It’s long been a trope among the antivaccine movement that vaccines cause cancer, dating back to long before the pandemic. During the COVID-19 pandemic, however, this old trope mutated into a new, more aggressive form with claims arising in 2022 (as best as I can tell) that COVID-19 vaccines can cause what antivaxxers dubbed “turbo cancers,” cancers seen mainly in younger people that are unusually aggressive and more often fatal compared to “run-of-the-mill” cancers. Since then, the “turbo cancer” claim has, as cancers do, mutated and evolved to the point where there are some antivaxxers out there who, whenever they hear of a case of cancer in a celebrity, will blame the vaccine, regardless of the cancer type, how old the celebrity is, or whether the celebrity has even been vaccinated or not. In this, “turbo cancer” resembles “died suddenly,” in which it is always the COVID-19 vaccine to blame for the cancer or the sudden death, respectively, regardless of age or circumstances. COVID-19 vaccines seem to be magical that way, at least among conspiracy theorists. If you die and were ever vaccinated against COVID-19, it was the vaccines, regardless of how old you were or how bad your health might have been before death. If you develop cancer and were ever vaccinated against COVID-19, it was the vaccine.
But what about the disease itself?
COVID and cancer: Start with anecdotes
A more interesting question came up in an article in The Washington Post by Ariana Eunjung Cha entitled ‘Unusual’ cancers emerged after the pandemic. Doctors ask if covid is to blame. Of course, the idea that SARS-CoV-2 can cause cancer is somewhat more plausible than the claims that COVID-19 vaccines, in particular the mRNA-based COVID-19 vaccines, can cause “turbo cancer”; so I was interested right away in what sort of evidence is marshaled to support the idea. I was, however, tempted to adapt Betteridge’s Law of Headlines to this article, given the implied question—without an actual question mark!—and conclude that the answer is no, but I thought I’d go through the article, do some PubMed searches, and see what I thought about the state of the evidence regarding this question. Unfortunately (although not unexpectedly) the article starts out with some oncologists sharing anecdotes, because of course it does:
Kashyap Patel looked forward to his team’s Friday lunches. All the doctors from his oncology practice would gather in the open-air courtyard under the shadow of a tall magnolia tree and catch up. The atmosphere tended to the lighthearted and optimistic. But that week, he was distressed.
It was 2021, a year into the coronavirus pandemic, and as he slid into a chair, Patel shared that he’d just seen a patient in his 40s with cholangiocarcinoma, a rare and lethal cancer of the bile ducts that typically strikes people in their 70s and 80s. Initially, there was silence, and then one colleague after another said they’d recently treated patients who had similar diagnoses. Within a year of that meeting, the office had recorded seven such cases.
“I’ve been in practice 23 years and have never seen anything like this,” Patel, CEO of Carolina Blood and Cancer Care Associates, later recalled. Asutosh Gor, another oncologist, agreed: “We were all shaken.”
There was other weirdness, too: multiple patients contending with multiple types of cancer arising almost simultaneously, and more than a dozen new cases of other rare cancers.
Increasingly, Patel was left with an unsettling thought: Could the coronavirus be inflaming the embers of cancer?
I’ve discussed so many times over the years why “personal clinical experience” of physicians and other clinicians can be profoundly misleading. My favorite go-to examples have been pediatricians like Dr. Jay Gordon, who famously cited his “personal clinical experience” as the reason why he had long before come to the conclusion that vaccines cause autism. Worse, in Dr. Gordon’s case, he was absolutely convinced that his 30 years of “personal clinical experience” trumped all the scientific and epidemiological evidence that had consistently found (and consistently continues to find) no link between vaccines and autism. There are, of course, a lot of other examples, such as oncologists who come to believe that something or other causes cancer, such as wifi or cell phones, even though the evidence does not support that. I’m not saying that Drs. Patel and Gor fall into this category; I don’t know that yet. What I am saying is that we physicians frequently forget that we are human beings who are very prone to confirmation bias, recency bias, and selective memory, which is why careful epidemiological studies and randomized controlled clinical trials came into existence to determine causes of disease and identify treatments that work.
