Epidemiological and microbiological investigation of a large increase in vibriosis, northern Europe, 2018

Background Vibriosis cases in Northern European countries and countries bordering the Baltic Sea increased during heatwaves in 2014 and 2018. Aim We describe the epidemiology of vibriosis and the genetic diversity of Vibrio spp. isolates from Norway, Sweden, Denmark, Finland, Poland and Estonia in 2018, a year with an exceptionally warm summer. Methods In a retrospective study, we analysed demographics, geographical distribution, seasonality, causative species and severity of non-travel-related vibriosis cases in 2018. Data sources included surveillance systems, national laboratory notification databases and/or nationwide surveys to public health microbiology laboratories. Moreover, we performed whole genome sequencing and multilocus sequence typing of available isolates from 2014 to 2018 to map their genetic diversity. Results In 2018, we identified 445 non-travel-related vibriosis cases in the study countries, considerably more than the median of 126 cases between 2014 and 2017 (range: 87–272). The main reported mode of transmission was exposure to seawater. We observed a species-specific geographical disparity of vibriosis cases across the Nordic-Baltic region. Severe vibriosis was associated with infections caused by Vibrio vulnificus (adjOR: 17.2; 95% CI: 3.3–90.5) or Vibrio parahaemolyticus (adjOR: 2.1; 95% CI: 1.0–4.5), age ≥ 65 years (65–79 years: adjOR: 3.9; 95% CI: 1.7–8.7; ≥ 80 years: adjOR: 15.5; 95% CI: 4.4–54.3) or acquiring infections during summer (adjOR: 5.1; 95% CI: 2.4–10.9). Although phylogenetic analysis revealed diversity between Vibrio spp. isolates, two V. vulnificus clusters were identified. Conclusion Shared sentinel surveillance for vibriosis during summer may be valuable to monitor this emerging public health issue.


Introduction
The habitat of Vibrio bacteria is fresh and brackish water with moderate salinity. Nontoxigenic Vibrio cholerae, as well as several human pathogenic non-cholera Vibrio species, including Vibrio alginolyticus, Vibrio parahaemolyticus and Vibrio vulnificus, cause vibriosis after seawater exposure or consumption of contaminated seafood [1]. Clinical manifestations range from mild gastroenteritis and otitis to wound infections that may lead to severe necrotizing fasciitis and septicaemia with a potentially fatal outcome [2][3][4][5].
The Baltic Sea region is one of the areas where increasing numbers of cases related to Vibrio species causing vibriosis (VCV) have been reported in the last decades [6]. Several studies have shown how the occurrence of heatwaves, that lead to an increase of sea surface temperature, are linked with an increase of the number of reported vibriosis cases [4,[7][8][9][10][11][12].
For instance, 2006, 2010, and particularly 2014 (the warmest year in historical records to that date), were the years with a significantly warm summer in the Baltic Sea region, and were also the years with the largest number of vibriosis cases reported [6,11]. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 4 However, there is a notable gap in surveillance data of vibriosis [1,6], since it is not a notifiable disease in the majority of European countries. Therefore, the aim of this multicountry study was to describe the epidemiology of vibriosis cases in countries bordering the North and Baltic Seas area during the most recent exceptionally warm year of 2018 [13, 14], in order to investigate the extension of these infections in the study countries, map their genetic diversity, understand the predictors for developing severe vibriosis, and to propose recommendations for public health measures.

Study design and case definition
We conducted a retrospective cross-sectional study to analyse the epidemiology of VCV infections reported in 2018 in Norway, Denmark, Sweden, Finland, Poland, Estonia, and Latvia, further referred to as the study countries. Available data of vibriosis cases since the last warmest summer (2014) were used to contextualize the number of VCV infections of 2018.
A case of vibriosis was defined as a laboratory confirmed VCV infection from the study countries; those related to travel were excluded. For few cases (n=18) more than one Vibrio species were recorded concurrently in the same patient. In such cases only the species and sample type related to a more severe infection was included.

