This review primarily stemmed from my previous blog post in which I pointed crucial experimental flaws in the most recent study published by Lyons-Weiler and colleagues . The reply was swift and expected, a diatribe written on his own page that was vociferous, slanderous and completely inappropriate for a public statement.
I was surprised that a man claiming to be from science, and that should be “seasoned” by now about handling on critiques from reviewers would have shown an unmeasured tone in such official communication.
Surprised? I was not, and confirmed the reason why I use a pen name. Before submitting it as a post, I considered directly sending my letter to the editor of the journal, only to recuse from it by fear of retaliation and harassment from Lyons-Weiler or by some random anti-vaccine person to my institution.
In this delicate period of my academic career, such a decision can have devastating consequences on unfolding events related to my career. Shall I be in a better position, I would not have hesitated to send the letter to the editor.
Interestingly, such a slanderous attack of the messenger seems a staple amongst anti-vaccine crowds, but not expected from scientists. Yet, such slanderous behavior is not exclusive to the author. Recently, a letter to editor written by peers of the authors (in the name of Christopher Exley, Christopher Shaw, and Romain Gherardi as they co-authored the retracted counter-letter) was judged so slanderous against a scientist that raised important and valid concerns on a study published by Crepeaux and colleagues in “Toxicology” journal .
It is important to note to the reader not familiar with academic publishing that the retraction of a letter to the editor is so rare, that it suggests the extreme gravity of the slanderous claims made.
More recently, a similar behavior (albeit less slanderous) came from the recent retraction of the recent study by DeLong and colleagues , following critiques and concerns on PubPeer. Instead of addressing directly the stipulations raised by readers, the author underwent a series of diatribes while failing to address the major flaws of the paper.
During my professional development, learning to accept rejection and criticism from peer-reviews (be it a manuscript submitted to a journal or a grant proposal) is part of being a scientist. We do not mince words, and we will say bluntly (but with academic finesse) when a study or grant is not considered “competitive” and “compelling”.
We learn to undergo the stage of grievance and learn to deal with it several times a year. Yet, we also learn to never ever write a reply in the spur of the moment, in the full rage and anger. One day, my department chair told us (in one of these lab meetings) to never ever send an angry letter because you may regret (and will likely regret) it later.
Sending an angry letter will not change the editor or program officer to accept your paper or your grant proposal. Worse, it will backfire on you and reduce your chance of having your manuscript or grant proposal accepted in the future, as you will be labeled “that annoying scientist”.
Sure, write that angry letter. But don’t send it. Leave it to sediment overnight on your desk and come back to it the day after and re-read it. You will be surprised once the haze of anger is gone how callous and slanderous the letter was and failed to make a constructive and detailed counter-response to the reviewers.
This post is not intended to be a counter-response to Lyons-Weiler, that would be futile and sterile. However, I want to use this as a learning experience by using a study cited by the author  that Lyons-Weiler used against me as it claims it vindicates him (*spoilers* it does not, quite the opposite indeed. Read the rest of the post to have a detailed explanation) and discuss on why this study is indeed adding an additional stone in making the point of the overall safety of aluminum adjuvants in vaccines.
For those coming from his page, I want you to come with a neutral and objective mind and read carefully the following sections and follow me through my review. To the authors of this study, I would like to address you my fullest gratitude for your recent work published by your research group in helping advance the (very) complicated and punctuated literature on the pharmacokinetics of aluminum adjuvants over the last 40 years.
My apologies if I use your figures in this post, but I consider it essential to provide these to the readers that may not have access to your publication.
Before I go into a lengthy discussion of the study, I wanted to bring a bit of the background context about aluminum toxicity and where it originates. As a support, I will use the most recent technical report published by the American Academy of Pediatrics for the discussion of guidelines and major ideas  and remains discrete on detailed literature (which can be found on Pubmed for those interested).
The toxicity of aluminum became a concern about 50 years ago when patients suffering from severe kidney diseases (put on hemodialysis) were given aluminum-based phosphate binders to limit the absorption of phosphorus at the GI tract (as a measure to control the hyperphosphatemia associated with patients on dialysis).
