Does being exposed to a small quantity of SARS-CoV-2 virus result in less severe disease than being exposed to a large quantity?

The answer is relevant to how we should respond to COVID-19.

If high dose exposures are worse than low-dose exposures, then:

  • We should consider people who spend a lot of time with infected people, like healthcare workers and family members of COVID-19 patients, to be more at risk than people who get briefly exposed to the virus
    • Reopening lower-risk public spaces (like outdoor cafes or parks) may be low-risk compared to reopening spaces that involve a lot of close contact (like gyms and nightclubs)
    • We should prioritize PPE even more for people who regularly interact with COVID-19 patients
  • Low dose exposure to SARS-CoV-2 may produce immunity without producing serious illness.
    • If low dose exposure is safe and produces immunity, it may be good for people and available faster than a vaccine can be manufactured and approved.

Bottom Lines

People who have higher respiratory viral loads are significantly more likely to have severe COVID-19; the same pattern held in both the SARS and MERS coronavirus epidemics. More virus in the body does tend to correspond to more severe disease.

There’s far less data about how different forms of exposure correlate to disease severity, but there are a few studies pointing towards a greater chance of severe COVID-19 from household contacts than from travel, and one study indicates that mask usage increases the probability that a SARS case will be asymptomatic. These constitute weak evidence that larger doses of exposure to human coronaviruses cause more severe disease.

When human volunteers are experimentally exposed to viruses, some viruses cause more severe symptoms at higher doses, while some viruses don’t. None of the relevant studies were on coronaviruses, however.

In animal experimental exposures to virus, higher doses consistently cause more severe symptoms, including in experiments on the coronaviruses SARS, MERS, and PEDV.

The available evidence is highly incomplete, but tends towards the conclusion that lower dose exposure to COVID-19 should result in less severe disease.

Viral Load in COVID-19

Here’s a table of studies that compared viral load in mild and severe COVID-19 patients.

N Mean Viral Load, Mild Mean Viral Load, Severe Significance Location of Sample Context
15 Ct = 25 (n = 6, range 17-32) Ct = 27 (n = 9, range 19-33) n.s. Nasopharyngeal swab Private hospital, Mumbai, India[1]
92 Ct = 28 (n = 62, sigma = 0.5) Ct = 25 (n = 30, sigma = 0.5) p = 0.017 Sputum Zhejiang, China; severe patients older & with more comorbidities but not different exposure histories. [2]
96 4.1 log copies/mL (n = 22, sigma = 1.4 ) 5.1 log copies/mL (n = 74, sigma = 1.4) p = 0.03 Sputum (no difference between mild & severe pts in serum or stool) Zhejiang, China; severe patients more likely to be from Wuhan[3]
76 delta-CT = -1.25 (n = 30, sigma = 5.2) delta-CT = 4.48(n = 46, sigma = 3.0) p < 0.001 nasopharyngeal Nanchang, China; median time from disease onset = 4 days[4]
23 5 log copies/mL (n = 10, sigma = 2.2) 6 log copies/mL (n = 13, sigma = 3.0) n.s. oropharyngeal Hong Kong[5]; mild & severe patients had similar duration of illness at admission, age, sex, & comorbidities
12 Ct = 35.2 (n = 10, sigma = 5) Ct = 26 (n = 3, sigma = 5.1) p = 0.0177 nasal Guangdong, China [6]
18 Ct = 30.9(n = 12, sigma = 5.6) Ct = 30.3 (n = 6, sigma = 5.3) n.s. nasopharyngeal Singapore[7]
6 Ct = 39 (n = 2, sigma = 1.2) Ct = 39 (n = 4, sigma = 1.1) n.s. nasopharyngeal Chongqing, China; all patients had previously recovered from COVID-19 and had reoccurence of disease.[8]
94 Ct=30 (n = 76, sigma = 5) Ct = 30 (n = 18, sigma = 4) n.s Throat swabs Guangzhou, China [9]
11 6.39 log copies/mL (n = 5, sigma = 0.9) 6.15 log copies/mL (n = 6, sigma = 1.6) n.s. nasopharyngeal Hong Kong [10]; first 11 patients diagnosed with COVID-19
5 6.3 log copies/ 1000 cells (n = 2, sigma = 1.25) 5.3 log copies/1000 cells (n = 3, sigma = 2.4) n.s. nasopharyngeal France [11]; patients traveling from Wuhan

Do severe COVID-19 patients have higher viral loads than mild patients?

