Skip to main content

Clustering and superspreading potential of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections in Hong Kong

Our take —

This study, available as a preprint and thus not yet peer reviewed, suggests that there is potential in this setting for superspreading events of SARS-CoV-2, but potentially less than for SARS or MERS. Physical distancing and closures, particularly for high-risk social settings, may be needed in conjunction with rapid tracing and quarantine of confirmed contacts.

Study design

Case Series; Other

Study population and setting

Clusters of SARS-CoV-2 in Hong Kong were identified and described using contact tracing data as of April 28, 2020. Cases were laboratory-confirmed using RT-PCR. Clusters were defined as 2+ cases with known contact prior to, or following symptom onset, and were categorized based on travel history (solely imported cases, initiated by an imported case, initiated by a local case). Contact histories and symptom onset dates were used to identify probable transmission chains within clusters. The median serial interval, excluding asymptomatic cases, was calculated as the difference between onset dates of each infector-infected pair. The observed reproduction number was also estimated. Events were characterized as “superspreading” if there were 6-8 secondary cases.

Summary of Main Findings

A total of 1,038 cases were identified as of April 28, 2020. Just over half of all cases of SARS-CoV-2 in Hong Kong (539/1038) were associated with at least one of 135 identified clusters. Median cluster size was two individuals; 61% of clusters were made up of only imported cases (82/135), 22% of clusters were initiated by an imported case (30/135), and 17% were initiated by a local case. Within the 53 clusters initiated by a local or imported infection, 245 of 329 were linked and among infector-infected pairs, and the median serial interval was 4 days (IQR 3-8 days). The number of secondary transmissions was higher in social settings (e.g. bars, restaurants) than in family or work settings (p<0.001). The estimated reproductive number was 0.58 (95% CI: 0.45-0.71). Between 5 and 7 probable superspreading events were identified with substantial overdispersion in transmissibility (k=0.45, 95% CI 0.30-0.72).

Study Strengths

Analysis of contact tracing data and all laboratory-confirmed cases of SARS-CoV-2, including asymptomatic cases and contacts, allowed for the characterization of all clusters in Hong Kong, as well as the calculation of the serial interval, reproduction number, and overdispersion in transmissibility.


Estimates of the reproductive number and overdispersion of transmissibility are dependent on complete data, and any incomplete data could bias estimates. It is possible, for example, that there were other transmission events that may have been missed. Additionally, Hong Kong was able to rapidly suppress transmission and control the epidemic, and results from this study should not be applied directly in other settings.

Value added

This study provides information on the reproductive number, serial interval, and the potential for superspreading events in Hong Kong.

This review was posted on: 24 July 2020