The return of the Medical Hypotheses anti-vaccine howler

When it rains it pours, as they say.

Yes, sometimes there’s so much going on that I can’t possibly blog about it all, particularly now that I’ve cut back a bit. This week seems to be turning into one of those weeks. Yesterday, I couldn’t resist having a bit of fun with the grande dame of the anti-vaccine movement, Barbara Loe Fisher when she released a seriously hypocritical and silly press release whining about how mistreated she thinks her organization, the Orwellian-named National Vaccine Information Center (NVIC) has been because the American Academy of Pediatrics (AAP) had the audacity to–gasp!–write a letter to Delta Airlines complaining about its airing of an NVIC video about the flu vaccine that was cleverly packed with subtle misinformation topped off with a link to the NVIC website and its cornucopia of anti-vaccine canards. However, to have that bit of fun, I had to put off blogging about another hilariously off-base bit of pseudoscience from the anti-vaccine movement, this time courtesy of everyone’s favorite promoter of quackery on the Internet (with the possible exception of Mike Adams), “Dr.” Joe Mercola. So what’s Mercola up to this time? This time, he’s flogging another bit of anti-vaccine pseudoscience on his blog in a post entitled Is Alarming Rise in Autism Linked to 1988 Event? I could just answer Mercola’s question by saying no, but I was curious what he was talking about, which is always my downfall (or at least the downfall of some of my neurons, which inevitably undergo apoptosis when exposed to the burning stupid that the anti-vaccine movement is capable of). I had to suffer; so I figured I’d make my readers suffer through it too, the difference being that I can usually make it entertaining.

So let’s see what Mercola is talking about:

In 1988, the first conjugate vaccine was approved for use in the U.S.

It was intended to protect infants and young children against Haemophilus influenzae type b (Hib); a bacterial infection that can lead to pneumonia, infections of your blood, joints, bones, and pericardium.

Historically, it has also been a leading cause of bacterial meningitis.

Since that time, the vaccine has been approved in most developed countries, including Denmark and Israel where the vaccine was added to their national vaccine programs in 1993 and 1994, respectively.

Starting in the late 1980’s, there was a marked increase in the reported prevalence of autism spectrum disorders among children in the U.S.

A similar increase was seen in Denmark and Israel.

Do I smell a post hoc ergo propter hoc fallacy here? Is there a correlation between global warming and the decreasing number of pirates in the world? But rather than wasting too much time with Mercola’s credulous regurgitation of this alleged study, let’s find the study itself. This isn’t too difficult because, unlike a lot of cranks, Mercola is actually kind enough to provide a link to the study, which is by someone named Brian J. Richmand and entitled Hypothesis: Conjugate vaccines may predispose children to autism spectrum disorders.

“Hypothesis”? As Scooby-Doo would say, “Ruh-roh, Raggy.”

It looks as though this particular “study” is in the crank journal Medical Hypotheses (MH). It’s been a while since I’ve written about this particular journal; so let’s take a brief trip down memory lane. MH has “distinguished” itself from other medical and scientific journals in its willingness to publish things such as a “hypothesis” by autism quacks Mark and David Geier that provided the basis for their use of chemical castration to treat “vaccine-induced” autism, a “hypothesis” claiming that antiperspirants cause breast cancer, a “hypothesis” by prominent anti-vaccine activists that thimerosal in vaccines causes autism, as well as other hypotheses claiming that masturbation is a treatment for nasal congestion or that high-heeled shoes are linked to schizophrenia. Finally, MH went one bizarre hypothesis too far when it published an article by arch-HIV/AIDS denialist Peter Duesberg that was so beyond the pale that it was too much for Elsevier, the publisher of MH, which finally told then-editor Bruce Charlton that enough’s enough. After Elsevier declined to renew Charlton’s contract, I was curious what would become of MH, given that MH’s–shall we say?–openness to virtually any hypothesis, no matter how wacky and how without a basis in science derived mainly from Charlton’s editorial philosophy. Don’t get me wrong. As I’ve pointed out, I always thought that there was a place in the medical literature for a journal devoted to exploring highly speculative hypotheses, but the problem with MH is that it became a magnet for cranks and, because it was listed in PubMed, it was easy for cranks to represent it as a legitimately peer-reviewed journal and cite its articles even though Charlton admitted that it wasn’t peer-reviewed.

