And now death by Gardasil? Again, not so fast…

I guess this is in effect part two of yesterday’s post. Regular daily readers (and you are a regular daily reader, aren’t you?) will remember that yesterday I commented on the recent uptick in anti-Gardasil vaccine rhetoric coming from the antivaccine crank blog Age of Autism and other sources, in the process deconstructing speculation masquerading as a case report allegedly indicting the quadrivalent HPV vaccine as a potential cause of premature ovarian failure in a 16 year old Australian girl. The article was so bad and so biased that I couldn’t believe BMJ Case Reports published it in the form it did. It’s almost enough to make me take back all those nice things I said about the BMJ in the wake of its publication of Brian Deer’s expose of Andrew Wakefield’s fraud, but in reality the publication of one bad piece of antivaccine propaganda doesn’t invalidate all the good the BMJ did by commissioning Brian Deer to lay the results of his investigation on the line.

You’ll also recall that I noted another article that’s popping up on antivaccine websites. Not content with trying to scare parents and girls with stories of Gardasil robbing young women of their womanhood, antivaccinationists now want to scare parents with stories of Gardasil robbing girls of their lives. So it was that the antivaccine propaganda blog touting an article by the latest scientists to have drunk the pseudoscience Kool Aid that is antivax, Lucija Tomljenovic and Christopher A. Shaw. AoA was joined by the rabidly anti-Gardasil antivaccine group (IN)SANE Vax, Inc., the blog for the antivaccine movie The Greater Good, Gaia Health, and others. That’s no surprise, because we’ve met Dr. Shaw before. He was a key player in that particular bit of propaganda, likening vaccines to part of a toxic soup of chemicals that contributes to autism. We’ve also met Tomljenovic before. Along with Shaw as her co-author, she wrote one of the silliest attempts at blaming vaccines for autism that I’ve seen in a long time (and I’ve seen many, many very silly attempts to blame vaccines for autism over the years). Basically, Tomljenovic and Shaw tried to “prove” that the rising autism prevalence is due to the use of aluminum adjuvants in vaccines and failed utterly, hilariously confusing correlation with causation in a textbook example of how not to draw inferences from statistical data.

Finally, we’ve also met Shaw before when he testified before an investigative committee looking into the death of an 18-year-old woman named Jasmine Renata in New Zealand. Basically, Jasmine’s mother was convinced that her daughter had died because of the HPV vaccine rather than a much more plausible and likely cause, an undiagnosed heart conduction defect. He claimed that he found HPV DNA associated with aluminum in Jasmine’s brain, which would be a real pharmacokinetic and stoichiometric feat if true given the tiny amount of HPV DNA found in the HPV vaccine. It’s such a small amount that it takes a remarkably sensitive (and possibly nonspecific) nested quantitative real time PCR technique promoted by a discredited pathologist named Sin Hang Lee working with (IN)SANE Vax to detect it. So annoyed was Dr. Shaw at my criticism of his testimony that he even briefly showed up in the comments of this post.

All of this brings us to the latest spew by Tomljenovic and Shaw, which was published in a journal I had never heard of, Pharmaceutical Regulatory Affairs, and entitled Death after quadrivalent human papillomavirus (HPV) vaccination: Causal or coincidental? Guess which side Tomljenovic and Shaw want to persuade you to accept as the correct side of this question? (Hint: It ain’t the science-based side.)

Basically, this article consists of two case reports. Case 1 is very clearly Jasmine Renata, whose case I discussed in depth about three months ago. The case description matches almost exactly:

A 19-year-old female without a relevant medical history and taking no drugs expired in her sleep, approximately 6 months after her third and final qHPV vaccine booster and following exacerbation of initial vaccination-related symptoms. She had last been seen alive by her parents the previous evening. Her symptoms started after the first qHPV injection when she developed warts on her hand that persisted throughout the vaccination period. In addition, she suffered from unexplained fatigue, muscle weakness, tachycardia, chest pain, tingling in extremities, irritability, mental confusion and periods of amnesia (memory lapses). The autopsy was unremarkable and failed to determine the exact cause of death. Internal examination revealed some minor changes involving the gallbladder and the uterine cervix (both of which on further examination by microbiological studies and histology revealed no significant disease). After a full autopsy no major abnormality was found anatomically, microbiologically or toxicologically that might have been regarded as a potential cause of death. Histological analysis of the brain hippocampus, cerebellum and watershed cortex allegedly revealed no evidence of neuronal loss or neuroinflammatory changes. However, the autopsy report did not specify which immune antibodies and stains were used for histological investigations.

