Modernizing Variolation

I have a four-year-old daughter who started her first year of preschool in early September, 2025. Prior to preschool enrollment she had been taken care of by parents and daytime nannies. During this time she was ill only once, testing positive for COVID at about 2.5 years old. The first 6 weeks or so of preschool, spanning September and the first couple of weeks of October, were relatively uneventful. The weather was warm and the children spent a substantial amount of time outdoors.

But I had been warned by many other parents that the first year in preschool can bring seemingly unending plagues of illness and grief. And so it began….

In mid October the morning weather turned colder and the children began to spend more time inside. This is the earliest part of flu season in the Northeast. In preparation for traversing this gauntlet of pestilence, I was vaccinated to the hilt, receiving every vaccine available to me, including the current influenza injectable vaccine and the relatively new RSV vaccine. Unfortunately, rhinoviruses, adenoviruses, and parainfluenza viruses cause a majority of respiratory infections, but there are no vaccines against any of them due to rapid mutation and the vast number of variants.

While I don’t think we ever got the flu, the next several months in our home were dominated by a series of respiratory illnesses (at least 6) and one bout of norovirus.¹ Never in my life have I been so sick, so often. The six-week-long case of RSV that I had from mid January through February (2026) was probably the most serious illness I’ve ever had²—despite having received the RSV vaccine just a few months before, in early October (this was the only illness I suffered among those I was vaccinated against). In late January mucus completely clogged my nose and obstructed my lungs, and stayed that way for about two months. The hearing in my right ear largely disappeared with a severe blockage of my right eustachian tube. Even now in mid-May, it remains partly blocked, so my hearing isn’t fully restored.

During these several months of almost non-stop illness and dysfunction I began to expand on the germ of an idea I’d been incubating for almost a year, and I became increasingly serious about using my daughter’s illness as a source of material for immunization. This wasn’t a ridiculous or wild idea; I was inspired by a long history of people’s successful efforts to counter the debilitating and deadly effects of smallpox and other pathogens through keen observation and clever mitigation techniques.

The variolation innovation

Widespread use of vaccines is relatively new, but modern vaccines were preceded by centuries of progressively less dangerous precursors. In fact, one approach—a technique called variolation—was not only among the first recorded methods to prevent or reduce severity of an infectious disease (smallpox), it was the forerunner of modern vaccination.

The variolation procedure involved exposing a person to material taken from a smallpox patient (or recently variolated person), with the hope that the resulting infection would be mild yet provide robust immunity to the virus. Often, this was done through a process called inoculation—by rubbing smallpox scabs or fluid from pustules into scratches or incisions in the skin of the recipient. While the smallpox virus was normally spread through the air and caused a severe infection starting in the lungs, infection by inoculation through the skin usually led to a milder case that still induced a protective immune response. This is because pathogens have evolved to attack through specific routes of infection. They often enter cells through specific molecular receptors on a cell’s surface. Most such receptors are absent from many other cell types. As a result, viruses have limited repertoires of attack via atypical routes. When they are placed outside of their normal “tropism” (specific cell type and tissue environments), they are often attenuated and more easily battled by the immune system.

Some of the first reported uses of variolation in China (between the 15th and 17th centuries) describe people taking smallpox scabs, grinding them up, and insufflating (or snorting) them into the nose. Since the respiratory tract is the primary smallpox infection site, one might be concerned that this approach might be likely to result in full-blown disease. This might have been one driver of reasonable attempts to give a smallpox dose via other routes, e.g. oral or dermabrasion.

There must have been considerable experimental testing of a range of variations before mature techniques became more widely used, and by the 17th century, mild smallpox cases were selected as donors. It was also determined that fresh scabs were more likely to lead to a full-blown infection, so a preferred method was to use scabs that had been left to dry for days or weeks. Parts of India, Africa, and Europe developed their own variolation approaches, all of which appear to have reduced mortality rates considerably relative to natural smallpox infection.

