More Gardasil fear mongering: A “critical review” of HPV vaccination that lacks critical thinking

Here we go again.

Antivaxers don’t like vaccines. This, we know. They blame them for everything from autism to autoimmune diseases to diabetes to sudden infant death syndrome. They even sometimes claim that shaken baby syndrome is a “misdiagnosis” for vaccine injury. However, there are two vaccines that stand out above all as the objects of antivaccine scorn. the first, of course, is the MMR vaccine. That’s on Andrew Wakefield., of course, who almost singlehandedly popularized the fear that the MMR vaccine causes autism. The second most hated vaccine (by antivaxers) is Gardasil or Cervarix, both vaccines against the human papilloma virus (HPV), the virus that causes genital wart and certain subtypes of which greatly predispose to cancer. The reason for the extreme negative reaction to HPV vaccines is not because they are thought to cause autism, given that they are not administered until girls are approaching pubert. Rather, it is because they are falsely considered to be unnecessary. HPV is sexually transmitted disease, which makes the cervical cancer HPV vaccines prevent a largely sexually transmitted disease. Many religiously inclined antivaxers falsely think that HPV vaccination will encourage promiscuity, even though the evidence is pretty strong that it doesn’t. Nor does it cause premature ovarian failure or sudden death, as antivaxers frequently claim. Nor is it contaminating your precious bodily fluids with HPV DNA.

In which Orac learns of a “horrifying” review article about HPV vaccination

So it was with little surprise at all that yet another allegedly systematic review article attempting to demonstrate that HPV vaccines are not safe. (They are.) I’ve been down this road before, but I do tend to take a look when new examples of these sorts of studies roll around because I never know what I am going to find and also know that these sorts of studies tend to get shared in a rabidly viral fashion on social media by antivaxers. If I can get my take on them out before that happens, then at least there’s something out there to combat the misinformation. I didn’t quite get there in that our old buddy “Dr. Jay” Gordon posted a link to Twitter:

On the other hand, Dr. Jay did tip me off to the study. So I suppose I should thank him for that. In any case, let’s first look at how the article is being reported. The link above is to European Pharmaceutical Manufacturer (epm) a magazine that really should know better:

The study, performed by researchers from Mexico’s National Institute of Cardiology and published online in Clinical Rheumatology, identified randomised trials of HPV vaccines in PubMed and reviewed post-marketing case series highlighting serious adverse events of the vaccines.

It was found that in the majority of the randomised studies identified the control groups were not administered with an inert placebo but rather an aluminium placebo. According to a report published by Collective Evolution using an aluminium based placebos would offer up much different comparative results than when administering a pharmacologically inert placebo.

When comparing the rates of serious adverse events, the researchers found that in two of the largest randomised trials, there were more occurrences in the HPV treated groups than in the aluminium placebo groups. Additionally, comparing girls vaccinated with the 4-valent HPV vaccine and those receiving the 9-valent dose, the latter group had more serious systemic adverse events. The researchers also revealed a ‘worrisome’ quotient of the number needed to vaccinate/number needed to harm in the nine-valent HPV vaccine.

Hoo-boy. I see tropes. I see antivaccine tropes. So. Many. Tropes. Also, if you link to Collective Evolution, you lose. It’s a wretched hive of scum and quackery. Maybe not as wretched as, say Natural News, but it’s pretty bad.

Not surprisingly, it didn’t take me long to find this article by antivaccine crank Robert F. Kennedy, Jr., Vaccine Manufacturers and FDA Regulators Used Statistical Gimmicks to Hide Risks of HPV Vaccines:

A new study published in Clinical Rheumatology exposes how vaccine manufacturers used phony placebos in clinical trials to conceal a wide range of devastating risks associated with HPV vaccines. Instead of using genuine inert placebos and comparing health impacts over a number of years, as is required for most new drug approvals, Merck and GlaxoSmithKline spiked their placebos with a neurotoxic aluminum adjuvant and cut observation periods to a matter of months.

