Sunday, July 12, 2009

The MMR-Autism Claim and Bad Science: Part I

Recently, some participants on Dr. Bob Sears The Vaccine Book Discussion Forum asked about the validity of his references that show a possible link between the MMR vaccine and autism. The list appears in The Vaccine Book: Resources pp. 257-259 (2007). So why would Dr. Bob believe that these studies support Wakefield's findings? Well, if you just read abstracts and conclusions, then you could easily make that mistake. But the devil's in the details and in this case, the details are extremely shoddy methods and scientific dishonesty. There are 7 'studies' and we will discuss these in a 2-part series, as they appear in the book and closely examine the methods.

The first is the original article by Wakefield et al. (1998) and this is Dr. Sears' summary:

This is the landmark study that began the whole debate. Wakefield found twelve children who developed regressive autism during the second year of life (after the MMR vaccine was given) who also had intestinal inflammatory disease consistent with a viral infection, probably measles related. See page 94 for further discussion.

The information on this paper is exhaustive and you can start by reading the numerous comments from the original Wakefield paper. Ten of the original 13 authors listed a Retraction of the Wakefield et al. paper and this is an excerpt:

We wish to make it clear that in this paper no causal link was established between MMR vaccine and autism as the data were insufficient. However, the possibility of such a link was raised and consequent events have had major implications for public health. In view of this, we consider now is the appropriate time that we should together formally retract the interpretation placed upon these findings in the paper, according to precedent (Zhang et al.).

So several things happened prior to and after that retraction and the most comprehensive explanation can be found on the website of journalist Brian Deer. While he is a journalist, his reports have held up to scrutiny and are, in part, the foundation for the General Medical Council's (GMC) disciplinary proceedings against Dr. Andrew Wakefield and Professors Simon Murch and John Walker-Smith. It is also a plain-English summary. Additionally, compelling evidence of both scientific fraud and abysmal methodologies came to light during the Omnibus Autism Proceedings (OAP), specifically by the testimonies of Dr. Nicholas Chadwick and Dr. Stephen Bustin. Dr. Chadwick was Andrew Wakefield's graduate student at the time of the now infamous MMR-autism study and could not find measles virus in some samples from work prior to the MMR-autism study. He notified Dr. Wakefield of this, who proceeded to report them as positive anyhow via a paper co-authored with Kawashima et al.; Dr. Chadwick asked that his name be removed from any publication that included these results. Incidentally, Dr. Wakefield was not asked to testify at the OAP.

An explanation of Dr. Bustin's testimony can be found here which also discusses the difficulties with the results found in the next paper to be discussed. In short, Wakefield et al. and subsequent studies seemingly supporting Dr. Wakefield's findings did not find what they claim they found. The researchers didn't take any of the simple measures to ensure the quality of their work and verify their findings. This was a scientific fraud shot-heard-around-the-world.

The next paper is by Uhlmann et al. (2002), which includes Wakefield as a co-author and this is Dr. Sears' summary:

This eleven-doctor group in Ireland did intestinal biopsies on a large group of children with autism and found that of the ninety-one who showed ileal lymphonodular hyperplasia and enterocolitis, seventy-five had confirmed measles virus infecting their intestinal linings. Only five of the seventy healthy control patients tested showed the measles virus. This study offers more support for Dr. Wakefield's claim of a link between measles virus, inflammatory bowel disease and autism.

Before we get to the methods, let's briefly jump to the conclusions. O'Leary's lab found that 82.4% of the children with developmental disorders had measles vaccine virus in their guts. This, by itself is astonishing and should be questioned. By these reported numbers, an investigator that wishes to verify these study results would only have to have a sample size of 6 autistic children (95% confidence interval) to detect a significant number of children with measles vaccine virus as compared to controls; this will become more relevant later.

Now for the methods; there was a control group and a developmentally delayed group but no effort was made to actually match the groups or describe their differences other than developmental status. But one of the most important and fatal flaws with the study, was the molecular tests that they carried out. Again, the full testimony of Dr. Stephen Bustin is available. The Uhlmann group, working in Dr. John O'Leary's now-defunct Unigenetics laboratory in Ireland did not describe their controls in a manner that could control for contamination or other investigators could parse. They also used source material that would have been degraded, and thus rendered useless by the methods they were using to detect measles vaccine virus sequences. The primers that are used to detect a particular sequence of interest, in this case, measles vaccine virus, were mismatched which increases the likelihood that a non-target sequence could be detected. But the most crucial flaw in this methodology is that the investigators never sequenced their results. Sequencing maps out the nucleotide arrangement for the target that was amplified; you then compare that to a DNA library such as GenBank which is a repository for millions of sequences. This would have told them (as it did Dr. Chadwick) that what they were getting was not measles vaccine virus but something else.