It would appear that the editorial standards under the new regime are the same as they were under the old regime. In other words, meet the new boss, same as the old boss. At least this article would seem to argue that that’s the case. Let’s go to the abstract:

The first conjugate vaccine was approved for use in the US in 1988 to protect infants and young children against the capsular bacteria Haemophilus influenzae type b (Hib). Since its introduction in the US, this vaccine has been approved in most developed countries, including Denmark and Israel where the vaccine was added to their national vaccine programs in 1993 and 1994, respectively.

There have been marked increases in the reported prevalence of autism spectrum disorders (ASDs) among children in the US beginning with birth cohorts in the late 1980s and in Denmark and Israel starting approximately 4-5 years later. Although these increases may partly reflect ascertainment biases, an exogenous trigger could explain a significant portion of the reported increases in ASDs. It is hypothesized here that the introduction of the Hib conjugate vaccine in the US in 1988 and its subsequent introduction in Denmark and Israel could explain a substantial portion of the initial increases in ASDs in those countries. The continuation of the trend toward increased rates of ASDs could be further explained by increased usage of the vaccine, a change in 1990 in the recommended age of vaccination in the US from 15 to 2 months, increased immunogenicity of the vaccine through changes in its carrier protein, and the subsequent introduction of the conjugate vaccine for Streptococcus pneumoniae.

Although conjugate vaccines have been highly effective in protecting infants and young children from the significant morbidity and mortality caused by Hib and S. pneumoniae, the potential effects of conjugate vaccines on neural development merit close examination. Conjugate vaccines fundamentally change the manner in which the immune systems of infants and young children function by deviating their immune responses to the targeted carbohydrate antigens from a state of hypo-responsiveness to a robust B2 B cell mediated response. This period of hypo-responsiveness to carbohydrate antigens coincides with the intense myelination process in infants and young children, and conjugate vaccines may have disrupted evolutionary forces that favored early brain development over the need to protect infants and young children from capsular bacteria.

First off, this is the first time I’ve seen an anti-vaccine advocate try to link the Hib vaccine to autism. Maybe I live a sheltered life (although I doubt it, given how much I dive into the writings of anti-vaccine advocates and various anti-vaccine websites). Be that as it may, the Hib vaccine is an incredible vaccine that has made an enormous impact in a brief 20 years or so. Before the Hib vaccine, pediatricians would dread seeing children in the emergency room with Hib because it could cause such serious disease and even death. Hib was a pediatrician’s nightmare. Since the Hib vaccine hit the market, incidence and morbidity from Hib have plummeted, and most pediatricians who trained after the early 1990s have never seen a case of bacterial meningitis, pneumonia, or epiglottitis due to Hib. This is the good that the Hib vaccine has done.

As for this “paper,” well, let’s just say that it’s rampant speculation with no evidence based on confusing correlation with causation. Actually, it doesn’t even make a particularly convincing case for a correlation. In the process, Richmand abuses evolution as a way of “explaining” why conjugated vaccines like the Hib vaccine could cause autism. As lawyers would say, Richmand uses a lot of facts not in evidence to try to convince the gullible that his hypothesis (and it’s really a stretch to call it that, given that real scientific hypotheses are generally supported by enough data to make them plausible and worth expending the time and resources to test). Why the lawyer crack? Because it appears that Richmand is not a scientist but rather a lawyer.

As is pointed out, the Hib vaccine was first approved in the U.S. in 1988 for infants starting at age 15 months, and in late 1990 was approved for use in infants aged two months. The vaccine was later approved in Denmark in 1993 and Israel in 1992, later becoming part of the recommended slate of vaccines administered through Israel’s national health care system in 1994. One can’t help but sense a bit of cherry picking of nations here. What about the many European nations other than Denmark that introduced the Hib vaccine into their childhood immunization schedules around the same time? If I were an anti-vaccine advocate and wanted to make a powerful case for a correlation between a vaccine like Hib and any condition, be it autism, neurodevelopmental problems, or whatever, I’d look at data from as many nations as I could. It’s not just the U.S., for instance, that has data regarding autism prevalence for birth cohorts from that period of time. Instead, Richmand picks three countries. One also can’t help but notice that two of these countries are ones where large studies were performed to demonstrate no correlation between thimerosal-containing vaccines and autism, in particular the Danish studies, which were among the earliest large, well-designed studies that failed to find a correlation between thimerosal and autism.