Data source and collection
Each country used different data sources including compulsory comprehensive passive surveillance systems for vibriosis (Sweden, Finland, Poland, and Estonia), national laboratory notification databases (Denmark and Latvia) or nationwide surveys to public health microbiology laboratories (Norway) ( Table S1). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 5 The reporting criteria varied between countries that had a surveillance system in place in 2018 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Epidemiological investigation and statistical analysis
We described the epidemiology of vibriosis cases reported in 2018 in the study countries per country and as total counts. Data presented included the sex-ratio, notification rate per 100,000 inhabitants, median age, distribution of cases across age groups, season and identified VCV (Table S2). Case numbers were presented by country and by region (NUTS3) while seasonality was further depicted by month of infection. Severity of infection was described by age group, and month of infection. Association of sex, age group, season, and VCV with developing severe vibriosis was further analysed by estimation of crude odds ratios (OR) and 95% confidence intervals (CI) by univariate logistic regression analysis. Adjusted OR (adjOR) with 95% CI were estimated in a multivariable analysis. Binary outcome was severe/non-severe vibriosis.
Data analysis was performed using Stata version 15.0 (2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LP. USA). Categorical variables were described as proportions with 95% CI and were compared using chi-squared test. Continuous variables were described using mean and standard deviation or median and range, and were compared using t-test or non-parametric Wilcoxon rank-sum test. Trends were assessed using a nonparametric test across ordered groups. Observations with missing values for variables under comparison were excluded from the respective analysis.
We used an alpha level of 0.05 for all statistical tests. Stata outputs of p-values p<0.000 are reported as p<0.001. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Sampling of VCV isolates, MLST and WGS analyses
The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 7 We collected available clinical VCV isolates in 2018 from the national public health institutes or regional laboratories, and complemented them with available clinical (2014-2017) and environmental (2018) isolates (Table S3). DNA was extracted and sequenced using standard operating procedures and Illumina sequencers. WGS raw files are available at the European Nucleotide Archive (https://www.ebi.ac.uk/ena) under study project accession number PRJEB43461. Accession numbers of all sequenced isolates are listed in Table S4.
Raw reads from each country were analysed together using a common pipeline for species identification, MLST, and phylogenetic analyses. We used BBmap (version 38.69) to clean the raw reads and FastQC (version 0.11.8) to generate quality reports of samples. Additionally, we used Kraken2 (version 2.0.8_beta) to confirm the species and Shovill (1.0.9) to assemble (using SPAdes version 3.13.1) the genomes.
We used Parsnp (v1.2) and a neighbour-joining algorithm to build the phylogenetic trees, and Snp-dists (0.7.0) to calculate the single nucleotide polymorphism (SNP) distance between isolates. A cluster was defined as two or more isolates within 30 SNPs difference.
An in-house pipeline was used for sequence mapping, generation of consensus sequences, alignment calculation, and SNP filtering (exclusion distance = 300). We used R package ggtree [16] to visualise the phylogenetic trees generated by the in-house pipeline (https://github.com/folkehelseinstituttet/Vibrio-Project).

Results
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Descriptive epidemiology of vibriosis cases
In 2018, 445 non-travel related cases of vibriosis were reported in the study countries, which was the highest single year case number compared to the four previous years (n=610) ( Figure 1A, Table 1 and S2).  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The vibriosis notification rates ranged between 0.5 in Finland and 2.9 per 100,000 inhabitants in Denmark. Due to limited number of cases (n=6), the notification rate was not calculated for Poland and Estonia. Latvia reported no cases (Table S2). The majority of the cases were male (n=277, 62.2%) ( Table 1) and the highest number of cases was reported in the age group 65-79 (n=109, 24.5%) followed by age groups 5-14 (n=91, 20.4%) and 45-64 years old (n=83, 18.7%) ( Figure 1B, Table 1).
We observed a difference in the proportions of species affecting each age group. The proportions of V. vulnificus and V. parahaemolyticus infections followed an upward trend . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Geographic distribution of vibriosis cases
The geographic distribution of the vibriosis cases differed between Vibrio species (Figure 2). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 11 and V. parahaemolyticus infections, mainly occurred in the coastal regions around the connecting sounds between the Baltic and the North Sea, particularly Oslo fjord in Norway, southwest Sweden and eastern Denmark ( Figure 2D).

