Skip to main content

Household secondary attack rate of COVID-19 and associated determinants in Guangzhou, China: a retrospective cohort study

Our take —

This contact tracing study in Guangzhou, China found that the infectivity of the virus was the same between the incubation period and the symptomatic periods. Attack rates among susceptible persons increased with age. The study expands the literature on household secondary attack rates; however, the quantification of infectvity from asymptomatic cases was a limitation.

Study design

Retrospective Cohort

Study population and setting

Researchers in Guangzhou, China used contact tracing data reported to the Guangzhou Municipal Centers for Disease Control reported between January 7 and February 18, 2020. This included 349 confirmed COVID-19 cases that formed 195 clusters of 215 index cases and 134 secondary or tertiary cases. The sample also included 1,964 uninfected contacts from these cases.

Summary of Main Findings

The attack rate for household contacts from index cases was 12.4% when defined based on close relatives versus 17.1% when household contacts were defined based on residential address. The household attack rate was lower among contacts <20 years of age (OR: 0.23, 95% CI: 0.11, 0.46) and contacts aged 20-59 years (OR: 0.64, 95% CI: 0.43, 0.97) compared to those 60 years or older. Overall, the secondary attack rates were not statistically different the incubation period and symptomatic period though the attack rate was non-statistically significantly lower during the symptomatic period (OR: 0.61 [95% CI: 0.27,1.38]). The estimated effective R (average number of individuals infected by a case given the proportion of Guangzhou that were susceptible) was 0.5 (95% CI: 0.49, 0.74).

Study Strengths

The study provides further evidence of the occurance of early transmission prior to symptom onset.


Although the researchers included asymptomatic cases in the analysis, there were only a few of these individuals and a proxy date for symptom onset was used, which may not be a valid comparison to symptomatic cases.

Value added

This was one of the first studies to utilize contact tracing data to estimate infectivity during the incubation period and contributes knowledge concerning basic epidemiologic features of coronavirus (e.g. effective reproductive number) in the context of social distancing and masking wearing.

This review was posted on: 1 August 2020