In 1777 George Washington mandated inoculation of the U.S. Continental Army prior to their confrontation with British troops during the Revolutionary War. British troops were often already immune to smallpox, having been exposed in crowded cities, whereas American soldiers from rural areas were largely non-immune, leading to high mortality and low morale. Washington’s mandatory inoculation order was a highly guarded secret to prevent the British from discovering the vulnerability of his army during recovery. Washington’s decision was extremely rational and highly successful, reducing the mortality of smallpox in the army from the prevailing 16% down to about 0.33%.³

The advantages of modernized variolation

Over the last 150+ years traditional variolation practices have been supplanted completely by the more convenient and safer practice of vaccination. As a result, there has been little interest in improving variolation methods. Nevertheless, the basic premise of variolation remains sound and it provides significant advantages over typical vaccination practices, including potentially low cost and immediate access at or near the site of an outbreak. There is little doubt that vaccines remain a better solution than any form of variolation. Modern vaccination practices involve controlled dosage, high immunogenicity, well-studied adjuvants, preventive dosing far in advance of a pathogen encounter, and other attractive features.

However, immediate access is a critical advantage for any pathogen countermeasure, especially during a serious outbreak. With the rise of the FDA and other regulatory bodies throughout the world—and the resulting complex and slow processes of clinical trialing and regulatory review—that critical speed advantage has been lost. Modern vaccines are wonderful, lifesaving inventions—as are properly functioning drug regulatory agencies—but a protracted regulatory process ensures that vaccines are not available when and where they are needed most.⁴ Pandemics might be caused by pathogens with the potential to severely impact or shut down normal supply chains and basic infrastructure. In cases where the pathogen is an engineered bioweapon with high virulence, countermeasures would ideally be created and deployed as quickly as possible—meaning days, or a few weeks at most. Delays will be extremely costly in terms of human suffering and death.

Another major advantage of MV is that it should provide protection against specific variant pathogens currently in circulation. Most vaccines for common infectious diseases—e.g., all of those for influenza and SARS-CoV-2—are highly specific for previous year’s variants, but by the time the vaccines are given, those variants often have all but disappeared. Even worse, by the time many people are infected, new variants have arisen that are at least partly resistant to the vaccine. Every year such delays cause countless millions of cases of serious illness and thousands of deaths, and are a main reason why influenza vaccines in the US are typically less than 50% effective. In the past 15 years, flu vaccines have only been more than 50% effective twice—the same number of times they have been less than 30% effective.

Another advantage of MV over typical vaccination practices is that it is much more similar to natural infection—minus the actual infection and illness. There are many unreasonable objections to vaccines and vaccination practices, but some are very reasonable and based on many decades of scientific evidence. Therefore, those who are hesitant or even unwilling to take certain vaccines likely will regard MV as a superior option. Many years of notable missteps by public health leadership in the US have eroded confidence in vaccines. Confidence is also diminishing because of the many highly publicized flaws of vaccines. Despite the sub-optimal state of current vaccines, important vaccine research and development efforts continue to be underfunded. Furthermore, the politicization of vaccine market dynamics combined with extreme inefficiencies of vaccine regulatory approvals have reduced vaccine choice and diversity in many countries, including the US.

As one noteworthy example, today, more than six years after the start of the pandemic, there still isn’t a single FDA approved mucosal vaccine against SARS-CoV-2, and none appears to be poised for approval. Furthermore, of the hundreds of SARS-CoV-2 vaccines developed and tested during the pandemic, only four were given emergency use authorization (EUA), including two highly similar mRNA vaccines: one from Moderna and another from BioNTech and Pfizer. EUA for the Johnson & Johnson vaccine was revoked by the FDA after about two years. This leaves people in the US with a choice between one subunit vaccine (Nuvaxovid) and two mRNA vaccines, none of which triggers a mucosal immune response strong enough to prevent infection or spread.⁵ The same weakness exists for vaccines against the most serious and common respiratory viruses currently in circulation, including influenza, Respiratory Syncytial Virus (RSV), Rhinovirus/Enterovirus (significant cause of “common cold”), Human Metapneumovirus, and others. MV allows for administration in multiple ways that at least have the potential to trigger mucosal immunity.

There are other advantages to variolation, but these were enough to convince me that I had to try to modernize variolation, to create a timely and convenient pathogen countermeasure that essentially anyone could use.

