Non-pharmaceutical interventions and the emergence of pathogen variants

Abstract Non-pharmaceutical interventions (NPIs), such as social distancing and contact tracing, are important public health measures that can reduce pathogen transmission. In addition to playing a crucial role in suppressing transmission, NPIs influence pathogen evolution by mediating mutation supply, restricting the availability of susceptible hosts, and altering the strength of selection for novel variants. Yet it is unclear how NPIs might affect the emergence of novel variants that are able to escape pre-existing immunity (partially or fully), are more transmissible or cause greater mortality. We analyse a stochastic two-strain epidemiological model to determine how the strength and timing of NPIs affect the emergence of variants with similar or contrasting life-history characteristics to the wild type. We show that, while stronger and timelier NPIs generally reduce the likelihood of variant emergence, it is possible for more transmissible variants with high cross-immunity to have a greater probability of emerging at intermediate levels of NPIs. This is because intermediate levels of NPIs allow an epidemic of the wild type that is neither too small (facilitating high mutation supply), nor too large (leaving a large pool of susceptible hosts), to prevent a novel variant from becoming established in the host population. However, since one cannot predict the characteristics of a variant, the best strategy to prevent emergence is likely to be an implementation of strong, timely NPIs.

characteristics to the wildtype. We show that, while stronger and timelier NPIs generally 23 reduce the likelihood of variant emergence, it is possible for more transmissible variants 24 with high cross immunity to have a greater probability of emerging at intermediate levels of 25 NPIs. However, since one cannot predict the characteristics of a variant, the best strategy to 26 prevent emergence is likely to be implementation of strong, timely NPIs. 27 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

28
During the COVID-19 pandemic, a wide range of non-pharmaceutical interventions, 29 including mask wearing, quarantine, isolation, and lockdowns, have been used by 30 governments around the world to suppress virus transmission. Although considerable 31 efforts have been made to understand how such interventions affect transmission, much 32 less attention has been paid to their effects on pathogen evolution. While vaccines are well-33 known to affect virus evolution, non-pharmaceutical interventions also influence mutation 34 supply and the strength of selection, and hence play a key role in the emergence of novel 35 variants. Here, we use a relatively simple mathematical model to explore how non-36 pharmaceutical interventions during an epidemic affect the emergence of a novel variant of 37 a pathogen. We show that in general, it is important to implement effective transmission-38 reducing public health measures in a timely manner to prevent the emergence of novel 39 variants, which may be more transmissible, more deadly, or able to escape immunity. 40 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 28, 2021.

41
Understanding the impact of interventions for infectious disease control on pathogen 42 evolution is a major challenge at the interface of public health and evolutionary biology [1-43 3]. During the COVID-19 pandemic, a wide range of non-pharmaceutical interventions (NPIs) 44 have been implemented across the world, including social distancing, improved hygiene 45 practices, lockdowns, quarantining exposed or isolating infected individuals, and contact 46 tracing. Such measures have been credited with reducing cases and bringing epidemics 47 whose transmission rates are both equally affected by NPIs, as determined by the 88 parameter (0 ≤ ≤ 1). We assume that, for each strain, the baseline transmission rate is 89 ! ( ∈ { , }), the infection fatality rate (IFR) for single infections is ! (i.e. a proportion ! 90 of individuals die from infection, on average), and the average infectious period is 1/ . 91 Following a primary infection, an individual is assumed to be fully immune to the strain by 92 which they were infected, and to have acquired partial cross-immunity, , to the other 93 strain (0 ≤ ≤ 1). Thus, when = 0 there is no cross-immunity between strains and when 94 = 1 there is full cross-immunity. Immunity is assumed to act to reduce a host's 95 susceptibility to reinfection [20]. We assume that the variant differs from the wildtype at a 96 single genetic locus, and that mutations occur between strains at rate , leading to 97 coinfections by both strains. The IFR for coinfections is assumed to be the average of the 98 At time, , we use the following notation: ' is the number of hosts that are susceptible to 101 both the wildtype and the variant (initially all individuals are susceptible to both strains), ' ! 102 is the number of primary infections by strain , ' !( is the number of secondary infections by 103 strain following recovery from primary infection by strain , ' is the number of 104 coinfections, ' ! is the number of individuals who have recovered from a primary infection 105 by strain (and are immune to strain ), and ' the number of individuals who have 106 recovered from both strains (and are therefore immune to both strains). In the absence of 107 NPIs, the per-capita force of infection on fully susceptible hosts for strain is denoted ' ! = 108 . We assume that NPIs are triggered (and remain in 109 place for the rest of a simulation) when the total disease prevalence, ' ?
CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) We run simulations of the model using the direct method version of the Gillespie stochastic 120 simulation algorithm [23] using a population size of = 100,000 and initial condition 121 Simulations are terminated when the number of hosts infected reaches 0. 131 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