Patients were given aluminum-based phosphate binders at a very high dose (1800-3600 mg daily, with 1 mg=1000 µg) for a very long period. Although the short-term effect of such treatment did not show any serious adverse reaction, the long-term effect of such treatment rapidly materialized by the onset of osteomalacia (bone weakening) occurring in patients, especially in the pediatric population.
This also was paralleled by the development in middle-to-late age patients a form of dementia called “dialysis dementia” or “dialysis encephalopathy”, a condition that was particularly documented from the mid-70s up to the late-90s. These conditions were enough to incentivize research on assessing the toxicity of aluminum in medical products.
This led to the assessment of neurotoxicity of aluminum in premature infants through the work of Bishop and colleagues , by assessing the toxicity of aluminum in medical products aimed for total parenteral nutrition (IV bags, back then aluminum was a common contaminant). The work of Bishop was so important that it was certainly used by the Food and Drug Administration (FDA) to set the limit of aluminum in TPN products (IV bags used for intravenous nutrition) to no more to 4-5 µg/kg/day.
The report is also clear about the safety of aluminum in vaccines, citing the literature that the vast majority of the literature failed to demonstrate possible causality between aluminum adjuvants and chronic neurological diseases, autism, primary ovary insufficiency and autoimmune diseases (the report further points to concerns of ASIA by the lack of proper criteria for diagnosis, as these criteria are “extremely vague and general”).
As a measure of precaution, patients with chronic kidney disease (CKD) are routinely monitored for aluminum levels in plasma (the acellular fraction of blood commonly used in pharmacokinetics and therapeutic monitoring) . In patients with CKD, a plasma concentration below 20 µg/L is considered as within the normal value (Note: this value is obtained from direct measurement in a biological sample, not relying on Clark’s formula).
Taken this into consideration, my statements follow the consensus and follow the recommendation of the highest authority. I personally consider that anyone questioning the consensus has an obligation to address their concerns by providing scientific evidence that surpasses the existing literature (in which the consensus is built on) with studies superseding it by their quality and quantity.
Pharmacokinetics of aluminum
In my previous post, I have mentioned the importance of assessing the level of aluminum in plasma as an indication of relative exposure. Despite the claim from the author that Clark’s formula prevails, I will repeat it again – outside of the determination of therapeutic dosing for a pediatric population, this formula has no relevance and merit when we discuss the pharmacokinetics of aluminum.
From the initial study by Priest (looking at the pharmacokinetics (PK) profile of aluminum following a single IV bolus of 26Al in human healthy volunteers)  to the most recent iteration by Weisser , one thing comes in mind about the PK of aluminum – it is complicated, very complicated.
Too complicated to be simplified down to a simple equation like Clark’s formula.
This is about how the PK of aluminum looks like:
For those that are not familiar, you can see that the plasma (blood) compartment is at the center of the diagram. This is where all the absorbed aluminum ends up, this is where the distribution in the peripheral tissues occur (bone, muscles, and liver), and this is also the compartment by which aluminum will be ultimately be eliminated by urinary excretion.
You will also note that only two species of aluminum are discussed here: the transferrin-bound and citrate-bound. Transferrin is a protein involved in the transport of iron (Fe) inside our body, such iron is commonly found as Fe2+ found (and in a minor species as its ferric form Fe3+).
Citrate is the salt form of citric acid. It is a polyacid formed by three carboxylate (COOH) groups that will become negatively charged (COO-) if the pH of dissolution is higher than the pKa of the group. The pKa of citrate are 2.92, 4.28 and 5.21 at 25º Celsius.
At physiological pH (7.4, which is the pH found in blood and tissues), citrate is found in the vast majority (>99%) into its complete ionized form (3 COO-). This allows free aluminum (Al3+) and free aluminum only (not the salt microparticles) to bind to these two molecules (transferrin/citrate) via electrostatic (ionic) interactions.
Note that Lyons-Weiler does not mention the presence of aluminum in particulate forms or entrapped into immune cells (e.g. macrophages). This diagram contradicts some claims made by some random person on the internet who claims that aluminum circulates inside the body as nanoparticles or macrophage bound.
