Talk:Paracetamol

Archives (Index)

1 2 3 4 5

This page is archived by ClueBot III.

Baccarelli 2025 review

@Asto77: In your edit with the edit summary "absolute madness to delete the 2025 review source and rely on 2018 sources", this review has been heavily criticized on the following grounds: The source in question (“Evaluation of the evidence on acetaminophen use and neurodevelopmental disorders using the Navigation Guide methodology”, PMID 40804730) was published in Environmental Health in August 2025 and uses a systematic review methodology to examine potential links between prenatal acetaminophen exposure and neurodevelopmental disorders (NDDs), including autism and ADHD, in children.

Conflicts of Interest

• Andrea Baccarelli (Harvard T.H. Chan School of Public Health) served as an expert witness for plaintiffs in lawsuits concerning acetaminophen and neurodevelopmental disorders before completing this review. He was compensated for this role, reportedly earning about $150,000 for more than 200 hours of work. While the paper does disclose Baccarelli's expert witness involvement, it frames the conflict as a potential perception issue rather than a substantive financial conflict. However a federal judge explicitly criticized his testimony as unreliable and biased, noting that Baccarelli selectively emphasized studies supporting his thesis and downplayed those that did not.
• Beate Ritz (UCLA) has a lengthy history as a paid expert witness for plaintiffs in environmental litigation, including glyphosate and paraquat lawsuits, often working for law firms that specialize in plaintiff-side environmental claims. Her prior litigation work is not fully disclosed in the conflict of interest statement for this paper.

Other Concerns

• The review presents itself as NIH-funded due to ongoing grants the authors have, but reports suggest those grants are unrelated to the topic of acetaminophen and neurodevelopment, and the NIH did not fund this specific review. The claim of NIH support is therefore potentially misleading.
• Commentators and the judiciary have criticized the scientific rigor and methodology of the review, including its selective citation practices and the use of the “Navigation Guide” approach, which is uncommon in pharmacoepidemiology and seen by critics as a poor fit for this evidence base.
• The work has been referenced in policy discussions, including at the White House, but experts and even some of the study's authors (notably Bauer) caution that the evidence is not definitive and that further research is necessary.

Reliability Assessment

• The review compiles a large body of evidence and rates the studies by risk of bias and methodological strength, but the principal conflicts of interest (notably Baccarelli’s paid expert testimony and Ritz’s extensive litigation support) weaken confidence in its objectivity.
• Judicial and critical scientific commentary indicates the possibility of confirmation bias and methodological cherry-picking. Funding claims are potentially misleading, and not all conflicts, especially those related to litigation, are disclosed transparently.

In summary: The review should not be considered a fully neutral or definitive authority due to significant, partially undisclosed conflicts of interest among the lead authors, judicial findings of bias in associated expert testimony, and credible criticism of methodological choices and funding transparency.

References

Boghog (talk) 11:51, 25 September 2025 (UTC)

I found this source cited elsewhere, and have removed it. It is not a systematic review, it does not follow the methods. It frankly makes up a new method that is purely original research, does not characterize it as such, and is therefore neither systematic or a review. It is truly the worst study I've ever come across, and I'm staggered about this whole ordeal. It doesn't apply GRADE despite saying it does, and it doesn't even apply the weaker Navigation Guide that it says it does. It literally says that it intentionally selectively weighs down articles in a completely random manner. The absence of dose-response gives one study a negative score!!! I don't even know what this is. WP:MEDRS & WP:MEDINDY totally invalidate it. CFCF (talk) 11:58, 25 September 2025 (UTC)

Really appreciate all the truly fascinating details here. Re: the research summary sentence in this article (& autism & causes of autism, would something like

“Large population-based studies indicate that prenatal paracetamol use is not clearly associated with autism, ADHD, or intellectual disability, with sibling-controlled analyses suggesting prior associations were likely due to confounding.”

be a cleaner summary?

CFCF TranquilityBanquet (talk) 09:57, 26 September 2025 (UTC)

Also wondering if the section should be moved to uses rather than being in adverse events

CFCF TranquilityBanquet (talk) 10:15, 26 September 2025 (UTC)

Thanks, @TranquilityBanquet. I think it makes sense in the adverse events section here, but it may be less glaring to put it in the other section. I agree about including the interpretation that it is likely due to confounding. I do however think it may make sense to discuss what the nature of the confounding is. Many readers will not be able to interpret that confounding means, especially if they are driven here due to the media hype. While it may be too much detail to go into exactly what the confounders are (and perhaps speculative), stating that common causes may drive both maternal paracetamol use and autism in the child would be useful (i.e. there is potentially not only confounding, but also mediation, but that is too complex to include). CFCF (talk) 10:22, 26 September 2025 (UTC)

Hmm… what about this?

