Study population and setting
The goal of this study was to estimate the origin and spread of SARS-CoV-2 in Europe prior to spring 2020 border closures. This study applied a phylodynamic model with geographic structure to SARS-CoV-2 genomes from Hubei, China, and France, Germany, Italy, and a composite of 16 European countries. All genomic sequences were obtained from GISAID, a global science initiative providing open-access to genomic data. A random selection of 10 sequences from Hubei collected on or before January 23, 2020 were selected to represent the virus origin. Additionally, all available sequences from France, Germany, and Italy collected on or before March 8, 2020 were considered to understand emergence in Europe; and sequences from 16 other European countries collected pre-March 8 were down-sampled to scale with COVID deaths and outbreak size in each country. The study estimated the rate of new cases arising from migration compared with the rate of new cases arising from within region transmission within the European regions. The authors combined case count data and estimates for migration and transmission rates to provide a timeline of early SARS-CoV-2 introduction and spread before border closures were implemented in Europe.
Summary of Main Findings
Findings from this study suggest that the predominant lineage (A2a viruses) spreading in Europe prior to border closures had a most recent common ancestor in Italy between mid-January and early February 2020. The results suggest that this was likely seeded by an earlier transmission event in either Hubei, China, or Germany. These findings suggest that SARS-CoV-2 was introduced from Hubei into France, Germany, Italy, and other European countries approximately two to four times each before March 8, 2020, and that there were approximately 18 estimated introductions from Italy to other European countries from these samples. Prior to the first border closures in Europe, the authors estimated that the rate of occurrence of new cases from within-country transmission was within the bounds of the estimated rate of new cases from migration. Shortly after the first evidence of sustained within-region transmission in Italy, outbreaks in the rest of Europe emerged.
This study linked a model of viral evolution to describe the mutational process of SAR-CoV-2 to an epidemiological model describing the transmission process, and was able to construct a timeline of early transmission events prior to border closures in Europe.
There are several limitations of this study, which the authors acknowledge and attempted to mitigate. Data from early cases identified represented those that were identified by ministries of health, and therefore were not a random sample. However, the model assumed uniform and random sampling of the total infected population in each region and therefore may be subject to bias. The analysis did not consider introduction of cases from any non-European regions other than Hubei; therefore, authors would not have been able to capture transmission from Hubei to Europe through an intermediate country of region. Under-sampling and biased sampling limit the inferences about migrations, and therefore migrations estimated in this analysis give a sense of general dynamics and the number of total migrations rather than provide precise estimates.
This study provides evidence suggesting the presence of sustained community transmission within European countries prior to border closures. These results suggest that border closure policies alone would not have prevented continued outbreaks within European countries.
This review was posted on: 22 March 2021