If the Clark equation was solely enough to explain the plasma concentration of aluminum in the body, I would have expected to see it mentioned in this article originated by the same author that Lyons-Weiler claimed he vindicates his work. I would naively ask him “where, in this diagram, does the Clark formula fits?”
The 2019 study
I will try to make the most concise and succinct as possible, as it a dense but nice piece of work done by the Weisser group. I personally see it as a continuation and complementation of the initial work done by Flarend and colleagues , trying to reach the gaps and address the limitations of that study.
I have to say, the experimental design is exemplary as seen by Table 1:
We have six groups used in this study: two experimental (Alhydrogel® and AdjuPhos®, two aluminum adjuvants commonly used in preclinical studies, and referred to as pAH and pAP. I will not discuss these groups as I consider not relevant for the translational purpose of the study) groups, three “clinical” (aluminum hydroxide (V1), aluminum phosphate (V3) or combined (V2) and a saline “placebo” group (Note: Avers love to claim that safety studies with saline placebo controls do not exist. Well, here is one).
The number of animals is, in my opinion, a bit low for the power of analysis (N=8-10/group being more common, especially if you have high interindividual variability, as seen with the standard deviation in the following figures). The administration route is adequate, the multiple injection sites are justified by the use of a volume (0.5mL) clinically adequate but can be challenging for the animal welfare (this can be an important volume for a 250g animal that can interfere with the animal well-being).
The amount of aluminum is clearly defined (0.6 mg of aluminum is 0.6 mg of aluminum injected, not 0.6 mg of aluminum salt). The minor concern is the relative dose of aluminum injected (1.4-2.3 mg/kg) which seems higher than I would expect when compared to a 2-months immunization schedule (~1mg for a baby that weighs 5 kg, this would account about 0.2 mg/kg).
We can argue about the allometric scale (and I will let experts in that commenting) but we can say the author is playing pretty safe margins here (about 10x a dose higher than what baby would get). The time of observation is also pretty impressive (80 days maintained in cages).
Considering the average lifespan up to 28 months , this is a pretty long-term study of aluminum, considering the significant costs of animals housing per diem in any AAALAC- or FELASA-certified laboratory.
The first relevant data coming in is…..of course plasma concentrations (again, no estimation of it by Clark’s formula), we have raw experimental data provided here:
The first thing we can see is the extreme fluctuation of aluminum plasma levels over time and between the different groups. We can see the important variability between animals (within the same group), as seen by the standard deviation (SD) error bars (“whiskers”). A quick look seems to suggest that rats showed an average plasma concentration fluctuating between 10-30 µg/L (saline) which are concentration not unheard of in humans (<10-20 µg/L is considered ”normal” in humans).
Because the analysis of such data is complicated, it requires the use of mathematical tools that are far more complicated than the use of the Clark formula.
One mathematical tool used in PK (and that I teach my students) is the use of the area under the curve (AUC) using the trapezoidal method. This method calculates the surface area under the curve and allows us to objectively quantify and compare groups.
It is the method of choice for comparing administration routes and to determine the bioavailability of a drug between an extravascular versus IV route (absolute) or between extravascular routes (relative), including different formulations for the same route or generic versus branded drugs (bioequivalence studies).
We can see that the AUC values obtained for both V1 and V3 were almost identical to the saline group, as the author reported no statistical differences (as marked by the * symbol). The only concern was for V2 (the combo group) which showed some elevation of the AUC compared to saline.
This data is additional evidence that aluminum adjuvants injected IM is not increasing aluminum plasma level compared to the saline group over an 80 days period. This data clearly contradicts the claim made by Lyons-Weiler and has a higher degree of evidence compared to his estimated model, and suggests flaws in his mathematical model.
Since Lyons-Weiler claims his model is valid, it was on him the burden of proof to validate his model with experimental data (which could have been easily performed, considering the presence of a physician in his study, the drawing of blood from patients on different schedule and the relatively routine analysis of aluminum in biological fluids by reputable biomedical labs).
The second table provides an extensive measurement of aluminum in animals at endpoint (80 days) in different tissues including brain tissues. The outcome is pretty interesting, as it is not exactly matching the previous data.