In fairness, however, one must also emphasize that the very first step in identifying associations is often “personal clinical experience” or personal observations of a potential link. Such observations in and of themselves are almost never sufficient to conclude that there is a link between an environmental factor (such as a virus and viral infection) and a disease, but they can be hypothesis generating observations that lead to more rigorous observations that can either confirm or fail to confirm the observed association scientifically. Because associations between environmental exposures and diseases are rarely as strong and clear-cut as, for example, tobacco smoking and lung cancer, it usually takes multiple large epidemiological investigations to confirm or refute the hypothesized link between an exposure and a disease and to quantify any dose-response effect. I will also note that, in the case of the pandemic, finding or refuting a link could be very difficult given potential confounding factors given the disruptions in screening and treatment that occurred in 2020 and well into 2021, combined with trends that predated the pandemic of more younger people being diagnosed with, for example, colorectal cancer, and the expected increase in cancer diagnoses that will occur over the next 20 years just because of the aging of the population.
One thing that I always do whenever I encounter articles like this is to Google the names of the experts used as sources, as in the past I have occasionally found some glaring failures to recognize that the source interviewed was a crank. I found no evidence of this for Drs. Patel or Gor, both of whom appear to be well-respected oncologists. That being said, I was curious about the example of cholangiocarcinoma. It is indeed an uncommon cancer, although there are other risk factors besides age for this cancer, including the most potent risk factor of all primary sclerosing cholangitis (which is associated with a 10-15% lifetime risk of cholangiocarcinoma), inflammatory bowel disease, cirrhosis or chronic liver disease due to viral hepatitis or chronic alcohol use, and a few others, none of which were mentioned. Moreover, one case of a cancer in a 40-something year old that is normally diagnosed in people over 70 does not a link make, and nothing is mentioned about whether the other cases were in younger people or in people in the age range more commonly associated with cholangiocarcinoma.
COVID and cancer: Cautionary notes
The WaPo article notes, as anyone discussing this topic must note, that the hypothesis that viral infection can cause cancer is not new. Many are the examples of cancers whose risk factors include viral infection, in particular cervical cancer, which is caused by certain strains of the human papilloma virus (HPV). Similarly, viral hepatitis, particularly due to hepatitis B and C, is a major cause of liver cancer worldwide, and infection with Epstein-Barr Virus (EBV) is associated with various forms of cancer, including Burkitt lymphoma, immunoblastic lymphoma, nasopharyngeal cancer, and stomach (gastric) cancer. Indeed, the WaPo article notes that 15-20% of all cancers worldwide likely originate from infections like these and other infections.
That being said, however, WaPo notes:
But there is no real world data linking SARS-CoV-2 to cancer, and some scientists remain skeptical.
John T. Schiller, a National Institutes of Health researcher and pioneer in the study of cancer-causing viruses, said pathogens known to cause cancer persist in the body long-term. But the class of respiratory viruses that includes influenza and RSV — a family that counts the coronavirus as a member — infects a patient and then typically goes away instead of lingering and is not believed to cause cancer.
“You can never say never, but that sort of … virus does not suggest being implicated in cancers,” Schiller said.
Count me in with the skeptics…for now. That doesn’t mean that I can’t be persuaded, but it does mean that the evidence that I’m seeing so far does not convince me. Let’s take a look, and maybe you’ll see why, even though in fairness there could be other mechanisms beyond just a direct transformation of cells into cancer:
David Tuveson, director of the Cancer Center at Cold Spring Harbor Laboratory and former president of the American Association for Cancer Research, said there’s no evidence the coronavirus directly transforms cells to make them cancerous. But that may not be the full story.
Tuveson said a number of small and early studies — many of which have been published within the past nine months — suggests that coronavirus infection can induce an inflammatory cascade and other responses that, in theory, could exacerbate the growth of cancer cells.