”Large population-based studies indicate that prenatal paracetamol use is not clearly linked to autism, ADHD, or intellectual disability, and studies comparing siblings suggest that earlier reported links were likely due to other factors, not paracetemol itself.“

My sense is that “stating that common causes may drive both maternal paracetamol use and autism in the child would be useful” is more explaining the implications of the reviews rather than letting them speak for themselves so to speak.

Re: uses—-it seems like a more natural fit there to me, I’m just not sure how to place, especially since pain and fever are their own sections. Maybe a “Uses in pregnancy” subsection? Its location in adverse effects seems to imply that there are known adverse effects—which, while there is still certainly room for debate, is certainly not the consensus view at present.

Its location in the Uses section also would make more sense in the current framing—-i.e. “here’s how it is to be used.” We could start with a sentence about it being firstline for pain and fever during pregnancy and then get into technicalities TranquilityBanquet (talk) 10:53, 26 September 2025 (UTC)

I agree with you, the only thing would be to not only explain that the studies are on siblings, but that they are very large. The Swedish study is >2.4 million children, which in itself makes it notable enough that it can be cited despite being a primary source per MEDRS. The same is true for the Japanese study which is >217,000 births. Both use national data from registers that are basically only available in the Nordic countries/South Korea/Japan - making the quality exceptional.

The reason this matters is pretty evident in Baccarelli, where they (intentionally?) ignore basic statistics, ranking a study of 996 individuals as high-quality. This is a particularly egregious deviation from both GRADE and Navigation Guide (which is one part of why it's not even a systematic review - they are entirely unsystematic and misrepresent their methodology - and deviate completely from both the methods they say they use and from any reasonable scientific practice [as described in part also in that great STAT article that was linked above Trump’s Tylenol warning cited a Harvard dean’s research. But a judge called his shifting conclusions ‘unreliable’]). A basic understanding of statistics will tell you that at ~3% prevalence, the precision and power in that study (even if it does measure fetal chord blood - i.e. a highly precise exposure classification) is horrible. Using the systematic review methods that they purport to use, it is "Very low quality" + "Very low certainty" evidence, while the other studies are probably the highest possible quality and certainty that could reasonably be attained. This is a major problem in recent times, and was so during Covid, where articles say they are systematic review, but really aren't. This just happens to be an especially egregious case, which is why it so clearly fails WP:MEDRS/WP:MEDINDY/WP:COISOURCE. CFCF (talk) 11:17, 26 September 2025 (UTC)

Very helpful and appreciated comments. I’ll update the reviews sentence summary in the three articles (are there any others?) in accordance. TranquilityBanquet (talk) 18:37, 26 September 2025 (UTC)

Worth including the research on use in pregnancy and possible links to asthma/wheeze or weaker ties to atopy or male reproductive effects? And more clear statements in general about its safety during pregnancy (WHO, EMA, etc.) Currently seems too focused on autism.

Asthma seems the most likely from the available research.

https://academic.oup.com/ije/article/45/2/512/2572607

https://www.sciencedirect.com/science/article/abs/pii/S0301054617300058?utm

https://pubmed.ncbi.nlm.nih.gov/30260196/

https://www.frontiersin.org/journals/toxicology/articles/10.3389/ftox.2022.884704/full?utm TranquilityBanquet (talk) 19:17, 26 September 2025 (UTC)

I am hesitant, because while the Norwegian study is sizable, I find it isn't enough to merit inclusion as a primary study according to WP:MEDRS. As for the systematic review you link there, it is not very high quality and it quite pronouncedly describes the biases of the underlying studies as considerable. It does suggest there is no publication bias, but there are also few articles overall. It could merit a very brief mention and careful mention, but no more. CFCF (talk) 18:31, 28 September 2025 (UTC)

The comment raises important concerns, but it risks sliding into wholesale delegitimization, which does not align with how Wikipedia should weigh reliable sources.

On conflicts of interest:

– It is true that both Baccarelli and Ritz have acted as expert witnesses. However, Environmental Health published the disclosure, and the paper underwent peer review. In environmental epidemiology, it is not unusual for high-profile experts to be engaged in litigation. The mere fact of being compensated for expert testimony does not automatically invalidate subsequent peer-reviewed research.