Researchers from Mexico’s National Institute of Cardiology pored over 28 studies published through January 2017—16 randomized trials and 12 post-marketing case series—pertaining to the three human papillomavirus (HPV) vaccines currently on the market globally. In their July 2017 peer-reviewed report, the authors, Manuel Martínez-Lavin and Luis Amezcua-Guerra, uncovered evidence of numerous adverse events, including life-threatening injuries, permanent disabilities, hospitalizations and deaths, reported after vaccination with GlaxoSmithKline’s bivalent Cervarix vaccine and Merck’s quadrivalent or nine-valent HPV vaccines (Gardasil and Gardasil 9). Pharmaceutical company scientists routinely dismissed, minimized or concealed those injuries using statistical gimmicks and invalid comparisons designed to diminish their relative significance.

I note that the article linked to under “invalid comparisons” was published by antivaccine activists Lucija Tomljenovic, Judy Wilyman (yes, that Judy Wilyman), Eva Vanamee, Toni Bark, and Christopher A Shaw. (More on why the comparisons are not invalid later, when I get into the article itself.)

Roll the tape: A biased, crappy “review” article designed to demonize HPV vaccines

So let’s look at the review article, which is by Manuel Martínez-Lavín and Luis Amezcua-Guerra at the National Institute of Cardiology in Mexico City. So, the first thing I wondered is: WTF is a vaccine article doing being published by Rheumatology? The second thing I wondered was: The National Institute of Cardiology in Mexico? Doing a vaccine study? It certainly is odd. The next thing I wondered was this: Who the heck are Manuel Martínez-Lavín and Luis Amezcua-Guerra? The name Manuel Martínez-Lavín seemed to ring a bell. I seem to recall having heard it before. Fortunately, almighty Google helped out. All I had to do was to search each name with “HPV,” “Gardasil,” or “Cervarix.” I immediately discovered that Martínez-Lavín is not exactly what one would call an…unbiased…source. Let’s just put it this way. When you’re approvingly cited by the antivaccine website SaneVax, a site that’s dedicated to fear mongering about Gardasil. It also turns out that Martínez-Lavín is not above doing highly dubious surveys without controls. From these, I learned that Martínez-Lavín apparently believes that HPV vaccines cause fibromyalgia, despite the lack of evidence that it does anything of the sort. Somehow I missed this horrible study, but fortunately Skeptical Raptor and Dr. Jen Gunter did not. Now, I must realize that I didn’t find much about Luis Amezcua-Guerra, but it’s not hard to see how biased the article by Manuel Martínez-Lavín and Luis Amezcua-Guerra is, even before looking at it in detail.

Let’s look at the first trope, which is a common anti-Gardasil trope that it is an inappropriate control to compare an aluminum-containing vaccine like Gardasil to an aluminum adjuvant without the actual antigens from the vaccine. The argument is that the best control should have been normal saline; i.e., an inert control. This is a profoundly ignorant argument when you have an intervention known to be safe based on many studies in many vaccines over the years (like aluminum adjuvants). When you have such an ingredient, then if you want to determine whether or not a vaccine containing that ingredient works and is safe, an excellent way to do it is to compare it to a control containing everything in the vaccine except the antigens that produce the immune response. In other words, the adjuvant-only control is a very good control. Channeling antivaccine tropes aplenty, the authors of the review try their best to convince you that the real reason this control was chosen in so many studies of Gardasil and Cervarix was to hide adverse events due to these vaccines. Of course, the existence of long term studies (like this one) comparing HPV vaccines to saline placebo controls rather undermines this particular antivaccine talking point. Basically, we have evidence from both studies comparing HPV vaccines to adjuvant-only controls and to saline controls showing that HPV is both effective and safe.

Again, go back to Lucija Tomljenovic’s paper cited earlier. That’s exactly the main claim in the paper. Let’s just put it this way, if an antivaccine crank like her makes an argument, that’s a pretty good indication that’ what’s being argued is pure, grade A BS. Of course, this is pure, grade A BS independent of Tomljenovic’s use of it. A good way of looking at it is that Tomljenovic uses it because she is an antivaxers and antivaxers gravitate to claims about vaccines that are grade A unadulterated BS. Also look at it this way: If Martínez-Lavín and Amezcua-Guerra think this is a such a compelling argument that they make it the centerpiece of their “analysis” of the randomized controlled trials (RCT) of Gardasil and Cervarix, it starts to make me question everything else in the paper. Adding to that impression is my perusal of the references, which include the works of such antivaccine “scientists” as, yes, Tomljenovic, Yehuda Shoenfeld (who just had a paper retracted), Deirdre Therese Little (who is on the board of advisors of an conservative Catholic Australian antivaccine group that preaches that Gardasil leads to promiscuity and who promotes the false message that Gardasil causes premature ovarian failure). Indeed, the paper by Little cited was deconstructed by yours truly when it came out.