Modernizing variolation

There are a few major advantages that a fully modernized form of variolation will have over traditional and dangerous variolation practices of the past. Some of these come directly from modern vaccine research, including scientifically validated approaches to pathogen inactivation, and a wide selection of adjuvants (compounds that are added to vaccines to amplify the immune response) to ensure sufficient immune stimulation. I started with the assumption that a pathogen either needs to be inactivated or administered in a tissue that is outside of its normal range or tropism (similar to the administration of smallpox via a dermal rather than respiratory route), or both. I researched the biomedical literature on topics ranging from variolation, to viral titers in nasal mucus, to modern methods for viral inactivation and for the creation of inactivated whole virus vaccines.

The production of the majority of inactivated whole virus vaccines use formaldehyde (the liquid version is formalin), a crosslinking agent. However, inactivation is slow, antigenic structure (structure of the virus that elicits an antibody response) is only partially preserved, and T-cell response is dramatically reduced. Plus, my main interest is the creation of protocols that others will be able to implement quickly and easily. Since formaldehyde is toxic, concentrations sufficient to inactivate a virus rapidly would have to be removed prior to self-administration. Therefore, I turned to other methods.

My first (failed) attempt: UV inactivation

The first inactivation method I decided to test is ultraviolet light (UV). UV light with a peak wavelength at about 260 nm (within the UV-C range of 100 nm to 280 nm) has been firmly established to damage both RNA and DNA, which is the genetic material of all known pathogens. Relative to formaldehyde, inactivation is fast, antigenic structure is not quite as well preserved, but T-cell response remains largely intact when compared to native virus. I purchased a few different types of low-pressure mercury vapor lamps, which emit at a peak of 253.7 nM. As chief scientist of Radvac, I already had on hand the other equipment I would need, including nitrile gloves, masks, and mucus sample collection tubes. I constructed some prototype.irradiation devices containing sample reservoirs in close proximity to the UV source. Irradiation times for inactivating most pathogens are well established, so I used these guidelines to irradiate samples for about 10 minutes using a 3W bulb at a distance of about 1 inch—which is overkill for most viruses.⁶ Nasal mucus from an infected person is often opaque and UV 253.7 nm transmission is greatly reduced. Therefore, the sample needs to be diluted and clarified prior to irradiation. My first two trials of the UV inactivation technology were challenging, including sample handling, sample dilution and clarification, and uniform irradiation.

When my daughter’s nose began to run and her voice became a bit hoarse, I collected nasal mucus samples from her twice over the course of the two days (corresponding to symptomatic days 3 and 4). I attempted collection on day 2 but the method I tried failed (I placed a 5 ml tube with a 14 mm opening under her nostrils). Successful collection on days 3 and 4 resulted from having her blow her nose into my gloved hand (I wore a mask and nitrile gloves on both hands).

I used the UV method twice over the course of days 3 and 4 to irradiate samples prior to nasal self-administration. The evening of day 5 (about 30 hours after the second nasal administration) I began to detect early symptoms and subsequently developed an approximately 9 day long respiratory illness. It is possible that I lost the natural infection vs. MV race, or that the complexity of the UV inactivation procedure exposed me to the pathogen, or that the UV procedure wasn’t fully effective and I self-administered active pathogen. Reflecting on the fact that I am a trained scientist with decades of experience working at a lab bench, and on the equipment requirements and procedural challenges of the UV inactivation approach, I decided that if MV is to be useful to non-experts, I had to seek simpler means for inactivation.

Inactivation with hydrogen peroxide

I next tested hydrogen peroxide. Although there isn’t an approved, whole virus vaccine inactivated by hydrogen peroxide, there are many publications showing that treatment of many viruses with 3% hydrogen peroxide (3% and up to 6% are sold for home use) inactivates them completely in 1 to 2 hours. These inactivated viruses elicit strong immune responses when paired with adjuvants. This research shows that hydrogen peroxide inactivation is not only sufficiently rapid, antigenic structure is reasonably preserved and is superior to UV-C (although possibly not as good as formaldehyde), and T-cell response remains largely intact. A relatively new hydrogen peroxide-based inactivation protocol shows very high antigenic retention, or even amplification, but it is much more complex; nevertheless, it suggests that other modifications to the simple base protocol might yield improvements.⁷ My initial tests with hydrogen peroxide were far easier than those with UV, and no specialized equipment was required.