This is a flawed proxy for the question of initial exposure, since patients are usually not tested for COVID-19 until at least several days after infection. Higher viral load could indicate that the virus replicated faster inside the body, or simply that the patients were later in their course of disease, rather than that the initial exposure was lower.

Still, the correlation between viral load and disease severity is suggestive as to whether there’s a dose-response effect in COVID-19.

Viral load can be measured with Ct, the number of PCR amplification cycles necessary before viral RNA is detectable. Lower Ct numbers mean exponentially more virus. It can also be measured by delta-Ct, the difference between the number of cycles to detection in a control sample (without virus) and the test sample. Finally, it can be measured by the absolute concentration of viral particles per mL. These numbers cannot be directly compared because the sensitivity of RT-PCR equipment varies and isn’t always given in the papers.

For all studies measured in Ct value, mean Ct score is 29.6 for mild cases and 26.1 for severe cases; this is a significant effect, p = 0.0004.

For the 4 studies measured in absolute concentration, mean concentration was 4.7 log copies/mL for mild cases and 5.3 log copies/mL for severe cases, which was not statistically significant, possibly because the sample size was too small.

Also note that the two studies which looked at sputum samples both found significant effects, and sputum samples reliably contain more virus RNA than nasopharyngeal samples. This suggests that viral load in the lungs is more predictive of the amount of lung damage & thus the severity of respiratory illness than viral load in the upper respiratory tract.

Severe COVID-19 patients have significantly higher initial viral loads than mild COVID-19 patients.

Exposure Effects in COVID-19

Another question we might ask is whether people with extensive exposure to the virus had more severe illness than people with transient or low-dose exposure.

In a study of 36 children with COVID-19, there was no significant difference in disease severity between those who had close contact with a family member with COVID-19 and those who had traveled to an area affected by COVID-19.[12]

In 1663 patients with COVID-19 in Wuhan, China, there was no significant difference in severity between those who did and did not have a family member with the disease or exposure to the Wuhan seafood market, but were significantly (p<0.0001) less likely to have severe disease if they were healthcare workers compared to the general population. On the other hand this may be because healthcare workers are younger (most of the patients were retirees.)[13]

In 568 COVID-19 patients in Wuhan, China, there was a nonsignificant trend (p = 0.06) for severe cases to be more likely to have had household exposure to the virus, while non-severe cases were more likely to have had hospital exposure to the virus.[14]

In another study of 487 patients with COVID-19 in Zhejiang Province, China, patients with severe disease were significantly less likely to have traveled to an affected area (49.0% vs. 65.1%, p = 0.027) but significantly more likely to be part of a family cluster of three or more patients (10.2% vs. 2.5%).[15]

Overall it doesn’t seem that the relationship between disease severity and magnitude of exposure to infected people has been much studied in COVID-19.

However we do have two Chinese studies indicating that severe cases of COVID-19 are more common in those who had frequent long-term exposure to other patients (i.e. sharing a household with many other patients) and less likely in those who have had transient contact with patients (as in travel to an epidemic-affected province.)

Viral Load in Other Human Coronaviruses


In 75 SARS patients, there was no difference in the rate of positive viral RNA samples at diagnosis between patients who later developed ARDS and those who did not.[16]

In 133 SARS patients, initial viral load was significantly (p = 0.025) associated with developing ARDS, and with shorter survival time (p = 0.006).[17]

In 12 SARS patients, viral RNA levels were 30x higher in patients who required ICU admission than in those who didn’t (p < 0.005).[18]

In 154 SARS patients, viral load in nasopharyngeal aspirated was associated (p < 0.01) with diarrhea, oxygen desaturation, mechanical ventilation, and death. Death was 54x as likely in patients positive for viral RNA than negative. [19]


In 17 MERS patients, viral load was nonsignificantly (p = 0.06) higher in severe than mild cases.[20]

In 21 MERS patients, blood positivity for MERS RNA was associated with much lower survival (p = 0.017) and higher rates of needing mechanical ventilation (p = 0.003). Higher respiratory viral load was not associated with survival.[21]

In 14 MERS patients, viral load was not significantly associated with mortality but plasma and respiratory viral loads were significantly higher in severe than mild cases.[22]

In 102 MERS patients, respiratory viral titers were significantly (p = 0.0087) higher in patients who died than patients who survived.[23]

The evidence seems very consistent that_ higher viral load correlates with more severe and deadly disease_ in both SARS and MERS.

