Differential aerosol shedding of SARS-CoV-2 Delta and Omicron variants during respiratory activities

Kwok Wai Tham (  bdgtkw@nus.edu.sg ) National University of Singapore Kai Sen Tan National University of Singapore Sean W. X. Ong MOHH https://orcid.org/0000-0002-8570-436X Ming Hui Koh National University of Singapore Douglas Jie Wen Tay National University of Singapore https://orcid.org/0000-0002-0523-3490 Daryl Zheng Hao Aw National University of Singapore Yi Wei Nah National University of Singapore Mohammed Ridzwan Abdullah National Centre for Infectious Diseases Justin Jang Hann Chu National University of Singapore https://orcid.org/0000-0002-1673-6819 Vincent Chow National University of Singapore Paul Tambyah National University of Singapore


Background
Periodic emergence of novel variants of concern (VOCs) has confounded efforts to eliminate SARS-CoV-2 transmission, due largely to changes in transmissibility, virulence, and immune evasion. Transmission may occur across a range of distances via respiratory particles including ner aerosolsalthough the precise extent and proportion of aerosol transmission remain uncertain. The rapid spread of more recent VOCs such as the Delta and Omicron variants has raised the question as to whether aerosol transmission is more e cient with these variants, as this would have a signi cant impact on public health interventions.
We have previously demonstrated that SARS-CoV-2 virus and its viral components can be shed via aerosols emitted from activities such as breathing, talking and singing (1), while others have managed to culture viable virus in similar ne aerosol samples in a small number of subjects (2). Both studies were conducted early in the Delta wave and before the Omicron wave which has rapidly spread globally, and there are limited data examining the impact of these variants on the extent of aerosol shedding and transmission. In this follow-up study, we employed the same sampling methodology to collect aerosol samples from patients infected with Delta and Omicron VOCs, and to evaluate viral loads and infectivity during different respiratory activities.

Methods
We recruited patients with SARS-CoV-2 infection con rmed by reverse transcription-polymerase chain reaction (RT-PCR) and hospitalized at the National Centre for Infectious Diseases. There were no speci c inclusion criteria, but we attempted to select patients as early into the infection phase as possible and with an index nasopharyngeal swab PCR cycle threshold (Ct) value of < 25.0. Patients requiring supplemental oxygen or with severe disease were excluded as they were likely to be un t for completion of study procedures. Patients with co-infections were also excluded to minimize confounding effects from other pathogens. Using the Gesundheit G-II exhaled breath collector (3), we collected aerosol samples from participants while they were breathing, talking, and singing without wearing a mask as previously described (1). We further assessed the impact of mask-wearing by repeating the talking segment with subjects wearing a standard surgical mask (ASSURE, Pharmex Healthcare, Singapore). Aerosols were collected as fractions of two sizes, namely coarse (> 5 µm) and ne (≤ 5 µm). We also collected nasal swabs prior to and on the same day of sampling. Samples were processed in a biosafety level-3 laboratory for virus culture and RT-qPCR (for N1 gene targets) to assess infectious viral load and viral RNA load, respectively. Samples were inoculated onto VeroE6 TMPRSS2 cells (for both passages of 14 days each) which represent the more suitable cell line for SARS-CoV-2 culture (4). Whole viral genome sequencing was conducted by the National Public Health Laboratory for VOC identi cation. Due to the relatively small sample size, statistical tests of signi cance were not performed, and our results are reported descriptively. This study was approved by the National Healthcare Group Domain Speci c Review Board (reference number 2020/01113). All study participants provided written informed consent.

Results
Nine patients were recruited -4 were infected with Delta VOC and 5 with Omicron VOC. Table 1 summarizes the aerosol test results, together with participant demographics, vaccination status, and symptoms on the day of sampling. Viral RNA was detectable in seven (77.8%) participants in at least one aerosol fraction including all patients with Omicron. Fine aerosols exhibited higher positivity (77.8% vs 44.4%) as well as greater viral loads (median 774.6 vs 354.2 copies, summating all respiratory activities) compared to coarse aerosols. The range of viral RNA copies per respiratory activity (99.2 to 5964.9 copies) was similar to our previous study (63.5 to 5821.4 copies) which included individuals infected with wild-type SARS-CoV-2 as well as variants including Alpha, Beta, and Kappa and one Delta (1). All viral cultures of aerosol samples were negative after two consecutive passages. However, cytopathic effect was observed for viral cultures of nasal swab samples from participants 5 and 6 (Supplementary Figure S1).