It was also noted in Dr. Bustin's audit of Unigenetics operating procedure for the Uhlmann et al. paper, that measles is an RNA virus and there was no RNA in many of their samples yet they used them and counted them as positives. Also, in some cases, they skipped the reverse-transcription (RT) step which converts the RNA to DNA and miraculously still found positives, which would be impossible and is therefore yet another crucial flaw in this experiment. The laboratory was located adjacent to a plasmid room. Since PCR is prone to contamination, all measures must be taken to ensure clean runs and control for contaminants. Plasmids are a source of this and since many people were walking in and out of the rooms, plasmids could easily make their way into reaction vessels. Again, sequencing would have shown them that, yet they refused to utilise this very basic (and inexpensive) technique. Unigenetics gave their samples to another investigator to replicate (as part of an investigation that was going on in the U.K.) and he could not get the same results with the proper extraction and PCR methods. Now Dr. Bustin was under oath to provide his professional observations of his audit of Unigenetics and accordingly, did not speculate about the Uhlmann et al. study. We are not under those constraints and will daresay the extraordinary lopsidedness of their results, does point to scientific fraud.

Two groups of investigators attempted to replicate Uhlmann et al.'s study and could not get the same results. The first study is by Afzal et al. (2006) who replicated the parameters of Uhlmann et al. BUT with more sensitive molecular methods, proper controls and quantification of starting RNA. They took blood samples from 15 children with autistic disorders that had the MMR jab, 10 of them reportedly regressed in the second year of life and after the MMR jab. They specifically examined leukocytes (white blood cells) because if there is measles virus replicating in their guts as Wakefield et al. contends, then evidence of the virus is going to be found elsewhere. Incidentally, Dr. Afzal asked Wakefield/Uhlmann/O'Leary for their samples to be tested but were refused, which in itself is highly unusual in the scientific community. Afzal et al. did not detect any measles virus, vaccine or wild-type in the children. Although their sample size was small (n=15), given the 82.4% of children with measles vaccine virus present that Uhlmann et al. claimed to have found, they should have detected 6 positives.

A relatively small number of patients was examined in this study. However the reported frequency of positive samples in the largest positive study is about 82% (Uhlmann et al., 2002). While it is impossible to prove a negative, it is clear that this figure is not consistent with zero positive from 15 as here. In fact if the null hypothesis is that 82% of patients have measles sequences, it would be inconsistent with fewer than 6 positives from 15 samples at the 95% confidence level. The failure to find any positives means that the incidence is far lower than this, and possibly zero.

The second study is by D'Souza et al. 2006 and followed the Uhlmann et al. and Kawashima et al. protocols even more closely but with, again, better precision and optimisation of their methods. They used peripheral blood mononuclear cells (PBMCs) which are a perfectly acceptable surrogate for gut biopsies or cerebral spinal fluid. They also had a larger study group with controls and their characteristics included gastrointestinal complaints and the majority receiving 2 doses of MMR:

Fifty-one of 54 children from the ASD group and 31 of 34 from the control group had received MMR vaccine. From the ASD group, 2 children had not been vaccinated, and 1 child’s status was unknown. From the control group, vaccination status was unknown for 3 children. The majority of children with ASD (n = 43; 84%) and the developmentally normal children (n = 28; 90%) had received 2 doses. The Quebec immunization schedule recommends a 2-dose MMR schedule at ages 12 and 18 months.14 Only 1 (control) child had received 3 doses. Gastrointestinal symptoms, including constipation, cramping, diarrhea, and abnormal stools, were reported in 43 (79.6%) of 54 children from the ASD group and 11 (32.3%) of 34 children from the control group. Regression was reported in 20 (37%) of 54 subjects in the ASD group. The time between vaccination and blood analysis was similar between the 2 groups. The mean time (±SD) between vaccination and blood sampling for the ASD group was 30.4 (±15.9) months and 26.6 (±15.3) months for the control group (2-tailed t test, p = 1.056, not significant).

And their results:

The majority of samples generated positive results in the Uhlmann assays. However, melting curve analyses revealed that only a proportion of these PCR products had the correct melting temperature (ie, identical to the positive MV control). Furthermore, only 2 products generated by the Uhlmann nucleoprotein assay, 3 by the Uhlmann fusion assay, and 4 by the Uhlmann hemagglutinin assay had the correct amplicon size when separated by gel electrophoresis. We successfully obtained sequences for 7 of these 9 products, but none were MV RNA. No specimen was positive using the Kawashima primer-based assays or our own probe-based fusion gene assay.

This is in line with Afzal et al.'s results and again, since almost 80% of the children had gastrointestinal complaints, a minimum of 6 positives should have been detectable if present.

The third paper in the first part of this series is by Thompson et al. (1995) also includes Wakefield as a co-author and this is Dr. Sears' summary:

This group of doctors studied thirty-five hundred adult patients with inflammatory bowel disease and other chronic intestinal conditions who had received the single measles vaccine in 1964 as part of the vaccine safety trials. When compared to a control group of eleven thousand adults who didn't get the measles vaccine, the vaccinated people were three times more likely to develop inflammatory bowel disease. This study predates Dr. Wakefield's work and is consistent with his concern about a link between the measles vaccine and inflammatory bowel disease.