Exposure Effects in Other Human Coronaviruses

Among 80 health care workers in Singapore exposed to SARS, 45 were serologically positive for the virus. Of those 45, healthcare workers were significantly more likely to be asymptomatic (p = 0.025) if they used masks.[24]

During the SARS epidemic in Hong Kong, there was no significant association between the risk of death and any disease source (household, hospital, community-acquired, airplane, or none of the above.)[25]

In 1649 SARS patients in Beijing, contact with a SARS patient prior to illness was not a significant predictor of mortality.[26]

Healthcare workers with confirmed MERS infection were younger, more likely to be female, and less likely to have comorbidities disease than other cases of MERS. They also had significantly (p < 0.001) lower risk of death and higher chance of being asymptomatic.[27]

MERS patients have a higher risk of mortality (HR = 2.9, p= 0.001) from hospital-acquired infections as opposed to other routes of infection such as household contacts or camels; even after adjusting for age and comorbidities.[28]

Healthcare workers may be exposed to larger doses of SARS or MERS, but it doesn’t show up in increased risk of death or severe disease for healthcare workers, possibly because healthcare workers tend to be younger than other patients.

Hospital-acquired MERS infections are more deadly than other sources of infection, but this may be confounded by the fact that patients who are already in hospitals for other reasons tend to be sicker.

The one piece of evidence from SARS and MERS that points to low-dose exposure being safer is that wearing masks is more likely to result in asymptomatic SARS infection than not wearing masks.

Dose-Response Effects in Other Viruses: Human Studies

Humans have been experimentally exposed to viruses (usually with milder effects than SARS or MERS) in a few challenge trials; this can help us ascertain whether there’s a dose-response effect to the quantity of initial viral exposure on the severity of the disease symptoms.

Adenovirus Type 4

In 16 military recruits, the probability of infection varied with the dose of adenovirus they were exposed to, but the probability of illness did not.[29]


In 127 adult volunteers inoculated with various doses of ECHO-11 virus, there was a dose-response relationship with infectivity, but dose had no effect on the severity of upper respiratory symptoms in the infected group. On the other hand, high-dose cases were significantly more likely than low-dose cases to have lower respiratory symptoms (cough, sore throat, laryngitis). Prior challenge reduced symptoms upon rechallenge, and more for the high-dose-prior-exposed than the low-dose-prior-exposed group.[30]


Across 36 studies with different doses and strains, there was no association of higher doses of influenza with higher chance of the subjects becoming ill.[31]


In a study of 57 adults infected with different doses of Norwalk virus, higher doses were associated with significantly (p =0.001) faster time before symptom onset and longer (p = 0.04) duration of illness, as well as a dose-response relationship in probability of infection.[32]

Respiratory Synctial Virus

In 35 adult volunteers, there was no association between the dose of RSV administered and the viral load or the probability of infection. Symptom severity, however, as well as cytokine levels, correlated closely with viral load, both across patients and within patients over time.[33]

Volunteers infected with low-dose RSV did not develop illness (0/16), while volunteers infected with high-dose RSV did develop colds (6/17). Illness was independent of the amount of viral shedding.[34]

In 36 volunteers infected with high or low dose RSV, higher doses were associated with higher risk of infection but not higher risk of symptomatic illness.[35]


In 155 young adult volunteers infected with different doses of rhinovirus, daily symptom scores were consistently higher in those exposed to higher doses. Higher antibody titers and higher doses of virus were associated with higher rates of infection.[36]


In a study of 62 adult volunteers experimentally infected with different doses of rotavirus, there was a dose-response relationship between the concentration of virus and the probability of infection, but there was no relationship between the virus dose and the probability of experiencing symptoms.[37]

Whether higher initial doses of virus correlate with more severe symptoms or higher probability of having symptoms at all, seems to depend on the virus. There is no relationship between dose and symptom severity in human volunteers exposed to adenovirus, influenza, and rotavirus, but there is a relationship in echovirus, rhinovirus, and norovirus, and the results are ambiguous in respiratory synctial virus.

