Anthony Fauci's July 2026 Senate testimony proved the truth about masks, social distancing and vaccines, and his silence under the Fifth Amendment is itself the confession

FactCheck.org became the fourth fact-checker on this claim in four days, publishing at 21:53 UTC, and it obtained the one thing none of the previous three had: an answer from the office making the claim, given after the number had been taken apart. Asked to defend the 82 percent figure, Senator Ron Johnson's communications director Grace Carnathan did not defend it. She said the question is not what the actual percentage of miscarriage risk is, that we will probably never know because of what she described as grossly inadequate and negligent, if not corrupt, safety surveillance, and that the question should instead be why federal health officials withheld information on safety signals. Read alongside the rest of this file, that is the whole pattern in one quotation. The statistic was the reason the texts mattered; with the statistic gone, the texts still matter, and now nothing about them can be checked. FactCheck.org also fills in the exchange the senators excerpted. Their video set a single January 2021 text against public comments Fauci made between 3 February and 30 August 2021, and FactCheck.org reviewed those public statements and found they hewed fairly closely to what he had written privately. The day after the line about fever, Fauci wrote in the same thread that more than 10,000 pregnant women had been vaccinated following the emergency use authorisation and that no issues had arisen, adding that animal toxicity studies did not raise any red flags. Victoria Male, an associate professor of reproductive immunology at Imperial College London, told FactCheck.org that in January 2021 it was sensible to be cautious about a theoretical possibility, that subsequent evidence found no increased risk of pregnancy loss, and that current data continue to show no increased risk of miscarriage or other pregnancy complication associated with COVID-19 vaccination. Recorded as reporting, not adopted; this file accuses nobody of anything, takes no position on any medical question and offers no guidance. source →

Snopes rated the claim False on 14 August, making three fact-checkers on the same figure in four days, and the version it examined is not the version the other two examined. That is the thing to log. Lead Stories on 11 August and PolitiFact on 12 August both addressed an arithmetic claim: whether any study shows an 81 or 82 percent miscarriage rate. Snopes addressed a different sentence, that the released texts prove Fauci covered up data showing such a rate. Between Tuesday and Friday the claim stopped being about a number and became an accusation of concealment, and the posts Snopes quotes carry it to its conclusion, one reading 'They said the Covid vaccine had a 12.6% miscarriage rate for pregnant women but it was actually 82%' and calling for Fauci and his co-conspirators to be jailed. Recorded as what posts said, not adopted, and this file continues to accuse nobody of anything. The pattern is the point and it is the same one this file has tracked from the start. A misread statistic will be corrected and survive; converted into a coverup, it no longer depends on the statistic at all, because the evidence for concealment is the same text message either way and no arithmetic can touch it. Snopes reports the underlying figure comes from a misinterpretation of a 2021 study and that studies have consistently found miscarriage rates among vaccinated women in line with the general population. This file takes no position on any medical question and offers no guidance. source →

PolitiFact rated the 82 percent figure False on 12 August, and its check supplies two things this file did not have. The first is provenance at both ends. The claim is now being made directly by Senator Ron Johnson, in a video posted to Fox News Digital on 10 August, rather than by commentators repeating him; and his spokesperson Grace Carnathan told PolitiFact that he was quoting a 2022 open letter by Dr James Thorp, an obstetrician who PolitiFact says has argued that COVID-19 vaccines increase miscarriage risk. Asked for additional data, Carnathan pointed back to the Fauci text messages. The study figure is offered as what makes the texts significant, and the texts are offered as the support for the study figure, with nothing cited outside the pair. The second thing is the part that closes the argument, and it is not a rebuttal so much as a completion. Johnson's objection to the original denominator is correct as far as it goes: 700 of the 827 completed pregnancies involved women vaccinated when they were already in the third trimester, so 827 was not a useful denominator. The study's own authors said exactly that. In a September 2021 follow-up on the same registry they acknowledged the problem, surveyed 2,456 people vaccinated shortly before conception or before 20 weeks' gestation, adjusted for the timing of vaccination, and reported a cumulative miscarriage risk of between 12.8 and 14.1 percent. Johnson carried out the first half of a correction the researchers had already published in full, and stopped at the point where dividing by 127 returned a number. PolitiFact adds a March 2026 revisit of the same registry covering 12,097 women vaccinated before 20 weeks, which found a cumulative risk of 13.5 percent. It also gives the text exchange more fully: Vivek Murthy had asked about the timing of vaccination during pregnancy, and Fauci's first answer was that there was no theoretical reason to favour early vaccination over late, with the line about miscarriage coming after it. This file continues to take no position on any medical question and offers no guidance. source →

