key: cord-0765819-pqn0qg3f
authors: Duchene, S.; Featherstone, L.; Freiesleben de Blasio, B.; Holmes, E. C.; Bohlin, J.; Pettersson, J. H.- O.
title: The impact of early public health interventions on SARS-CoV-2 transmission and evolution
date: 2020-11-18
journal: nan
DOI: 10.1101/2020.11.18.20233767
sha: b074ad566c04afb4f5e749a9a80b2b377e59c5c7
doc_id: 765819
cord_uid: pqn0qg3f

Many countries have attempted to control COVID-19 through the implementation of non-pharmaceutical interventions. However, it remains unclear how different control strategies have impacted SARS-CoV-2 virus transmission dynamics at the local level. Using complete SARS-CoV-2 genomes, we inferred the relative frequencies of virus importation and exportation, as well as virus transmission chain dynamics in Nordic countries - Denmark, Finland, Iceland, Norway and Sweden - during the first months of the pandemic. Our analyses revealed that Sweden experienced more numerous transmission chains, which tended to have more cases, and were of longer duration, a set of features that increased with time. Together with Denmark, Sweden was also a net exporter of SARS-CoV-2. Hence, Sweden effectively constituted an epidemiological and evolutionary 'refugia' that enabled the virus to maintain active transmission and spread to other geographic localities. This analysis highlights the utility of genomic surveillance where active transmission chain monitoring is a key metric.

Since its initial description in December 2019, SARS-CoV-2, the causative agent of 48 COVID19 1,2 , has rapidly led to an unprecedented global health crisis. The pandemic has 49 caused tens of millions of infections and over 1.3 million deaths worldwide and continues to 50 accelerate, imposing a significant impact on health care systems, societies and the global 51 economy. Countries are continuously struggling with how to effectively counteract the 52 pandemic, balancing the protection of health with social and economic considerations. In the 53 absence of therapeutics and available therapies and vaccines, efforts have centred on so-54 called 'non-pharmaceutical strategies', particularly initial short-term large-scale restrictions to 55 population movement (e.g. 'lock-downs'), increased testing and various levels of social 56 distancing. Analysis of the local epidemiological consequences of different control strategies 57 provides key information on the most effective approaches to reduce the rate of virus 58 transmission within communities. businesses continued almost as usual 3 . In contrast, Norway and Denmark enforced a more 67 invasive population movement restriction that included enforced home office for workers in 68 the public sector, home-bound schooling, targeted private sector close downs, as well as 69 closed international borders for non-residents. Iceland, a relatively small homogenous island 70 population (of ~350,000) never initiated a population movement restriction as Norway and 71 Denmark, but rather focused on large-scale testing and contact tracing to limit virus spread 72 within the community. In relation to population size, Sweden has had a higher number of 73 COVID19-related cases and deaths, than all other Nordic countries 4,5 , with a total of around 74 1450/58 cases/deaths per 100,000 people in Sweden, compared to around 951/13 in 75 Denmark, 428/5 in Norway, 321/7 in Finland, and 3 in Iceland, as of 30 th of October 2020 6 , 76 and predominantly occurring between April and May 2020. 77 78 Although the relative 'success' of COVID-19 control measures are normally gauged in the 79 number of cases and deaths at the country level, it is also the case that intervention and 80 mitigation strategies may lead to marked differences in transmission dynamics among 81 populations, which may in turn impact the evolution of the virus. Using a comparative 82 analysis of genome sequence data we addressed whether the different approaches to 83 COVID-19 control employed by the Nordic countries resulted in differences in virus 84 transmission dynamics and the relative frequencies of virus importation/exportation during 85 the first seven months of the pandemic. Accordingly, we compiled a representative data set 86 of SARS-CoV-2 virus genomes and performed a phylo-epidemiological study to identify any 87 differences in transmission chain dynamics between these countries.

We conducted a set of analyses on transmission chains of sampled genomes, also known as 91 "transmission lineages" 7 , which are defined as monophyletic groups of at least two genomes 92 sampled from a Nordic country. The estimated time of emergence of sampled transmission 93 chains provides information about the onset of community transmission, although such an 94 association is sensitive to sampling bias, particularly the time-scale of country-specific 95 sequencing efforts. Hence, we estimated that the first sampled transmission chains emerged 96 . CC-BY-NC-ND 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) preprint 3 from around January in Sweden (15 th January, 95% confidence interval, CI: 30 th December -97 16 th February) to late February in Norway (25 th February, 95% CI: 5 th February -7 th March) 98 (Table 1 ). In Norway the average date of emergence of all transmission chains was around 99 early June (2 nd June, 95% CI: 24 th May -13 th June), over a month later than in other 100 countries (Table 1 ). This likely reflects that there are fewer genome sequences from between 101 February and May -that is, during the first phase of the pandemic -than between June and 102 October in this country (Supplementary Figure S1 (Table 1 ). In particular, the H t -index for Sweden was 9 (95% CI: 8-9), 138 whereas those for other countries were 8 for Denmark (95% CI: 4-9), 7 for Finland (95% CI: Table 1 ).

For all non-Nordic countries, the number of genomes per country was subsampled via cd-hit 243 v.4.8.1 (https://github.com/weizhongli/cdhit) to include 99% of the total diversity per country, 244 apart from USA for which 97% clustering was applied. All genome sequences were aligned 245 . CC-BY-NC-ND 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-NC-ND 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) preprint

The copyright holder for this this version posted November 18, 2020. ; https://doi.org/10.1101/2020.11.18.20233767 doi: medRxiv preprint 

A new coronavirus associated with human respiratory disease in China

Genomic characterisation and epidemiology of 2019 novel coronavirus: 303 implications for virus origins and receptor binding

The first eight months of Sweden's COVID-19 strategy and the key 305 actions and actors that were involved

Lockdowns and COVID-19 Deaths in Scandinavia

Magnitude, demographics and dynamics of the effect of the first wave of 309 the COVID-19 pandemic on all-cause mortality in 21 industrialized countries

WHO Coronavirus Disease (COVID-19) Dashboard

Preliminary analysis of SARS-CoV-2 importation & establishment of UK 313 transmission lineages

A phylodynamic workflow to rapidly gain insights into the dispersal 315 history and dynamics of SARS-CoV-2 lineages

Spread of SARS-CoV-2 in the Icelandic Population

Estimating the effects of non-pharmaceutical interventions on COVID-321 19 in Europe

Genomic epidemiology reveals transmission patterns and 323 dynamics of SARS-CoV-2 in

Managing COVID-19 spread with voluntary public-327 health measures: Sweden as a case study for pandemic control

Spatio-Temporal Mutational Profile Appearances of Swedish SARS-CoV-2 330 during the Early Pandemic

IQ-TREE 2: New Models and Efficient Methods for Phylogenetic 332 Inference in the Genomic Era

Exploring the temporal 334 structure of heterochronous sequences using TempEst (formerly Path-O-Gen)

Fast Dating Using Least-Squares Criteria 337 and Algorithms