Mayo Clinic, nference AI analysis finds no real-world link among COVID-19 vaccines and brain blood clots

COVID-19 vaccine
Using nference’s AI software to parse the Mayo Clinic health records, researchers found 10 cases of brain blood clots among more than 132,000 COVID-19 vaccine doses. (solarseven/ iStock/ Getty Images Plus)

After a strong start to the U.S. vaccination campaign, the rate of daily shots has significantly slowed in recent weeks—despite nationwide offers of free beers, doughnuts and million-dollar lottery checks as inoculation incentives.

The plateau is largely driven by lingering uncertainty about the long-term effects of the three available COVID-19 vaccines, from Pfizer and BioNTech, Moderna and Johnson & Johnson. A new study from nference and the Mayo Clinic, however, erases at least one of those concerns by finding no links among any of the three vaccines and cerebral venous sinus thrombosis, the rare brain clots known as CVST.

The study results were published this week in the Journal of Stroke and Cerebrovascular Diseases. The FDA's paused and unpaused of the use of J&J's vaccine in April, after reports that at least eight patients developed blood clots in their brains, including one woman who died from the condition.

The FDA ultimately concluded that the risk of developing CVST from any of the COVID vaccines was too low to permanently stop using the J&J shot, and now the Mayo Clinic and nference confirm that there is no statistically significant link between CVST and those vaccines—as well as other, non-COVID shots.

RELATED: Mayo Clinic, nference to expand data partnership spanning COVID-19 and more

nference used its artificial intelligence software to analyze the health records of Mayo Clinic patients who had received any of the three vaccines. The software combed through electronic health record data, lab test results and clinical notes to tally any instances of CVST that were reported within 30 days of vaccine administration.

Out of a combined 132,916 doses, only 10 people were diagnosed with CVST in the month after—all of whom had received Pfizer and BioNTech's mRNA vaccine—amounting to just 0.0019% of the records included in the analysis.

And while those 10 cases did appear within a few weeks of vaccination, an equal number of CVST cases were reported in the 30 days leading up to vaccination, indicating that the post-shot blood clots were likely not linked to the vaccine, according to researchers.

“This study comes at a critical moment in the pandemic and provides independent evidence that it is generally safe to be vaccinated against COVID-19," Venky Soundararajan, co-founder and chief scientific officer of nference, said in a statement. "Our real-world, independent analysis, conducted with Mayo Clinic physicians, proves that anecdotal reports about the dangers of CVST as a result of COVID-19 vaccines are not generalizable."

“The real threat is the continued global emergence of SARS-CoV-2 variants, and continued mass vaccination is humanity's best bet to eradicate COVID-19 across our planet,” Soundararajan added.

In total, the study looked at 771,805 vaccinations—including 10 FDA-approved, non-COVID shots—spanning 266,094 Mayo patients and dating as far back as January 2017, and found no statistically significant differences in CVST rates for any of them.

RELATED: JPM: Mayo Clinic, nference launch ambitious data platform initiative

The analysis is the latest product of a longstanding AI and analytics-focused partnership between nference and Mayo Clinic. Their collaboration began in 2017, with the launch of drug repurposing biotech Qrativ.

Since then, the pair have joined forces to build an AI-powered platform to turn millions of tissue slides and years’ worth of unstructured physician notes into a searchable annotated database. And earlier this year, they announced plans to expand that project into research projects in digital pathology and heart rhythm diagnostics.