No, COVID vaccines still aren’t giving people cancer

The "turbo cancer" scare is back, this time with oncologists attached. The evidence hasn't budged.

A viral claim has been resuscitated. You may have seen it early in the COVID vaccine days: the claim that the vaccine is linked to the big C. This was back when mRNA technology seemed new, despite decades of research already behind it. The scariest version even had a name, turbo cancer, for tumors supposedly appearing or roaring back at unnatural speed. I'd hoped the claim was behind us. It isn't. It's back with new faces attached, including actual oncologists, the very people who treat cancer for a living. There's even a new study being used to support it. Look closely, though, and it falls apart.

The full post is shared on the Unbiased Science social media accounts.

Three sources are driving the current wave. Angus Dalgleish, a British oncologist, says he’s watched patients relapse after repeat doses and thinks the vaccine suppresses the T-cells that keep cancer in check. Wafik El-Deiry, a US cancer researcher, published a review in January that gathered mostly single-patient case reports and said they add up to a cancer “signal,” while conceding in the same paper that it can’t show the vaccine caused anything. When other scientists picked it apart, he called the criticism censorship and pointed to a cyberattack that took the journal’s site down as proof he was being silenced. Underneath both is a large South Korean study reporting higher cancer rates among people vaccinated against COVID-19. It’s the one piece of population-scale data the claim has, the citation nearly every scary post leans on, so it’s the one we need to look at closely.

First, though, the anecdotes. A sharp oncologist noticing a pattern in their own patients is worth taking seriously (that’s often how real signals first surface). But noticing is where a question starts, not where it’s answered. The answer needs a comparison group, something no single doctor’s caseload can provide, and neither can a pile of case reports. Without one, you can only show that two things happened close together, not that one caused the other.

An oncology practice makes that especially hard, and not through any fault of the doctor. It’s simply where recurrences, delayed diagnoses, and heavy monitoring all cluster, the hardest possible setting to eyeball whether something new is happening. That’s where most of these case reports came from, and case reports are what El-Deiry’s review was built on. Let’s discuss…

The study everyone is citing

A paper ran in Biomarker Research last year. It’s big: the study covered more than eight million people in the Korean national health system, comparing cancer diagnoses in vaccinated versus unvaccinated people after statistical matching. At one year, the vaccinated group showed higher rates of six cancers, the biggest bumps in prostate, lung, and thyroid. At a glance, alarming. Yet even the study’s own authors are careful here, as stated in their Supplementary Materials: “Although our data provide insights into COVID-19 vaccination–associated cancer risk, our findings do not establish causal relationships.”

Read it closely and it comes apart in three critical ways:

1. The timing doesn’t work. This is a one-year window, and solid tumors don’t run on that schedule. A prostate or lung cancer found within twelve months of a vaccine was already growing well before the needle went in. A one-year study is highly unlikely to detect a cancer caused by the vaccine. It can only catch one that was already there and finally got found.

2. More looking finds more. Vaccinated people were plugged into the medical system, seeing doctors and getting scans more often. Look harder at a group and you are more likely to find something that was already there. The pandemic layers a second distortion on top of that first one. Routine screening shut down for a stretch, then bunched up during catch-up, and the vaccinated group, already more plugged into the system, was the most likely to get swept into that wave. Compressed diagnoses in the more-screened group can look like a spike that was never really there. Nothing new was created, it was just found late. Lockdowns also shifted risk factors. Alcohol consumption rose and physical activity dropped, though those move cancer over years, not the months this claim is about. Then look at which cancers light up: thyroid, prostate, breast, the ones we screen and scan for most. That matters because of a quirk called overdiagnosis. Scan enough people and you turn up small, real cancers that were never going to grow or cause harm, the kind a person would have died with, not from. Thyroid is the classic example: as neck imaging spread, diagnoses climbed for years while thyroid deaths stayed flat. More cancers were found, but no more people died, because the extra ones were never dangerous. What those cancers share isn’t biology; it’s that they’re the ones most shaped by screening and incidental detection.

