Comparative Aerosol and Surface Stability of SARS-CoV-2 Variants of Concern

SARS-CoV-2 transmits principally by air; contact and fomite transmission may also occur. Variants of concern are more transmissible than ancestral SARS-CoV-2. We found indications of possible increased aerosol and surface stability for early variants of concern, but not for the Delta and Omicron variants. Stability changes are unlikely to explain increased transmissibility.

For aerosol stability, we directly compared the exponential decay rate of different SARS-CoV-2 isolates (Table) by measuring virus titer at 0, 3, and 8 hours; the 8-hour time point was chosen through modeling to maximize information on decay rate given the observed 3-hour decay. We performed experiments as single runs (0-to-3 or 0-to-8 hours) and collected samples at start and finish to minimize virus loss and humidity changes from repeat sampling. We conducted all runs in triplicate. To estimate quantities of sampled virus, we analyzed air samples collected at 0, 3, or 8 hours postaerosolization by quantitative reverse transcription PCR for the SARS-CoV-2 envelope (E) gene to quantify the genome copies within the samples. To determine the remaining concentration of infectious SARS-CoV-2 virions, we titrated samples on standard Vero E6 cells. To check robustness, we also titrated the samples on 2 Vero E6 TMPRSS2-expressing lines, yielding similar results (Appendix). We estimated exponential decay of infectious virus relative to the amount of remaining genome copies to account for particle settling and other physical loss of viruses, although we also estimated decay rates from uncorrected titration data as a robustness check, which yielded similar results (Appendix).
We recovered viable SARS-CoV-2 virus from the drum for all VOCs (Figure 1, panel A). The quantity of viable virus decayed exponentially over time (Figure 1,   Omicron. That is, initial spike protein divergence from WA1 (heuristically quantified by the number of amino acid substitutions) appeared to produce increased relative stability, but further evolutionary divergence reverted stability back to that of WA1, or even below it (    Next, we investigated surface stability of VOCs compared with the ancestral variant on polyethylene. Again, all variants exhibited exponential decay (Figure 2, panel A). We found a half-life of 4.82 (4.23-5.49) hours for WA1, similar to our previous estimates ( Figure 2 In both aerosol and surface results, the posterior 95% credible intervals for most ratios overlap 1. Experimental noise could possibly explain the apparent trend toward increased stability for B.1, Alpha, and Beta, although the clear bulk of posterior probability mass indicates greater half-lives. Conversely, the posterior ratios indicate clearly that Delta and Omicron are not markedly more stable than WA1 and might be less stable (particularly Omicron and particularly on surfaces).

Conclusions
Several studies have analyzed the stability of SARS-CoV-2 on surfaces or in aerosols in a Goldberg rotating drum (3)(4)(5)(6)(7). Most have focused on the duration over which infectious virus could be detected. In this study, we paired a model-optimized experimental design with Bayesian hierarchical analysis to systematically measure virus half-life across 6 SARS-CoV-2 variants and directly estimate relative half-lives with full error propagation. We found a small initial increase in aerosol stability from ancestral WA1 to the B.1, Alpha, and Beta variants, with some statistical uncertainty. However, we found that Delta has a half-life similar to that of WA1 and that Omicron likely has a shorter one. In surface measurements, the VOCs followed the same pattern of relative stability, confirming that the overall stability of SARS-CoV-2 variants is determined by similar factors in aerosols and on surfaces (8). Divergent results on the aerosol and surface stability of VOCs have been reported (7,8).
Our study suggests that aerosol stability is likely not a major factor driving the increase in transmissibility observed with several VOCs (9,10). The early rise in stability for B.1 and its descendants Alpha and Beta might have arisen incidentally from selection for other viral traits that favored higher transmission. Epidemiologic and experimental studies suggest that the window for transmission is typically relatively short (<1 hour), and thus a modest change in aerosol half-life would not have discernible epidemiologic effects (11). However, in specific contexts of enclosed spaces, it will remain vital to understand the temporal profile of transmission risks after the release of aerosols containing SARS-CoV-2 from an infected person. We conducted our experiments under laboratory conditions using tissue culture media, so biological factors potentially affecting decay (e.g., airway mucins and other components of airway-lining fluids) were not considered. Novel approaches studying aerosol microenvironments have reported initial rapid loss of SARS-CoV-2 infectiousness in the seconds after aerosolization (12); our work only addresses SARS-CoV-2 decay and stability over longer timescales, after the initial deposition loss has occurred.
Whereas evolutionary selection for previous variants favored high transmission among immunologically naive humans (13), since late 2021, global population-level selection has favored antigenic change (14) and the consequent ability to transmit among nonnaive persons. Our findings suggest that increased transmissibility through antigenic evolution might come at a tolerable cost to the virus in environmental stability. Overall, the differences in environmental stability among different VOCs, in aerosols or on surfaces, are unlikely to be driving variant population-level epidemiology.