How Vaccines Might Improve Long Covid

A working hypothesis from an expert immunologist

Akiko Iwasaki


A female clinical pharmacist with Seattle Indian Health Board (SIHB) administers a shot of the Moderna COVID-19 vaccine, to frontline workers at the SIHB on December 21, 2020 in Seattle, Washington.
Photo: Karen Ducey/Getty

It’s clear that vaccines have helped some people with long Covid with their symptoms. While the numbers are still small, these are encouraging signs. What follows is my hypothesis as to how vaccines might improve long Covid.

Back when I first learned about long Covid in June 2020, I proposed three possible mechanisms that might be causing it: 1) a persistent viral reservoir; 2) “viral ghost,” which are fragments of the virus (RNA, proteins) that linger after the infection has been cleared but are still capable of stimulating the immune system; and 3) an autoimmune response induced by the infection. Of course, other mechanisms may also contribute.

Since then, many studies have provided support for all three of these mechanisms. Research has shown that viral reservoirs are present in tissues, viral RNA is found in non-respiratory tissues and is associated with inflammation, and diverse autoantibodies are detected in some Covid patients.

The three mechanisms of long Covid I proposed above are not mutually exclusive, and all three may benefit from the vaccines. If the first is true, vaccine-induced T cells (immune cells that attack and kill infected cells) and antibody responses may be able to eliminate the viral reservoir. If the second is true, vaccine-induced immunity may be able to eliminate the viral ghost if such viral components are associated with the spike protein, which the virus uses to gain entry into cells. If the third is true, the vaccine might divert autoimmune cells, as I will describe below.

I suspect that people with long Covid have varying degrees of all three mechanisms taking place. Thus, long Covid consists of multiple types of diseases. By understanding which mechanism(s) are causing long Covid within each person, suitable treatment can be given.

To determine which mechanism(s) are responsible for vaccine-mediated improvement in long Covid, we can design a trial in which long haulers are given one of four vaccines.

Another possible way in which vaccines can help long Covid symptoms is through stimulation of innate immune responses, caused by the adjuvant component of the vaccines. Vaccines induce T and B cell responses. This requires two components, antigen (the stuff detected by T and B cell receptors) and adjuvant (the stuff that triggers innate immune responses).

Adjuvants stimulate the immune system by mimicking features of pathogens. Many vaccines are simply antigen combined with added adjuvant (like alum) — mRNA vaccines do not require extra adjuvant because the RNA itself serves as the adjuvant that stimulates the innate immune system.

Adjuvants can only induce transient inflammation. It could be that short-term inflammation caused by the adjuvant might somehow divert the leukocytes (immune cells) causing long Covid. If this is the case, the beneficial impact of vaccines may not be long lasting, because innate immune signals are temporary.

It is also possible that adaptive immune responses induced by vaccines divert the leukocytes causing long Covid. For example, antibody or T cell responses directed against a new antigen, regardless of what it is, might skew the immune system to shift its attention to the new antigen — though it is not clear for how long.

To determine which mechanism(s) are responsible for vaccine-mediated improvement in long Covid, we can design a trial in which long haulers are given one of four vaccines:

1. mRNA SARS-CoV-2 spike vaccine (the Covid-19 vaccine)

2. mRNA irrelevant antigen vaccine (e.g. Zika virus vaccine)

3. An empty mRNA vaccine (not coding for protein)

4. Placebo (saline)

The outcome of the trial can tell us both what may be driving long Covid as well as which therapies are likely to work best. Such a trial may be difficult for various reasons but is worth considering, as is the possibility of an animal model to try first. If only the mRNA coding for the spike protein (the Covid-19 vaccine) but not for an irrelevant protein (Zika virus) or empty mRNA helps people recover from long Covid, this implies the viral reservoir and viral remnant scenario may be the basis for long Covid.

If an empty mRNA vaccine (not encoding for a protein) improves long Covid, even transiently, such data imply that innate immune stimulation is able to reprogram the leukocytes causing the symptoms. If so, we can think of stimulating the innate immune system using empty mRNA or viral RNA mimic, or giving interferon-beta therapy periodically to improve symptoms. Interferon-beta therapy is also used to treat multiple sclerosis patients.

Lastly, my hope is that the data from such a trial might also inform treatment of chronic fatigue syndrome (ME/CFS), because people suffering from ME/CFS might have similar underlying causes of disease and therefore might benefit from similar treatments.

This piece originated from my February 28 Twitter thread.



Akiko Iwasaki

Immunologist working on viral infectious diseases and vaccines. Professor at Yale. Investigator of Howard Hughes Medical Institute. Love teaching and learning.