Not surprisingly, this isn’t even a very good systematic or “critical” review. It’s definitely not a meta-analysis, although our intrepid authors do try to make it sound like one by “reanalyzing” data from the papers they want to try to refute. Most of them didn’t show any difference in adverse events in placebo or HPV vaccine group. There were two, however, that, according to the authors, did:

Two of the largest HPV vaccine randomized trials did find significantly more severe adverse events in the tested vaccine group vs. the comparator group: The 4-year interim follow-up VIVIANE study safety analysis compared 2881 healthy women older than 25 years injected with the bivalent HPV vaccine vs. 2871 age-matched women injected with aluminum placebo [29]. As expected in large randomized trials, both groups displayed remarkably similar baseline characteristics. General solicited symptoms during the 7-day post-vaccination period occurred more often in the HPV vaccine group (65%) than those in the control group (58%). Our calculated 2 × 2 contingency table p value was <0.01. Vaccine-related general solicited symptoms during the 7-day post-vaccination period were also more frequent after HPV vaccination (41%) than those after placebo injection (36%) p < 0.001. Fourteen deaths occurred in the vaccine group vs. three deaths in the control group (p = 0.012 by Fisher’s exact test). None of the deaths were believed to be related to vaccination. One less death was reported in the 84-month follow-up VIVIANE study, a woman diagnosed with breast cancer 6 months after the third dose of the vaccine [35]. Even after this correction, the death rate difference (13 vs. 3) remains significant (p = 0.021).

This is just plain silly. The authors are doing 2×2 contingency tables because they didn’t have the raw data, while the authors of the original VIVIANE study did and did much more sophisticated analyses. Here’s what the the actual VIVIANE study says about this:

Solicited injection-site symptoms occurred in more women in the vaccine group than in the control group (table 4). General solicited symptoms during the 7-day post-vaccination period occurred slightly more often in the vaccine group than in the control group. The incidence of unsolicited symptoms, serious adverse events, medically significant conditions, new-onset chronic disease, and new-onset autoimmune disease was similar in both groups, and pregnancy outcomes did not differ between groups (table 4). 17 deaths occurred, 14 (<1%) of 2881 women in the vaccine group and three (<1%) of 2871 in the control group; none of the deaths were believed to be related to vaccination. The independent data monitoring committee did an unblinded review of all deaths; the causes of death were very variable and no cluster of disease type was noted (appendix p 4). The mean time between the last vaccination and death was 682 days (SD 321) in the vaccine group and 496 days (424) in the control group (range 67–1191 days for both groups), suggesting no temporal relation between vaccination and death.

In other words, the difference was mainly more injection site problems, which would be expected for an active vaccine being compared to just the adjuvant. One would expect more inflammatory reactions. As for the deaths, they were analyzed by an independent data monitoring committee and showed no pattern that suggested a link to the vaccine. Taking a look at the list in the appendix should tell you that it’s unlikely there was a link (click to embiggen):

For instance, there are three suicides, one homicide (it strains credulity, even for antivaxers, to think that HPV vaccination can cause one to become a murder victim), two cases of breast cancer (both women were in their late 40s), a case of drug hypersensitivity and renal failure. You get the idea.