In any case, it’s interesting to note that there are no graphs and no statistics presented to demonstrate that there even is a correlation between increases in autism prevalence and the introduction of the Hib vaccine into the U.S., Denmark, and Israel, but if you really want to know how bad Richmand’s speculations are, look no further than this passage:

Using regression analysis, McDonald and Paul then sought to determine (at a 95% confidence interval) the existence of a change-point in the cumulative incidence for autism disorder in the birth cohorts covered in each of the 3 selected studies, as well as for their worldwide data set. A change-point was calculated for both the California and Danish studies at 1987.5 (95% CI: 1987- 1988) and for the combined worldwide data set at 1988.7 (95% CI: 1985.9-1991.8). No change-point could be calculated for the Kohoku Ward, Japan data set, but the incidence of autism disorder increased steadily for the 1988 through 1996 birth cohorts. Inasmuch as data collection was not started in the Kohoku Ward study until 1988, the absence of a change-point in the Kohoku Ward data is consistent with the findings of the 1987.5 change-points for the California and Denmark data sets and the 1988.7 change-point for the worldwide data set.

It should also be noted that the change-point calculated by McDonald and Paul for the California data set for autism disorder is consistent with the previously discussed findings regarding combined ASDs among US school-aged children based on administrative data. Also of interest with regard to this hypothesis are comparative graphs presented by McDonald and Paul showing the rise in rates of autism disorder in the US and Denmark. Although McDonald and Paul calculated an identical change-point for both countries, these graphs indicate that the initial increase in autism disorder rates in Denmark was relatively modest and the preponderance of the increase in Danish autism disorder rates lagged the increase in the California study by several years.

I thought this study sounded familiar; so I did a quickie search of my blog archives, and, lo and behold, it turns out that I’ve blogged about this study before. Basically, it’s a pretty lousy study that used questionable statistical methodology in order to identify what its authors considered to be “change points” or inflections in the incidence curve over time using three studies, one in Japan, one in the U.S., and one from Denmark. An anti-abortion group jumped all over it and claimed to correlate various “change points” with the introduction of vaccines containing “fetal parts.” (I kid you not; I really can’t make stuff like this up.) Of course, there is a problem. In the U.S. study the change point identified occurred in 1987, before the introduction of the Hib conjugate vaccine into widespread use in the U.S.. Ditto Denmark, where the change point was also 1987 and, supposedly the change point for the world was 1988, although I never figured out how one could say that from just three studies in three nations. Be that as it may, this study is pretty darned thin gruel to conclude anything about autism incidence and whether there were “environmental triggers” introduced 23 years ago that caused inflections in the autism incidence curve.

As for the rest of the arguments Richmand makes, he really stretches to try to make a correlation. The funny thing is, though, that even if the McDonald study were correct, it would undermine Richmand’s hypothesis because his hypothesis would predict that autism incidence would take off two or three years after the introduction of the Hib vaccine in a nation for infants (i.e., the dose at two months). That would mean, in the U.S. at least, late 1993 at least. In any case, no such increases appear to correlate very closely with the introduction of Hib in any of the countries examined other than in a vague, general way. More importantly, the introduction of the broadened diagnostic criteria in the DSM-IV introduced in 1994 (in the U.S.) and 1996 (in Europe) rears its ugly head again for those determined to confuse correlation with causation; i.e., anti-vaccinationists. To paper over that hole in his hypothesis, Richmand starts blaming efforts to increase the immunogenicity of the vaccine through the 1990s and changes in time of the protein carrier used in the vaccine and then the approval of the pneumococcal vaccine in 2000.