141
We consider a scenario in which the variant is initially rare relative to the wildtype. The 142 variant is fitter than the wildtype when ℛ ' & > ℛ ' % , which requires: 143 Hence, a variant is always fitter when it is at least as transmissible as the wildtype and there 148 is not full cross immunity ( < 1). However, a less transmissible variant may also be fitter 149 provided the wildtype has already infected a sufficient proportion of the population and 150 there is not full cross immunity. Clearly, when < 1 the fitness of the variant will increase 151 relative to the wildtype as the pool of susceptible hosts for the latter is depleted. Note that 152 NPIs do not affect whether a variant is fitter than the wildtype since there is no differential 153 effect on transmission. NPIs will, however, affect mutation supply (and hence the 154 appearance of the variant) and the rate at which the variant can spread (establishment). mutation rate is very high, it is unlikely that the variant will appear before the wildtype is 160 driven extinct. Thus, the probability of a variant emerging is close to 0 whenever > 1 − CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.27.21257938 doi: medRxiv preprint regardless of transmissibility or cross immunity (Fig. 2g- the probability of a variant emerging depends on the relative transmissibility of the variant, 163 the strength of cross immunity, and the strength of NPIs. When cross immunity and NPIs are 164 both relatively weak, there is a high probability of variants emerging (Fig. 2g-i). This is 165 because there is a high mutation supply (weak restrictions) to generate the variant and a 166 large pool of susceptible individuals to exploit. 167 168 When the variant is less or equally as transmissible as the wildtype, there is an inverse effect 169 of cross immunity and NPIs on the probability of the variant emerging: weaker NPIs require 170 stronger cross immunity, and vice versa, to prevent the variant emerging ( Fig. 2g-h). When 171 the variant is sufficiently more transmissible, however, it has a high probability of emerging 172 even when cross immunity is strong (Fig. 2i). Most notably, the variant is more likely to 173 emerge for intermediate NPIs when cross-immunity is very high or complete ( ≈ 1). This is 174 because when NPIs are weak ( ≪ 1), there is a large outbreak of the wildtype, which leads 175 to a high mutation supply but also rapidly depletes the pool of hosts for the variant due to 176 cross-immunity (Fig. 3a, d). Thus, while a variant is likely to appear, it is unlikely to spread 177 widely in the population due to herd immunity [24]. When NPIs are stronger (but not too 178 strong) the wildtype causes a smaller outbreak, leading to a lower build-up of cross 179 immunity in the population while still maintaining a sufficient mutation supply for the 180 variant to appear with high probability (Fig. 3b, e). This creates just the right conditions to 181 allow the variant to sweep into the population. As a result of this phenomenon, it is possible 182 for total deaths (for both strains) to peak at intermediate NPIs when the variant is both 183 more transmissible and more deadly than the wildtype (Fig. 4). 184 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.   . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.27.21257938 doi: medRxiv preprint DELAYED NPIS ( = 0.2) 205 We now consider the case where NPIs are delayed until 20% of the population is infected 206 (Fig. 5). There are three notable effects of delaying NPIs. First, since NPIs are delayed, a non-207 negligible number of deaths now occur over the full range of NPI strengths (Fig. 5d-f). 208 Second, as the initial mutation supply is no longer curtailed by NPIs, variants can emerge 209 during the early stages for > 1 − # / # " (Fig. 5i). Third, more transmissible variants which 210 experience sufficiently high cross immunity with the wildtype ( ≈ 1) are no longer likely to 211 emerge at intermediate NPIs (compare Fig. 2i and Fig. 5i). This is because cross immunity 212 accumulates in the population to prevent the more transmissible variant from establishing. 213 Consequently, there is no effect of varying the relative mortality on total deaths when NPIs 214 are delayed for a more transmissible variant with high cross immunity (Fig. S1). 215 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

