Friday, September 17, 2010

Too Low - 1588 Reports of Miscarriage after H1N1 Vaccination

Three vaccination opponents put together a paper, including Gary S. Goldman, Ph.D.  It was first made public as an exhibit  in one of the truly stupid lawsuits filed by vaccination opponents against the FDA.  It was filed after the judge dismissed the case before trial as  additional information filed in support of a motion for reconsideration.

The exhibit starts with a lie. Under the heading Summary Statement  is:
Based on analysis of data from two different sources, the 2009-A-H1N1 inactivated-influenza vaccination program contributed to an estimate 1588 miscarriages and stillbirths among women 17 to 45 years of age.
There is nothing in the paper that provides any evidence that this is true.  Everyone would agree that the number of reports of adverse events after vaccination received in the passive (someone has to report) database, VAERS, isn't accurate. There have been varying estimates of how many reports should have been filed if everyone reported events. The Exhibit states:
CDC studies have reported that VAERS typically captures from 1% to 10% of the actual cases (typically the number of minor adverse reactions are reported at the lower percentage; while the reporting of more significant or major adverse events occurs at the higher percentage (page 3). 
No source is provided for this claim. But let us assume that it is correct. What can the Exhibit legitimately conclude?   Drum roll......using statistical methods, they determine that number of reports of miscarriage and stillbirth after 2009 H1N1 vaccination would have 1588 if all cases had been reported.  That's it.  Nothing More.

Does the number, 1588, and the method used provide any evidence of a negative correlation of vaccination with miscarriages and stillbirths (vaccination lessened the chances of having a miscarriage or stillbirth)?  No.  Does the paper provide any evidence of a positive correlation of vaccination with miscarriages and stillbirths? Does the paper provide any evidence that vaccination caused miscarriages and stillbirths?  No.

All it does is conclude there would have been 1588 reports if all women with miscarriages and stillbirths after 2009 H1N1 vaccination had reported them.  Hardly earthshaking. Except that the number is too low.

And 1588 is bizarrely low. Some simple arithmetic shows this."The overall miscarriage rate is reported as 15-20%, which means 15-20% of recognized pregnancies result in miscarriage"  In 2007,  4.26 million births were recorded in the United States in 2007.   Using 15%, that's 600,000 miscarriages annually. Monthly that's 72,000. The vaccination campaign was about 2 months long. So there were 144,000 miscarriages.  Let us assume that 10% of pregnant women were vaccinated and miscarried.  That's 14,400.   You only get near 1588 if you assume 1% of pregnant women were vaccinated.

However, the CDC's estimated 38% of pregnant women were vaccinated. 38% of 144,00 is 55,000, the number of women who were vaccinated and had miscarriages.   Obviously, there are some serious timing issues and an expert is needed to determine if the 55,000 is high or low when that is taken into account. But it can't be anywhere near 1588.  And even if 1588 is correct, it says nothing about correlation, let alone causation.

How did they come up with 1588 reports that should have been made? It is a capture-recapture statistical study on reports of miscarriage after vaccination with the 2009 H1N1 vaccine. Capture-recapture is easiest to understand with birds. You capture a bunch of birds, band them and release them.  Then you capture a bunch of birds in the same area very shortly after --- some of whom will have been captured before. With this information, you can calculate how many birds are in the area (the population).To use this method, you have to show that the two or more samples are taken one after another using the same methods.

The population number is merely how many birds are in the area.  Similarly, this Exhibit calculates that the true population of reports should have been around 1588.  They use two sets of data.  One is reports to VAERS, where anyone can report any adverse event after vaccination.  These reports are put up on a publicly available database.  If the initial report is categorized as a miscarriage that category doesn't change even if the facts in the initial report itself prove that it has nothing to do with a miscarriage.  For example, ID 362942-1 "denied pregnancy at time of vaccination. 12/29/2009 OB records prenatal visits. Patient pregnant and received live vaccine, no side effects noted."  counts as a miscarriage report.  Their search of VAERS returned 178 reports categorized as miscarriages and stillbirth.

The second set of data is reports of miscarriages and stillbirths reported to the National Coalition of Organized  Women, a group that is opposed to influenza vaccination and actively opposed H1N1 vaccination. They invited reports of adverse event after H1N1 vaccination.  They received 70 reports.

Seven reports were made to VAERS and NCOW. With this information, they calculated that proper number of reports ranged from a low of 946 to a high ot 3587 and with a number around 1588 as most likely.

How do we know if reports to VAERS and reports to NCOW qualify for a capture-recapture study? Because Gary S. Goldman, Ph.D says so:
Gary S. Goldman, Ph.D, author of various peer-reviewed medical journal publications, has verified the capture-recapture  (C-R) figures published in the NCOW report.  Dr. Goldman previously worked for 8 years as a Research Analyst for the L.A. County Department of Health Services in an epidemiological study project funded by the CDC."  (bottom).
It seems to me that the two samples are very, very different.  But an expert needs to determine this.

But in any case, there's nothing in the paper that provides any useful information to the issue at hand, whether H1N1 vaccination had any effect on the rate of miscarriages and abortion

Legal Sidelight
This paper is an Exhibit in the District Court for the District of Columber for one of the truly stupid lawsuits vaccination opponent filed against the FDA. This one, filed in 2007, seeks to stop the use of thimerosal in vaccines.

