Shingles Vaccines, Dementia, and the Limits of Skepticism
When does healthy scientific skepticism become dismissing good evidence?
In medicine, we often have disagreements about data, treatments, and the meaning of studies. This is a bit disconcerting for many in the public, because people want certainty. And in medicine, as in science, we have to be uncomfortable with uncertainty. The only people who are certain are fool, not scientists.
You will hear scientists talk about probabilities, you will hear hucksters tell you in confidence - hence the term “con men.”
Below, I am outlining a disagreement about the studies, pointing out that the vaccine against Shingles can decrease the risk of dementia. We don’t have an answer here, but this is an honest debate between two doctors (myself and Dr Prasad) about this data.
Can the Shingles Vaccine Prevent some dementia?
Last week, physician and medical commentator Dr. Vinay Prasad published an article questioning the growing body of research suggesting that shingles vaccination may reduce the risk of dementia.
If you haven’t heard of Dr. Prasad, you’re not alone. He is an oncologist and hematologist who has built a large following by criticizing what he considers weak medical evidence, unnecessary treatments, and the tendency of medicine to adopt interventions before they are adequately proven. On many issues, I’ve found his insistence on rigorous evidence refreshing. Medicine needs thoughtful skeptics.
The issue isn’t whether the shingles vaccine prevents shingles—that question was answered years ago. Shingrix is remarkably effective at preventing shingles and the debilitating nerve pain that can follow.
The question is whether preventing shingles also lowers the long-term risk of dementia.
Dr. Prasad argues the evidence isn’t convincing.
I differ with him - and here is why.
His Title Sets the Tone
His article is titled:
“Shingles Vaccines and Dementia: The Emperor Has No Clothes.”
That’s a provocative title.
It suggests researchers studying this question have collectively fooled themselves—that the apparent benefit is little more than an illusion.
That is a serious claim, and it deserves equally serious scrutiny. So let us go through his reasoning and my counters.
His First Argument: These Are “Just Observational Studies”
This is probably his central point.
He argues that observational studies cannot establish causation, and that people who choose to be vaccinated may simply be healthier than those who decline vaccination. This is called the “healthy user bias.” And it is a real bias. People who are healthier tend to take their medications, get vaccinated, see their dentist, watch their blood pressure, and maintain themselves reasonably well. Meaning, if we look at populations that take care of themselves, there are many reasons they maintain healthier and longer lives.
A broad brush approach claiming that those biases incorporate all observational studies is a mistake in evaluation of studies. Simply put, not all observational studies are equivalent, and not all should be dismissed merely because they were observational studies. And this came from one of those observations that was stunning.
The Observation That Started the Debate
In 2013, Wales introduced a shingles vaccination program. Because vaccine supplies were limited, an arbitrary birth-date cutoff determined eligibility.
If you had already turned 80, you were too old to qualify and would never receive the vaccine through the program.
If you were 79, you were eligible.
Think about what that means. Two people born only days apart lived in the same country, used the same National Health Service, saw the same physicians, had access to the same hospitals, and received the same home and community care. They were virtually identical in every important respect.
One difference separated them. One group became eligible for the shingles vaccine. The other did not. Then you follow the patients, in this case, they were followed for seven years.
Epidemiologists call this a natural experiment. Instead of patients choosing whether to be vaccinated—which always raises the question of whether healthier people are simply more likely to get vaccinated—the government’s eligibility rule effectively assigned people to one group or the other.
Those who became eligible for the shingles vaccine developed dementia about 20% less often than those who were never eligible. That finding didn’t prove the vaccine prevents dementia.
But it was far too large—and the study design far too elegant—to ignore. Naturally, it also attracted critics, like Dr. Vinay Prasad.
Dr. Prasad’s Critique
Dr. Prasad looks at these studies and isn’t convinced.
His first criticism is that these aren’t randomized clinical trials. They’re observational studies. People who get vaccinated may be different from people who don’t in ways we can’t completely measure. Maybe they’re more health-conscious. Maybe they exercise more. Maybe they see their doctor more often.
In a way, they are randomized - but randomized by age.
The Randomized Trial Isn’t the Only Tool We Have
I suspect that part of the difference between Dr. Prasad and me comes from our professional backgrounds.
Dr. Prasad is an oncologist. Oncology lives and dies by randomized clinical trials. If you’re deciding whether a new chemotherapy drug extends survival by three months, a randomized trial is the gold standard. It should be. Patients deserve that level of evidence before we expose them to drugs with significant toxicity.
Randomized trials are a remarkable tool. Not every important question in medicine can be answered with a randomized trial.
As a surgeon, I couldn’t randomize patients to receive an operation. We have choices with operations, but doing a sham operation, a controlled trial, is just something we cannot do. We could randomize who gets which surgery for a condition if we have two operations. But that is also difficult to do. If a patient needs weight loss surgery, the patient wants input into whether they get a gastric bypass or a sleeve, for example. But we can follow those groups, those who had a bypass and those who had a sleeve, and see how well they did over several years after surgery. We could also put them in the same classes for nutritional counseling and exercise, and follow all those variables. And indeed, we do that. But those are not randomized trials, they are not blinded trials, but they are important trials and observational.
In culinary medicine, I can’t lock thousands of people into identical kitchens for twenty years to determine whether olive oil, legumes, or fish reduce dementia. Nutrition researchers are constantly criticized for relying on observational evidence, but much of what we know about healthy dietary patterns comes from the convergence of epidemiology, physiology, metabolic studies, and long-term follow-up—not from perfect randomized trials.
Medicine uses many tools. Randomized trials. Observational studies. Natural experiments. Basic science. Pathology. Mechanistic studies. Genetics.
The art of medicine is knowing when the pieces begin to fit together.
