There exists a relationship between dose and infection, but there is no association between dose and disease severity for COVID-19. Non-pharmaceutical interventions, such as masks, may limit the inoculum dose from an exposure, thereby reducing the risk of infection, but they have little, if any, impact on COVID-19 severity.
A multi-disciplinary team of scientists with expertise in infectious disease epidemiology, infection prevention and control, animal research, occupational health, and virology examined the evidence for links between severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) dose, infection, and coronavirus disease 2019 (COVID-19) outcomes. This research was published in Clinical Infectious Diseases (CID) on October 15, 2021 (https://doi.org/10.1093/cid/ciab903). Motivated by reckless hypotheses and unsubstantiated theories that non-pharmaceutical interventions (NPIs) such as wearing face masks would lower infectious dose and protect the wearer from severe disease, we reviewed the animal, clinical, and epidemiologic literature for evidence of an association between dose, infection, and disease severity outcomes. Here, we present the most relevant findings of our CID review and provide the audience with an overview about the topic.
Viral dose and viral load
Viral dose or inoculum refers to the amount of virus received by a susceptible individual resulting in infection. Viral load, although often used interchangeably with viral dose, has a different meaning. Viral load refers to the amount of viral nucleic acid in a clinical specimen and does not necessarily reflect infectious virus nor does it relate to infectious dose.
Disease severity, applied to COVID-19, refers to the spectrum of clinical progression after infection. COVID-19 severity is often recommended to be classified into asymptomatic, mild, moderate, severe, and critical. However, distinct disease outcomes are often used given the availability of data. These proxies include hospitalisation, intensive care unit admission, mechanical ventilation, and mortality.
Our review of the clinical literature revealed that host factors, including older age (especially over 65 years), male sex, cardiometabolic comorbidities such as diabetes, hypertension, cardiovascular disease, chronic obstructive pulmonary disease, and others, immunocompromised status, smoking, and pregnancy are associated with severe COVID-19 outcomes. Being younger and without comorbidities does not guarantee mild disease outcomes, however. Host genetics and susceptibility are thought to play an important role by correlating with effective and early immune responses. Furthermore, long-term effects following SARS-CoV-2 infection are not negligible; the so-called post-acute sequelae of COVID-19 remain elusive and requires further studies.
Animal and human data suggest that some SARS-CoV-2 variants may be not only more transmissible but also possibly more pathogenic. The Delta variant, recently responsible for infection surges across the world, has been associated with higher risk of hospitalisation than other variants and appears to be more pathogenic in animal studies.
Studies in young non-human primates (NHPs), usually performed at high doses, result in respiratory outcomes that rapidly resolve and rarely lead to severe morbidity or mortality. A small number of studies in NHPs, hamsters, and some transgenic mice models have replicated severe disease due to age, sex, and comorbidities. Studies in transgenic mice may not be relevant to human disease, however.
One hamster study suggests the median infectious dose is very low (5 TCID50), which complicates dose response studies in laboratory animals and makes it unlikely that dose-response will be detectable in humans in real-world settings. To date, there are no reliable dose-response studies of relevant outcomes in NHPs.
Transmission by contact, fomites, and aerosols has been demonstrated in NHPs, rodents, and ferrets, but quantitative dose-response has not been studied for any of these routes. Although transmission has been demonstrated from infected to naïve animals, it is challenging to quantify dose-response in such experiments.
That infection occurs more frequently indoors, particularly in crowded and poorly ventilated spaces, is suggestive of dose-response for SARS-CoV-2, where infection is less likely at lower airborne particle concentrations. Lacking quantitative methods for sampling culture-based viral exposures, epidemiologic studies rely on proxies such as proximity, route, duration, and number of contacts, among others, to infer exposure. Most of these are difficult to measure and rely on either periodic observation or personal recall.
We explored epidemiologic studies in healthcare workers, where exposures are frequent and confounding factors are more likely to be controlled. However, the range of infection incidence in healthcare settings is very wide most likely due to differences in locations, exposures, community infection rates, and control measures. The most consistent risk factors for infection were being a nurse or working in a hospital non-emergency setting, lack of personal protection equipment, inadequate handwashing, caring for patients with COVID-19, and being present during intubation. All of these are suggestive of a link between exposure, dose, and infection.
On the other hand, rates of hospitalisation, severe COVID-19, and mortality are relatively low in healthcare workers. In the US, healthcare workers generally have less severe disease and are less likely to die than all patients with COVID-19, even if they experience unprotected or repeated exposures. This may be due to the relatively young age or baseline health status of most healthcare professionals. One study found higher rates of COVID-19 in older healthcare workers, which is consistent with findings in the general population, suggesting that disease severity is associated with host factors rather than dose.
In summary, we found some evidence of a relationship between dose and infection from animal experimental and human epidemiologic studies, but minimal data to support a relationship between SARS-CoV-2 dose and COVID-19 severity. Rather, host responses associated with age, sex, and cardiometabolic comorbidities are the primary determinants of disease outcomes. Highest risk exposures are those that involve closer contact with infectious people or longer time spent in environments with high respiratory particle concentrations from infectious individuals. Both aerosol-producing behaviours (e.g., coughing, sneezing, speaking, singing, shouting, breathing) and aerosol-generating medical procedures contribute to these concentrations and exposures.
Disease severity depends primarily on host biological and physiological responses and defence mechanisms but may also be impacted by environmental factors such as access to medical care and healthcare system capacity. Viral dose seems to contribute little, if any, to how disease develops once infection gets started following the interaction of the host with a dose above the infectiousness threshold.
NPIs such as mask wearing, physical distancing, and lockdowns may be effective at limiting dose thereby reducing the likelihood of infection, but will not have an impact on disease severity, which can range from asymptomatic to critical or even fatal. In contrast to vaccines, which do reduce severity of infection, no NPI guarantees an individual will experience asymptomatic or mild disease. Thus, recommending masks or other NPIs relies on the primary established benefit of reducing exposure rather than as a means of preventing severe outcomes. Messages that fail to differentiate these concepts may be dangerous.
We acknowledge the original work on the topic, entitled “SARS-CoV-2 Dose, Infection, and Disease Outcomes for COVID-19 – A Review,” which was published in Clinical Infectious Diseases (https://doi.org/10.1093/cid/ciab903) by L.M. Brosseau, K. Escandón, A.K. Ulrich, A.L. Rasmussen, C.J. Roy, G.J. Bix, S.V. Popescu, K. Moore, and M.T. Osterholm. Relevant references for the statements in this summary can be found in the original paper.