Epidemiologic Evidence

Most reports of presymptomatic (9–12), asymptomatic (13–15), or a combination of presymptomatic or asymptomatic SARS-CoV-2 transmission (16,17) were from China (Table 1). Presymptomatic or asymptomatic primary patients were typically exposed to SARS-CoV-2 during travel from Wuhan or another city in Hubei Province, China (9–16). One couple was exposed during a mass gathering in Shanghai for the Chinese Spring Festival (17). Reported cases of infected persons who transmitted the virus to others while presymptomatic or asymptomatic have occurred within families or households (9–11,13–17), during shared meals (10,12), or during visits with hospitalized family members (9,13). An inherent confounder to these reports from China is the inability to entirely rule out alternative SARS-CoV-2 exposure in the community early in the outbreak, when transmission in the community may have been undetected.

However, cases of presymptomatic transmission have been reported from other countries before widespread community transmission occurred. A report from Germany documented infection of a German businessman after exposure to a mildly symptomatic colleague visiting from China (18). Before becoming symptomatic, this businessman exposed 2 other colleagues who subsequently received a COVID-19 diagnosis but did not have contact with the primary patient from China or any other known source. A report from Singapore described 7 COVID-19 clusters resulting from presymptomatic transmission; presymptomatic primary patients varied from persons with travel from high-incidence countries to persons exposed in the local community (19). All primary patients experienced distinct periods of initial exposure and presymptomatic close contact with secondary patients who had no other known exposure risks. The incubation periods for presymptomatic primary patients with distinct exposures ranged from 3 to 11 days; for presymptomatic primary patients with travel history to an area with active transmission, the time from last exposure to symptom onset ranged from >2 to >9 days.

Virologic Evidence

Currently, SARS-CoV-2 infection is primarily diagnosed by detection of viral RNA via reverse transcription PCR (RT-PCR) or by viral culture and demonstration of cytopathic effect (20). Although RT-PCR identifies viral RNA and cannot determine whether infectious virus is present, infectiousness can be inferred from cycle threshold (C_t) values. The RT-PCR C_t value represents the number of PCR cycles required to detect SARS-CoV-2 RNA; lower values indicate higher viral load and imply higher infectiousness (20–22). The exact RT-PCR C_t values associated with the presence of infectious SARS-CoV-2 is unknown, but infectious virus has been isolated from a specimen with an RT-PCR C_t of 34 (23).

Four reports documented the presence of SARS-CoV-2 RNA with lower C_t values in samples collected from persons in whom symptoms of COVID-19 never developed (24–27) (Table 2). Two reports described specimens with low RT-PCR C_t values among presymptomatic and asymptomatic residents of a nursing home identified as part of the same outbreak investigation (23,28). Among these reports, RT-PCR C_t values for SARS-CoV-2 RNA in asymptomatically infected persons ranged from 14 to 40 (23–27). The study with data on presymptomatic infected patients reported an average RT-PCR C_t value of 24 (range 15–38) (23). Two reports described culture of infectious virus from persons with asymptomatic (24) and presymptomatic (23) SARS-CoV-2 infection. Although these reports did not identify actual virus transmission while presymptomatic or asymptomatic, the low RT-PCR C_t values (i.e., high viral load) and ability to isolate infectious SARS-CoV-2 provide plausible virologic evidence for SARS-CoV-2 transmission by persons not demonstrating symptoms.

Modeling Evidence

Two studies used models to estimate the serial interval (time between symptom onset in a primary patient and the secondary patient) (29,30) (Table 3). They estimated the serial interval of COVID-19 to be 4 days, which is shorter than the estimated median incubation period for COVID-19 of 5 days (31). One report suggested that up to 13% of infections may be transmitted during the presymptomatic period of illness (29). These studies relied on reports of primary and secondary cases and may be limited by recall bias; secondary patients are more likely to remember proximal exposures, biasing results toward a shorter serial interval.

Two models attempted to estimate the number of infections caused by asymptomatic, presymptomatic, or mildly symptomatic infected persons (30,32). These models varied widely; 1 model suggested that up to half of infections were transmitted from infected persons who were presymptomatic (33), and another suggested that up to four fifths of infections were transmitted by persons with no symptoms or mild symptoms (32). Both models suggested that a large number of persons with asymptomatic or mildly symptomatic infections were not detected by the health system and that these persons meaningfully contributed to ongoing community transmission (32,33). Although models are highly dependent on the assumptions built into them, these models suggest that the speed and extent of SARS-CoV-2 transmission cannot be accounted for solely by transmission from symptomatic persons.

Each of the epidemiologic, virologic, and modeling studies described has limitations. However, in the aggregate, these diverse studies suggest that SARS-CoV-2 can be transmitted by persons with presymptomatic or asymptomatic infection, which may meaningfully contribute to the propagation of the COVID-19 pandemic. This literature summation was conducted to support changes in CDC recommendations to reduce the risk for asymptomatic transmission and was not a systematic review. These conclusions are drawn from the literature available at the time and may change, given the rapidly evolving nature of the evidence base for asymptomatic transmission.