H1N1 vaccine and miscarriages: More dumpster diving in the VAERS database

Antivaccinationists just love the Vaccine Adverse Event Reporting (VAERS) database. As in love it to death. As in “can’t get enough of it.” The reason, of course, is that VAERS is a lot like an unmoderated discussion forum or, at best, a minimally moderated forum. Anyone can say anything they like. The reason is that it is a database to which anyone can add an entry, and there is only minimal effort to determine whether a given purported adverse reaction to vaccines being entered is actually the least bit likely to be even related to vaccines. Indeed, it took a man claiming that his child was turned into The Incredible Hulk before it caught the attention of VAERS staff. Moreover, it’s a database that’s been warped by antivaccine litigation, as unscrupulous lawyers encourage parents to make reports to the database implicating vaccines for their children’s autism. No wonder that antivaccine mavens like Mark Geier so love to dumpster dive in the VAERS database for spurious correlations that they can use to try to implicate vaccines as a cause of autism or other conditions, and just last week I saw that the antivaccine tradition continues as a truly incompetent researcher named Stephanie Seneff followed Mark Geier into the dumpster that is the VAERS database. Truly, for antivaccinationists, the VAERS database is a gift that keeps on giving.

And so it was just last month, when a guy named Gary Goldman decided to follow the Geiers and Seneff dumpster diving in VAERS, although he didn’t do it for correlations suggesting a correlation between vaccines and autism. Instead, he looked at a specific vaccine and tried to correlate this vaccine to miscarriages. Like the many incompetent and pseudoscientific “investigators” before him, Goldman completely misunderstood the nature of the VAERS database, particularly how it is about as excellent an example of the old computer principle of GIGO (“garbage in, garbage out”) that exists. The result is a paper published in Human & Experimental Toxicology, a rag of a journal that has published nonsense of this sort before (in fact, nonsense of this type by the very same author), entitled Comparison of VAERS fetal-loss reports during three consecutive influenza seasons Was there a synergistic fetal toxicity associated with the two-vaccine 2009/2010 season?

I could save you the trouble and just point out that the answer to Goldman’s question is no, but, then, that’s not how I roll. If I had to read the study, then you’ll have to read my discussion of the study, as bad as it is. Well, actually, no one has to read this blog, and you can stop any time you like, but you and I both know that you will almost certainly follow along. Such is the power of Orac.

First, I can’t resist citing the article that first pointed me to this particular study, a post on the antivaccine website Vactruth entitled 4,250% Increase in Fetal Deaths Reported to VAERS After Flu Shot Given to Pregnant Women:

Documentation received from the National Coalition of Organized Women (NCOW) states that between 2009 and 2010 the mercury-laden combined flu vaccinations have increased Vaccine Adverse Events Reporting Systems (VAERS) fetal death reports by 4,250 percent in pregnant women. Eileen Dannemann, NCOW’s director, made abundantly clear that despite these figures being known to the Centers for Disease Control (CDC), the multiple-strain, inactivated flu vaccine containing mercury (Thimerosal) has once again been recommended to pregnant women as a safe vaccination this season.

Outraged by the CDC’s total disregard for human life, Ms. Dannemann accused the CDC of ‘willful misconduct,’ saying that they are responsible for causing the deaths of thousands of unborn babies. She stated that the CDC deliberately misled the nation’s obstetricians and gynecologists and colluded with the American Journal of Obstetrics and Gynecology (AJOG) to mislead the public by advertising the flu vaccine as a safe vaccine for pregnant women when they knew fully well that it was causing a massive spike in fetal deaths.

Wow. That’s a serious charge. Apparently so serious that Ms. Dannemann wrote a letter to Joe Mercola about it. Yes, that’s exactly the person to whom I’d go if I thought I had found a horrific threat to public health, well Joe Mercola and an antivaccinationist who defiles my home state, Mary Tocco:

To emphasize their point, on October 28, 2010, NCOW requested that Dr. Rene Tocco present their data at the CDC headquarters in Atlanta, Georgia. The CDC’s Dr. Shimabakuru gave a presentation on significant adverse reactions to the H1N1 vaccine, such as cases of Guillane-Barre Syndrome, which appeared to have risen three percent, claiming it as an insignificant signal.

