A boatload of fail: Were two horrendously bad zombie “vaxed/antivaxed” studies retracted—again?

[Editorial update: I woke up this morning to find out that the answer to my question in the title is almost certainly yes. The post has been quickly altered to reflect that. See below.]

Believe it or not, I overlooked something in yesterday’s post about a putrefying, rotting mess of a “vaccinated versus unvaccinated” study carried out by an Andrew Wakefield fanboi named Anthony Mawson that purported to have found that vaccinated children have a much higher prevalence of neurodevelopmental disorders (NDDs) and and diseases not preventable by vaccines than unvaccinated children. I’ll refer you to yesterday’s post for the details of what, exactly was wrong with the study: no well-defined population, no clear inclusion and exclusion criteria, no sampling frame, no power calculation, not attempt to account for obvious confounding factors, no verification of vaccination status or medical diagnoses, and more. Add to the horrendous methodological flaws in the study the fact that this study was funded by two antivaccine organizations and published in a bottom-feeding predatory open access journals, and you’re looking at a boatload of fail. Yet, none of this has stopped antivaccine groups from flogging this study as their Holy Grail of studies, a “vaxed/unvaxed” study that shows that unvaccinated children are healthier than vaccinated children.

So what did I overlook? Well, thanks to The Gnat (and if you don’t know who The Gnat is, just look at the comments of yesterday’s post) I realized that I missed half the fail. Here’s what I mean:

Yes, that’s the conspiracy crank site InfoWars, and yes InfoWars is eating this study up, interviewing Celeste McGovern, who is promoted as a “vaccine expert.” I had never heard of her before, but it turns out that she writes for Claire Dwoskin’s Children’s Medical Safety Research Institute (CMSRI), one of the two antivaccine groups that funded Mawson’s study. (The other was Generation Rescue.) As expected, she is antivaccine to the core.

What I learned, though, from following The Gnat’s link is that there was a second paper published in the very same bottom-feeding predatory open access journal by Mawson using the very same survey dataset, Preterm birth, vaccination and neurodevelopmental disorders: a cross-sectional study of 6- to 12-year-old vaccinated and unvaccinated children. Clearly, Mawson believes in the concept of “minimal publishable unit” (or MPU, as we call it) in which you divide the data into as many papers as you can and still get manage to get published, hence his publishing two papers instead of one.

Amusingly, when I clicked on that link, I got a “not found” error message, but as of yesterday the paper was still in Google Cache, and I saved a copy for your edification. [Mawson et al] This led me to check the link to the other paper, and guess what? It was gone yesterday, too. So I went to the Google Cache. I didn’t have to “save” it, though. Age of Autism is hosting it. Thanks, antivaccine cranks.

I was amused that the Journal of Translational Science appeared to have taken the papers down, perhaps retracting them. Wow. If being retracted by a Frontiers journal is a dis, this is the the black hole of dis. [UPDATE: Retractionwatch reports that both papers have indeed been retracted again.]

Of course, what matters is the quality of the study. I laughed as I typed the last sentence because, come on. It’s an analysis from the same dataset that spawned the first risibly incompetently awful study. When the dataset itself is this badly flawed, the best adage to apply is GIGO, or “garbage in, garbage out.” Also like the previous study, the authors tip their hand early, revealing their antivaccine viewpoint:

Preterm birth (defined as birth occurring before 37 completed weeks of gestation) is known as a major risk factor for neurodevelopmental deficits, including cerebral palsy, intellectual disability, cognitive and speech delays, motor deficits, and visual impairment associated with retinopathy of prematurity. In particular, preterm birth is the leading cause of neurodevelopmental disorders (NDD) and disability, including the development of autism spectrum disorder (ASD) [1- 3], but the underlying mechanisms are not well understood. Preterm infants receive the same doses of the recommended vaccines on the same schedule as term infants in order to protect them from several infections [4-7]. However, the possible role of vaccination in the development of NDD in premature infants has not been assessed, partly because pre-licensure clinical trials of pediatric vaccines have routinely excluded ex-preterm infants, and because of the assumed overall safety of vaccinations [8-15].

