par for the course

[Guest Guest_linda]

New info (JAMA) on a major study that didn't support a link to autism. As usual, the study smells. . .

 

The Money Link

 

"Association Between Thimerosal-Containing Vaccine and Autism"

 

To the Editor: In their article on the association between

thimerosal-containing vaccines and autism, Dr Hviid and colleagues1

acknowledged their affiliations with Statens Serum Institut, Copenhagen,

Denmark, but did not disclose that the institute is a for-profit,

state-owned enterprise with roughly $120 million in annual revenue.

According to its 2002 Annual Report,2 vaccines represent approximately one

half of Statens Serum Institut's revenues and more than 80% of its profits.

Furthermore, Statens Serum Institut manufactured the now discontinued

monocomponent pertussis vaccine that contained thimerosal under

investigation in their study. They were also the providers of diphtheria and

tetanus components of a major thimerosal-containing diphtheria and tetanus

toxoids and acellular pertussis vaccine (DTaP) vaccine sold in the United

States.3

 

 

Letters Section Editor: Stephen J. Lurie, MD, PhD, Senior Editor.

JAMA. 2004;291:180.

 

 

- Bernard Rimland, PhD Autism Research Institute San Diego, Calif

 

1. Hviid A, Stellfeld M, Wohlfahrt J, Melbye M. Association between

thimerosal-containing vaccine and autism. JAMA. 2003;290:1763-1766.

 

 

The Bad Science Link:

 

To the Editor: Dr Hviid and colleagues1 found no increase in relative

risk of core autism from thimerosal in vaccines using the Danish autism

registry. Denmark removed thimerosal from infant vaccines in mid-1992. The

findings of Hviid et al are based on finding fewer older (born 1990-1992)

thimerosal-exposed children than younger (born 1992-1996) unexposed children

in the 2000 registry year. However, a sizable percentage of autism cases,

skewing toward older children, are lost from the registry each year. Thus,

the authors' finding is likely to be biased due to incomplete recordkeeping.

For instance, the 1995 registry2 contains 97 cases among 5- to

9-year-olds. This same cohort, as it grows older, becomes the 10- to

14-year-old cohort in the 2000 registry, where its number has decreased to

75 children, a decline of 22 cases or 23% of the original 1995 group. Hviid

et al stated that virtually all cases in their autism group were accurately

diagnosed, and thus it is unlikely that cases were removed due to subsequent

discovery of misdiagnosis and reclassification. Autism is a lifelong

disorder with near-normal lifespan,3 and few registry cases are in older age

groups likely to die. Therefore, virtually any case entered into the

registry should remain there. That some do not suggests administrative

error.

I calculated the extent of record loss for the 1991-2000 span studied

by Hviid et al. For each year, I added the number of newly enrolled cases

for that year to the number of previous year's cases. I compared this total

to the number of cases actually recorded in the registry for that year. For

4 of the years, the proportion lost amounts to one fourth of the cases. For

the 2000 registry year, 23% of the cases from the previous year are missing.

Cumulatively, 815 cases were dropped between 1991 and 2000, more than the

total number remaining in 2000.

Removed cases accumulate each year, so for any given registry year,

proportionately more removed cases fall into older age groups, because with

each successive year, the removed cases get older. The effect is a bias

toward more accurate counting of younger age cohorts while undercounting

older ones. The relative risk and conclusions of Hviid et al are predicated

on finding fewer cases in the older thimerosal cohort and more in the

younger nonthimerosal groups. This is an untenable approach given the

recordkeeping problem, and thus Hviid et al should either adjust their 2000

data for record loss or use an alternative methodology.

-Sallie Bernard, BA Safe Minds (Sensible Action for Ending Mercury-Induced

Neurological Disorders) Aspen, Colo

[Guest June]

Thanks for the info Linda. It's pretty shocking to see how researchers are willing to put their name to bad studies. There must be a payoff somewhere. The problem is the public only sees the headlines, and the same goes for doctors.

June