Discussion
In this study, we found that patients infected with Delta and Omicron variants generated detectable viral particles in respiratory aerosols, similar to other ndings in patients with previous SARS-CoV-2 variants. Consistent with previous reports, ne aerosols accounted for a larger proportion of the total viral load, raising the possibility of transmission over greater distances. The negative viral cultures from aerosols may continue to re ect the technical di culties in viral aerosol sampling as opposed to true non-infectivity of exhaled viral particles (5). However, viable virus was cultured from nasal swabs of two Omicroninfected participants, which may corroborate what others found where Omicron is concentrated in the upper airway in animal models (6).
In this pilot study, we observed similar viral loads generated from aerosols of Delta and Omicron variants during talking with and without surgical masks. This nding may be attributed to the relatively small sample size and single mask type studied, and limited by the inability to quantify viable viruses. Other modelling studies have also shown less e cient reduction of infectious aerosols as well as leakage from surgical masks (7,8). Furthermore, our sampling method induces suction of aerosols from the facial region, whether emitted through or leaked around the mask. Future studies should systematically investigate the effects of various mask types on infectious respiratory aerosol production. These will be critical to help craft public health policies to reduce aerosol transmission including respiratory protection, social distancing, ventilation and air ltration and disinfection.
The signi cantly enhanced transmissibility of Delta variants was hypothesized to be related to increased respiratory viral loads (9). which was thought to lead to increased aerosol transmission via greater respiratory shedding (10). On the other hand, emerging data shows that respiratory viral loads are comparatively lower for the Omicron variant, despite its greater transmissibility (11)(12)(13). Our preliminary data suggest that while Omicron infected patients have similar individual viral loads with Delta and other variants, Omicron virus were detected more consistently in respiratory aerosols even in vaccinees -albeit limited by the small sample size and inability to adjust for confounders such as day of illness and clinical symptoms. Our data indicated no difference in the magnitude of aerosol emissions between the variants tested suggesting that it is not aerosol emissions per se that accounts for the dominance of the Omicron variant over delta. Rather, the more consistent presence of Omicron variant RNA in aerosols from all ve fully vaccinated patients studied, suggests that perhaps better adaptation to different hosts, enhanced receptor binding or greater immune and vaccine evasion, may explain its increased transmissibility (14)(15)(16)(17). vaccinated patients with the Omicron variant including one subject with a booster could generate detectable viral RNA copies in ne respiratory aerosols unlike two of our Delta patients. This suggests that vaccination alone is insu cient to stop secondary spread, as was seen in reports of secondary household transmission from vaccinated individuals infected with both variants (18)(19)(20). Nonetheless, broad community vaccination with effective vaccines will remain a key public health intervention to mitigate the impact of future waves on healthcare systems as they remain effective in preventing severe disease.
As observed previously, we also noticed a signi cant divergence of viral RNA copies between patients, with participants 6 and 9 accounting for 74.4% of the total detectable viral aerosols. This is consistent with the "super-spreader" phenomenon and inter-individual heterogeneity observed in COVID-19 transmission (21), where a minority of infected patients contributes to a majority of secondary transmission (22). This also needs to be further explored.

Conclusion
Our preliminary study underscores the transmission potential of Delta and Omicron variants via ne and coarse aerosols from infected patients during normal respiratory activities. The more consistent detection of Omicron viral RNA despite vaccination and minimal symptoms may contribute to its higher transmissibility. More detailed studies exploring molecular, physical, and host responses to these variants and sub-variants (20) will be critical to elucidate the mechanisms underpinning their increased transmissibility, and to formulate targeted public health interventions.