The release has now been used to revive a five-year-old arithmetic error, and it has been checked. On 10 August the political talk show host Katie Pavlich posted on X that the vaccine was effectively an abortion shot 'according to new documents just released by Senators Johnson and Paul', citing an 82 percent miscarriage rate said to be hidden in the raw data of the 2021 Shimabukuro study and an 81 percent rate said to come from Pfizer's post-authorisation reporting. Lead Stories rated both figures false on 11 August. On the first: Table 4 of the study, published in the New England Journal of Medicine in April 2021 from CDC surveillance data, records 104 spontaneous abortions before 20 weeks among 827 completed pregnancies, which is 12.6 percent. The 82 percent is produced by discarding the 700 participants vaccinated in the third trimester and dividing by the remaining 127. That is not a rate. The registry was tracking outcomes across everyone enrolled rather than asking whether each participant was at risk of miscarriage at the moment she was vaccinated, and removing a group after their outcomes are known changes the population being measured. Incomplete follow-up compounds it, because pregnancies that end early resolve quickly while healthy pregnancies from the same window had not yet reached term when the data was cut. On the second figure, Lead Stories found something simpler. Pfizer's report covered 270 pregnancies, of which 238 had no outcome recorded; of the remaining 32, 28 were reported losses. The 81 percent is said to come from 28 divided by 32, except that 28 divided by 32 is 87.5 percent, so the number does not match its own stated calculation, quite apart from the fact that excluding 238 unresolved pregnancies cannot produce a rate. Lead Stories notes it first addressed this claim in July 2021, and points to a February 2023 paper pooling 21 studies across roughly 150,000 pregnant women. This file takes no position on any medical question and offers no guidance; what it records is that the documents released last week are being used to relaunch a miscalculation that predates them by five years, and that the second of the two figures does not even survive its own arithmetic. source →

The silence has ended and the same operation is being run on the words. On 10 August Senators Ron Johnson and Rand Paul published the first batch of text messages from Fauci's former government phone, which the Department of Health and Human Services produced to the subcommittee on 5 August. The senators say the device holds more than 34,000 messages and 522 voicemails, that only three contacts are named, and that it is too early to say whether anything has been deleted; those are their characterisations, not established findings. Eleven texts were published. The one now circulating is a single clause. In an exchange dated 25 January 2021, Vivek Murthy asked whether there was any data or theoretical reason to prefer vaccinating early rather than late in pregnancy. Fauci answered that there was none, and then added that since many people have significant cytokine storm and fever after the second dose, this theoretically could be associated with miscarriage in the first trimester; Rochelle Walensky replied that it was a good point. The clause travelling on social media is that second sentence with the first one removed. The next day, in the same release, Fauci wrote that any vaccine with limited pregnancy data should be carefully considered and that one must weigh the potential risks against the benefits, noting that COVID-19 infection could be especially serious in pregnancy, and that more than 10,000 pregnant people had been vaccinated under the emergency authorisation with no issues arisen. CNN sets the documented record around it: the CDC formally recommended vaccination in pregnancy on 11 August 2021 after analysing data and finding no increased risk of miscarriage, and the American College of Obstetricians and Gynecologists still recommends it as of this spring. This file offers no view on any medical question and no guidance to anyone. What it rates is a reading method, and the method has not changed: when he said nothing, the nothing was the confession; now that eleven of 34,000 messages are public, one clause out of a two-clause answer is. source →

A fourth Fauci check landed this evening and it is the cleanest demonstration yet of the pattern logged earlier today. Snopes rated False the claim that Fauci co-authored a paper finding that masks caused most deaths in the 1918 pandemic. The paper is real: 'Predominant Role of Bacterial Pneumonia as a Cause of Death in Pandemic Influenza', published in The Journal of Infectious Diseases on 1 October 2008, drawing on lung tissue from 58 autopsies plus pathologic and bacteriologic data from 109 published reports covering 8,398 more. It concluded that most deaths likely followed from secondary bacterial pneumonia caused by bacteria that normally live in the nose and throat, entering the lungs along the path the virus had cleared. Fauci's own words in the 2008 NIAID release were that the virus landed the first blow while bacteria delivered the knockout punch. Snopes searched the paper for 'masks' and 'face coverings' and got no results; the subject does not appear in it. Two dates make the point. The claim is old, addressed by Reuters in 2021 and by the Associated Press in 2023, when Fauci called the mask suggestion nonsense with absolutely no evidence behind it. And the current wave began on 30 July 2026, the day after the Senate hearing. The hearing did not produce this claim. It gave people a reason to post it again, which is what a retrieval system does. source →

