HSV-1, Cold Sores, and Alzheimer's: The APOE4 Connection and the Antiviral Question

In a 1997 study, people who had both herpes simplex virus type 1 (HSV-1, the common cold-sore virus) in their brain and at least one copy of APOE4 had roughly 17 times the odds of Alzheimer’s disease, while either factor on its own barely moved the needle. That single finding launched almost three decades of research and one obvious carrier question: if a virus and your gene team up to raise risk, could an antiviral pill break up the partnership? In early 2026 a randomized trial answered that question directly, and the answer is not the one most carriers were hoping for.

What the evidence actually shows

The original signal came from Ruth Itzhaki and colleagues in The Lancet. Looking at brain tissue, they found that the combination of HSV-1 plus an APOE4 allele carried an odds ratio of about 16.8 for Alzheimer’s (95% confidence interval 3.61 to 77.8). APOE4 on its own, in people without detectable brain HSV-1, sat at an odds ratio of just 1.67, not statistically distinguishable from no risk at all. Their blunt conclusion was that neither the virus nor the gene alone raised risk much; it was the pair together that did the damage. Note the enormous confidence interval on that 16.8, a tell that the sample was small. This is a striking hypothesis-generator, not a settled number.

The interaction has held up in better-designed work since. A 2020 French study (Linard and colleagues, in Alzheimer’s & Dementia) followed older adults and measured antibodies to HSV-1 over time. Among APOE4 carriers, signs of recent viral reactivation tracked with sharply higher dementia risk: elevated IgM antibodies came with a hazard ratio of 3.68 (95% CI 1.08 to 12.55) and high IgG levels with 3.28 (95% CI 1.19 to 9.03). In people without APOE4, there was no significant link at all. So the pattern repeats: the virus seems to matter for the brain mainly in carriers.

Then comes the part that makes this more than an academic curiosity. A 2021 Swedish nested case-control study (Hemmingsson and colleagues) looked at HSV-1-positive people and asked whether those who had ever filled an antiviral prescription (acyclovir, valacyclovir, and the like) went on to develop Alzheimer’s less often. They did: an odds ratio of about 0.29, which works out to roughly 70% lower odds. On its face, that looks like a green light.

It is not, and here is why. That last study is observational, and the gap it is measuring is tiny in absolute terms: only 2.3% of the dementia cases had taken an antiviral, versus 7.6% of the dementia-free controls. People who get treated for cold sores differ from people who do not in ways that are hard to fully scrub out. To actually test whether the drug changes the disease, you need a randomized trial. In February 2026 we got one.

The VALAD trial (Devanand and colleagues, published in JAMA) randomized 120 people with early, symptomatic Alzheimer’s and HSV antibodies to either valacyclovir or placebo for 78 weeks, then measured cognitive change on the standard ADAS-Cog scale. The valacyclovir group did not do better. They did slightly worse: they declined by about 10.9 points versus 6.9 in the placebo group, a 3.9-point gap favoring placebo (P = 0.01). The authors concluded the drug was not effective and should not be used to treat early Alzheimer’s in this population. That is the cleanest causal evidence we have, and it is firmly negative.

Why the gene and the virus interact

The leading explanation rests on what APOE actually does, and there is one analogy that ties it together: think of HSV-1 as a houseguest who never fully leaves, and APOE4 as both a too-easy front door and a sloppy cleanup crew.

HSV-1 is a master of hiding. After the first infection (often in childhood), it retreats into nerve cells and goes dormant, then reactivates under stress, illness, or aging, which is the cold sore you get on your lip. The worry is that the same reactivation can happen along nerves that reach the brain. Here APOE enters twice. The ApoE protein is a lipid carrier that the virus appears to exploit to slip into cells, and the E4 form may make that entry easier, so the virus gets in more readily. APOE4 is also the version that clears cellular debris and tamps down brain inflammation poorly. So in a carrier, each reactivation may both let more virus in and leave a messier, more inflamed aftermath. Repeat that cycle over decades and you get exactly the smoldering neuroinflammation that feeds amyloid and tau, the hallmarks of Alzheimer’s. (For the wider picture of what the E4 protein gets wrong in the brain, see how APOE4 affects the brain.)

That is a plausible, mechanistically coherent story. It is also why the negative trial is so instructive: a believable mechanism is a reason to run the experiment, not a substitute for running it.