We did not observe differences in plasma levels between V1 and V3 compared to saline (based on the AUC values). We would assume that no differences in the brain tissue content would occur. Yet, we observed an amount twice higher than saline control. Dry weight, wet weight, geometric means….it is unequivocal. The standard deviations are also tight. We cannot argue with the numbers, these are factual. But it is important to put them into context and have logical and critical thinking.
Here is I think we reach the limitations and some contradictory results from this study, especially when we compare to the study coming from the same lab that performed, using aluminum citrate injected in rats IM at a dose of 0.36 mg/kg (which matches at 1.8 mg Al for a 5 kg baby) .
In that study, the authors found no differences in Al content between citrate and control groups, both at a wet weight and dry weight (with 95% confidence intervals pretty large). I think having plasma sampling timepoints more frequent within the first 24 hours would have allowed them to capture the peak concentration (Cmax) and maybe allow a certain correlation between differences in AUC during the first 24 hours. What does the author say?
I think it is pretty clear here, this increase is unlikely to be significant and may be inherent from the limitations from the experimental design (in particular with small sample sizes). That’s likely a vindication of the author on my side, and also a refutation for those promoting the idea that aluminum in the brain by macrophages. If only Lyons-Weiler read the whole study instead of cherry-picking the sentences targeted at me.
Another issue that I mentioned about Lyons-Weiler studies (and Exley, as seen in his recent commentary published in J Trace Elements ) is the biggest mistake someone can ever make about pharmacokinetics – making the claim that extravascular injections are strictly identical to IV injection, with the premise that 100% of a dose is circulating in the body within minutes of injection.
That would require an absorption process reaching 100% within minutes. Again, Weisser’s experimental design was adequate and well-thought as she measured the amount of Al remaining at the injection site by 80 days.
As we can see, V1 (hydroxide) still holds 65% of the amount injected after 80 days, whereas V3 (AP) has less than 20% remaining at the injection site. These data nicely support the previous work of Flarend, as he reported values falling roughly in the same ballpark if extrapolated to 80 days (AP was absorbed in a faster extent than AH in its 28 days timepoint).
This again supports my claim that aluminum has such a slow release to be considered harmful. It also suggests that the slow release from the injection site is unlikely to create a spike over the basal aluminum levels coming from food/drink intake and from inhalation.
I will not finish the rest of the study, as the remaining figures (Figure 3 and Figure 4) are either addressing data already presented (Figure 3 is another representation of the data shown in Table 2; whereas Figure 4 discuss about the relationship between the amount released at injection site and AUCplasma and bone concentration).
The only point of agreement between Lyons-Weiler and me will be that this is a very good study presented by Weisser. She has been doing a remarkable effort to bring momentum into the field that has been on-off (likely due to low priority in funding such programs, Weisser is a recipient of the German Ministry of Health).
Initially presenting as the “silver bullet” that killed my rebuttal, I have to admit that after reviewing it (and hopefully readers will agree with me), this study brought by Lyons-Weiler ricocheted and lead to a serious self-inflicted wound.
I just wish instead of writing his diatribe following my rebuttal, he took the time to read this study carefully in an unbiased manner. Unfortunately, based on his quote, I am afraid he was more interested to cherry-pick a sentence to point it as a straw in my eye, but could not see the beam in his.
With all honesty, I would say this study sides with me and even vindicate my rebuttal for several reasons:
- Nowhere in her recent study, Weisser used Clark’s equation touted by him in his study. The simple reason is that this equation has no place in PK studies, especially when dealing with such a complicated compound.
- Unlike what Lyons-Weiler wants us to believe, you cannot count aluminum as a whole, and you have to separate the dissolved fraction (Al3+, that is bound to transferrin and citrate) from the solid (salt) fraction. Weisser made it clear in her 2017 study that the only forms that matter when it comes to aluminum PK is the dissolved fraction of it. This is some basic concept we learn in HS chemistry about the reactivity of chemicals between their different states.