It remains unclear how well any of these observations translate to COVID-19. Unfortunately there were no studies comparing the effect of dose on the symptoms of any of the mild human coronaviruses.

Dose-Response Effects in Other Viruses: Animal Studies


Exposure to 100 plaque-forming units of Ebola virus in the nose was lethal to macaques; but exposure to 10 plaque-forming units caused no clinical symptoms or detectable antibodies or viral RNA.[38]


Eastern Equine Encephalitis Virus is lethal in high doses in macaques (6/6 animals died) but less severe in low-dose exposure (2/6 animals died and clinical scores were lower). [39]


Serial dilutions of the baboon herpesvirus HPV2 inoculated into mice showed a dose-response curve with higher doses resulting in higher rates of infection, CNS symptoms, and death.[40]


There was a dose-response relationship between initial inoculation dose, probability of infection, and probability of death, in H5N1 and H7N1 influenza in turkeys, chickens, and ducks.[41]

Probability of infection, probability of clinical signs, and survival time all varied dose-dependently by the inoculation dose of H5N1 in pigeons.[42]

A H0N1 strain of influenza in mice had a dose-response relationship for both infectiousness and mortality; it was more deadly when introduced to the respiratory tract than intranasally.[43]

A PR8 strain of influenza has a dose-response relationship with serum viral load, weight loss, clinical score, and mortality. The low dose and high dose groups had similar antibody and leukocyte recruitment levels as the high dose, but no mortality. [44]


In transgenic mice, there was a dose-response effect in lethality for exposure to MERS virus: it killed 50% of the mice at a dose of 10 TCID50, compared to 25% at 1.25 TCID50 and 100% at 100 TCID50.[45]


PEDV is a coronavirus that causes diarrhea in pigs. It has a dose-response relationship to symptoms: 0.056 TCID50 caused diarrhea in 25% of piglets while 0.56 TCID50 and higher caused diarrhea in 100% of piglets.[46]


Both mice and guinea pigs exposed to SARS virus had a higher probability of increased rectal temperature at higher doses of virus exposure. For guinea pigs, ID50 = 5.47 log CPE50, or 50% of guinea pigs were infected at 5.47 * log (the dose at which 50% of cells died).[47]


Simian immunodeficiency virus, closely related to HIV, had a dose-response relationship with the probability of infection in macaque monkeys, but steady-state (2-week) viral load among those infected did not correlate with initial dose. There was no correlation between survival time and dose.[48]

Low-dose challenge with SIV successfully produced “silent infection” (no symptoms or viral RNA but virus-specific T cell proliferation.) However, when these monkeys were exposed to a high dose of SIV, they were not immune but became infected and developed AIDS-like disease at the same rate as naive monkeys.[49]

Dose-response relationships between virus inoculum and infection rate or death seem to be common in animal studies, and consistently, including in coronaviruses, higher doses cause more severe symptoms.

However, in at least one case (SIV) a dose of virus that was low enough to produce asymptomatic infection did not produce immunity to future exposures, so we can’t assume that low-dose exposure always brings immunity.


[1]Shah, Sweta, et al. “Initial Observations with Molecular Testing for COVID-19 in a Private Hospital in Mumbai, India.”

[2]Yu, Xia, et al. “SARS-CoV-2 viral load in sputum correlates with risk of COVID-19 progression.” Critical Care 24.1 (2020): 1-4.

[3]Zheng, Shufa, et al. “Viral load dynamics and disease severity in patients infected with SARS-CoV-2 in Zhejiang province, China, January-March 2020: retrospective cohort study.” bmj 369 (2020).

[4]Liu, Yang, et al. “Correlation Between Relative Nasopharyngeal Virus RNA Load and Lymphocyte Count Disease Severity in Patients with COVID-19.” Viral Immunology (2020).

[5]To, Kelvin Kai-Wang, et al. “Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study.” The Lancet Infectious Diseases (2020).

[6]Zou, Lirong, et al. “SARS-CoV-2 viral load in upper respiratory specimens of infected patients.” New England Journal of Medicine 382.12 (2020): 1177-1179.

[7]Young, Barnaby Edward, et al. “Epidemiologic features and clinical course of patients infected with SARS-CoV-2 in Singapore.” Jama 323.15 (2020): 1488-1494.

