20-05-2025
The Strange Link Between Cold Sores and Alzheimer's Disease
This transcript has been edited for clarity.
Welcome to Impact Factor , your weekly dose of commentary on a new medical study. I'm Dr F. Perry Wilson from the Yale School of Medicine.
Two-thirds of you reading this will know the feeling. It starts with a numb, tingly, feeling in the lip. A day or so later, some redness, some swelling, and then, yup, a cold sore. It's a little frustrating, maybe a little embarrassing, but you wait it out for a few days and it goes away. No big deal, right?
Except for the fact that multiple studies suggest that cold sores might increase your risk for Alzheimer's disease.
Cold sores come from a viral infection, specifically herpes simplex virus 1 (HSV-1). There are multiple herpesviruses, which are all DNA viruses and include HSV-2 which causes the sexually transmitted infection; though, to be fair, both HSV-1 and -2 can lead to both types of infections. Varicella — the virus that causes chicken pox and shingles, Ebstein-Barr virus, CMV — are all herpesviruses. If you're human, you have almost certainly been infected by at least one. In any case, HSV-1 is one of the most common viral infections in the world. It's estimated that about two-thirds of the adult population are infected.
Unlike other viruses, such as flu or coronavirus, herpesviruses are incredibly difficult to completely fight off from your body. They get around immune surveillance by hiding out in the nucleus of other cells as just an innocuous bundle of DNA. This latent phase is asymptomatic. It lies dormant until, for reasons that are still not entirely clear, the DNA bundle loosens a bit and the cellular machinery turns those instructions into the proteins that make up new virus particles and boom — outbreak. The immune system gets revved up, the outbreak is contained, and the cycle repeats.
What does this all have to do with Alzheimer's disease? I was inspired to dig into this a bit because of a study appearing this week in BMJ Open , which suggests that HSV-1 infection nearly doubles the risk of Alzheimer's.
Let me run through the study's findings and then we can figure out if this makes any sense at all.
Researchers used the IQVIA PharMetrics Plus database to conduct the study. This is basically a large administrative claims database that covers much of the United States. It basically aggregates all the billing codes for medical care and medications from a bunch of commercial insurers; there are more than 200 million individuals represented in the file.
From those, they found 344,628 individuals who were diagnosed with Alzheimer's disease. For controls, they identified another 344,628 individuals with the same age, gender, region of the country, date of entry into the database, and — to account for contact with the medical system — the number of inpatient and outpatient visits.
Despite that, the groups were not exactly comparable. The individuals who would go on to develop Alzheimer's disease had a greater number of comorbidities, for example.
But the kicker of the study — the headline — is this finding. People with Alzheimer's disease were twice as likely to have HSV-1 compared with the controls. After accounting for the differences between them, infection with HSV-1 increased the odds of subsequently developing Alzheimer's disease by 80%.
Did you catch the problem with this graph? Take a look at the Y-axis. That's on the percentage scale. Sure, the people who went on to develop Alzheimer's disease had double the rate of HSV-1 infection, but the raw number is 0.44% vs 0.24%. Didn't I tell you at the beginning that about two-thirds of us are infected with HSV-1? That's quite a bit higher than 0.44%. What is going on here?
Welcome to the world of administrative data. The problem here is that the researchers could only identify people with HSV-1 based on some provider diagnosing them with HSV-1. More than that, entering a billing code for HSV-1.
Have you ever had a cold sore? Do you know whether your doctor added that to your medical history and billed insurance for it? Probably not. So we're missing an enormous number of infections here, and that calls the whole conclusion into question.
Now, you might say, sure, doctors aren't diagnosing the vast majority of HSV-1 cases, but surely this is true both for people who go on to develop Alzheimer's and for those who don't, and therefore the inference is valid. Maybe. But I'd feel better if we were talking about missing something like 10% of diagnoses instead of 99% like we are here.
I don't want to discount this too much, though. The paper has some other interesting findings. For instance, there was also a higher rate of HSV-2 and varicella infection among those who developed Alzheimer's disease; those are the other herpesviruses that infect nerve cells. There was no difference in rates of cytomegalovirus infection — another herpesvirus, but one that infects monocytes instead of nerve cells.
But let's say we believe the link between HSV and Alzheimer's, what can we do about it? The authors hypothesized that, if HSV is causative of Alzheimer's, treatment with antivirals would reduce the risk of Alzheimer's disease. And since prescription information was present in the dataset, they could model this.
Sure enough, those treated with antivirals were less likely — about 17% less likely — to develop Alzheimer's disease. This is interesting to me. In general, when you look at people who are treated for a condition, you can assume they had a more severe form of the condition (short of the treatment being done in the context of a randomized trial). Basically, people who get treated tend to be sicker than people who don't get treated, and so, in general, you see worse outcomes in the treated group — a stubborn problem in observational data called confounding by indication.
Here, we see the opposite, which adds some weight to the argument. So, despite the poor capture of HSV-1 infections, the link could be real. Some other studies support this hypothesis.
Alzheimer's disease is characterized by amyloid plaque deposition in the brain. Some mouse studies have shown that HSV induces the formation of amyloid plaques as an immune response and impairs the mouse's cognitive ability.
This study prospectively studied 1000 Swedish older adults over time and measured antibodies to HSV: 82% of people had those antibodies which comports with what we would expect. Still, those with the antibodies had about twice the risk of developing dementia as those without.
The authors of the paper in BMJ Open suggest 'antiherpetic therapies as potentially protective for AD-related dementia.' That feels like a bit of a leap to me at this point, and I will point out that this paper was funded by Gilead Sciences who have quite a few antivirals on the market and a new anti-herpetic drug that has recently completed phase 1a testing— so… grains of salt.
Still, for those who suffer from cold sores, a study like this may push you a bit towards treatment, at least during an outbreak. Short-term valacyclovir is relatively safe and reduces the duration of the cold sore by about a day, which is nice. But if it reduces your risk of dementia as well, well, it might be a no-brainer.
