Video Abstract
Increasing vaccination of pregnant women makes it important to assess safety events potentially linked to prenatal vaccination. This study investigates the association between prenatal tetanus, diphtheria, acellular pertussis (Tdap) vaccination and autism spectrum disorder (ASD) risk in offspring.
This is a retrospective cohort study of mother-child pairs with deliveries January 1, 2011 to December 31, 2014 at Kaiser Permanente Southern California hospitals. Maternal Tdap vaccination from pregnancy start to delivery date was obtained from electronic medical records. A diagnosis of ASD was obtained by using International Classification of Diseases, Ninth and Tenth Revision codes. Children were managed from birth to first ASD diagnosis, end of membership, or end of follow-up (June 30, 2017). Cox proportional hazards models estimated the unadjusted and adjusted hazard ratios (HRs) for the association between maternal Tdap vaccination and ASD, with inverse probability of treatment weighting to adjust for confounding.
Women vaccinated were more likely to be Asian American or Pacific Islander, be nulliparous, have a higher education, receive influenza vaccination prenatally, and give birth at term. ASD was diagnosed in 1341 (1.6%) children, and the incidence rate was 3.78 per 1000 person years in the Tdap exposed and 4.05 per 1000 person years in the unexposed group (HR: 0.98, 95% confidence interval: 0.88–1.09). The inverse probability of treatment weighting–adjusted analyses revealed that prenatal Tdap vaccination was not associated with an increased ASD risk (HR: 0.85, 95% confidence interval: 0.77–0.95).
Prenatal Tdap vaccination was not associated with an increased ASD risk. We support recommendations to vaccinate pregnant women to protect infants, who are at highest risk of death after pertussis infection.
Comments
Prenatal Tdap and ASD study is flawed and must be retracted
Becerra et al.(1) recently published a study on prenatal Tetanus, Diphtheria, Acellular Pertussis (Tdap) Vaccination and Autism Spectrum Disorder (ASD).
As described previously(2), ignoring causal mechanisms result in flawed studies.
This Tdap ASD study(1) was based on Sanofi Pasteur's Adacel Tdap vaccine alone and on data collected at a few Kaiser facilities in Southern California alone.
The type and quantity of target proteins in Adacel are “Each 0.5 mL dose contains 5 Lf tetanus toxoid (T), 2 Lf diphtheria toxoid (d), and acellular pertussis antigens [2.5 mcg detoxified pertussis toxin (PT), 5 mcg filamentous hemagglutinin (FHA), 3 mcg pertactin (PRN), 5 mcg fimbriae types 2 and 3 (FIM)]” (3).
GSK’s Boostrix Tdap vaccine package insert says (4)“Each 0.5-mL dose is formulated to contain 5 Lf of tetanus toxoid, 2.5 Lf of diphtheria toxoid, 8 mcg of inactivated PT, 8 mcg of FHA, and 2.5 mcg of pertactin (69 kiloDalton outer membrane protein).”
Mostly the same proteins but somewhat different quantities in each vaccine.
The authors' ASSUMPTION is that the target proteins in Adacel are the potential cause of autism and therefore the subject of their study. If their assumption is true, target proteins being standardized, Adacel would be somewhat representative of all Tdap vaccines (although the protein quantities differ) and thus would be a candidate to perform the study as they did.
However, a major, widely prevalent ASD biomarker (70-75% of patients test positive) is the folate receptor alpha antibody (FRAA)(5). These FRAA bind with higher affinity to bovine FRA than human FRA(6). These FRAA are induced directed against bovine FRA but cross-react and bind to human FRA in the choroid plexus, block folate uptake and cause ASD(5). The majority of FRAA are of the IgG4 subclass(6). IgG4 antibodies are known to cross the placenta. Maternal FRAA have thus been shown to cross the placenta, block folate uptake in the fetus and cause ASD.(7)
One or more of bacteria used to prepare the Tdap vaccine are grown in casein or casamino acids containing growth media. The Tdap vaccine therefore contains residual quantities of non-target proteins, bovine milk proteins in this case, used to derive the casein or casamino acids. The residual non-target bovine FRA protein, a bovine milk protein in the vaccines are therefore an obvious source for the induction of FRAA.