He has wondered whether it could be more akin to an environmental stressor — like tobacco, alcohol, asbestos or microplastics.
“Covid wrecks the body, and that’s where cancers can start,” Tuveson said, explaining how autopsy studies of people who died of covid-19 showed prematurely aged tissue.
It drives me crazy how, in 2024, reporters often still don’t include links to the relevant studies cited either as part of the article or by scientists interviewed for the article in the online version of news articles. Drives. Me. Crazy. There is no excuse. I wasted a fair amount of time on PubMed trying to find out which studies Dr. Tuveson was referring to, and the best I could come up with were studies showing that the virus could induce cellular senescence in the lung; e.g., this one, this one, and a few others. Senescence is the property that most normal cells have that limit the total number of cell divisions that they can undergo, after which they cease to divide and become senescent. Whether this is what Dr. Tuveson was talking about or not, I don’t know and can’t tell; so I’ll move on. (End of rant, also given that there are links to most studies in this article; I guess I was just annoyed about this one because this is a fairly general claim presented without links to studies.)
I will, however, note that chronic inflammation is usually not the sort of stimulus that causes cancer rapidly. Most cancers associated with chronic inflammation are diagnosed many years, sometimes decades, after the chronic inflammation starts; e.g., Marjolin’s ulcers, which are chronic ulcers after burn injuries that can degenerate into skin cancer, for which the mean latency after injury is 20-30 years, although one must concede that cases have been observed as soon as a few years after injury. Basically, in considering this hypothesis for COVID-19, we run into the same problem that those suggesting that COVID-19 vaccines. cause cancer run into, the fact that the pandemic is only four years old and that even the strongest carcinogens, such as ionizing radiation, generally do not cause solid tumors until a decade or more after exposure, although some hematologic malignancies can arise as little as two years after exposure. That’s why, if there’s one thing that’s true in this reporting, it’s that it will take many years to confirm or refute a link between COVID-19 infection and various cancers.
COVID and cancer: Evidence and hypotheses
First, let’s look at the current mainstream hypothesis, namely that the increase in late stage cancers has been observed since the pandemic started is likely an artifact of delayed care and delayed screening, which the WaPo article acknowledges as a possibility:
Even as the first wave of the coronavirus pummeled the United States, public health officials predicted a surge of cancer cases. A Lancet Oncology paper analyzed a national registry showing increases of Stage 4 disease — the most severe — across many cancer types in late 2020. Baptist Health Miami Cancer Institute, UC San Diego Health and other large institutions have released data showing continued increases in late-stage cancers.
Xuesong Han, scientific director of health services research at the American Cancer Society and lead author of the Lancet Oncology study, attributed the jump to people delaying or skipping care because of fears related to the virus or because of economic reasons and also to cultural factors, language barriers and discrimination. But Han acknowledged that biological mechanisms of SARS-CoV-2, the virus that causes covid-19, could be at play.
“I don’t have the data to support this opinion,” Han said. “But it’s an important question to follow up on.”
This is much better in that the article includes links to the actual studies. In any case, these studies (plus one in JAMA Oncology published in February that examined a nationwide cancer database) suggest that the increases observed were due to underdiagnosis of cancers, particularly early stage disease, during the time period when clinics and hospitals were closed to nonurgent cases. I’m sure that this is at least part of the story, but it it the whole story? The evidence in the article is largely mechanistic and circumstantial, as WaPo described some of the research presented at a recent symposium about COVID-19 and cancer.
For example:
The group’s loosely affiliated members are launching research studies that are trying to piece together the puzzle of coronavirus infection, long covid and cancer.
Wallace — the University of Pennsylvania scientist considered a father of the field of human mitochondrial genetics, which explores the power plants that fuel human cells — is researching how covid affects energy production in cells and how that might influence cancer vulnerability.
Separately, biodata experts are sequencing the gene profiles of organs from people who succumbed to covid and underwent autopsies.