About a month after my somewhat disappointing experience with UV inactivation, my brother came to visit and stayed in our basement apartment. Soon after he arrived, my daughter became symptomatic with a respiratory illness, and within two or three days all four members of our home had it. I had not fully worked out a simple and dependable protocol, so I suffered through yet another illness without taking action.

But I was sick and tired of being sick and tired, and I was very frustrated that I had no reasonable countermeasures. I increased the intensity of my research and I worked out a simple hydrogen peroxide inactivation protocol. Based on initial research and experimentation, the next time a household member showed symptoms of a respiratory illness, I planned to use my still-evolving protocol to self-administer multiple doses of hydrogen peroxide inactivated pathogen.

Unfortunately, the next illness came fast on the heels of the previous one. At the end of April, just over 2 weeks after recovering from the previous illness, my daughter’s nose began to run and my brother became ill. It is difficult to tell which one was the index case in our house because my daughter was suffering from seasonal allergies at the time. Prior to my brother’s symptoms, her nose was running regularly, she was sneezing a lot and coughing occasionally, but her symptoms never got much worse. However, her runny nose did taper off after a couple of weeks, so I think she was the most likely source of the infection. Given this uncertainty, the timeline presented here corresponds to my brother’s illness, starting with Day 1, when he first reported his initial symptoms.

Timeline

Day 1. My brother began to sniffle and clear his throat.

Day 2. My brother’s symptoms worsened. My daughter’s nose continued to run and she sneezed frequently. Her symptoms did not appear to worsen over the coming days.

Day 3. In the evening, my wife started showing symptoms, but they were very mild. As soon as my brother and my wife started showing symptoms, I kept my distance from them. It was still unclear whether or not our daughter was ill, but I did not have the option of keeping my distance from her. Still, when I had to lift, carry, or hug her, I held my breath (she has a habit of sneezing or coughing in my face). When I drove her in the car, I lowered the windows a few inches. I took these minimal measures to give me some infection-free time to work out the details of the MV approach.

After my UV inactivation attempt ended in failure and illness I was apprehensive about trying MV again, but the few days ahead appeared daunting. As my wife grew increasingly ill, the inside of my home would reach peak infectiousness. With this sword of Damocles hanging over me, I worked as quickly as possible to finalize the first draft of my new protocols.

Day 4. I remained free of respiratory illness symptoms the entire day.

Like my brother, my wife developed a cough, general malaise, and a runny nose that lasted about 10 days. My brother’s symptoms worsened and he developed a fever. I quarantined him to the basement apartment of our house. He remained in quarantine until Day 10. I wore a mask or held my breath when I needed to go into the basement. My wife’s symptoms worsened and she developed a nasty-sounding, hacking cough. I kept a distance of several feet from her while she was obviously symptomatic.

I decided that I had to take a chance on my rudimentary and untested hydrogen peroxide inactivation protocol, and my equally untested nasal self-administration protocol. I started by collecting a sample from my brother. He had been symptomatic for four days and was experiencing much more severe symptoms than my wife, so he was the best sample donor candidate. I put on gloves and a mask, grabbed a collection tube, and descended the basement stairs to pay him a visit. I uncapped and handed him the 5 mL sample collection tube. He had some trouble depositing mucus directly from his nostrils into the 14 mm opening of the tube, but after some effort he was able to deposit about 0.5 mL into the tube. I capped the tube and immediately took it for a decontamination rinse with 70% isopropanol.

I had the tube sitting on my desk as I reviewed notes on my computer to help guide me in preparing the sample for self-administration. I had a rough protocol and some ideas for variations typed on the page, but I was in a bit of a rush because it was getting late in the evening. So I sketched an outline of the protocol in my head and gathered all the materials I thought I would need. (Soon you will be able to find a full, updated, and refined protocol on the Radvac website, so I won’t provide every detail in this account.)

