Intrinsic generation time of the SARS-CoV-2 Omicron variant: An observational study of household transmission

Abstract

Background: Starting from the final months of 2021, the SARS-CoV-2 Omicron variant expanded globally, swiftly replacing Delta, the variant that was dominant at the time. Many uncertainties remain about the epidemiology of Omicron; here, we aim to estimate its generation time.

Methods: We used a Bayesian approach to analyze 23,122 SARS-CoV-2 infected individuals clustered in 8903 households as determined from contact tracing operations in Reggio Emilia, Italy, throughout January 2022. We estimated the distribution of the intrinsic generation time (the time between the infection dates of an infector and its secondary cases in a fully susceptible population), realized household generation time, realized serial interval (time between symptom onset of an infector and its secondary cases), and contribution of pre-symptomatic transmission.

Findings: We estimated a mean intrinsic generation time of 6.84 days (95% credible intervals, CrI, 5.72-8.60), and a mean realized household generation time of 3.59 days (95%CrI: 3.55-3.60). The household serial interval was 2.38 days (95%CrI 2.30-2.47) with about 51% (95%CrI 45-56%) of infections caused by symptomatic individuals being generated before symptom onset.

Interpretation: These results indicate that the intrinsic generation time of the SARS-CoV-2 Omicron variant might not have shortened as compared to previous estimates on ancestral lineages, Alpha and Delta, in the same geographic setting. Like for previous lineages, pre-symptomatic transmission appears to play a key role for Omicron transmission. Estimates in this study may be useful to design quarantine, isolation and contact tracing protocols and to support surveillance (e.g., for the accurate computation of reproduction numbers).

Funding: The study was partially funded by EU grant 874850 MOOD.

Keywords: Bayesian inference; COVID-19; Contact tracing; Generation time; Omicron; SARS-CoV-2.

Figure 1

Illustrative example of a household cluster. A household with 5 members, of which #4 (asymptomatic) was infected outside the household (in the general community) and then transmitted to cases #5 and #3 (both symptomatic). Case #3 infected #2 while #1 remained uninfected. #3, #2 and #1 were vaccinated with 1 dose, 2 doses, and 2 doses + booster respectively. In the bottom part of the figure, we show examples of the temporal intervals of interest for this work. Note that for the household serial interval and the realized household generation time, the source of infection (whether from outside the household or from a household member, and, in the latter case, which household member) is also unobserved and needs to be probabilistically reconstructed. The intrinsic generation time is not displayed as it represents the distribution of generation times among infections occurring in the general population in a fully susceptible population.

Figure 2

Estimates of the generation time for the Omicron variant. A) Distribution of the intrinsic generation time; solid line: mean estimate; shaded area: 95% CrI; B) Distribution of the realized household generation time; bars: mean estimate; vertical lines: 95% CrI.

Figure 3

Estimates of generation times for the Omicron variant under different sensitivity analyses. A) Posterior distributions of the mean intrinsic generation time; B) Mean distributions of the intrinsic generation time. Point: mean value; box: interquantile range; whiskers: 95% CrI. The labels on the y-axis represent the performed sensitivity analysis to evaluate the robustness of baseline model results against different model assumptions where we consider: a) only households genotyped as Omicron; b) only household composed of unvaccinated individuals; c) an incubation period for Omicron with the same distribution as previous estimates for Delta (mean: 4.5 days; standard deviation: 2.1 days); d) a prolonged diagnostic delay for asymptomatic individuals (mean: 7.58 days, standard deviation: 1.61 days); e) a prolonged diagnostic delay for asymptomatic individuals (mean: 7.58 days, standard deviation: 1.97 days); f) the possibility of false negative tests; g) a halved transmissibility for asymptomatic individuals; h) a halved transmissibility for vaccinated individuals; i) a scenario where any effort to quarantine positive cases would not impact the force of infection from outside the household; j) previous infection from other variants provides no cross-immunity against Omicron infection.