224
The emergence of novel variants depends on the interaction between mutation supply and 225 the strength of selection, both of which are influenced by NPIs. Here, we have shown how 226 NPIs and the stage of an epidemic at which they are triggered affect the emergence of novel 227 variants with a range of life-history characteristics. Although stronger, earlier 228 implementation of NPIs generally reduces the likelihood that a novel variant will emerge, 229 more transmissible variants that exhibit a high degree of cross immunity with the wildtype 230 may be most likely to emerge when NPIs are enacted early but are of insufficient strength to 231 drive the wildtype extinct quickly (Fig. 2i, Fig. 3b). This echoes a theoretical result for 232 vaccination, which suggests that imperfect vaccination may provide the optimal conditions 233 for vaccine-escape variants to emerge [8,25]. This is because imperfect vaccination can 234 allow a significant mutation supply while also exerting selective pressure for vaccine-escape 235 variants. 236 237 Here, we found a similar result for NPIs. When NPIs are weak, a large outbreak of the 238 wildtype can occur. This allows a variant to appear (high mutation supply) but prevents it 239 from spreading widely due to the accumulation of cross immunity in the host population 240 (the variant cannot establish). When NPIs are at an intermediate level, however, there may 241 be sufficient mutation supply to allow the variant to appear but insufficient accumulation of 242 cross immunity from the wildtype. The higher transmissibility of the variant then facilitates 243 its establishment. If the variant is also sufficiently more virulent, then it is possible that this 244 will increase the overall number of deaths (Fig. 4). These patterns at high cross immunity 245 disappear when NPIs are delayed, but in general the variant is able to emerge over a wider 246 set of conditions compared to when NPIs are introduced quickly. 247 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.27.21257938 doi: medRxiv preprint While it is possible that intermediate strength or timely NPIs may occasionally lead to more 249 negative outcomes than weaker or delayed NPIs as described above, we note that this 250 requires the variant to have a specific set of characteristics. Since it is challenging to predict 251 the phenotypic characteristics of novel variants, we contend that the optimal strategy to 252 prevent variants of concern arising is almost always to ensure that NPIs are strong and 253 implemented in a timely fashion. If this is done, then the mutation supply is constrained, 254 preventing novel variants from appearing in the first place. In order to explore the potential effects of NPIs on the emergence of pathogen variants, we 268 made several simplifying assumptions in our modelling approach. First, we assumed that 269 there is no differential effect of NPIs on the wildtype and the variant transmission rates. 270 While this is often likely to be true, it is possible that NPIs may affect some variants more 271 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 28, 2021. Population heterogeneity and contact structure both affect pathogen transmission, and so 295 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.27.21257938 doi: medRxiv preprint are also likely to affect the initial spread of variants. We would therefore expect the 296 emergence probability to increase in some cases and decrease in others. However, the 297 qualitative patterns shown here are unlikely to change. Finally, increasing the genetic and 298 phenotypic landscape will tend to make it more difficult for variants to emerge. This is 299 because a greater number of mutations, some of which may be deleterious due to epistasis, 300 are likely to be required to generate the variant. One could crudely model this by reducing 301 the mutation supply in our model to mimic the lower rate of accumulating multiple 302 mutations, which would quantitatively, but not qualitatively, change our results. 303

304
In conclusion, NPIs have significant impacts on the emergence of novel variants by 305 mediating both the mutation supply and the strength of selection. Although stronger NPIs 306 generally reduce the probability that a new variant of concern will emerge, there are 307 circumstances -namely, when cross immunity is high and the variant is more transmissible 308 -where NPIs of intermediate strength can increase variant emergence, potentially leading 309 to a higher level of mortality. However, this requires a very particular set of circumstances 310 and one cannot predict where a variant will emerge in phenotype space (i.e. its 311 transmissibility, virulence, and level of cross immunity). The optimal strategy to prevent 312 variants emerging is therefore to ensure that NPIs are sufficiently strong to drive the 313 wildtype extinct, thereby cutting off the mutation supply. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 28, 2021. ; https://doi.org/10.1101/2021.05.27.21257938 doi: medRxiv preprint