- Lyons-Weiler wanted us to believe that there is no need for blood/plasma samples, as only a theoretical model would suffice. Weisser has again supported me by putting the plasma compartment as its central place in the pharmacokinetics of aluminum. This study also reinforces that experimental data primes over a theoretical model, that any theoretical model has at some point to be validated by experimental data. This is crucial information that Lyons-Weiler should present to validate his model if he wants to be taken seriously by the scientific community.
- The study shows no increase in aluminum plasma levels (using a dose 10x the clinical one), contradicting his model and his alarmists’ claim and eventually the whole need of a “Vaccine-Friendly Plan”. What is the benefit of a delayed schedule as the issue of aluminum is non-existent? This study nicely complemented the work of Flarend, supports the initial conclusion by Mitkus by providing experimental evidence about his model.
- The PK of aluminum is indeed complicated, and maybe more than a 2-compartments and requires the need of additional experimental studies to help in the development of a mathematical model that best fits its fate inside the body.
- Brain accumulation is not a concern, and clearly suggest that alternative hypothesis touted by some (“macrophages” hypothesis promoted by Gherardi and colleagues) is at best a marginal route of entry.
- Aluminum hydroxide is marked by its very slow release (and absorption) from the injection site, contradicting the claim made by Lyons-Weiler in 2018 that suggested an immediate availability of aluminum in its Al3+ form :
I think there is a valuable lesson to take from here. I could have angrily reacted to his tirade by writing an inflamed and malevolent reply in the midst of my gut reaction. That would have been sterile and betray my ethics as an academic scientist.
Instead, I took it as an opportunity to set a case based on facts and experimental evidence. He used this study as a prop to claim vindication to my previous post.
I think we all agree that I did an honest and detailed review of the paper, and the data not only did not rebut my claims but also certainly turned against his own study. This will be my last interaction on this paper and replies to Lyons-Weiler.
Unless you have bad faith, you have to agree that such a study initially touted by Lyons-Weiler as vindicating him and destroying my previous post is short of doing any of these. Indeed, this study aligns perfectly with the scientific consensus (and therefore validates my claims since I have been following the consensus) and put Lyons-Weiler in a very delicate position as it refutes basically his two studies [1, 12].
At this point, having Lyons-Weiler flip this paper from being lauded to being ostracized would be a very uncomfortable move for him and would further ruin his reputation.
Unlike what Lyons-Weiler wants us to believe, there are no merits of Clark’s equation in PK/TK studies (which is requires much more sophisticated mathematical tools) and therefore completely undermines the need for a “vaccine friendly plan” or any alternative schedule delaying immunizations for safety purposes.
I am not asking for anything from him, I am not asking for apologies from him. I am just asking him and his peers to take this study as an example of how science must and should be done. I am asking their sponsors that if they think vaccines present safety issues to ask for proposals showing such a detailed and well-conceived experimental method, coupled with the utmost scientific rigor. Until then, “vaccine safety” studies as reported by anti-vaccines will be doomed to be mocked, flagged and eventually retracted for serious flaws by the scientific community.
Finally, if Lyons-Weiler thinks he is still right, I will ask him one favor – please ask Dr. Weisser to review this post and communicate the official correspondence between them that my review was completely wrong and that his study is sound and adequate, as Dr. Weisser has much better credentials and expertise than me to review his study.
Since Lyons-Weiler made the claim, the burden of proof shall hold upon him solely.
This article is by VaultDwellerSYR, a pseudonym used by a faculty member of a School of Pharmacy within a large medical school. They have significant research and publications in the effect of certain chemicals on the brain. Although we are opposed to all arguments from authority, the author has a substantial record of actual, published research in this field of brain cell biology and biochemistry.
The author has stated that he has no conflict of interest to disclose.
- McFarland G, La Joie E, Thomas P, Lyons-Weiler J. Acute exposure and chronic retention of aluminum in three vaccine schedules and effects of genetic and environmental variation. J Trace Elem Med Biol. 2019 Dec 5;58:126444. doi: 10.1016/j.jtemb.2019.126444. [Epub ahead of print] PubMed PMID: 31846784.