That’s why I don’t dismiss the shingles studies simply because they aren’t randomized trials. Instead, I ask a different question.
Taken together, do the epidemiology, the biology, and the clinical observations point in the same direction?
Increasingly, they do.
Dementia Diagnosis and Time
Prasad’s second point is that dementia isn’t like measuring blood pressure or cholesterol. It develops gradually, the diagnosis can be delayed, and not everyone is evaluated in exactly the same way. If one group has more contact with the healthcare system than another, you could end up diagnosing more dementia simply because you’re looking harder.
In this case, these are older patients at high risk of dementia. One in five people at age 85 develop dementia. If an intervention delays that, and we follow a large group of these patients (age 79 and 80) over seven years, we will have enough data to give us great deal of certainty. This is what we found.
His conclusion is that the excitement over shingles vaccination reducing dementia has gotten ahead of the evidence.
To me, asking these questions is where the scientific discussion begins—not where it ends.
Why I Think the Evidence Is Stronger
If this were just another retrospective review from an insurance database, I’d probably be far more skeptical. But it isn’t.
The Welsh study took advantage of a natural experiment created by the vaccine rollout itself. Another group, the Danish registry study, comes from a healthcare system that follows essentially the entire population. Everyone has access to the same healthcare. Vaccinations are recorded. Hospitalizations are recorded. Dementia diagnoses are recorded. Researchers aren’t piecing together fragments from multiple insurance companies—they’re looking at an entire nation’s experience.
The question isn’t whether a study is perfect. The question is whether it’s good enough to make us pay attention.
When multiple studies, from different countries, using different methods, continue to point in the same direction, scientists should become more interested, not less.
Then there’s the biology.
Varicella-zoster (the chickenpox and Shingles virus) isn’t just a virus that causes a painful rash. It infects nerve tissue, remains dormant for decades, and can reactivate later in life. Shingles does more than cause the most painful rash you ever felt. Shingles virus can cause encephalitis, vasculopathy, stroke, cranial nerve palsies, and long-lasting nerve pain. Basically, it can turn your brain into mush, so you not only have a painful rash whose pain may never go away, but also major brain damage.
So when researchers find that vaccination also reduces the risk of dementia, my first thought is, “That actually makes biological sense. Now let’s see if the evidence continues to support it.”
One comment from Dr. Prasad caught my attention. He jokes that the reported effect works “faster than Tums,” implying that the benefit appears implausibly quickly.
I think that overlooks an important feature of the population being studied. As I mentioned previously, these were adults in their late seventies and early eighties—an age when dementia rises sharply. You don’t need to wait decades to see this effect.
In statistics, one of the biggest challenges is avoiding a Type II error—failing to detect a real difference simply because the study is too small or too few outcomes occur. That’s called being underpowered. These studies weren’t underpowered.
They followed very large populations in an age group where dementia occurs frequently enough that researchers could meaningfully compare vaccinated and unvaccinated groups.
Why I Find This Plausible
Perhaps the reason I see these studies differently is because of the lens through which I’ve spent my career looking at disease.
Chickenpox is a herpes virus, and even that cold sore is more than just a cold sore. I’ve never thought of herpes viruses as “just a rash.” As a physician, I’ve seen what herpes simplex virus can do when it reaches the brain. HSV encephalitis is one of the most devastating infections we treat. Patients can suffer permanent memory loss, personality changes, seizures, and profound cognitive impairment. The virus has an affinity for nervous tissue, and when it causes inflammation there, the consequences can be life-changing.
Varicella-zoster virus belongs to that herpes virus family. It doesn’t simply cause a painful rash. It establishes lifelong latency in nerve tissue. It reactivates years later. It causes post-herpetic neuralgia, vasculopathy, encephalitis, myelitis, cranial nerve palsies, retinal disease, and stroke. Throughout my career, I’ve watched shingles devastate older adults, particularly those in long-term care. It can turn an independent person into someone living with chronic pain for months—or years.
That’s why I received Shingrix (the Shingles vaccine). Not because I thought it might prevent dementia. I got it because I never wanted shingles. If you’ve cared for patients with post-herpetic neuralgia, you don’t need much convincing.
Now we’re seeing evidence that preventing shingles may also delay or reduce the risk of dementia. Do we know that for certain? No. Do I think the evidence is becoming increasingly persuasive? Yes.
And if protecting myself from one devastating neurologic disease also lowers my chances of another, that’s not the reason I rolled up my sleeve—but it’s certainly a welcome bonus.
Where Dr. Prasad and I Really Disagree
I don’t think Dr. Prasad and I actually disagree much about the data. We disagree about how physicians should think about evidence.
Should one observational study change medical practice? We both would agree that it should not. Observational data is what gave us the idea that smoking led to heart disease and cancer. First observed in the 1600’s by physicians, repeated by the AMA in the early 1900’s and then by the surgeon general in 1964. But that was many observations, as well as animal studies, but no randomized trial.
Should several elegant observational studies, performed in different countries, using different methods, all pointing in the same direction, combined with a biologically plausible mechanism, make us pay attention?
Absolutely.
Medicine isn’t practiced with a single tool. Randomized trials are extraordinary, but they aren’t the only way we learn. Surgeons know this. Nutrition scientists know this. Epidemiologists know this. Every specialty has learned that different questions require different tools.
To me, these studies have crossed an important threshold. They haven’t proven that shingles vaccination prevents dementia. They have convinced me that the question deserves to be taken seriously.
And that’s enough for me to be pleased that I received my shingles vaccine.
But this debate raises a much larger question—one that goes far beyond shingles.
How should doctors decide when evidence is “good enough?”
That’s what I’d like to explore next for our paid members. Thank you for reading this- I hoped this helped you see that in medicine we have are less certain than a huckster.
