For years, headlines have warned that vaping may cause heart attacks, stroke, liver disease, cancer and now even chronic kidney disease.
The claims are often sensational – and widely shared. Once they are published in peer-reviewed journals, they can look like hard facts.
But a closer look at the evidence reveals a more complicated story.
Some of the most alarming studies linking vapes to serious disease have later been retracted. Others have not been formally withdrawn, but have been heavily challenged by outside researchers who say the methods are so weak that the conclusions cannot be trusted. Official records show that at least five vaping-risk papers have been retracted since 2020.
That does not mean vaping is risk-free, but it does mean a growing number of high-profile claims have turned out to rest on shaky foundations.
The latest flashpoint centres on a 2025 paper published in BMC Public Health, which reported that vape use was linked to chronic kidney disease (CKD) in a “dose-dependent” way. The study analysed data from 872 adults and concluded that people who vape had “2.50-fold higher odds of CKD” after adjusting for various factors.
At face value, the findings look definitive. But within weeks of publication, other researchers began questioning whether the results could be trusted at all.
A study under the microscope
Arielle Selya, a behavioural scientist who has spent years analysing vaping research, said she and her colleagues were immediately sceptical. They attempted to replicate the study using the same publicly available dataset – but say they were unable to reproduce even the most basic numbers.
In a formal letter to the journal, they wrote that they were “unable to even approximately replicate the most basic numbers” reported in the paper. The original study identified 63 cases of CKD among 188 vape users – around 34 per cent. Selya’s team found just 17 cases in a slightly larger sample, equivalent to 8.4 per cent.
They warned that the analysis appeared to include “hallucinated cases of CKD” and called on the journal to either verify the data or retract the paper. The concerns did not stop there. The critics argued the study failed to properly account for smoking history – a key issue in almost all research on vaping and health.
The same problems, again and again
According to Selya, these kinds of issues are not unusual. “You’ve hit the nail on the head,” she said. “Of the hundreds of studies I’ve seen on e-cigarette use and some health outcomes, the vast majority have at least two critical flaws: not fully accounting for smoking history, and unknown directionality of the association.”
The problem is simple, but often overlooked. Most people who vape are current or former smokers. If a study does not properly capture their full smoking history – how long they smoked, how much, and when they quit – it can end up attributing smoking-related disease to vaping instead.
“The result is that lingering or ongoing effects from cigarette smoking are being blamed on vaping,” Selya said.
Timing is another major issue. Many studies are cross-sectional, meaning they look at people at a single point in time. That makes it difficult to determine what came first.
“In fact, in the current population, it’s more likely that any given smoking-related disease developed well before initiating vaping,” she said.
That opens the door to what researchers call reverse causality – where people develop smoking-related illness, switch to vaping, but the data is interpreted the other way around.
When flawed studies make headlines
Clive Bates, a long-time advocate for tobacco harm reduction, says these problems are well known – but often ignored.
“Confirmation bias is at the heart of it,” he said. “They find what they want to find, and there are many ways to do that.”
He points to a familiar pattern of errors: “confusing correlation with causation, ignoring smoking history, reverse causation (when people turn to vapes because they are ill), attributing harms to vapes that arose before the person started vaping.”
Despite these limitations, studies can still generate headlines suggesting vaping causes serious disease.
“A scare story works well in the media,” Bates said. “It gives an action-seeking politician something to do that doesn’t cost too much; it gives academics a reason to exist, and attracts funding.”
Retractions are the visible tip of the iceberg
The most visible sign of problems in the research is the number of studies that have been formally retracted.
In 2020, a widely reported paper linking vaping to heart attacks was withdrawn after it emerged that many of the heart attacks in the dataset had occurred before participants started vaping.
In late 2022, a study suggesting higher cancer risk among vape users was later retracted over concerns about its methodology and data processing.
In 2023, a paper linking smoking and vape use to chronic liver disease was retracted.
More recently, a 2022 study linking vaping to stroke was retracted after the journal identified “several major errors in the data analysis,” including “impossible sample sizes” and uncertainty over whether vaping occurred before the stroke.
And in 2026, a systematic review linking vaping to cancer was pulled after an investigation found “multiple serious flaws that materially affect the reliability of the findings and conclusions.”
These cases are not identical, but they share a common thread: strong claims that did not survive closer scrutiny.
Selya says the retractions may only scratch the surface. “The disconcerting thing is, some of the papers that were retracted were no worse than hundreds of similar papers that still remain on the record,” she said.
Why corrections come too late
Even when flawed studies are eventually challenged, the correction process is slow – often far slower than the spread of the original claim.
“It’s always encouraging to see the scientific record being corrected like this,” Selya said. “But… the damage is often already done by the time that happens.”
Bates agrees. “Journal editors seem extremely reluctant to correct or retract papers,” he said. “It can take years, but it usually doesn’t happen at all.”
In the meantime, early findings can shape public perception, media coverage and even policy decisions. Bates said: “A lie gets halfway around the world before the truth has a chance to get its pants on.”
What better research would look like
Selya says many of the problems could be addressed with relatively simple improvements in study design. She argues that researchers should adjust for cumulative smoking history – not just whether someone is a current or former smoker – and should analyse results separately by smoking status.
“If e-cigarettes have a causal effect on the health outcome, then this should show the same effect regardless of if the person never smoked, formerly smoked, or currently smokes,” she said.
She also calls for better data on timing, specifically whether disease developed before or after someone started vaping. “These may not prevent every kind of bias, but they would fix the large majority of issues I see each week in the new literature,” she said.
Real-world consequences
The stakes are not just academic. Selya says misunderstandings about vaping risk are widespread, not only among the public, but also among healthcare professionals.
“A large majority of the general public, of people who smoke, and even of healthcare providers don’t have correct understanding of the continuum of harm across tobacco products,” she said.
She added that flawed research has likely contributed to that confusion. “The alarm about the harms of vaping has likely caused much more real harm with respect to scaring people who smoke away from trying e-cigarettes or scaring people who already switched back to cigarettes.”
Bates makes a similar point. “More people will continue to smoke, more people will revert to smoking, and more people will take up smoking instead of vaping,” he said. “Nothing good comes from misleading people about safer alternatives to cigarettes.”
A system under strain
Both experts say the issue goes beyond individual studies, and to deeper problems in how research is produced and published. Selya points to publication pressures, data limitations and funding incentives that prioritise certain types of findings.
Bates sees a wider structural problem. “… It is endemic in research where there is a political dimension and powerful interest groups,” he said.
In a field where the stakes are high and the evidence is still evolving, the biggest headlines don’t always rest on the strongest science.
Disclosure: Arielle Selya is an employee of Pinney Associates which consults to Juul Labs on tobacco harm reduction. She also serves as a scientific advisor to the Global Forum on Nicotine. Her opinions here are her own and do not reflect those of her clients or employers.