The numbers in one place

Finding	Figure	What kind of evidence
Brain HSV-1 + APOE4 vs neither	OR ~16.8 (CI 3.61-77.8)	Small case study, 1997
APOE4 alone, no brain HSV-1	OR ~1.67 (CI 0.21-13.4)	Same study; not significant
Recent HSV-1 reactivation in carriers (IgM)	HR 3.68 (CI 1.08-12.55)	Prospective cohort, 2020
Ever used an antiviral (HSV-1+ people)	OR ~0.29, ~70% lower	Observational, confounded
Valacyclovir vs placebo in early AD	3.9 points worse on ADAS-Cog (P=.01)	Randomized trial, 2026

What is contested, and what is not

What is solid: APOE4 and HSV-1 interact, and the interaction is largely carrier-specific. That has replicated across decades and study designs.

What is shaky: everything about treatment. The “antivirals lower risk” signal is observational and small, exactly the kind of association that randomized trials so often fail to confirm. And VALAD has a real limitation worth naming, treating people who already have Alzheimer’s is not the same as preventing it in healthy carriers. It is conceivable that antivirals started decades earlier, before symptoms, would behave differently. But that is a hope, not data, and the only randomized evidence in hand is negative (and slightly harmful). Until a prevention trial in healthy carriers exists, the honest reading is: do not treat the antiviral idea as established. This is a good place to practice reading a study like a skeptic: a strong mechanism plus an observational association is precisely the setup that fools people.

What to actually do

The disappointing-but-honest bottom line: there is no evidence-based antiviral move here for an APOE4 carrier. Do not start acyclovir or valacyclovir to prevent or treat Alzheimer’s. The one randomized trial said it does not help and may hurt.

What is reasonable:

• Treat cold sores as cold sores. If you get frequent or severe oral or genital herpes outbreaks, treating them for their own sake is standard care. Just do not frame it as dementia prevention, because that benefit is unproven.
• Pivot to the lever that does have stronger evidence: the shingles vaccine. Shingles comes from a different herpes-family virus, but multiple studies link vaccination to lower dementia risk, and the evidence there is more robust than anything on HSV-1 antivirals. This is the genuinely actionable cousin of the cold-sore story. See the shingles vaccine and dementia risk.
• Put your energy where the returns are proven. The biggest, best-supported risk reductions for carriers come from the basics done consistently: blood pressure, exercise, sleep, hearing, and diet. The virus angle is interesting; these are decisive. See how APOE4 affects the brain and exercise and APOE4.

Common questions

I get cold sores and I carry APOE4. Should I be worried? Not in a way that should change your daily life. Most people carry HSV-1 (it is extremely common), and carrying it is not a diagnosis. The interaction raises risk statistically across a population; it does not predict your individual fate. Focus on the proven levers.

If antivirals lowered risk in one study, why not just take them to be safe? Because the better evidence says they do not work for this. The “lower risk” study was observational and could not separate the drug from the kind of person who takes it, and the randomized trial that could separate them found no benefit and a small worsening. “To be safe” is not a free action when the controlled data point the wrong way.

Could a prevention trial in healthy carriers still turn out positive? Possibly. Treating established disease is a different test than preventing it, so the door is not fully closed. But that is a reason to wait for evidence, not to act ahead of it. Association plus a believable mechanism is not causation, and the one experiment designed to prove a treatable causal link came back negative.

The gene-virus link is real and fascinating, but the actionable payoff today is a vaccine, not an antiviral. This is general education, not medical advice.

Sources

• Itzhaki RF, Lin WR, et al. Herpes simplex virus type 1 in brain and risk of Alzheimer’s disease. The Lancet, 1997. PubMed
• Linard M, et al. Interaction between APOE4 and herpes simplex virus type 1 in Alzheimer’s disease. Alzheimer’s & Dementia, 2020. PubMed
• Hemmingsson E-S, et al. Antiviral treatment associated with reduced risk of clinical Alzheimer’s disease. Alzheimer’s & Dementia: TRCI, 2021. PMC
• Devanand DP, et al. Valacyclovir Treatment of Early Symptomatic Alzheimer Disease: The VALAD Randomized Clinical Trial. JAMA, 2026. PubMed
• ApoE and HSV-1 entry / biomarker context. Scientific Reports, 2024. Article