Due to the relative simplicity of the protocol, even the first time through went fairly smoothly—despite the fact that I was partly making up or modifying the protocol as I went. I mixed the sample with about an equal volume of 3% hydrogen peroxide, mixed it well, and spread it out onto a small, clean ceramic plate. There were a few small lumps in the sample, so I used a fork from our kitchen to homogenize the sample by mashing and mixing it on the plate, until it was spread in a fairly uniform layer about 1 to 2 mm thick. I put the plate on a table in a secure location and left the sample to dry overnight.

Day 5. I remained free of respiratory illness symptoms the entire day. I kept my distance from others in the home as much as possible until I administered the MV dose.

In the morning I retrieved the plate. The sample had dried to a dull, thin film on the plate. I used a single-edged razor blade to scrape it off the plate and to chop the sample into a granular powder.

Prior to self-administration of the inactivated sample, I cleaned out my nose by repeatedly blowing it into a tissue and alternating that with a couple squirts of saline nasal spray up each nostril. I repeated that cycle a few times. One last time I blew my nose to evacuate all contents. My left nostril was very clear and open; my right nostril remained partly closed. I rolled up a clean piece of paper into a tube a bit over 3 inches (8 cm) long and about ¼ inch (6 mm) in diameter. I snorted half the sample into the clear nostril first, then snorted the remainder into the other nostril. Time of self-administration was about 10:30 am.

I felt a mild discomfort when the sample hit the inside of my nostrils. Therefore, I decided to follow administration with a chaser squirt of sterile saline nasal spray. I went about my day and paid attention to my nose and sinuses. There was no discomfort and no obvious change in nasal health. After a few hours it seemed like the procedure was safe so I delivered a collection tube to my brother and asked him to deposit a second sample. Once again he was able to collect about half a milliliter, and I repeated the entire sample preparation procedure. This time, I dried the sample for several hours and administered the sample at about 10 pm in the evening. Time between doses: about 11.5 hours.

Day 6. I remained free of respiratory illness symptoms the entire day.

I woke up with moderate nasal congestion, which is not uncommon for me during spring allergy season. I felt mostly fine and normal throughout the day, although the nasal congestion persisted and was probably a bit heavier than usual.

Day 7. Today, nasal congestion was very heavy. I felt mentally and physically sluggish, and mildly feverish, consistent with respiratory illness symptoms. Nevertheless, I didn’t feel too bad physically (70ish percent), and the weather was nice, so I spent part of the day outdoors. By the time I went to bed I still felt slightly unwell. I was increasingly certain that I had become infected.

Day 8. To my surprise, I woke up feeling better (I felt maybe 90ish percent), and felt increasingly normal as the day progressed. Nasal congestion remained moderately heavy.

Day 9. I woke up feeling normal. Nasal congestion levels returned to normal. No symptoms of respiratory illness. My brother says he is feeling better but my wife remains symptomatic, so I am not out of the woods.

Day 10. I woke up feeling normal. I remained free of respiratory illness symptoms the entire day. My brother is mostly recovered and has emerged from his quarantine. My wife remains symptomatic, so I remain cautious.

Day 11. I woke up feeling normal. I remained free of respiratory illness symptoms the entire day. My wife feels better. I assume she is not contagious. I think I have avoided infection.

Day 12. My wife is feeling well but has a lingering cough, which is normal for her. I am convinced that I have avoided infection.

My interpretation of the results

I administered the MV doses on the morning and evening of Day 5, and on Days 6 and 7 I thought I was becoming sick, but by Day 8 I began to feel better. It is impossible to say whether or not the MV self-administration protected me against the pathogen in our home. I was vigilant about social distancing, but I was in the same home as two (probably three) overtly sick people. I have tried social distancing within my home many times before and have failed most of the time.