And a University of Colorado team is studying whether covid reawakens dormant cancer cells in mice. Their provocative findings, according to a preprint report released in April, showed that when mice that were cancer survivors were infected with SARS-CoV-2, dormant cancer cells proliferated in the lungs. They saw similar results with the flu virus.
The preprint is interesting, but largely a mouse model, and I’ve learned over the years to be fairly skeptical of mouse models, at least initially. Sometimes they translate to humans; often they don’t. In fairness, the WaPo article does mention this, before moving on to other articles. Again, it’s all circumstantial:
Other studies offer telltale clues about the link between viruses and cancer.
Pathologists from the University of Arkansas for Medical Sciences reported in 2021 in the journal Communications Biology that SARS-CoV-2 proteins fueled the replication of a herpesvirus considered one of the major viruses leading to cancer. Other studies have implicated the coronavirus in helping to stimulate dormant breast cancer cells.
A paper published in 2023 in the journal Biochimie explored mechanisms the coronavirus could exploit to aggravate several forms of cancer, including lung, colorectal, pancreatic and oral. Researchers suggested the most likely pathway was through disrupting the body’s ability to suppress tumors, but researchers acknowledged a lack of direct evidence to support the theory.
First off, I’m not particularly impressed with the Biochimie review article, as it really represents a lot of speculation. Particularly weak, in my opinion, was the speculation of a link with breast cancer based on, well:
The hyperglycosylated S protein of the SARS-CoV-2 gamma variant can downregulate E-cadherin expression and upregulate the levels of N-cadherin and of the Zinc finger protein SNAI1 (SNAIL), resulting in increased EMT in breast epithelia cells. This, as with NF-κB pathway activation, can cause metastasis of breast cancers [174]. In another study the correlation between inhibiting the SARS-CoV-2 receptor, ACE2, and breast cancer has been illustrated [175]. COVID-19 treatment strategies which target ACE2 contributes to a dysregulated immune system which favors cancer progression. This pro-tumorigenic state is dominated by immunosuppressive T regulatory (Treg) cells and anti-inflammatory cytokines, such as IL-10. This immunosuppressive state in group 1 or luminal A breast cancer may enable tumor cells to proliferate and metastasize to distant organs [175].
The other article on breast cancer cited comes from 2020 and is interesting, but again quite speculative, as is the article on SARS-CoV-2 reactivating replication of herpes virus, which is even more speculative. None of this means that there might not be a link between SARS-CoV-2 infection and cancer, but it’s not particularly persuasive, at least not yet, to me. Other speculation in the Biochimie review is more suggestive, such as the ability of the virus to cause pulmonary fibrosis, which is a risk factor for lung cancer.
Given all the speculation about mechanism in this paper, I was surprised to see that nothing by Dr. Wafik El-Deiry was mentioned. He’s featured in this blog before because he has been a bit too ready to believe the “turbo cancer” narrative and has blurred the line, probably unintentionally (although I’m still not sure) between his research suggesting that spike protein from the virus can interfere with the activity of p53, a tumor suppressor gene so important that it has been called the “guardian of the genome” and claims that the vaccine causes cancer. Indeed, I’m kind of surprised that he wasn’t interviewed for this article. But I digress.
I’ll also note that during my PubMed searches I found one interesting analysis published less than two months ago examining the question of whether there is an association between the three types of COVID-19 exposures (critically ill COVID-19, hospitalized COVID-19, and respiratory syndrome infection) and 33 different types of cancers in the European population using a Mendelian randomization (MR) model and found:
The results of the inverse-variance-weighted model indicated that genetic liabilities to critically ill COVID-19 had suggestive causal associations with the increased risk for HER2-positive breast cancer (odds ratio [OR] = 1.0924; p-value = 0.0116), esophageal cancer (OR = 1.0004; p-value = 0.0226), colorectal cancer (OR = 1.0010; p-value = 0.0242), stomach cancer (OR = 1.2394; p-value = 0.0331), and colon cancer (OR = 1.0006; p-value = 0.0453). The genetic liabilities to hospitalized COVID-19 had suggestive causal associations with the increased risk for HER2-positive breast cancer (OR = 1.1096; p-value = 0.0458), esophageal cancer (OR = 1.0005; p-value = 0.0440) as well as stomach cancer (OR = 1.3043; p-value = 0.0476). The genetic liabilities to SARS-CoV-2 infection had suggestive causal associations with the increased risk for stomach cancer (OR = 2.8563; p-value = 0.0019) but with the decreasing risk for head and neck cancer (OR = 0.9986, p-value = 0.0426). The causal associations of the above combinations were robust through the test of heterogeneity and pleiotropy. Together, our study indicated that COVID-19 had causal effects on cancer risk.