Any Tdap vaccine administration, including prenatal Tdap vaccine administration could therefore cause the induction of maternal FRAA. These residual proteins however, are contaminants that are by definition, NOT standardized, NOT regulated and NOT characterized/reported by the vaccine maker.
Consider the H1N1 vaccine in Europe. The vaccines contained standardized quantities of the target hemaggluttinin (HA) proteins. The vaccines also contained residual quantities of unregulated contaminant non-target proteins - H1N1 nucleoproteins. The contaminant quantity varied from vendor to vendor due to their process. The result was disastrous. The Pandemrix vaccine induced numerous cases of narcolepsy. The Arepanrix and Focetria vaccines performed better. So even though these vaccines were considered equivalent, the outcomes were devastatingly different. The contaminant H1N1 nucleoproteins induced cross-reacting antibodies directed against human hypocretin receptors due to molecular mimicry, thus inducing narcolepsy.(8)
Residual bovine casein, a bovine milk protein, measured in just five samples of Adacel showed a twofold difference in quantity (8-17ng/ml) (9). Similarly, there is a huge variation in residual ovalbumin in chick egg derived influenza vaccines 0.3-38.3ug/ml(10). This is to be expected because the quantity of non-target proteins in vaccines are unregulated.
The residual quantity of bovine FRA in Adacel stock used at Kaiser facilities (sources of study data) are an unknown. Whether that residual quantity of bovine FRA is representative of the quantity in worldwide Tdap vaccines is an unknown. Therefore, the study results cannot be claimed to be representative and applicable to all Tdap vaccines worldwide. The authors' fundamental ASSUMPTION does not hold true. For this reason, their conclusion that prenatal Tdap vaccines in general are not associated with autism is not supported by the evidence.
Given the above, the authors should retract the article.
If the authors had researched the details I have provided, it would have been obvious to them that they need a completely different study type and study design to investigate this matter.
Even in this study, the authors describe that follow-up was on average 6 months less (inexplicably) in the vaccinated vs. unvaccinated group. Therefore a critical period where ASD diagnosis is most likely to occur, was unaccounted, introducing bias. This could have changed the outcome of the study and show that prenatal Adacel was indeed associated with autism.
Further, as noted in the beginning, 75% of ASD patients test positive for FRAA. Therefore, maternal FRAA induced ASD may only be 25% of the cases. Then there is also non-FRAA mediated, maternal autism related (MAR) autism autoantibody related mechanisms(11). So if accounted properly, even a small number of cases diagnosed during that 6 month period could have altered the outcome of the study.
References
1. Becerra-Culqui TA, Getahun D, Chiu V, Sy LS, Tseng HF. Prenatal Tetanus, Diphtheria, Acellular Pertussis Vaccination and Autism Spectrum Disorder. Pediatrics. United States; 2018 Aug 13;e20180120.
2. Arumugham V. Epidemiological studies that ignore mechanism of disease causation are flawed and mechanistic evidence demonstrates that vaccines cause autism [Internet]. 2017. Available from: https://doi.org/10.5281/zenodo.1041905
3. Pasteur S. Adacel Package Insert [Internet]. 2005. Available from: https://www.fda.gov/downloads/biologicsbloodvaccines/vaccines/approvedpr...
4. Glaxo Smith Kline. Boostrix Package Insert [Internet]. 2005. Available from: https://www.fda.gov/downloads/BiologicsBloodVaccines/UCM152842.pdf
5. Frye RE, Sequeira JM, Quadros E V, James SJ, Rossignol D a. Cerebral folate receptor autoantibodies in autism spectrum disorder. Mol Psychiatry. 2012;18(3):369–81.
6. Ramaekers VT, Sequeira JM, Blau N, Quadros E V. A milk-free diet downregulates folate receptor autoimmunity in cerebral folate deficiency syndrome. Dev Med Child Neurol. 2008;50(5):346–52.