No, it isn't Dr. Sears and this is why. This paper is merely an attempt to validate results from a previous paper by Lewin et al. (1995) of which (surprise surprise) Wakefield is a co-author of.

Groups were recruited from different study populations with no attempt to match or control for differences that would also contribute to or explain prevalence of intestinal bowel disorders (IBD). Information obtained, such as measles vaccine history and IBD history from the control group (not supposedly vaccinated for measles) was sloppy and relied heavily upon self-reporting. Essentially comparing apples to oranges. This introduces considerable bias potential because there is no standardisation of data collection regarding exposures and disease outcomes so the investigators didn't have all of the information that could be potential confounders for the measles jab. And this is the case as you will see.

Their methods for measles detection (as explained in the Lewin et al. paper) were prone to cross-reactivity with other antigens so they couldn't actually claim that they found measles without other molecular detection methods for viral presence. Other investigators had failed to find evidence of measles virus in gut biopsy samples of IBD patients using much better detection methods, if one could be bothered to read the comments published in the Lancet about this paper. They are really quite helpful, particularly if one is unsure of how to read and understand the methods of a controversial study. We highly recommend doing this.

Another problem with this study is that the authors postulated that measles infection, therefore attenuated, live, measles vaccine is a risk factor for IBD (which includes Crohn's disease and ulcerative colitis). Thompson et al. found a significantly lower risk of IBD in the unvaccinated control group as compared to the vaccinated group. This is incongruous with their hypothesis because 85% of the unvaccinated control group reportedly had natural measles infections versus no natural measles infections in the vaccinated group. If measles infection is associated with a higher risk of IBD you would expect to see higher numbers in the control group. This just further demonstrates the selection bias and slipshod methodology used to ascertain this association. A criticism of this study (published in the comments) deigned to point out that:

The relative risk of measles vaccinees developing inflammatory bowel disease compared with unvaccinated individuals (relative risk 2.5-3.0) should be in perspective: Crohn’s disease is more common in subjects whose first house had a hot water tap (odds ratio 5.0).

In other words, there is a stronger association between those that had hot water in their house as children and IBD, than those that received the single measles jab. It is interesting to note that even Wakefield himself has backed away from this work (Chadwick et al. 1998).

These papers are not independent replications of Wakefield's 1998 Lancet paper other than all being slapdash works that range from appalling incompetence to seemingly outright fraud. They and their supporting works all have Wakefield as a co-author on them. We will review the next 4 studies Dr. Sears claims 'support' Wakefield's work listed in The Vaccine Book in the next coming weeks.

ETA 07.19.09 for accuracy.


  1. Wow - great post, thanks! I'll be sure to link to it next time I write something about Wakefield.

  2. Thanks skepticat; it seems we have a few things in common, one of them being an esteemed opinion of 'chiroquacktors' (love that term).


  3. I think this is a great post. Thanks for taking the time.

    Why use the Unigenetics if it were "defunct"? (Hornig et. al, 2008)

    As to Thomspon et. al. - I think you've unknowingly explained why there might be a difference between natural vs. attentuated infection. Perhaps the host is incapable of sufficiently clearing the pathogen...

  4. Anon,
    Why use the Unigenetics if it were "defunct"? (Hornig et. al, 2008)

    It wasn't but Dr. O'Leary (the owner of the former Unigenetics lab) did participate in this study via his academic/clinical lab in Coombs Women's Hospital. It is important to note that he had to adhere to a strict set of protocols and these results were confirmed by 2 other labs that participated in this study and were blinded to the samples and each other's results.

    As to Thomspon et. al. - I think you've unknowingly explained why there might be a difference between natural vs. attentuated infection. Perhaps the host is incapable of sufficiently clearing the pathogen...

    Perhaps I am misreading your comment but his hypothesis was based upon the observation by a Swedish study that postulated that wild-type measles infection was a risk factor for Crohn's disease. There is nothing to indicate that the attenuation of the measles virus affects viral clearance by the host, on the contrary, the virus is more efficiently cleared by the host without measles pathogenesis. And, since they used poor detection methods and appeared to be trying to acquire their results to fit their hypothesis, rather than the more rigourous method of disproving the null, the Thompson et al. paper does nothing to validate the MMR-autism claim.

  5. This is such a useful post as I find many vaccine critics are still claiming that Wakefield's original findings demonstrate a compelling correlation whether or not causation can be concluded. They haven't kept up with the latest research or OAP bombshells and are unaware that the findings were fraudulent to begin with. Chalk up another embarrassment for Dr. Bob and thanks for another detailed and patient post that should be enlightening to anyone with an interest in the issue.