Starting with a weak to nonexistent association, what does an MH author do? He gets out a shovel and digs deeper, this time invoking immunology to try to explain why conjugated vaccines are different (and presumably cause teh autism). He goes on and on about differences in B-cell and T-cell responses elicited by carbohydrates:

Although conjugate vaccines are truly life saving, these vaccines by-pass important immunological regulatory mechanisms and, consequently, deserve careful retrospective scrutiny. In the absence of conjugate vaccines, B cells with an affinity for carbohydrate antigens are carefully regulated in order to prevent autoimmunity [30-32]. Because antibodies produced against carbohydrate antigens (including the Hib capsule) are often intrinsically autoreactive with self-carbohydrates, differentiation of these B cells into B2 cells may lead to autoimmunity [30-32]. Antibody to self-carbohydrates has been associated with several autoimmune disorders, including systemic lupus erythematosus [30,31], myocarditis and rheumatic heart [36], Sydenham’s chorea [37], and pediatric neuropsychiatric disorder associated with Streptococcus (PANDAS) [38]. Unlike antibodies produced by B2 cells, the shorter-lived, lower affinity antibodies produced by B1 and MZB cells are less likely to lead to autoimmunity [30,36].

Of course, no strong evidence is presented that any of these immunological changes have anything to do with the pathogenesis of autism, but none of this stops Richmand from invoking an evolutionary basis for these immune responses and doing more handwaving about how they might be relevant to autism:

The hypo-responsiveness of the immune systems of infants and young children to capsular bacteria is unexpected from an evolutionary perspective. In addition to the inability of infants and young children to mount an effective TI immune response to capsular bacteria, IgM and IgG2 do not readily cross the placenta [5,40], suggesting that fetuses also may not be well protected from capsular bacteria. The vulnerability of infants, young children, and possibly fetuses to capsular bacteria would appear to create a significant evolutionary disadvantage.

These multiple limitations on the presence of antibody to carbohydrate antigens in fetuses, infants, and young children raise the question of whether there is a countervailing evolutionary advantage to these limitations. One possible explanation is that antibody to carbohydrate antigens may be cross-reactive with neural glycoproteins between protection against capsular bacteria and the need for neural development.

This possibility is consistent with the fact that myelination occurs in stages from infancy through at least early childhood [34], which appears to coincide with the period during which the immune system is hypo-responsive to carbohydrate antigens. Moreover, the period during which the immune system is least responsive to carbohydrate antigens corresponds to the period of most intense myelination. Viewed in this context, it is possible that not only is antibody to carbohydrates deleterious to neural development in infants and young children, but their negative effects are enhanced by the robust TD immune responses induced by conjugate vaccines. The balance struck through evolution may have been significantly disrupted through the introduction of conjugate vaccines for infants and young children (Chart 2).

Even if all of this were true, it would say nothing about whether Richmand’s hypothesis were reasonable, because for Richmand’s hypothesis to be worth exploring, there would have to be a real correlation between the introduction of the Hib vaccine and an increase in the incidence of autism. Quite simply, he doesn’t make the case that there is.

Now, I know what you’re probably thinking. You’re probably thinking: This is MH. What did you expect? Fair enough. I was more interested in whether MH had cleaned up its act after the ouster of Bruce Charlton as its editor. Clearly, it hasn’t, and the new regime has apparently happily published more fodder for the anti-vaccine movement, so much so that an outrageously credulous accompanying editorial by Noel R. Rose of the Johns Hopkins University School of Medicine. Dr. Rose appears to be a bit on the edge as far as anti-vaccine beliefs go, at least if his editorial Conjugate vaccines and autism is any indication. Either that, or Dr. Rose is so open-minded that he’s allowed his brains to fall out, given that in his editorial he praises Richmand as an example of noncientist “original thinkers who do not have the standard diplomas, writing that he has “raised significant questions” and arguing that the “merits of this thoughtful hypothesis should be carefully weighed.” To me, the only significant question that Richmand’s article raised was how on earth MH could keep publishing such tripe after Elsevier supposedly cleaned house among its editorial staff and instituted a system of peer review and the only merit of this hypothesis that should be “carefully weighed” is whether a printout of the article would make a good lining for a bird cage. After all, if this article by Richmand is any indication, MH continues be–shall we say?–“open” to pseudoscience and willing to provide nonscientists with anti-vaccine proclivities a forum that provides them with a patina of seeming respectability plus fodder for cranks like Joe Mercola to write posts castigating vaccines. Same as it ever was.