists involved in the study left CDC to work for a pharmaceutical company called GlaxoSmithKline (GSK). In addition to the CDC scientists there were also researchers from three health maintenance organizations (HMOs) who participated.

Dr. Thomas Verstraeten, the researcher who left CDC in 2001 to work for GSK, worked at CDC during the critical time when the study was designed and the data were analyzed. As a result, the journal Pediatrics, which published the study on November 4, 2003, listed Dr. Verstraeten’s affiliation as the CDC. Pediatrics, like many professional journals, has a policy of identifying an author’s affiliation at the time a study was conducted (rather than their affiliation when the article is published). This article was thoroughly peer reviewed before it was published. However, to avoid any perceived conflict of interest, CDC should have assured that Dr. Verstraeten’s current employment status, as well as his status when the work was carried out, were both disclosed in the journal article. Once Dr. Verstraeten’s began working at GSK, his involvement in this study was limited to reviewing drafts of the manuscript.


  1. Did the study find any evidence that indicated that exposure to thimerosal in vaccines was associated with autism?

No. In one of the first analyses of the data there was a weak result that found a possible increased risk for autism among children who received vaccines at the highest level of thimerosal (> 62.5ug) by the third month of life, but this result was not statistically significant and was later found to have been based on incorrect data. Scientists have a responsibility to assure that all findings and results are based on accurate data—especially before publishing them.

These early findings were not supported by later analyses of additional cases with more accurate data which showed that children, who received vaccines containing thimerosal, either by three months of life or by seven months of life, were not at increased risk for developing autism. There was no suggestion of an increased risk for autism even among those children who received vaccines with the highest levels of thimerosal by seven months of age (i.e., those receiving 162.5 ug, 175 ug, or more than 175 ug thimerosal by 7 months of age). These results, however, cannot be considered definitive since the study was not specifically designed to assess a complex condition such as autism but to guide the development of follow-up studies at CDC. These studies will investigate more rigorously possible associations between thimerosal in vaccines and a number of neurodevelopmental disorders.


  1. The research in the VSD study began in 1999. Did the results or findings of the study change over time?

Yes, thanks to suggestions from other scientists, researchers, and organizations, improvements were made in the databases, research methods, and statistical procedures used to analyze the data. It is accepted and sound scientific practice, especially with complex and important research issues, to seek and use the advice and recommendations from both internal and external reviewers to strengthen a study as much as possible before publishing a final paper. In this case, four major improvements were made after the initial findings were presented in 1999:

  • The methods used to analyze the data were refined and improved based on expert input from inside CDC and from outside CDC.
  • Errors in the data were corrected (e.g., mistakes in medical records, errors regarding the thimerosal content of certain vaccines) to make the results more accurate.
  • Suggestions by reviewers were incorporated into the analysis to address and reduce potential biases (e.g., differences in health care seeking behavior among parents that could lead to some children going to the doctors more often and therefore being diagnosed more often).
  • More children with diagnoses of interest were identified as the study progressed and the children at the HMOs became older.


  1. There were two phases to this study. Why was a second phase added?

In science, it is important to be able to repeat (that is, replicate) research findings in order to show that they did not occur by chance alone. If you obtain the same results more than once, using similar methods but a different dataset, you can have more confidence in your findings and be better able to make strong conclusions. In a review of preliminary analysis results by a panel of CDC scientists, the suggestion was made that it would be important to attempt to replicate the results in an independent data set. This is why a second phase of the study was conducted using data from a third HMO.

Due to the smaller number of children in the third HMO, researchers could only assess the most common outcomes associated with thimerosal in the first part of the study. No association was found with autism in the first part of the study, and it was not investigated in the second part of the study. The third HMO pioneered the development of electronic medical records and had the required computerized databases (vaccinations and health care encounters) similar to those that had been used in the two VSD HMOs in Phase I of the study. This long history of active use of the databases for multiple needs provided many opportunities for feedback and continued improvements in data quality. These qualities in turn made this HMO a good candidate for the Phase II study when the need arose.


  1. Can people outside CDC use the VSD data to conduct their own research?

Yes. In 2002, the CDC and the participating HMOs established a data sharing program to make the VSD database available to any member of the public who follows the established procedures for the Vaccine Safety Datalink Data Sharing program. External researchers can submit research proposals to conduct new studies of vaccine safety or to reanalyze study-specific datasets from published VSD studies. External researchers have the opportunity to submit research proposals for new vaccine safety hypotheses that include any or all of the data variables available in the VSD. Data are extracted from the entire VSD database and provided to external researchers in response to specific Institutional Review Board (IRB)-approved research protocols.

In compliance with federal regulations, access by external researchers to the VSD data files or to datasets from VSD published studies requires review and approval by the appropriate IRBs of the relevant HMOs. The IRBs have the responsibility to protect the confidentiality and privacy of their members’ medical records and to adhere to the rules and regulations of their specific institution; therefore, each of the HMO IRBs must review any request for access to the VSD data files that contain information on its HMO members.

Upon approval by the relevant IRB protocol, datasets are prepared according to the external researcher’s approved research protocol. External researchers access the approved VSD data at the Research Data Center (RDC) located at the National Center for Health Statistics (NCHS). The RDC, established in 1998 by NCHS, is a research facility located at the CDC’s NCHS facility in Hyattsville, Maryland.

Further details about the VSD data sharing process can be found in "Guidelines for Data Sharing Program for External Researchers:
Access to CDC’s Vaccine Safety Datalink Data


  1. Do any of the currently routinely recommended childhood vaccines contain thimerosal as a preservative?

Today, with the exception of some influenza (flu) vaccines, none of the vaccines routinely used in the U.S. to protect preschool children against 12 infectious diseases contain thimerosal as a preservative. These include vaccines that protect children from the measles, mumps, rubella, chickenpox, hepatitis B, diphtheria, tetanus, pertussis (whooping cough), Haemophilus influenzae type b (Hib), polio, influenza (flu), and pneumococcal disease. No serious harmful effects have ever been reported or found with these doses of thimerosal in vaccines. In addition, the measles-mumps-rubella (MMR), polio, varicella (chickenpox) and pneumococcal conjugate (PCV) vaccines never contained thimerosal.


  1. Are there any vaccines recommended for preschool children that contain thimerosal as a preservative?

Yes. Recently the Advisory Committee on Immunization Practices recommended flu vaccine for children 6 to 23 months of age and all children at high risk for complications from influenza. Although some flu vaccines contain the preservative thimerosal, the amounts found in these vaccines are considered safe for use in children. For parents, physicians, and other healthcare providers who have concerns, there is some flu vaccine available that does not contain thimerosal as a preservative.

There is no indication that the small amounts of thimerosal found in some flu vaccines have any association with any harmful health effects on children of this age.


  1. Has there been an independent report on the possible association between exposure to thimerosal as a preservative in vaccines and serious health outcomes?

Yes. One was conducted by the Institute of Medicine (IOM) in 2001. ((Immunization Safety Review: Thimerosal-Containing Vaccines and Neurodevelopmental Disorders 2001).


National Immunization Program (NIP)
NIP Home | Contact Us | Help | Glossary | About | Accessibility

This page last reviewed and modified on April 22, 2005


Department of Health and Human Services
Centers for Disease Control and Prevention
CDC Home
  |  CDC Search  |  CDC Health Topics A-Z