The authors look at another randomized double blind RCT of Gardasil that contrasted the 9-valent dose versus the quadrivalent formulation. Basically, instead of four serotypes, there are nine in the newer vaccine. I’m fed up with this “review” already; so I’m just going to cite the actual RCT:

The recipients of the 9vHPV vaccine were more likely than the recipients of the qHPV vaccine to have adverse events related to the injection site (90.7% vs. 84.9%), with the most common events (incidence ≥2%) being pain, swelling, erythema, and pruritus (Table Adverse Events.); more than 90% of these events were mild to moderate in intensity. Events of severe intensity were more common in the 9vHPV group. The frequency of systemic adverse events was generally similar in the two groups — 55.8% in the 9vHPV vaccine group and 54.9% in the qHPV vaccine group. The most common systemic adverse events related to vaccination (incidence ≥2%) were headache, pyrexia, nausea, dizziness, and fatigue. Less than 0.1% of participants discontinued study vaccination because of a vaccine-related adverse event. All the serious adverse events are listed according to system organ class in Tables S6 and S7 in the Supplementary Appendix. Pregnancy was reported in 1192 participants in the 9vHPV group and 1129 participants in the qHPV group, and information on outcomes was available for approximately 85% of these pregnancies (Table S8 in the Supplementary Appendix). The proportions of participants with live births, difficulty with delivery, spontaneous abortions, and late fetal deaths were similar in the two groups. Congenital anomalies were reported in a total of 32 infants and 9 fetuses (20 in the 9vHPV group and 21 in the qHPV group). No congenital anomaly was reported in the case of pregnancies with an estimated date of conception that was within 30 days before or after any vaccination (these pregnancies represented approximately 8% of the total number of pregnancies with known outcomes).

So basically, there were more injection site problems in the 9vHPV group, which is not surprising for a more immunogenic vaccine, and there were systemic symptoms like headache, pyrexia, nausea, dizziness, and fatigue, but clearly not that bad if only 0.1% of the participants stopped the three dose series as a result. Not understanding the concept of number needed to treat with respect to vaccines, the authors plunge boldly ahead:

The average number needed to vaccinate with the 9- valent dose to prevent one episode of the pre-specified primary end-points that would not otherwise have been prevented by the 4-valent immunization is 1757 with 95% CI ranging from 131 to infinity

That’s actually pretty damned good for a vaccine compared to another. If you vaccinate millions of women the advantage of the 9vHPV vaccine would become apparent in the form of thousands of additional cases of HPV prevented. Also remember that the 4vHPV contains the four most common types of HPV associated with cervical cancer. Adding five more types would, by definition, add less common types of HPV and produce less benefit.

And the “cover-up” continues

Finally, the authors think that the post-marketing studies are “covering up” adverse events from HPV vaccines. In their table (Table 2), they cite a whole pubnch of studies, including the one by Martínez-Lavin that I cited above that was nothing more than a questionnaire-based study in which he claims to have found an association between HPV vaccination and fibromyalgia in 45 women. They also cite—you guessed it—Tomljenovic twice, including three papers claiming to find the “ASIA syndrome” post-vaccination. It’s a syndrome so vaguely defined and (of course) not accepted by anyone other than Yehuda Shoenfeld and his fellow travelers as a legitimate medical syndrome. They also include dubious papers claiming to link Gardasil to premature ovarian failure.

The authors then go on to cite a number of postmarketing studies looking at HPV vaccine adverse events. A lot of these studies found increased incidence of syncopal symptoms (like dizziness and occasionally passing out) after vaccination. Of course, that’s why we insist that children receiving any vaccination, but Gardasil in particular (given the propensity of girls of the age receive Gardasil have syncopal episodes after blood draws and injections anyway), be observed for a while after receiving the dose. It’s not a reason not to vaccinate. Injections of all types are associated with fainting, as the authors of several of the papers note and the authors of this review do not.

A much better review of postlicensure studies of HPV vaccines by Vichnin et al found that only “syncope, and possibly skin infections were associated with vaccination in the postlicensure setting” and that serious adverse events, “such as adverse pregnancy outcomes, autoimmune diseases (including Guillain-Barre Syndrome and multiple sclerosis), anaphylaxis, venous thromboembolism and stroke, were extensively studied, and no increase in the incidence of these events was found compared with background rates.”

Overall, this is a terrible systematic review. It is clearly designed to make HPV vaccination look as bad as the authors can make it look by playing up known adverse events due to HPV vaccines, such as syncope, claiming that adverse events are vastly underreported, and citing papers by antivaccine cranks as “evidence” that there are all sorts of horrible things caused by Gardasil that “They” don’t want you to know about. Not surprisingly, it’s spreading in the antivaccine crankosphere. Surprisingly, I haven’t seen it on Natural News yet. It’s coming, though. I’m sure of it.