– Judicial criticism of testimony in a legal case is not the same as scientific peer review. Importing courtroom assessments into an encyclopedic evaluation risks conflating distinct contexts (legal vs. scientific).

On methodology:

– The Navigation Guide is a systematic review framework specifically developed for environmental exposures. While it is less common in pharmacoepidemiology, “uncommon” is not synonymous with “invalid.”

– Other meta-analyses often cited (e.g., Masarwa 2018, Ji 2020) also have limitations and potential biases. No synthesis of this evidence base is without caveats. Using the Navigation Guide to rate study quality and evidence strength is a legitimate, though debatable, choice.

On funding statements:

– Referencing ongoing NIH grants may appear ambiguous, but it does not amount to misconduct. It is common practice to acknowledge institutional funding that indirectly supports researchers’ time and work.

On weight in the article:

– No one is suggesting the 2025 review is the “final word.” It is, however, a recent, peer-reviewed synthesis that has entered the policy and scientific discussion, including citations in public debates.

– The appropriate approach on Wikipedia is not to exclude it, but to present it with balance: noting its conclusions, while also documenting the criticisms, conflicts of interest, and contrasting evidence (such as the large 2024 Swedish cohort study and regulatory positions from EMA, MHRA, WHO, etc.). Fedifra92 (talk) 07:43, 1 October 2025 (UTC)

Repeating what CFCF wrote above the methods used by Baccarelli 2025 review appears to be seriously flawed: It is not a systematic review, it does not follow the methods. It frankly makes up a new method that is purely original research, does not characterize it as such, and is therefore neither systematic or a review. It is truly the worst study I've ever come across, and I'm staggered about this whole ordeal. It doesn't apply GRADE despite saying it does, and it doesn't even apply the weaker Navigation Guide that it says it does. It literally says that it intentionally selectively weighs down articles in a completely random manner. The absence of dose-response gives one study a negative score!!! I don't even know what this is. CFCF

Boghog (talk) 07:58, 1 October 2025 (UTC)

Summary of the methodological criticisms of the Baccarelli 2025 Review:

Issue	Baccarelli 2025 Review	Critiques	Sources Supporting Critiques
Exposure assessment	Mostly retrospective/self-report	High recall bias; imprecise
Confounder adjustment	Limited in most included studies	Fails to address familial and genetic confounding
Study heterogeneity	High; varying outcome/exposure measures	Low validity of pooled results
Analytical framework (Navigation Guide usage)	Adapted from environmental health	Not validated for medication/epidemiology evidence
Dose–response / Exposure quantification	Largely missing or poor	Weakens any causal claims
Conflicts of interest	Declared, but disputed by critics	Bias heightened by litigation and payment history

References

Boghog (talk) 08:52, 1 October 2025 (UTC)

Fever as a confounder

Maternal fever and infections are major confounding variables in paracetamol use during pregnancy and autism risk.

• Fever itself is linked to higher autism risk in offspring, independent of any medication use.
• Because paracetamol is commonly used to treat fever, observed associations between paracetamol and autism may partly reflect the effects of the fever or underlying illness rather than the drug itself.
• Epidemiologists describe this as *confounding by indication*—the condition for which the medication is taken can bias results unless carefully controlled.

Adjustment in recent studies

• Recent large, well-controlled studies (including sibling-controlled designs) that adjust for fever and related factors do not support a causal link between paracetamol use in pregnancy and autism.
• Regulatory and expert reviews (e.g., WHO, EMA, national drug safety agencies) stress that risk assessments must account for the underlying condition—especially fever—rather than focusing solely on paracetamol exposure.

Given this evidence, most current high-quality sources conclude that fever is a key confounder. Any discussion in the article should reflect that the apparent association between paracetamol use in pregnancy and autism risk is likely due, at least in part, to the underlying fever or infection rather than the medication itself. Boghog (talk) 16:56, 1 October 2025 (UTC)

References

Vandalism from XBaller89

Take note of the persistent vandalism in History section TranquilityBanquet (talk) 08:54, 26 September 2025 (UTC)

Panadol topic has to be in Wikipedia

like Nurefen. The informations about both in the google page are not accurate. I stick on with Wikipedia ~2026-13075-44 (talk) 06:35, 17 May 2026 (UTC)

Is there any information in regards to the conformation to the USP, BP and EP in this topic — Preceding unsigned comment added by ~2026-13075-44 (talk) 06:42, 17 May 2026 (UTC)

First tested on Omkar hazare

First paracetamol was tested on Karmala King Omkar hajare ~2026-38306-38 (talk) 14:45, 5 July 2026 (UTC)