One possibility is that MV administration on Day 5 triggered strong innate immunity⁸, and that was the source of the extreme congestion and ill feelings over the following days. Maybe that alone protected me during the peak infectious period. If this is the case, adaptive immunity might have been unnecessary for protection in this situation, and the MV approach is a complicated way to achieve innate immune protection. Far easier ways exist, including the application of triple antibiotic ointment (e.g. Neosporin and generics) into the nostrils.⁹

Nevertheless, it is also possible that I had encountered a variant of the pathogen previously and the MV procedure began to elicit adaptive memory by activating memory T cells and B cells. If the prior variant wasn’t much different from the current culprit, and the past encounter was relatively recent, this response can take as little as a single day. If this is what happened, in addition to mounting an innate immune response, my body began to mount a so-called secondary adaptive response that contributed to the ill feeling and malaise I felt on Day 7.

It is unsurprising that MV has the potential to trigger a potentially protective innate immune response, but this protection will be relatively short-lived. If a serious pathogen is in circulation and deployments of FDA approved vaccines won’t occur in the foreseeable future, then it is extremely important that MV research and development lead to a protocol capable of stimulating longer-lasting adaptive immunity. Now that I have a pilot MV protocol to follow, next time there is an illness in our home I will collect a sample explicitly for the purpose of identifying the pathogen. I can assess my immune profiles before and after infection. I will bank blood and saliva samples in our lab freezer, and test and compare those to post-MV samples. I also plan to include adjuvants (double or triple antibiotic ointment and carbomer) as a regular part of future MV dosing regimens.

1. We assume it was norovirus because my wife and daughter were very ill with a gastrointestinal bug, but I had no symptoms despite all three of us being in very close contact without masks, including frequent and extended car trips. According to my whole genome sequence data, I am fortunately completely resistant to norovirus infection.

2. Believably diagnosed through a hospital serological test.

3. See page 53, Bayne-Jones, Stanhope. The evolution of preventive medicine in the United States Army, 1607-1939. Office of the Surgeon General, Department of the Army, 1968.

4. This is a very complex issue but I am not criticizing the FDA. The public is largely to blame for any irrationality displayed by the FDA. I think it is safe to assume that the FDA cannot be made highly efficient, and that highly effective and extremely safe but novel vaccines will not given approval when (within days of an outbreak) they are needed most.

5. Although these injectable vaccines trigger strong systemic antibody responses, they elicit weak mucosal immune responses. Mucosal immunity is present at all external surfaces and orifices, and limits or prevents entry of pathogens into the body at sites of initial infection (skin, nasal and respiratory tracts, gastrointestinal tract, reproductive and urinary tracts). It is in part defined by and differentiated from systemic immunity by immunoglobulin A (IgA) antibodies and specific T-cells, whereas systemic immunity antibodies are immunoglobulin G (IgG) and a different repertoire of T-cells, although there is some overlap and coordination between these two systems.

6. Adenovirus has a double-stranded DNA genome and is well known to require a high dose of UV irradiation, establishing an upper end for irradiation. One reason for this is because the complementary strand enables efficient repair of the damaged strand by host DNA repair systems. Ideally, a pathogen is irradiated with sufficient energy to induce enough damage that DNA repair systems cannot repair the damage.

7. This approach uses much lower concentrations of hydrogen peroxide in combination with copper and and the DNA-binding copper chelator methisazone, better preserving antigenic structures, resulting in higher antibody titers. Amanna, I.J., Thomas, A., Engelmann, F., Hammarlund, E., Raué, H.P., Bailey, A.L., Poore, E.A., Quintel, B.K., Lewis, A.D., Axthelm, M.K. and Johnson, A.L., 2024. Development of a hydrogen peroxide-inactivated vaccine that protects against viscerotropic yellow fever in a non-human primate model. Cell Reports Medicine, 5(7).

8. Innate immunity is the body’s rapid, non-specific first line of defense, which uses inflammation to increase blood flow and deliver immune cells to the affected area, and fever to create an inhospitable environment for pathogens and boost the overall immune response. Adaptive immunity is a targeted defense system that relies on antibodies to bind and neutralize specific foreign invaders, and T cells to directly destroy infected cells and coordinate the overall immune attack.

9. Mao, Tianyang, et al. “Intranasal neomycin evokes broad-spectrum antiviral immunity in the upper respiratory tract.” Proceedings of the National Academy of Sciences 121.18 (2024): e2319566121. https://www.pnas.org/doi/pdf/10.1073/pnas.2319566121