It is increasingly recognized that coronaviruses can cause major emerging viral disease threats, with the respiratory syndromes SARS and MERS being two recent examples, and two coronaviruses now endemic in humans (229E and OC43) have emerged from animals within the past few hundred years1. The outbreak of the coronavirus SARS-CoV-2 started in December 2019. On the 30 January 2020, the World Health Organization declared this event a Public Health Emergency of International Concern. The reported cases and deaths of COVID-19 already exceed those of SARS or MERS. Here we highlight some of the key recent findings related to this global epidemic.

SARS-CoV-2 can be readily cultured from clinical specimens, and viral isolates are now available in mainland China2 and elsewhere, including in our own laboratory (Fig. 1). SARS-CoV-2 is genetically similar to other coronaviruses in the subgenus Sarbecovirus, a clade of betacoronaviruses formed by the coronavirus that causes SARS (SARS-CoV) and other SARS-CoV-like coronaviruses found in bats3,4. Recombinations between coronaviruses are common, and SARS-CoV is believed to be a recombinant between bat sarbecorviruses. Interestingly, the whole genome of SARS-CoV-2 is highly similar to that of a bat coronavirus detected in 2013 (>96% sequence identity)4, which suggests that the immediate ancestor of SARS-CoV-2 has been circulating in bats for at least several years.

Fig. 1: An electron microscopy image showing SARS-CoV-2 isolated at The University of Hong Kong.
figure 1

Provided by John Nicholls (Department of Pathology).

Full genome analyses of the virus2,3 indicate that this epidemic was caused by a single zoonotic introduction and that the virus is relatively stable, genetically, in humans3. The first human cluster was reported in association with exposure to a seafood market2,5 that is known to sell live wild game animals for consumption. It is possible that the zoonotic transmission of SARS-CoV-2 might involve an intermediate host (or hosts), as was observed in the SARS epidemic. However, some of the earliest cases had no epidemiological exposure to this market5. It is therefore not yet clear whether the initial zoonotic jump occurred directly from bats to humans or whether an intermediate mammalian species was involved. Identification of the antecedent zoonotic source is relevant because further zoonotic transmission events may well occur unless the transmission pathways of the initial zoonotic event are identified and interrupted.

Previous research on several SARS-CoV-like bat coronaviruses demonstrated that some of these viruses can use the human receptor ACE-2 for infection. The SARS-CoV-2 spike protein is predicted to be structurally similar to that of SARS-CoV3 and, indeed, it can be bound by a monoclonal antibody that is specific for the spike of SARS-CoV6. Although variations in key residues that are essential for binding to ACE-2 were found in the spike of SARS-CoV-2, this novel virus is experimentally capable of using human, swine, bat and civet ACE-2, but not mouse ACE-2, for entry4. The spike of SARS-CoV-2 can also theoretically interact with ACE-2 from other animal species7.

In initial clinical reports on 99 patients confirmed as being infected with SARS-CoV-2, symptoms of fever and cough were commonly seen (>80%). Shortness of breath (31%) and muscle ache (11%) were also seen in patients8. In contrast to patients infected by human coronaviruses that cause the common cold, runny nose and sore throat were less common (≤5%) in hospitalized patients but may be more common in milder illness (discussed below)9,10. In the hospital-based case series, radiological evidence of bilateral (75%) or unilateral (25%) pneumonia was seen, sometimes with evidence of multiple mottling and ground-glass opacities. 17% of the patients developed acute respiratory distress syndrome that sometimes led to multiple organ dysfunction and death. Approximately 75% of the patients required supplemental oxygen, and 13% required mechanical ventilation. The age of affected patients ranged from 21 years to 82 years, with 67% of them being >50 years of age and 51% having underlying co-morbidities. The clinical presentations and progression were broadly similar to those in patients with MERS or SARS8.

Recent data from case clusters suggest that the overall clinical spectrum of this disease can be more heterogeneous9,10. Upper respiratory symptoms such as sore throat and nasal congestion, as well as diarrhea, may be seen in milder cases. Radiological evidence of pneumonia may be seen even in asymptomatic infections. These clusters also suggest that older age is associated with more-severe disease, with young adults and children having progressively less-severe disease9. An age-associated increase in disease severity was also observed in SARS.

Lower respiratory specimens (e.g., sputum) appear to have a higher viral load than that of upper respiratory specimens (e.g., throat swab)9. Viral RNA was also detected in blood and stool specimens11,12, but it is not known whether these non-respiratory samples are infectious or not. Given that fecal samples from patients with SARS were infectious in some instances (e.g., the Amoy Gardens incident in Hong Kong), precautions against fecal–oral transmission are advisable.

Apart from the early cases2,5, subsequent human infections were caused by sustainable human-to-human transmission. Using the first 425 confirmed cases in Wuhan, Li et al. estimated that the mean incubation period of infection with SARS-CoV-2 was 5.2 days (95% confidence interval (CI), 4.1–7.0), with about 95% of the cases developing symptoms within 12.5 days of an exposure5, justifying the current recommendations of a 14-day period for medical observation or quarantine. The reproductive number (R0; the number of secondary cases expected in a completely susceptible population) and the epidemic doubling time were estimated to be 2.2 (95% CI, 1.4–3.9) and 7.4 days (95% CI, 4.2–14), respectively. Studies from others also have led to broadly similar figures13. These are comparable to those observed during the SARS epidemic. However, transmission of SARS-CoV-2 can occur from patients with mild disease9,10. Whether transmission can occur during the late incubation period remains controversial10. This is in sharp contrast to the transmission pattern observed during SARS, for which transmission rarely occurred until after the 4–5 days after symptom onset. Taken together, these findings suggest that the public-health interventions that successfully interrupted the spread of SARS-CoV are unlikely to be as effective in the current outbreak.

Using data from the numbers of exported cases from Wuhan and data on travel patterns, Wu et al. estimated that there were >75,000 infected people in Wuhan between 1 December 2019 and 25 January 2020 (ref. 13). With the current trends and assuming a reduction in transmissibility due to interventions, they predicted the outbreak in Wuhan will peak in April 2020. They also predicted that this epidemic will continue to grow exponentially outside Wuhan. Their simulations further suggested that a 50% reduction in the transmission of this disease achieved through public-health interventions, but without a reduction in population movement, can dramatically delay the exponential growth of this disease for at least a few months. While implementation of aggressive disease-control measures such as school closure and social distancing may defer the establishment of transmission in countries at risk of disease importation, it is still unclear if global spread of this disease can now be prevented.

Although much has been learned in the past few weeks, a number of crucial knowledge gaps remain. These include the modes of transmission, the stability of the virus in environments, mechanisms of pathogenesis and effective treatments and vaccines. In the current circumstances, the most important question is that of disease severity. It is relevant to note that in the early stage of the 2009 H1N1 influenza virus pandemic, case-fatality estimates as high as 10% were reported. However, population-based age-stratified sero-epidemiological studies revealed that the true overall case fatality was about 0.001% (ref. 14). Thus, sero-epidemiology is needed for a reliable estimate of true disease severity. Past infection may also translate into population immunity, which are data that need to be accounted for in future transmission models of the virus. It is relevant to note that infection with MERS-CoV or MERS disease does not always lead to detectable antibody responses15. If SARS-CoV-2 infection has antibody-response kinetics similar to those of MERS-CoV infection, this may have implications for sero-epidemiology and the development of herd immunity. Thus, research on both the antibody kinetics and cell-mediated immune-response kinetics of SARS-CoV-2 is a priority.