[8]Wu, Jing, et al. “Clinical Characteristics and Outcomes of Discharged COVID-19 Patients with Reoccurrence of SARS-CoV-2 RNA in a County Hospital of Western Chongqing, China.” China (4/14/2020) (2020).

[9]He, Xi, et al. “Temporal dynamics in viral shedding and transmissibility of COVID-19.” Nature medicine (2020): 1-4.

[10]Lui, Grace, et al. “Viral dynamics of SARS-CoV-2 across a spectrum of disease severity in COVID-19.” The Journal of Infection (2020).

[11]Lescure, Francois-Xavier, et al. “Clinical and virological data of the first cases of COVID-19 in Europe: a case series.” The Lancet Infectious Diseases (2020).

[12]Qiu, Haiyan, et al. “Clinical and epidemiological features of 36 children with coronavirus disease 2019 (COVID-19) in Zhejiang, China: an observational cohort study.” The Lancet Infectious Diseases (2020).

[13]Xu, Xiao-Wei, et al. “Clinical findings in a group of patients infected with the 2019 novel coronavirus (SARS-Cov-2) outside of Wuhan, China: retrospective case series.” bmj 368 (2020).

[14]Li, Xiaochen, et al. “Risk factors for severity and mortality in adult COVID-19 inpatients in Wuhan.” Journal of Allergy and Clinical Immunology (2020).

[15]Shi, Y., et al. “Host susceptibility to severe COVID-19: a retrospective analysis of 487 case outside Wuhan.” (2020).

[16]Peiris, Joseph Sriyal Malik, et al. “Clinical progression and viral load in a community outbreak of coronavirus-associated SARS pneumonia: a prospective study.” The Lancet 361.9371 (2003): 1767-1772.

[17]Chu, Chung-Ming, et al. “Initial viral load and the outcomes of SARS.” Cmaj 171.11 (2004): 1349-1352.

[18] Ng, Enders KO, et al. “Quantitative analysis and prognostic implication of SARS coronavirus RNA in the plasma and serum of patients with severe acute respiratory syndrome.” Clinical chemistry 49.12 (2003): 1976-1980.

[19]Hung, I. F. N., et al. “Viral loads in clinical specimens and SARS manifestations.” Emerging infectious diseases 10.9 (2004): 1550.

[20]Oh, Myoung-don, et al. “Viral load kinetics of MERS coronavirus infection.” New England Journal of Medicine 375.13 (2016): 1303-1305.

[21]Kim, So Yeon, et al. “Viral RNA in blood as indicator of severe outcome in Middle East respiratory syndrome coronavirus infection.” Emerging infectious diseases 22.10 (2016): 1813.

[22]Min, Chan-Ki, et al. “Comparative and kinetic analysis of viral shedding and immunological responses in MERS patients representing a broad spectrum of disease severity.” Scientific reports 6.1 (2016): 1-12.

[23] Feikin, Daniel R., et al. “Association of higher MERS-CoV virus load with severe disease and death, Saudi Arabia, 2014.” Emerging infectious diseases 21.11 (2015): 2029.


[24] Wilder-Smith, Annelies, et al. “Asymptomatic SARS coronavirus infection among healthcare workers, Singapore.” Emerging infectious diseases 11.7 (2005): 1142.

[25]Leung, Gabriel M., et al. “The epidemiology of severe acute respiratory syndrome in the 2003 Hong Kong epidemic: an analysis of all 1755 patients.” Annals of internal medicine 141.9 (2004): 662-673.

[26]Liu, Min, et al. “Risk factors for SARS-related deaths in 2003, Beijing.” Biomedical and Environmental Sciences 19.5 (2006): 336.

[27]Elkholy, Amgad A., et al. “MERS-CoV infection among healthcare workers and risk factors for death: Retrospective analysis of all laboratory-confirmed cases reported to WHO from 2012 to 2 June 2018.” Journal of infection and public health (2019).

[28]Ahmed, Anwar E. “The predictors of 3-and 30-day mortality in 660 MERS-CoV patients.” BMC infectious diseases 17.1 (2017): 615.

[29]Couch, R. B., et al. “The minimal infectious dose of adenovirus type 4; the case for natural transmission by viral aerosol.” Transactions of the American Clinical and Climatological Association 80 (1969): 205.