Severity of Vibrio infections
The proportion of severe VCV infections increased significantly with increasing age (p<0.001) and it differed by VCV (p<0.001) ( Figure 3A and Table 2). We observed the highest proportion of severe infections for V. vulnificus (95.6%) and V. parahaemolyticus (65.2%), while these were lower yet substantial for non-toxigenic V. cholerae (33.0%) and V. alginolyticus (31.6%) ( Table 1 and 2). The exposure of these severe infections with non-toxigenic V. cholerae and V. alginolyticus was largely unknown (48% and 70%, respectively) or cases were exposed to seawater/bathing (48% and 25%, respectively).
In terms of age, these infections were shifted slightly towards the younger age group. On the contrary, more than 70% of the severe V. vulnificus and V. parahaemolyticus infections occurred in the age groups 65-79 and 80+, while it was 58% and 41% of the severe V.
cholerae and V. alginolyticus cases respectively that belonged to these age groups. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint All VCV were more frequently reported in summer, when the majority of cases occurred (n=326, 73.3%; ranging per species from 63.8% to 97.8%) ( Table 1). No difference in the seasonal distribution of vibriosis cases was observed between countries ( Figure 3B, Table   S2). According to our multivariable model, the likelihood of developing a severe infection was significantly increased among the elderly (65-79 years: adjOR=3.9; 95% CI: 1.

Microbiological and molecular investigations
We analysed whole genome sequences of 135 clinical Vibrio isolates isolated in 2018.
Additionally, we included 16 available clinical isolates from travel related cases, 14 clinical . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 13 isolates from 2014-2017 period and 13 Finnish environmental non-toxigenic V. cholerae isolates to investigate the genetic diversity of Vibrio in the study countries (Table S3).

Phylogenetic analysis
SNP analysis showed a high diversity of isolates for all species with several clusters of nontravel related cases (Figure 4 and 5, S1-S2).  Nine clusters with two or three cases each of non-toxigenic V. cholerae isolates (≤30 SNPs difference) were identified in Sweden (n=4), Sweden/Finland (n=4), and Finland (n=1) ( Figure   4). Cases whose isolates clustered were sampled close in time (median 7 days; range 2-86 days) but detailed information on place of infection was not available. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 14 Additionally, two clusters of V. vulnificus isolates with <10 SNPs difference were detected ( Figure 5): one cluster with nine isolates in Norway, where the cases had been infected within 40 days and about 60 kilometres apart, and one cluster with two isolates in Sweden, where the cases had been infected 30 days and about 55 kilometres apart.

MLST analysis
Within the 178 isolates included in this study, 20 groups of isolates with the same ST were identified. Of these, ten groups were pairs of isolates from a single country (Sweden, Norway or Finland), three were pairs from two countries (Denmark/Sweden, n=1, or Sweden/Finland, n=2), six included 3-4 isolates each, and the largest group of nine V.
vulnificus isolates (ST534) was detected in Norway ( Figure 4 and 5, Table S4). Finally, a single V. parahaemolyticus isolate from Norway, found in a gastrointestinal infection in spring of 2014, was identified as the pandemic ST3 (Table S4, Figure S3).