And now there’s Goldman’s “study.” I hesitate to dig in, as bad science always gives me indigestion, but it’s my chosen avocation to look at pseudoscientific claims such as this. In the introduction, Goldman spends considerable time discussing a study by Moro et al that looked at the question of whether there was an increase in reports of adverse events in VAERS associated with the H1N1 vaccine during the year when it was feared that there would be a pandemic (2009-2010 flu season). Contrary to what you might think, it was a study that failed to identify any concerning patterns of fetal loss associated with the H1N1 vaccines. Specifically, the authors concluded, “Review of reports to VAERS following H1N1 vaccination in pregnant women did not identify any concerning patterns of maternal or fetal outcomes.” Of course, Moro et al suffers from the same GIGO problem that Goldman’s paper does, but at least the authors recognized that VAERS is an early warning system and that it’s a system not designed to be an accurate reflection of incidence of various adverse events associated with vaccination. It also can’t demonstrate causality. Be that as it may, Moro et al identified features associated with spontaneous abortions and stillbirths, and found the usual suspects: advanced maternal age, smoking, history of intrauterine fetal death (i.e., previous stillbirth), and the like, leading Moro et al to observe about spontaneous abortion (SAB) after vaccination:

SAB is a relatively frequent event in pregnancy, with a rate as high as 22.4% in women aged 34 years old or older and 10.4% in women younger than age 25 years.27 Stillbirths occur at a background rate of 0.4% of all pregnancies or 6.22 per 1000 live births and fetal deaths.30 There is underreporting to VAERS in general, and the proportion of AEs following immunization among pregnant women that are reported to VAERS is unknown. Nonetheless, the reporting rates to VAERS for SABs and stillbirths after H1N1 vaccine was several orders of magnitude lower than the expected rates of fetal losses in the general population of pregnant women [27] and [30] during a time of heightened awareness about vaccine safety.

The VAERS data provide no indication that the occurrence of SABs and stillbirths following influenza vaccination is higher than in the general population.

None of this stops our intrepid antivaccine researcher from cherry picking statements from Moro et al, using them to imply that there is a horrible problem, and then diving right into the VAERS database in a swan dive into the muck, “bravely” asking the question of whether there’s some sort of “synergistic toxicity” between the H1N1 vaccine and the seasonal influenza vaccine that was given in addition to the H1N1 vaccine during the 2009-2010 flu season. Basically, Goldman looked at fetal loss reports in the VAERS database for the two-vaccine 2009/2010 influenza season compared with reports from the immediately prior (2008/2009) and subsequent (2010/2011) single-vaccine seasons. Goldman then used VAERS and another “independent” data source from a group called the National Coalition of Organized Women (NCOW), who administered an Internet survey:

An independent survey was conducted by the National Coalition of Organized Women (NCOW) via the Internet to serve as a second surveillance source for pregnant women suffering A-H1N1 fetal loss during the two-vaccine 2009/2010 influenza season. Eileen Dannemann, director of NCOW, oversaw this study and the data collected are summarized in the Results section. In response to a public service announcement delivered via several websites on the Internet, respondents contacted one of two study coordinators via phone or e-mail address. The respondents provided relevant details including (a) type of influenza vaccine received, (b) date of vaccination, (c) type of vaccine, (d) date of onset of symptom/symptoms, (e) date of SAB or miscarriage, (f) geographic location, (g) whether or not the AE was reported to VAERS, and (h) other miscellaneous comments.

There’s so much wrong here that it’s tempting to stop right here and point out that, no matter what Goldman found, it’s almost certainly a pair of fetid, stinking dingo’s kidneys. What Goldman has done is to take two sources, one a passive surveillance system prone to reporting biases (both over- and underreporting) that can’t be used to estimate incidence or prevalence, the other an internet survey performed by an advocacy group, and combine them. Somehow he expects that from these sources he can come up with a halfway reliable estimate of the incidence of spontaneous abortion and stillbirth after vaccines, compare it to data from other sources for total number of pregnancies, H1N1 and seasonal flu vaccine uptake, and then use this gmish of data sources to prove that the H1N1 vaccine adds some sort of synergistic toxicity to the seasonal flu vaccine. And, to top it off, he seems to think he can do it accurately for three different flu seasons. Surprise, surprise, he can’t! None of this stops Goldman from “finding” a massive increase in reports of fetal loss during the 2009/2010 flu season. Per Goldman, although there was a four-fold increase in pregnant women vaccinated in 2009/2010 compared to the previous year, there was a 43.5-fold increase in fetal loss reports. Of course, this all assumes that you can use VAERS or an Internet survey to produce a reliable estimate an actual population incidence of an adverse reaction.