“Assumed overall safety of vaccinations”? If that isn’t a dead giveaway for antivaccine views, I don’t know what is, particularly after pointing out that preterm infants receive the same doses of vaccines that full term infants do. The authors also tip their hand later in the paper:

While the safety of vaccines is officially assured, observational studies have involved only a limited number of vaccines and vaccine ingredients, and none has reported on the long-term outcomes of the present vaccination schedule [39], which has been expanded and accelerated in recent decades [40]. The current childhood vaccination program now includes 48 doses of vaccines for 14 diseases from birth to age 6 years compared to 3 vaccinations for 7 diseases in the 1970s [41].

And here:

Since special efforts are made to vaccinate preterm infants, the effects of prematurity are difficult to separate from those of vaccination. Given the benefits of vaccination, it has not been thought necessary to do so. On the other hand, vaccine safety assessment in preterm infants is particularly important due to the frequency of adverse events associated with prematurity itself [21]. Adverse cardiorespiratory events including apnea, bradycardia and desaturations (oxygen saturation <90%) are well documented following vaccination in many preterm infants, yet vaccination is strongly recommended regardless of such events, since the prevention of infection is considered paramount.

of course, it’s also not true that vaccinating preterm infants is just “assumed” to be safe. There is a lot of evidence that it is safe (e.g., this recent study), which is why the American Academy of Pediatrics recommends vaccination of medically stable preterm infants on the same schedule as full-term infants based on chronological age. I also note that the studies cited by Mawson point out that the episodes of apnea, bradycardia, and oxygen desaturation sometimes seen after vaccination of preterm infants are transient, do not have serious consequences, and don’t have a detrimental impact on the infants’ clinical course.

This also appears to be another study without a clear hypothesis, just like the last one. The closest I could find to a hypothesis was at the end of the introduction, where the authors state that their purpose was:

This paper presents additional results of a survey designed to compare the health outcomes of vaccinated and unvaccinated children educated at home, based on mothers’ anonymous reports on the birth histories and physician-diagnosed illnesses in their children. The analysis explores the possible role of vaccination in NDD among children born preterm.

So what is the hypothesis? That vaccines are dangerous to preterm infants? That seems to be about as specific a hypothesis as Mawson can come up with. In fairness, sometimes exploratory studies are a perfectly useful thing to do as a hypothesis-generating strategy, but this study is such a mess that it can’t really even be said to be doing that. In any case, I won’t dwell much on the methods, as I already discussed them in depth yesterday. All the issues about a biased sample of home schooled children, with the same confounders as before in being far less likely to vaccinate and veing likely to have significant differences in health-seeking behavior apply, as do the defect of relying solely on the responses of mothers without verifying diagnoses or vaccination status.

So, not surprisingly, this paper reports elevated prevalence of autism and other neurodevelopmental disorders attributable to vaccination, with odds ratios in the range of 3.7 to 5.2 depending upon the specific NDD, with an odds ratio for any NDD of 3.7. Not surprisingly, they found that preterm birth was associated with an odds ratio of 4.9 attributable to preterm birth. Whew, right? If they hadn’t found that, it would have been the first study almost ever not to find a correlation between preterm birth and NDD, a finding so well accepted that not finding it would be a serious red flag.

Now here’s the kicker. Mawson et al claim to find that all of the risk of NDD in preterm infants is due to vaccination. Just sit back and chew on that for a while as I list their key findings (if you can call them that):

  1. Preterm birth without vaccination (P/V-) was not associated with NDD.
  2. Term birth with vaccination (P-/V) was associated with a significant 2.7-fold increase in the odds of NDD.
  3. Preterm birth with vaccination (P/V) was associated with a significant 5.4-fold increase in the odds of NDD compared to the odds of NDD given term birth and vaccination (P-/V).
  4. Preterm birth with vaccination (P/V) was associated with a nonsignificant 12.3-fold increased odds of NDD compared to preterm birth without vaccination (P/V-) (not technically significant because no child in the sample with an NDD was both preterm and unvaccinated).
  5. Preterm birth with vaccination (P/V) was associated with a significant 14.5-fold increased odds of NDD compared to being neither preterm nor vaccinated (P-/V-).