[30]Saliba, Gilbert S., Sylvia L. Franklin, and George Gee Jackson. “ECHO-11 as a respiratory virus: quantitation of infection in man.” The Journal of clinical investigation 47.6 (1968): 1303-1313.

[31]Yezli, Saber, and Jonathan A. Otter. “Minimum infective dose of the major human respiratory and enteric viruses transmitted through food and the environment.” Food and Environmental Virology 3.1 (2011): 1-30.

[32]Atmar, Robert L., et al. “Determination of the 50% human infectious dose for Norwalk virus.” The Journal of infectious diseases 209.7 (2014): 1016-1022.

[33]DeVincenzo, John P., et al. “Viral load drives disease in humans experimentally infected with respiratory syncytial virus.” American journal of respiratory and critical care medicine 182.10 (2010): 1305-1314.

[34]Mills, John, et al. “Experimental respiratory syncytial virus infection of adults: possible mechanisms of resistance to infection and illness.” The Journal of Immunology 107.1 (1971): 123-130.

[35]Lee, F. Eun-Hyung, et al. “Experimental infection of humans with A2 respiratory syncytial virus.” Antiviral research 63.3 (2004): 191-196.

[36]Hendley, J. Owen, William P. Edmondson Jr, and Jack M. Gwaltney Jr. “Relation between naturally acquired immunity and infectivity of two rhinoviruses in volunteers.” Journal of Infectious Diseases 125.3 (1972): 243-248.

[37]Ward, Richard L., et al. “Human rotavirus studies in volunteers: determination of infectious dose and serological response to infection.” Journal of Infectious Diseases 154.5 (1986): 871-880.

[38]Mire, Chad E., et al. “Oral and conjunctival exposure of nonhuman primates to low doses of Ebola Makona virus.” The Journal of infectious diseases 214.suppl_3 (2016): S263-S267.

[39]Reed, Douglas S., et al. “Severe encephalitis in cynomolgus macaques exposed to aerosolized eastern equine encephalitis virus.” The Journal of infectious diseases 196.3 (2007): 441-450.

[40]Ritchey, J. W., et al. “Comparative pathology of infections with baboon and African green monkey α-herpesviruses in mice.” Journal of comparative pathology 127.2-3 (2002): 150-161.

[41]Aldous, E. W., et al. “Infection dynamics of highly pathogenic avian influenza and virulent avian paramyxovirus type 1 viruses in chickens, turkeys and ducks.” Avian Pathology 39.4 (2010): 265-273.

[42] Phonaknguen, Rassameepen, et al. “Minimal susceptibility to highly pathogenic avian influenza H5N1 viral infection of pigeons (Columba livia) and potential transmission of the virus to comingled domestic chickens.” Kasetsart J-Nat Sci 47 (2013): 720-732.

[43]Yetter, Robert A., et al. “Outcome of influenza infection: effect of site of initial infection and heterotypic immunity.” Infection and immunity 29.2 (1980): 654-662.

[44]Powell, Timothy J., et al. “The immune system provides a strong response to even a low exposure to virus.” Clinical immunology 119.1 (2006): 87-94

[45]Tao, Xinrong, et al. “Characterization and demonstration of the value of a lethal mouse model of Middle East respiratory syndrome coronavirus infection and disease.” Journal of virology 90.1 (2016): 57-67.

[46]Neumann, Eric J., and William F. Hall. “Disease Control, Prevention, and Elimination.” Diseases of Swine (2019): 123-157.

[47]Chepurnov, A. A., A. A. Dadaeva, and E. M. Malkova. “Symptoms of infection caused by SARS coronavirus in laboratory mice and guinea pigs.” Doklady Biological Sciences. Vol. 397. No. 1. Nature Publishing Group, 2004.

[48]Holterman, Lennart, et al. “The rate of progression to AIDS is independent of virus dose in simian immunodeficiency virus-infected macaques.” Journal of General Virology 81.7 (2000): 1719-1726

[49]Dittmer, Ulf, et al. “Repeated exposure of rhesus macaques to low doses of simian immunodeficiency virus (SIV) did not protect them against the consequences of a high-dose SIV challenge.” Journal of general virology 76.6 (1995): 1307-1315.