You can’t, but Goldman labors mightily to convince you that you can, although one can’t help but note that in the 2010/2011 flu season the H1N1 vaccine was combined with the seasonal flu vaccine. If there were “synergistic toxicity,” we would expect to see it in the 2010/2011 flu season. We don’t. Goldman tries to convince readers that there isn’t a massive reporting bias, despite all the hype and the antivaccine movement going into high gear by looking at the reports of anaphylaxis and noting that they were only elevated by around 10%. Of course, during hte 2009/2010 flu season, anaphylaxis wasn’t what the antivaccine movement stoked fears about. Miscarriages in pregnant women and Guillain-Barre syndrome were. For instance, this story and this video, both from the 2009 flu season:

Then there’s NCOW itself. Here’s an interview with Eileen Dannemann, in which NCOW is described as the “grandmother of many of the anti-GMO coalitions.” As I’ve pointed out many times before, anti-GMO often goes together with antivaccine, and it turns out that this rule of thumb applies quite well to NCOW. Together with Leland Lehrman, Dannemann co-founded the Progressive Convergence. Now click on the link for Progressive Convergence. What do you see? I see a banner urging readers to join the Vaccine Liberation Army, features an antivaccine flier. There’s even a whole section on vaccines that includes every major antivaccine trope in the book, including blaming vaccines for shaken baby syndrome (the vilest lie of all), links to studies by mercury militia founding members and antivaccine warriors extraordinaire, Mark and David Geier (just search this blog using “Mark Geier” for examples of how bad research can be), and a page with links to a wide variety of very bad antivaccine studies (are there any kind?), many of which I’ve blogged about on this very blog over the last eight years provided by Mark and David Geier themselves!

So let’s see. One of Goldman’s data sources is an report based on an Internet survey by a rabidly antivaccine group that frightens pregnant women about vaccines, in particular the flu vaccine, and even more particularly the H1N1 vaccine during the H1N1 pandemic in 2009-2010. It was, as far as I can tell, never published in the peer-reviewed literature, and was a very shoddy analysis. It’s also a data source in which he features prominently. Then there are whole Internet communities that were pushing the H1N1-SAB link, including on that “mother of all” antivaccine mother communities, Mothering.com, where claims that the H1N1 vaccine causes miscarriages were given wide publicity in the community. Moreover, miscarriages are, unfortunately, very common. Indeed, Steve Novella estimated:

There are about 4.2 million births a year in the US. About 15-20% of pregnancies result in spontaneous abortion (miscarriage or stillbirth). Even if we take the lower number, that’s 700,000 spontaneous abortions per year, or 58,000 per month. This means that over the two months of the vaccine program in 2009-2010, 116,000 women in the US would have had a spontaneous abortion. Half of them were vaccinated. Let’s further say that half of those vaccinated had their miscarriage or stillbirth after they were vaccinated – this leaves us with about 29 thousand woman who had spontaneous abortions following getting the H1N1 vaccine last season – just as the background rate.

So King’s estimates, based upon wild extrapolation from a few hundred reported cases, is still short of the background rate by an order of magnitude.

It’s the same principle by which antivaccinationists try to blame vaccination for autism. There is a large number of children every year in whom the first symptoms of autism are noticed after vaccination. Unless one controls for this background rate and demonstrates that there is a higher risk of autism after vaccination than just the background rate of diagnoses per time period normally observed in the population, you can’t say there’s an increased risk of autism after vaccination. The same holds true for miscarriages, but is even more difficult given the even larger number of miscarriages. After all, only around 1% of children are diagnosed with autism or autism spectrum disorder, while around 15-20% of all pregnancies end in miscarriage. Now, as then, Goldman and NCOW do not correct for this, and, quite frankly, even if they did you still can’t accurately estimate adverse reaction incidence from VAERS. Making such estimates rests on a chain of assumptions and the use of multiple data sets. Also, given that Goldman used two different data sources, one wonders if he corrected for double-counting (miscarriages counted both in VAERS and in NCOW’s report). If he did, I don’t see it.

The bottom line is that Goldman’s “study” is an absolutely atrocious bit of egregious antivaccine nonsense designed to frighten pregnant women into refusing the flu vaccine. It boggles that mind that such dreck can pass peer review and be published in an ostensibly peer-reviewed journal. That it was says a lot more about the journal than whether vaccines cause miscarriages. It also goes to show that there are journals out there that could easily go out of business and not be missed in the least—except by cranks and pseudoscientists.