3. The study’s own data sink the dose-response story. The sophisticated version of the claim isn’t that a vaccine builds a tumor from scratch in a year. It’s that it accelerates one already quietly there, every dose adding a push. That’s a dose-response claim. The study tested it directly, comparing vaccinated people who received more doses with those who received fewer. Overall cancer risk came out flat: a hazard ratio of 1.01, the confidence interval sitting right on top of no effect. The headline cancers (lung, colorectal, breast, thyroid) showed no increase with more doses. Leukemia went down. Gastric and pancreatic did tick up with more doses, but overall cancer risk didn’t budge. Plus, older, sicker people (some already living with cancer) were first in line for extra doses, so the most-dosed group started at higher baseline risk. A setup like that should have manufactured a dose-response on its own, and it still came back flat.

The study is on shakier ground than that, anyway. In October, the journal that published it, Biomarker Research, issued a formal expression of concern, alerting readers that problems with the paper had been raised and were under investigation.

The paper ran as a research letter rather than a full study, and its results fall straight into the multiple-comparisons problem: test dozens of cancer types and a few will clear the statistical bar by chance alone. The ones that did here were the common, heavily screened cancers, with one cluster of them pinned on the mRNA vaccines (primarily Pfizer-BioNTech) and a different cluster on the viral-vector vaccines (likely AstraZeneca or Janssen, though the paper doesn’t specify brands), and no biology tying either set together. Buried in the supplementary material, though not the main text, is the further concession that the one-year window is too short to judge cancer, and that reverse causation or surveillance bias can’t be excluded. Independent fact-checkers landed in the same place. Citing the study as proof means overruling the journal, the authors, and the fact-checkers all at once.

No mechanism that holds up, no signal

Step back from any one paper and the deeper problem is that there’s no demonstrated mechanism and no population-level signal.

The full post can be found on the Unbiased Science social media pages.

The vaccine’s mRNA works in the cytoplasm, is short-lived, and is cleared by the cell’s normal machinery. It’s not written into your DNA. People have proposed mechanisms anyway, such as the idea that the vaccine “exhausts” the immune system, allowing cancer to progress. But a proposed mechanism is a hypothesis, not a finding, and none of these has been shown to happen. After billions of doses, the cancer signal you’d expect from a real carcinogen hasn’t turned up. US cancer registries make it plain: incidence dipped in 2020 as the pandemic delayed diagnoses, then resumed the same trend it was on before the pandemic, and overall cancer death rates kept falling, with no surge tracking the vaccine rollout. That’s the pattern you’d expect if the vaccines weren’t a carcinogen, not what you’d expect if billions of people had just picked up a powerful new one.

There’s a fair limit to admit here. These vaccines are only a few years old, and a cancer with a decades-long fuse wouldn’t show up yet in anyone’s data, theirs or mine, so nobody gets to call the long game fully settled. But look at what “turbo cancer” actually claims: fast, aggressive tumors within weeks or months, dormant disease roaring back right after a dose. That’s a short-fuse claim, and short fuses are exactly what a few years of data can test, and the same registries already ruling out a surge would be showing exactly that: a spike in aggressive, fast-moving cancers. There isn’t one.

Through decades of research, we know that many types of cancer develop slowly, over decades, through multistep cellular changes that induce cancer cell growth. Colorectal cancer is the best-studied example, developing over roughly 40 years. Cancers linked to chemical carcinogens like asbestos or arsenic show a similar timeline. None of the six cancers flagged in the Korean study, thyroid, gastric, colorectal, lung, breast, prostate, are cancers known for rapid onset.

Two claims deserve their own answer, though. Dalgleish points to melanoma recurrence, and the Korean study flagged elevated pancreatic cancer with more booster doses. Neither is the smoking gun it looks like. Pancreatic cancer is already known for aggressive, fast-moving disease once it appears, that’s a feature of the cancer, not a new trigger. Melanoma recurrence is famously unpredictable, showing up anywhere from months to over a decade after apparent remission, which is exactly why survivors stay under long-term surveillance. A relapse near a vaccine dose tells you nothing about what caused it.