The first pages of the paper attempt to be a scientific paper. Then it turns into a diatribe against using thimerosal containing flu vaccine for pregnant women for the 2010-2011 flu season.

These lawsuits are almost impossible to win because the plaintiffs can't show standing  (is the type of case the court should decide and do you have sufficient interest so that you can sue) especially in a case where the remedy sought is against actions in the future.  Assuming you can get past the  standing issue, there's a really big problem --- the FDA has discretion in making its decisions.  So it isn't good enough for the court to go through all the evidence and decide that they would decide differently than the FDA and therefore overturn the FDA decision.  Rather the court has to decide that FDA didn't follow the rules it should have in deciding the issue and/or their decision was really wacko considering the evidence.

Having said all this, the case has never got past the issue of standing and the defendants motion to dismiss the case before trial was granted. You can read the July 2010 opinion here. Like Orly Taitz and the birthers, the vaccination opponents filed a bunch of papers asking the District Court judge to reconsider the decision.  As part of the motion for reconsideration they filed this Exhibit.

Mark Geier, like OrlyTaitz, provides comic relief.
The experience of Dr. Mark Geier, one of the named plaintiffs, directly contradicts both the plaintiffs' theory that mercury-free vaccines are not readily available and their theory that Mercury-Free members would be too confused to determine whether a particular vaccine contains mercury or not. See id., Ex. 6 (Decl. of Dr. Mark Geier). (page 17). 


  1. I like how you state that 1,588 miscarriages are "hardly earthshaking". My one miscarriage following the H1N1 shot was earth shaking to me. My ultrasound showed the baby was fine and healthy the day I got the vaccine. I got the vaccine and became so weak I could hardly walk for three days. I lost the baby and the final ultrasound showed it stopped growing within a day or two of getting the shot. The 1,588 are the miscarriages that occured following the shot. Not all shots are equal. If a multidose vial is not rolled properly the preservatives can be given out in a larger dose. This would account for some women being able to get the shot and not have a miscarriage. There's also the question of whether a few years from now will there be children with learning problems who's mothers were vaccinated while they were in utero. Maybe every one vaccinated did not have a miscarriage but it's still way too soon to say how the babies were affected. Before you spout out a crappy misinformed column maybe you should talk to someone actually affected by the shot. I know without a doubt the shot caused my miscarriage. It tore my life apart. One miscarriage is too many.

  2. One miscarriage is too many. Every day bad things happen to people. The isuue is whether more bad things happen after H1N1 vaccination than expected. The research to date has said no, with one exception. There may be excess cases of GBS at the rate of 1 in m million.

    I'm surprised you are so definitive about the cause of the miscarriage. Some of the miscarriages that happened soon after vaccination must result from other causes.

  3. Ultrasound itself, studies show, can contribute to miscarriage. The combination of vaccine and ultrasound could be causative.

    Refs: Yale Univ (2006), mouse study had difficulty gathering data at higher ultrasound exposure times, due to fetuses being absorbed back into mother, or miscarriages cannabalized, after ultrasound.

  4. Verstraeten) Thomas M.) MD) NIP, Division of Epidemiology and Surveillance, Vaccine Safety
    and Development Branch, Mailstop E-61, 770-639-8327.
    EIS Class Year of Entry: 1999
    No previous EIS Conference presentations
    Mackel Award consideration: No
    Number of abstracts submitted: 2, priority this abstract: 1
    Strong preference for poster presentation: No
    Thomas M. Verstraeten, R. Davies, D. Gu, F DeStefano
    Increased risk of developmental neurologic impairment after high exposure to thimerosal-containing vaccine in first month of life.
    Background: Concern has risen on the presence of the ethylmercury containing preservative
    thimerosal in vaccines. We assessed the risk for neurologic and renal impairment associated with past exposure to thimerosal-containing vaccine using automated data from the Vaccine Safety Datal ink (VSD). VSD is a large linked database from four health maintenance organizations in Washington, Oregon and California, containing immunization, medical visit and demographic data on over 400,000 infants born between '91 and '97.
    Methods: We categorizedthecumulative ethylmercuryexposure fromthimerosal containing vaccines after one month of life and assessed the subsequent risk of degenerative and developmental neurologic disorders and renal disorders before the age of six. We applied proportional hazard models adjusting for HMO, year of birth, and gender, excluding premature babies.
    Results: We identified 286 children with degenerative and 3702 with developmental neurologic disorders, and 310 with renal disorders. The relative risk (RR) of developing a neurologic development disorder was 1.8 ( 95% confidence intervals [CI] :1.1-2.8) when comparing the highest exposure group at 1 month of age (cumulative dose> 25 ug) to the unexposed group. Within this group we also found an elevated risk for the following disorders: autism (RR 7.6, 95% Cl = 1.8-31.5), non organic sleep disorders (RR 5.0, 95% Cl = 1.6-15.9}, and speech disorders (RR 2.1, 95% (1=1.1-4.0). For the neurologic degenerative
    and renal disorders group we found no significantly increased risk or a decreased risk.
    Conclusion: This analysis suggests that high exposure to ethyl mercury from thimerosal-containing vaccines in the first month of life increases the risk of subsequent development of neurologic development impairment, but not of neurologic degenerative or renal impairment. Further confirmatory studies are needed.