Case 2 is very almost certainly another cause célèbre in the antivaccine movement, Annabelle Morin of Quebec, who died in 2008 and whose parents are suing Merck for her death after apparently encountered Joe Mercola’s fear mongering about Gardasil online. The clinical history presented sounds very much like Morin’s:

A 14-year-old female with a previous history of migraines and oral contraceptive use developed more severe migraines, speech problems, dizziness, weakness, inability to walk, depressed consciousness, confusion, amnesia and vomiting 14 days after receiving her first qHPV vaccine injection. These symptoms gradually resolved. However, 15 days after her second qHPV vaccine booster she was found unconscious in her bathtub by her mother 30 minutes after she had entered the bathroom to have a shower. Emergency help was summoned and arrived quickly. Resuscitation efforts were attempted. The paramedic noted that the patient was found without a pulse. Upon arrival at the hospital and approximately 30 minutes later, the patient suffered cardiac arrest. Resuscitation was terminated approximately 40 minutes later and the patient was pronounced dead.

Given the resemblance, it must be Annabelle they’re writing about here.

The first thing I noticed here s that Shaw’s story seems to be…evolving. What do I mean? Simple. Go back and look at what he said about Jasmine originally. He claimed that he found HPV DNA in her blood and spleen after her death, saying to the inquest into Jasmine Renata’s death:

He said it was not the result of a natural HPV infection, most likely the DNA was bound to aluminium which was also found in Jasmine.

“The HPV gene is foreign DNA and its detection six months after injection is not normal,’’ he told the inquest.

He said the DNA may cause a reaction that could lead to lethal shock although it was not known if it caused her death but it needed further investigation.

He said it was not known if it was the cause of death but it needed further investigation.

So let’s see. Originally, Shaw was saying that somehow the aluminum adjuvant in the HPV vaccine somehow complexed with HPV DNA from the vaccine in order to work its evil effects. He said this even though, as I explained above, the amount of HPV DNA in a dose of Gardasil is so small that it takes ridiculously sensitive PCR techniques to detect it, and, even then, I’m not entirely convinced that what Lee found really was HPV DNA and not a contaminant. We’re probably talking nanogram, if not picogram, quantities of DNA. It defies basic chemistry and plausibility to propose that such a minuscule amount of DNA not packaged in a virus or other delivery vector could cause such a reaction. It makes one wonder, it does. Did Shaw realize that his original story didn’t pass the smell test? Did he realize that molecular biologists were rolling in the floor with laughter or snorting at him with contempt over this explanation? Did he actually take my post to heart? Think of it. His manuscript is listed as having originally been submitted on September 13, 2012 and accepted for publication on October 2, with a publication date of October 4. While noting that that’s a mighty fast turnaround time (which makes me wonder about the quality of the peer review for this journal), I also note that it’s also over a month after my post. Sure, I could have an overinflated sense of my own importance. It wouldn’t be the first time I’ve been accused of that. But, still, I wonder.

Whatever the reasons, Tomljenovic and Shaw are now claiming something somewhat different but only marginally more plausible. I’m not a neuroscientist, nor am I a pathologist, but one thing I can’t help but note is that neither Tomljenovic and Shaw are pathologists, either. In particular, neither of them are neuropathologists. If there’s one thing I know about neuropathology, it’s that it’s tricky. It’s not trivial to get antibodies to work properly on brain tissue, nor is it a trivial matter to interpret. None of this seems to have induced our not-so-dynamic duo to have recruited a neuropathologist to assist them with interpretation of the immunohistochemical stains of the brain sections. Were I a reviewer for this paper, I would have recommended not publishing it without a neuropathologist as an author. None of this keeps the authors from boldly proclaiming that neuroinflammation caused by the HPV-16L1 antigen is strongly implicated as the cause of these girls’ deaths through an autoimmune vasculitis due to the deposition of HPV-16L1

Not so fast, there pardner.

The first thing to consider is the biological plausibility of the HPV-16L1 protein somehow depositing in the brain vasculature in order to be able to cause an immune reaction. Each dose of Gardasil contains approximately 20 μg HPV-16L1 protein, which is injected locally intramuscularly. While this is clearly not difficult to detect the way the trace DNA left over in Gardasil is, it’s still by no means a lot of protein. Yet, according to our not-so-dynamic duo, this tiny amount of protein caused this:

In both cases, the autopsy revealed no anatomical, microbiological nor toxicological findings that might have explained the death of the individuals. In contrast, our IHC analysis showed evidence of an autoimmune vasculitis potentially triggered by the cross-reactive HPV-16L1 antibodies binding to the wall of cerebral blood vessels in all examined brain samples. We also detected the presence of HPV-16L1 particles within the cerebral vasculature with some HPV-16L1 particles adhering to the blood vessel walls. HPV-18L1 antibodies did not bind to cerebral blood vessels nor any other neural tissues. IHC also showed increased T-cell signalling and marked activation of the classical antibody-dependent complement pathway in cerebral vascular tissues from both cases. This pattern of complement activation in the absence of an active brain infection indicates an abnormal triggering of the immune response in which the immune attack is directed towards self-tissue.