7. Frye RE, Sequeira JM, Quadros E, Rossignol DA. Folate Receptor Alpha Autoantibodies Modulate Thyroid Function in Autism Spectrum Disorer. North Am J Med Sci. 2014;7(1):1–7.
8. Ahmed SS, Volkmuth W, Duca J, Corti L, Pallaoro M, Pezzicoli A, et al. Antibodies to influenza nucleoprotein cross-react with human hypocretin receptor 2 (ABSTRACT ONLY). Sci Transl Med. 2015;7(294):294ra105–294ra105.
9. Kattan JD, Cox AL, Nowak-Wegrzyn A, Gimenez G, Bardina L, Sampson HA, et al. Allergic reactions to diphtheria, tetanus, and acellular pertussis vaccines among children with milk allergy. J Allergy Clin Immunol. 2011;Conference(var.pagings):AB238.
10. Goldis M, Bardina L, Lin J, Sampson HA. Evaluation of Egg Protein Contamination in Influenza Vaccines. J Allergy Clin Immunol. Elsevier; 2016 Jan 9;125(2):AB129.
11. Arumugham V. Strong protein sequence alignment between autoantigens involved in maternal autoantibody related autism and vaccine antigens [Internet]. 2017. Available from: https://www.zenodo.org/record/1034571
RE: Age of ASD Diagnosis
The latest CDC ADDM Network information shows that the median age for a diagnosis of an autism spectrum disorder (ASD) is 52 months (4.33 years). Thus, the analysis by Becerra-Culqui is inadequate to pick up nearly 50% of the ASD cases in the sampling presented. The average age for follow-up in the vaccinated group was 3.85 years and the unvaccinated group was 4.44 years. Not only does this bias the analysis to find more diagnoses in the unvaccinated group, but it invalidates the analysis overall as many cases in both groups would simply not yet be diagnosed within the follow-up duration employed. A previous study (Geier et al. 2013 Transl Neurodegeneration 2:25) yielded calculated values 4.2 years and 1.54 years for the mean and standard deviation of age of ASD diagnosis (based on the CDC's Vaccine Safety Datalink). Applying these values, the Becerra-Culqui analysis would miss an estimated 59% of the ASD cases in the vaccinated group and 44% cases in the unvaccinated group, assuming a normal distribution of age of diagnosis.
RE: Analysis flaw: Prenatal Tetanus, Diphtheria, Acellular Pertussis Vaccination and Autism Spectrum Disorder
I was very curious to see the results of this analysis as there seems to be a suggestion that in utero exposure to inflammation may influence many health outcomes of children including autism. The issue I see with this analysis is that the follow-up for the unvaccinated group was significantly longer than the vaccinated group. Given that a significant portion of the autism diagnoses for the former occurred after 3 years of age, I don't see how anything can be concluded from the analysis done given this crucial difference between groups would influence how many diagnoses are made (the outcome of the analysis). The authors should consider adjusting for this important variable or using case-matched analysis so that both groups have the same length of follow-up.
RE: Average age of ASD diagnosis
Thank you for your comment. You are absolutely correct that the age of diagnosis is different between the two cohorts. The children in the vaccinated group seem to be diagnosed younger simply because of differential follow-up time between the vaccinated and the unvaccinated groups. The reason for this is that vaccination rates were lower in earlier years, thus the unvaccinated group are represented more in earlier birth cohorts and had the potential to reach an older age during follow-up. This is discussed in the fourth paragraph in the discussion section. For this reason we stratified the analyses by birth year to control for this issue, and the stratified results were consistent with overall results.
Average age of ASD diagnosis
I read your paper with interest. One comment that comes to mind is that it seems that Table 2 suggests that the average age of diagnosis may be different among the two cohorts, in particular the children of mothers who have been vaccinated seem to be diagnosed earlier than those of mothers who have not been vaccinated. Have you computed the average age of an ASD diagnosis for both cohorts? If so, could you please share it as I did not find this information in the paper?