- Crépeaux G, Exley C, Shaw CA, Gherardi RK. RETRACTED – Letter to the editor. Toxicology. 2017 Sep 1;390:159. doi: 10.1016/j.tox.2017.09.010. Epub 2017 Sep 18. PubMed PMID: 28928034.
- RETRACTED ARTICLE: [A lowered probability of pregnancy in females in the USA aged 25-29 who received a human papillomavirus vaccine injection]. J Toxicol Environ Health A. 2019 Dec 10;:1. doi: 10.1080/15287394.2019.1669991. [Epub ahead of print] PubMed PMID: 31821111.
- Weisser K, Göen T, Oduro JD, Wangorsch G, Hanschmann KO, Keller-Stanislawski B. Aluminium in plasma and tissues after intramuscular injection of adjuvanted human vaccines in rats. Arch Toxicol. 2019 Oct;93(10):2787-2796. doi: 10.1007/s00204-019-02561-z. Epub 2019 Sep 14. PubMed PMID: 31522239.
- Corkins MR. Aluminum Effects in Infants and Children. Pediatrics. 2019 Dec;144(6). doi: 10.1542/peds.2019-3148. PubMed PMID: 31767714.
- Bishop NJ, Morley R, Day JP, Lucas A. Aluminum neurotoxicity in preterm infants receiving intravenous-feeding solutions. N Engl J Med. 1997 May 29;336(22):1557-61. doi: 10.1056/NEJM199705293362203. PubMed PMID: 9164811.
- Schifman RB, Luevano DR. Aluminum Toxicity: Evaluation of 16-Year Trend Among 14 919 Patients and 45 480 Results. Arch Pathol Lab Med. 2018 Jun;142(6):742-746. doi: 10.5858/arpa.2017-0049-OA. Epub 2018 Mar 6. PubMed PMID: 29509029.
- Priest ND, Newton D, Day JP, Talbot RJ, Warner AJ. Human metabolism of aluminium-26 and gallium-67 injected as citrates. Hum Exp Toxicol. 1995 Mar;14(3):287-93. doi: 10.1177/096032719501400309. PubMed PMID: 7779460.
- Weisser K, Stübler S, Matheis W, Huisinga W. Towards toxicokinetic modelling of aluminium exposure from adjuvants in medicinal products. Regul Toxicol Pharmacol. 2017 Aug;88:310-321. doi: 10.1016/j.yrtph.2017.02.018. Epub 2017 Feb 22. PubMed PMID: 28237896.
- Flarend RE, Hem SL, White JL, Elmore D, Suckow MA, Rudy AC, Dandashli EA. In vivo absorption of aluminium-containing vaccine adjuvants using 26Al. Vaccine. 1997 Aug-Sep;15(12-13):1314-8. doi: 10.1016/s0264-410x(97)00041-8. PubMed PMID: 9302736.
- Weisser K, Göen T, Oduro JD, Wangorsch G, Hanschmann KO, Keller-Stanislawski B. Aluminium toxicokinetics after intramuscular, subcutaneous, and intravenous injection of Al citrate solution in rats. Arch Toxicol. 2019 Jan;93(1):37-47. doi: 10.1007/s00204-018-2323-8. Epub 2018 Oct 9. PubMed PMID: 30302509.
- Lyons-Weiler J, Ricketson R. Reconsideration of the immunotherapeutic pediatric safe dose levels of aluminum. J Trace Elem Med Biol. 2018 Jul;48:67-73. doi: 10.1016/j.jtemb.2018.02.025. Epub 2018 Mar 8. PubMed PMID: 29773196.
- Alvarado JC, Fuentes-Santamaría V, Gabaldón-Ull MC, Blanco JL, Juiz JM. Wistar rats: a forgotten model of age-related hearing loss. Front Aging Neurosci. 2014;6:29. doi: 10.3389/fnagi.2014.00029. eCollection 2014. PubMed PMID: 24634657; PubMed Central PMCID: PMC3942650.
- Exley C. An aluminium adjuvant in a vaccine is an acute exposure to aluminium. J Trace Elem Med Biol. 2020 Jan;57:57-59. doi: 10.1016/j.jtemb.2019.09.010. Epub 2019 Sep 18. PubMed PMID: 31561170.