Let’s just say that I’m not entirely convinced that this model did indicate causal effects on cancer risk for these tumors, although I was intrigued that only the HER2-positive subtype of breast cancer showed an association. I also note that the associations were generally quite small; for example, a 9.2% increased risk of HER2-positive breast cancer (for stomach cancer, 23.9% increased risk and for esophageal and colon cancer, much less than 1% increased risk) associated with severe infection. It also doesn’t make a lot of sense that less severe COVID-19 might associate with esophageal cancer. Moreover, “statistically significant” does not equal “clinically significant. For example, I’m not even sure why they reported barely statistically significant increased risks for certain cancers of around 0.05%-0.10%. Such tiny increases could easily be due to random noise and, even if valid, would not be clinically significant. Even so, though, the numbers for other cancers, such as the 23.9% increased risk of stomach cancer, are somewhat intriguing.
The WaPo article moves on to describe the sorts of research being done to clarify whether there is a link between COVID-19 infection and cancer:
From his practice in this Southern town, Patel is conducting his own research into what he has taken to calling “an unusual pattern” of cancers. He is driven by watching patients — especially younger ones — die so quickly.
He’s looking at potential correlations between long-covid markers and unusual cancers. He has collected data from nearly 300 patients and wants to create a national registry to analyze trends. So far, his office has logged more than 15 patients with multiple cancers, more than 35 patients with rare cancers and more than 15 couples with new cancers since the pandemic began in 2020.
Patel theorizes the effects of coronavirus infections could be cumulative in people infected multiple times. Pandemic-related stress may compound the threat, he said, by exacerbating inflammation.
If a link is established between the virus and cancer, he said, doctors might identify patients at greater risk and implement screenings earlier and even put some patients on anti-inflammatory drugs.
This is all well and good, but data from a small practice like this is unlikely to be particularly informative given the small numbers of patients enrolled. Determining associations and teasing out causation versus correlation require large databases; so a national registry would be better. However, we don’t need to reinvent the wheel, given that there are already huge cancer databases, such as the SEER Database and the American College of Surgeons National Cancer Database (NCDB), which already track a number of cancer risk factors—although, as I’ve learned over the years, a frustratingly small number in some cases—could be adapted to track infections in future patients entered into the database. I also note from some minor experience that producing and maintaining such databases requires a huge amount of resources and work.
The bottom line
One bottom line that I must emphasize before discussing the science regarding cancer and COVID-19 is that there is not a shred of even mildly compelling evidence that COVID-19 vaccines cause cancer, “turbo cancer” or otherwise, nor is there a plausible mechanism by which the vaccine could do this, not that these observations have stopped antivaxxers from citing this news article:
Leave it to the NY Post to add to its secondary reporting of a WaPo article a couple of paragraphs falsely trying to blame the vaccines, even though the WaPo article quoted a scientist explicitly saying that there’s no scientific reason to say it was the vaccines, while featuring a photo of a nurse taking care of a critically ill patient in an ICU. Antivaxxers gonna antivax, I guess.
But back to the question of COVID-19 infection itself and vaccines.