Discussion
Our study provides a detailed overview on the occurrence of vibriosis in the Nordic and Baltic Sea regions in 2018. In context of epidemiological and microbiological findings as well as conducted studies from 2014-2018 [11,12], our results highlight the importance of vibriosis as a concern to public health in this geographic area. Even though the data have been collected using different systems, the study countries reported similar patterns in is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 15 60 degrees North latitude in Finland, which, to the best of our knowledge, is the highest latitude northward at which V. vulnificus has been reported. These findings underline a concern about the spread of this pathogen due to seawater warming [18].
Our results from 2018 confirm this pattern with the majority of infections occurring in summer months. Additionally, in this study almost half of all infections reported in 2018 were categorised as severe infections that also mainly occurred during summer season.
Mild ear infections may have long reporting delays up to months until a patient seeks medical care [19,20] compared with rapidly developing severe blood or wound infections.
We observed a similar pattern for the cases in this study, which could explain why reporting of mild vibriosis stretched more into autumn and winter and reporting of severe infections concentrated in summer months. More accurate information on the probable infection date would be needed to confirm this hypothesis. The likely source of infection was available for a subset of cases suggesting that the mode of transmission was mostly through seawater rather than through consumption of contaminated seafood.
The majority of vibriosis cases in the study countries were domestic and males were more frequently affected than females, consistent with other reports [12,21]. Even though the majority of cases were present among adults, about a fifth of the detected cases were among children up to 14 years of age, who mostly had ear infections and mild vibriosis; severe infections on the other hand were found to be associated with increasing age. This is likely due to underlying conditions being overrepresented among elderly people. In addition to increasing age, we also found that being infected by V. vulnificus or V. parahaemolyticus was a risk factor for a more severe VCV infection likely due to the greater pathogenicity of these microorganisms [1,2]. On the other hand, despite V. cholerae and V. alginolyticus is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint 16 predominantly causing mild infections, in our study, about one third of cases infected by these species were sampled from blood/serum or wounds. Thus, in absence of systematic data on hospitalisation and symptoms, these infections were also considered as severe.
These cases were of a lower median age compared to V. vulnificus and V. parahaemolyticus infections, and the exposure was largely unknown with only some cases exposed to seawater/bathing. Given a substantial proportion of cases classified as severe, V. cholerae and V. vulnificus should therefore not be underestimated in vibriosis diagnosis, as was also pointed out previously [22].
We observed a geographic disparity in the distribution of VCV in the study countries. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint clusters of V. cholerae, containing two to three isolates each, without a clear epidemiological link. The same V. cholerae strains detected in one or more countries might be due to common exposure to contaminated seafood or environmental spread of clones through e.g. Some limitations apply to our investigation. There were differences in data sources and data availability between the study countries. Notification rates should therefore be compared carefully as vibriosis is not notifiable in all study countries or not for all species. Especially mild infections might have been reported with a delay and/or underreported. Conversely, in some cases a disease could have been misclassified as vibriosis when the identified Vibrio species were merely opportunistic microorganisms present at the site of infection. Case severity classification used in this analysis was not reported directly in any study country, but was inferred based on the sample type. Additionally, cases without known travel history were considered as non-travel related, which could have potentially led to an overestimation of vibriosis cases in the Baltic Sea region. Furthermore, the place of residence was used as proxy when place of infection was not available. Regarding the molecular findings, SNP analysis needs to be carefully evaluated since recombination is one of the major sources of genomic changes in Vibrio. Therefore, the removal of changes caused by recombination could have provided a better insight from the evolutionary perspective. Finally, laboratory . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. ; https://doi.org/10.1101/2021.11.19.21266449 doi: medRxiv preprint methodology, capacity and priorities to diagnose and report VCV infections likely differed among the study countries.
During our investigation, we have performed a systematic and consistent analysis of epidemiological data from different countries, and we have combined it with the genomic analysis of strains from cases to achieve a comprehensive understanding of the occurrence of VCV infections in this affected region. Despite the low incidence, severe VCV infections are clinically costly [33] and climate changing predictions as well as population and socioeconomic projections for the upcoming years suggest that they are likely to increase in the future due to more favourable growth conditions for VCV [18,34]. It is therefore of interest to detect and report the VCV infections in countries bordering the Baltic Sea and connecting regions to the North Sea to further monitor the situation, especially during summer heatwaves. Moreover, such surveillance would facilitate risk assessments and allow for targeted interventions, including risk communication to raise awareness between clinicians and populations at risk towards vibriosis. Thereby, countries without comprehensive surveillance could benefit from establishing or expanding dedicated surveillance systems to detect and prevent vibriosis cases. In particular, a shared sentinel system during summer months might be highly valuable. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted November 21, 2021.