Here’s the thing. The autopsy didn’t show any abnormalities in Jasmine’s brain that could account for her death. The same was true of Annabelle. If these girls had an immune-based vasculitis, it should have been visible as severe inflammation in the brain tissues on normal H&E sections (sections stained with the usual blue and pink dye and no special immunohistochemical stains). Then, the IHC would reveal the potential cause. For instance, in Case 1, Tomljenovic and Shaw note that the pathologist found no signs of neuronal loss or neuroinflammatory changes but that it wasn’t reported which antibodies were used. More than likely the answer is none; routine sections would just be H&E, with IHC reserved for cases in which an abnormality was noted on H&E. In Case 2, Tomljenovic and Shaw bemoaned how the pathologist didn’t use any specific antibodies for inflammatory markers but rather noted changes consistent with terminal ischemic-hypoxic encephalopathy. What that means is that the pathologist thought that the changes he observed were consistent with the brain’s having suffered a significant period of time with low blood flow, rendering it ischemic, something that could have happened if the girl had a very low blood pressure for a while before she died.

So what we have is a non-pathologist, who clearly doesn’t know how to interpret common pathological findings, throwing every inflammation-related antibody in the book at the brain section and then concluding that there is some sort of neuroinflammation because he sees staining in the blood vessels in the brain. There’s one huge problem, though, and Tomljenovic and Shaw basically admit it:

The obvious limitations of our study are that the tissues examined represent two individuals against which there were no control samples. For this reason, we could not obtain a quantitative measure of immunoreactivity. We aim in the future to further corroborate our findings by examining brain tissues from other cases of sudden and unexplained death following HPV vaccination, as well as control brain tissue from age-matched individuals who clearly died from nonvaccination related causes. Nonetheless, the marked resemblance in immunostaining patterns for all immunohistological markers in brain tissue specimens in the present two cases (i.e., compare Figures 1-4), as well as the similarity between their symptoms and those noted on VAERS reports related to post-HPV vaccination vasculopathies (some of which were medically ascertained cases; Tables 2 and 3), strongly support our present conclusions.

Uh, no they don’t. Tomljenovic and Shaw’s findings could just as well be nonspecific immunostaining. Not only didn’t they examine sections of age-matched normal brains as controls (the very minimum they could have done), but, as far as I can tell from this paper, they neglected to do some very basic controls for any IHC experiment. Specifically, I don’t see any reference to a “no primary antibody” control or an isotype control. In the absence of such controls, there is no compelling evidence that the staining that Tomljenovic and Shaw are reporting is anything other than due to nonspecific interactions between the antibody and cellular structures. This is a lot more common than one might think. Many are the physicians and scientists who have tried to do IHC for a research project and found that it’s nowhere near as cookbook as the antibody and reagent manufacturers would like you to think from their protocols. It’s one reason why IHC is often referred to as an art as much as a science. it often takes a lot of trial and error to get it right, even with commercial antibodies with seemingly well-defined protocols from the manufacturer.

Another thing they could have done to convince skeptics like me is something called an adsorption control Basically, that means incubating the primary antibody with the antigen it’s supposed to detect. If the interaction being observed is specific, then adsorbing it this way should eliminate the staining, because the specific antibody will complex with the antigen. Finally, there should be a positive control. For HPV proteins, this would mean perhaps cervical cancer lesions known to be due to HPV infection. I will admit that a lot of investigators don’t report all of these controls, but they usually report at least a couple of them, usually the isotype control and the no primary antibody control. In particular, for a novel or unusual finding (such as the finding of HPV-16L1 in the neurovasculature), more controls need to be done. Also, reviewers will give a skilled pathologist a bit more of the benefit of the doubt if he doesn’t present all controls, because, well, he’s a pathologist and does this sort of thing every day for a living in real patient samples. Tomljenovic and Shaw don’t rate that sort of benefit of the doubt. They should show their work, but do not. If I had to guess, what we’re seing in their IHC sections is nonspecific staining.

That having been said, I’m not a pathologist, either, although I’ve done a fair amount of IHC in my laboratory. The difference, of course, is that I’m not diagnosing patients. In any case, as Dirty Harry once said, “A man’s got to know his limitations.” I know mine. I know that I could be wrong. So I’d love it if any of my readers are pathologists, particularly neuropathologists, could comment. In particular, the choice of antibodies by Tomljenovic and Shaw seems to include antibodies that are mainly used for research and not routinely used in clinical specimens for actual diagnosis, which means even more that, even if what we’re seeing is not nonspecific, its significance is unclear. Still, given the tiny amount of HPV-16L1 in each Gardasil dose, I’d bet that the finding of this protein in the neurovasculature of these girls is almost certainly nonspecific staining.

Basically, Tomljenovic and Shaw have become the latest not-so-dynamic antivaccine duo trying to lend scientific credibility to quack antivaccine views. They’ve clearly been co-opted by the antivaccine movement and have become true believers, so much so that Shaw is in danger of throwing away whatever scientific credibility Shaw once had. I hope they like their new friends. If they continue on this path, they’ll soon be replacing their scientific colleagues for a bunch of cranks, just the way Andrew Wakefield, Mark and David Geier, and Dr. Sing Han Lee have. Soon all they’ll be good for is giving aid and comfort to antivaccine cranks by churning out crap studies that can be cited over and over and over again by antivaccinationists. Oh, joy.