January 24, 2012

PFCs and antibodies

This analysis accompanied a roundup which can be viewed here.

Title, Date of Publication & Journal

‘Serum Vaccine Antibody Concentrations In Children Exposed to Perfluorinated Compounds’ by Grandjean, P et al, published in JAMA  on Tuesday 24 January 2012.

Claim supported by evidence?

The paper provides limited evidence that PFCs are associated with a reduction in vaccine response, but provides no evidence of a causal relationship between the two.

Summary

• An observational study examining a possible link between antibody levels and PFC concentrations in children.
• Objective: to determine whether PFC exposure is associated with antibody response in childhood vaccinations.
• This is a prospective cohort study. Overall, the results appear to add some weight to the claim of association, but any associations observed may be partly due to ‘confounding’ effects of unmeasured variables

Study Conclusions

The paper’s main conclusion is “Elevated exposures to PFCs were associated with reduced humoral immune response to routine childhood immunizations in children aged 5 and 7 years.”

The paper does not conclude that elevated PFC concentrations reduce the effectiveness of vaccines.

It appears as though there may be an association between increased PFC and decreased antibodies (for diphtheria the relationship is stronger than tetanus)

The authors acknowledge that “overlapping confidence intervals and the lack of comparative toxicological studies prevent inference in regard to causal attribution”.

All results are reported in terms of a doubling in PFC concentration, but it is not clear whether such an increase is realistic.

Of the 80 results in Table 3, the biggest effect (highlighted in the press release) is the 39% decrease in diphtheria antibody concentration for a doubling in maternal PFOS. The size and statistical significance of the effect appears inconsistent with the other results, although the decrease seems consistent across types of PFC and age.

17 of the 80 results in Table 3 show a statistically significant decrease in antibody levels associated with doubled PFC concentrations. 2 of the 80 show a statistically significant increase. The remaining 61 are not statistically significant.

Strengths/Limitations

Using as much as possible of a whole birth cohort adds strength to the study, as it’s more likely to be representative.

Blood samples were collected, providing stronger evidence than if patient records had been used.

Some ‘confounders’ (age, sex, time since vaccine booster) were accounted for. However, other potentially important ones were not, e.g. weight.

PFC levels were not assessed at age 7; antibody levels at this age were compared with PFC levels at age 5. This provides high potential for confounding effects.

Conducting many statistical tests (80 in Table 3 alone) will always produce some false positive results. It is not immediately obvious that a distinct pattern can be seen, especially in terms of statistical significance, so these must be interpreted with caution.

Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research. A list of contributors, including affiliations, is available here.