My perspective after having read the WaPo article and done a perusal of the recent scientific literature is that it is certainly possible that COVID-19 infection could predispose to certain kinds of cancers. It might even reawaken dormant cancer cells in cancer survivors. However, the evidence thus far is only mildly suggestive, to me at least, and, truth be told, to most virologists and cancer biologists. It could well be a case of a confounder such as delays in cancer screening and treatment. Indeed a number of oncologists and cancer researchers predicted that delays in screening and treatment due to the early pandemic might result in an increase in cancer cases over the next few years, as the system caught up with the delays. Again, it will take years of research to tease out what is or is not happening with respect to cancer prevalence and whether it is linked to COVID-19 infection, and it will be difficult to separate effects of COVID-19 from confounders, such as the pre-pandemic increasing incidence of cancers (e.g., colorectal) cancer in younger people and the expected increase in cancer prevalence just because the median age of the population in developed countries is continuing to increase.
Unfortunately, WaPo couldn’t resist a very annoying journalistic trope. Remember how the article started with a group of oncologists sharing anecdotes and speculating over whether COVID-19 was causing an increase in incidence of rare cancers and multiple cancers? Well, the article is bookended with an anecdote:
Then there’s Bob and Bonnie Krall, a couple who in a 14-month stretch endured three types of cancer between them, despite neither having a family history or genetic predisposition.
You know where this is going next:
Cancers typically start in one part of the body and spread. It’s rare for discrete cancers to begin in different parts of the body during a short window of time. Patel said the Kralls and the 78-year-old had coronavirus antibodies “through the roof” in their blood although he’s not sure how that relates to cancer, if at all.
Patel met the Kralls in 2022 when Bob was diagnosed with chronic lymphocytic leukemia, a cancer of the blood and bone marrow. During one of her husband’s treatments, Bonnie mentioned she was dealing with her own health issues related to long covid, including stomach pain. Medications weren’t helping. Patel ran tests and discovered Bonnie had cancer, too. By the time she was scheduled for surgery a few weeks after diagnosis, the malignancy in her abdomen had grown three more centimeters and weighed 8.5 pounds. This year, Bob’s doctors found cancer in his lungs.
I hate to hear of cases like these, and I wish the Kralls nothing but good fortune (and, of course, response to therapy) in their battle with their cancers. Unfortunately, though, Cha used their case to make the speculation about COVID-19 and cancer feel much more believable to a lay public, more so than current scientific evidence would suggest:
Bob’s blood cancer, CLL, is considered rare with only four or five cases in every 100,000 people, but he has been surprised to find four of his neighbors and friends have it, too.
“It’s like a cold. It seems like everyone has it,” said Bob, 73, a flight instructor.
Even before Patel mentioned his research, Bonnie, 74, who used to work part-time as a front desk assistant and lives with her husband in Fort Mill, S.C., had wondered whether their cancers were related to the coronavirus. She was infected shortly before her cancer diagnosis.
“Maybe if we didn’t get covid …” she started before trailing off. Bob shrugged and finished for her. “Maybe we would have been better? Maybe we would have been worse.”
Notice the implication that “everybody’s getting” this particular rare cancer and it therefore is likely to be due to COVID-19. Maybe. Maybe not. We don’t even know that “everyone has it”—or even that way more people “have it” than before 2020. This is all purely anecdotal and a very annoying journalistic trope that I’ve seen over and over and over that, in the name of providing “human interest” to a story, actually undermines the reporting by presenting human beings who are suffering as examples suggesting a phenomenon that might very well not be real. The same thing was routinely done in stories about vaccines and autism. Of course, this is not a case of a hypothesis that’s been long falsified, like the claim that vaccines cause autism, but it does give undue credence to one side of the controversy.
Over the next few years, I fully expect the evidence base to grow and the science to be done that will clarify if there really is an association between cancer risk and COVID-19 infection, repeat COVID-19 infections, or long COVID. Right now, I am agnostic regarding such a link. It is not implausible that such a link might exist between COVID-19 and elevated risk of certain cancers. I’m just hoping that in a couple of years I can revisit this topic and have a more definitive answer one way or the other.