If individual cases can’t settle it, population-level data should be able to. The Korean study’s numbers can be explained by issues like timing, screening, and reverse causation. What can’t be explained is the silence in the monitoring systems built specifically to catch rare vaccine harms. No wave has shown up there, and that isn’t for lack of looking. COVID vaccines have been watched more closely than any vaccine in history. That monitoring caught rare harms within months of rollout: myocarditis in young men and a clotting syndrome with the adenovirus versions. A system sensitive enough to catch those would be very unlikely to miss a wave of aggressive cancers.

The claim hit mainstream news last year, when a speaker at a UK political conference tied Dalgleish’s version, with no evidence, to the cancers in the royal family. In response, the BMJ put it to cancer experts, who were blunt: there’s no medical evidence for turbo cancer, no plausible mechanism, and no rise in any cancer after the rollout once you track the numbers. One flagged the darker irony that the same mRNA technology getting blamed here is now behind some of the most promising cancer treatments in development.

The wrong suspect

Now ask what changed for the entire human population in 2020. Not the vaccine, which came later. A brand-new virus swept through nearly every country on earth. Most people caught it more than once, and it’s still circulating. If you’re going to look for something with the reach and the biology to plausibly touch cancer, the virus is a more sensible place to start than a vaccine that arrived afterward. That doesn’t mean it does. It means it’s a better place to look, and a few researchers have started looking there.

There’s preclinical work suggesting the infection itself, not the vaccine, can wake up dormant cancer cells. Mouse studies out of a University of Colorado lab (see here and here) made headlines for showing exactly that. Dr. Jenn Dowd went looking for the same pattern in people, and it wasn’t there. The human studies making the scary claim leaned on datasets that barely tracked who had actually been infected, or that just restated something we already knew, that COVID raises your overall risk of dying. Cancer deaths didn’t climb at the population level even after nearly everyone caught COVID. So where does that leave me? Unconvinced that either the shot or the infection is driving cancer in humans, though the latter is far more plausible (and that’s where a preliminary lab signal actually is). But more plausible is doing narrow work here. It means the question deserves more study, not that it’s closer to proven.

That points at a problem sitting underneath every study aimed at the vaccine. To pin cancer on the shot, you compare vaccinated people against unvaccinated people and watch for a difference. But by late 2022, nearly everyone had caught COVID at least once, vaccinated and unvaccinated alike. So the unvaccinated were never a clean, virus-free group to measure against. If the whole worry is spike protein or lingering inflammation, the unvaccinated got a bigger dose of both, straight from the infection. Most of these studies never measured prior infection well enough to pull the two apart. The big Korean one couldn’t. And when one study finally waited a realistic stretch before it started counting cancers, the association shrank or flipped.

That same reasoning, spike and inflammation as the culprits, turns on itself. Infection delivers more of both than the vaccine does, and vaccination lowers your exposure to them. The findings people are waving around to argue against the shot are an argument for getting it.

If you’re a survivor and a scary post has you second-guessing your next dose, the National Cancer Institute is blunt about this: There’s no evidence that COVID vaccines cause cancer, trigger a recurrence, or make an existing cancer more aggressive, and they don’t change your DNA. What they can do is make the lymph nodes under your arm swell and light up on a scan for a week or two, a normal immune response and not a returning tumor, though your care team should know about the vaccine before any imaging so nobody misreads it. That’s one benign, well-documented reason a post-vaccine scan can look alarming when nothing is wrong. The NCI’s guidance for people with cancer is to get vaccinated, because the virus is the real danger to them.

The myth carries a cost. mRNA is becoming one of the most promising tools we have for treating cancer, with more than 120 clinical trials (we recently posted about one) teaching immune systems to hunt down tumor cells. Scaring people off it protects no one. It pulls something useful away from the patients most likely to need it, and it spooks the immunocompromised and cancer survivors away from a vaccine that guards them against a virus that can seriously hurt them.

If a post like this rattled you, that reaction makes sense, but it doesn’t make the post right. The term “turbo cancer” is invented. The study everyone is citing argues against the scary version more than for it, and its own journal has attached an expression of concern to it. No proposed mechanism has held up, and the rapid cancers the story predicts haven’t appeared across billions of doses.

So go ahead and turbo-cancel this claim.

Stay curious,

Unbiased Science