Twelve days after the hearing, the measurable thing is what people are reading rather than what is being published, and it says the silence is working as a retrieval system. Of the thirteen fact-checks on Snopes' most-read list this morning, three concern Fauci. One of them dates from September 2020 and asks whether he criticised the China travel restrictions. Another, published 28 July, examines a claim that a pulmonary infarct recorded in his diary was a side effect of a COVID-19 vaccine, and Snopes rated it Mixture: the diary entry is real and dated June 2021, roughly six months after his first dose, and the CDC does not list pulmonary infarction as a common side effect of the mRNA vaccines used in the United States. The third is the sharpest, because it runs the other way. Snopes published a check on 2 August of the claim that the US president told people to inject bleach, and it exists in that form because the six-year-old rumour resurfaced after the Fauci hearing; Snopes rated it Mostly False, noting he asked the task force to look into injecting disinfectants and then said in the follow-up that cleaning products would be used to sterilise surfaces rather than injected. So a hearing that produced no answers is pulling a 2020 archive back to the top of the charts in both political directions at once. None of these checks is new, and that is the point of logging them together: the reading is current even though the material is not, which is what a claim looks like when it has stopped depending on events to sustain it. source →

The reasoning this file rates has now turned around and pointed the other way, which is worth recording because the answer does not change. Snopes reported on 7 August that posts defending Fauci had surfaced a claim that the US president invoked the Fifth Amendment more than 400 times in a 2022 deposition for his civil fraud case, put forward against Republican arguments that an innocent person would not have invoked it repeatedly. Snopes did not rate the claim. It found the figure supported by partial and secondhand evidence, resting on a judge's statement in a motion in that case, and said it could not access enough primary material to reach a verdict, which is a useful reminder that a counter-claim deployed in someone's defence is not exempt from checking either. On the substance, the position in this file covers both men identically. Invoking the privilege is not an admission by whoever does it, and reading a second person's invocations as proof of their guilt repeats precisely the error being objected to. The principle is not a scoreboard. It exists so that a refusal to answer cannot be counted as an answer, and it protects people one does not like on exactly the same terms as people one does. source →

The contempt process this file promised to track has moved, and the account needs updating rather than the verdict. On 6 August the Senate Homeland Security and Governmental Affairs Committee voted 8 to 5 along party lines to hold Fauci in contempt of Congress, with all Democrats opposing and two more opposing by proxy. Senator Paul told the panel that holding a witness in contempt is a serious thing and should be rare, but that the contempt power exists for precisely this circumstance. The usual next step would be a vote on the Senate floor; Paul told CBS News he intends instead to send the resolution directly to the Justice Department as a referral, arguing the floor vote is not required. An attorney for Fauci, David Schertler, called the vote a crude political stunt intended to punish his client for exercising his constitutional rights. Both characterisations are recorded here as statements, because that is what they are. NPR's account also sharpens something this file had left as a bare grey area: the January 2025 pardon covers conduct between 2014 and January 2025, and does not extend past that date. Paul argues the immunity means Fauci had no reason to fear testifying; the limit is also a concrete reason why invoking the privilege is not the empty gesture the viral claim treats it as. Nothing here changes the rating. A committee referral is not a charge, no prosecutor has acted, and no finding of wrongdoing exists against anyone. source →

Snopes published an investigation on 5 August into the one concrete, checkable allegation put to Fauci at the hearing, and could not rate it. Senator Josh Hawley accused him of using NIH staff on federal time to solicit cash prizes for himself totalling more than a million dollars, drawing on internal NIH emails released by Senator Rand Paul; Fauci declined to answer, citing the Fifth. Snopes reports that the emails, if legitimate, show staff assembling the required components of nomination packages after other scientists had already put Fauci forward, and that it found no evidence in them of him asking staff to solicit those nominations. The emails also indicate he worked with the NIH ethics office on which awards he could accept, and NIH employees may accept such prizes where they confer, in the rule's words, an exceptionally high honor in the fields of medicine or scientific research. Snopes has filed a public records request to see the originals and says it cannot rate the claim until it does. Fauci did not immediately respond to a request for comment and Hawley's office did not provide further documentation. Set that against what this file rates. The most specific accusation the hearing produced remains, a week on, unverifiable from the public record, because the underlying documents are a senator's release rather than an established one. An event that yielded no answers has not yielded a checkable allegation either. source →

Opening entry. The distinguishing feature of this claim is its multilingual spread: on 4 August the Brazilian fact-checker Boatos published five separate debunks in Portuguese, Spanish and English, with different payloads by market. The English versions concern masks and distancing, the Portuguese one attaches the hearing to Jair Bolsonaro's pandemic record, and the Spanish ones to vaccines and to hydroxychloroquine and ivermectin. All rest on the same impossibility: Fauci answered no questions on 29 July, so the hearing cannot have established the content of answers he did not give. The committee's contempt process was still running when this file was published and will be tracked here. source →