And there’s your huge red flag. This study is claiming to have found that there is no increased risk of NDD associated with preterm birth in unvaccinated infants, a finding so out of whack with a huge existing body of evidence linking preterm birth to elevated risk of NDD that it’s hard to believe. Then, in vaccinated children, preterm birth is a risk factor for NDD, with an odds ratio of 5.4. This is a result that, quite simply, does not make sense. The combination of these findings represents one reason why these results are suspect. Another reason why they’re suspect is small numbers. There were, for example, zero preterm infants who were unvaccinated with an NDD and only 8 infants who were not preterm who were vaccinated. Another problem is that Mawson does five pairwise comparisons but does not correct for multiple comparisons as he should have. At least one of the p-values would cease to be significant and another would be borderline statistically significant.

Based on gruel as thin or thinner than the first report, Mawson plunges deep into antivaccine speculation, suggesting that that preterm birth and vaccination are synergistic in causing NDDs and even speculating—without evidence—on a mechanism. He goes way beyond our reality, and not in a good way:

While additional studies are needed to verify and explain the present findings, a tentative hypothesis of the mechanisms linking preterm birth and vaccination with NDD is outlined as follows. Receipt of one or more vaccines could precipitate NDD in some preterm infants by exacerbating a preexisting inflammatory state associated with prematurity, leading to hepatic encephalopathy and hypoxic-ischemic brain damage. Impaired liver function is a predisposing factor for preterm birth [54,55] and the latter is associated with increased risks of hypoxic-ischemic brain injury [56]. A possible biochemical basis for vaccination-associated NDD in preterm infants could involve the spillage of a membranolytic biliary metabolite from the maternal liver into the circulation and its transfer to the fetus, contributing thereby to the pathogenesis of preterm birth itself [55] and possibly being further increased to neurotoxic concentrations by the impact of vaccination on the infant’s liver. Consistent with this hypothesis, liver dysfunction is reported as an adverse effect of vaccination [57] and as a feature of children with autism [58,59]. Furthermore, hyperbilirubinemia is associated with hypoxic-ischemic brain damage [60] and is a feature of the preterm infant as well as children with later-onset cognitive disorders and ASD [61,62].

This is nothing more than technobabble. For those of you not familiar with what that is, it’s a concept taken from Star Trek, particularly Star Trek: The Next Generation, in which technical-sounding verbiage is used to describe the solution or explanation of scientific, medical, or engineering problem. It’s also a wonderful example of how antivaxers start relating multiple biologic phenomena and observations together as an overall explanation for how autism and other NDDs could be caused by vaccines. Never mind that there is no good evidence that autism is caused by vaccines. Certainly Mawson’s two papers do not constitute anything resembling good evidence that vaccines cause autism.

I can’t help but be amused that these two papers have apparently been retracted (although I hasten to add that the only evidence that they have been retracted is that they no longer appear on the OAT journal website and that OAT has issued no statement that I am aware of at this time). I also can’t help but feel extremely pleasurable schadenfreude, because retraction by an even worse predatory journal than a Frontiers journal is a fate that Mawson richly deserves for conducting two such horrible studies at the behest of antivaccine activists. I also suspect that OAT will probably never issue a statement. Why? For a predatory journal to be able to keep fleecing its marks, it can’t have a lot of attention directed at how awful one of its papers is, regardless of how crappy all of its other papers are. When one paper is unlucky enough to attract attention (or two papers), better for the article to disappear. It couldn’t have happened to a nicer guy than Mawson, with the exception of Andrew Wakefield and any number of other antivaccine-sympathetic “scientists.”