Molecular epidemiology of invasive Group A Streptococcal infections before and after the COVID-19 pandemic in Switzerland

Group A Streptococcus, GAS, COVID-19, emm-type, FT-IR spectroscopy, whole 36 genome sequencing, WGS 37 . CC-BY-NC-ND 4.0 International licenseIt 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 (which was not certified by peer review) The copyright holder for this preprint this version posted April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 3

38 Group A Streptococcus (GAS, aka Streptococcus pyogenes ) poses a significant public health 39 concern, causing a diverse spectrum of infections with high mortality rates. Following the 40 COVID-19 pandemic, a resurgence of invasive GAS (iGAS) infections has been documented, 41 necessitating efficient outbreak detection methods. Whole genome sequencing (WGS) serves 42 as the gold standard for GAS molecular typing, albeit constrained by time and costs. This 43 study aimed to characterize the post-pandemic increased prevalence of iGAS on the molecular 44 epidemiological level in order to assess whether new, more virulent variants have emerged, as 45 well as to assess the performance of the rapid and cost-effective Fourier-transform infrared 46 (FT-IR) spectroscopy as an alternative to WGS for detecting and characterizing GAS 47 transmission routes. A total of 66 iGAS strains isolated from nine Swiss hospitals during the 48 COVID-19 post-pandemic increased GAS prevalence were evaluated and compared to 15 49 strains collected before and 12 during the COVID-19 pandemic. FT-IR measurements and 50 WGS were conducted for network analysis. Demographic, clinical, and epidemiological data 51 were collected. Skin and soft tissue infection was the most common diagnosis, followed by 52 primary bacteremia and pneumonia. Viral co-infections were found in 25% of cases and were 53 significantly associated with more severe disease requiring intensive care unit admission. 54 WGS analysis did not reveal emerging GAS genetic distinct variants after the COVID-19 55 pandemic, indicating the absence of a pandemic-induced shift. FT-IR spectroscopy exhibited 56

in differentiating genetically distant GAS strains, yielding poor overlap with 57 WGS-derived clusters. The emm1/ST28 gebotype was predominant in our cohort and was 58 associated with five of the seven deaths recorded, in accordance with the molecular 59 epidemiological data before the pandemic. Additionally, no notable shift in antibiotic 60 susceptibility patterns was observed. Our data suggest that mainly non-pathogen related 61 factors contributed to the recent increased prevalence of iGAS. 62 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 4

63 Group A Streptococcus (GAS, aka Streptococcus pyogenes) is a pathobiont that causes a wide 64 spectrum of infections, ranging from pharyngitis and superficial skin infections to severe 65 invasive infections, such as necrotizing fasciitis, pneumonia or meningitis associated with 66 high mortality and morbidity rates, with 500,000 estimated deaths annually worldwide [1-3]. 67 After a historically low incidence of GAS infections during the COVID-19 pandemic, an 68 alarmingly increasing incidence of invasive GAS (iGAS) infections has been reported, since 69 October 2022, in many European countries and in the USA, mostly among children under 10 70 years of age, including many fatalities [4-7]. Also in Switzerland, since November 2022, a 71 fourfold increase of the registered iGAS infections in children, including four deaths, has been 72 reported compared to the pre- COVID-19 pandemic era [8]. So far, there is no evidence for 73 the emergence of a more virulent clone [9, 10]. In the United Kingdom the pre-pandemic 74 predominant M1uk clone was detected in the majority of the GAS strains isolated during the 75 December 2022 outbreak [9]. The causes of the resurgence of iGAS remain somewhat 76 elusive. 77 The detection and management of outbreaks, particularly in the case of GAS, rely on 78 understanding transmission routes and genetic diversity. Molecular typing by whole genome 79 sequencing (WGS) or conventional (Sanger) sequencing of the hypervariable N-terminal 80 region of the emm gene, encoding the surface M protein [11], is crucial for epidemiological 81 studies and outbreak investigations, although it is time-consuming, expensive and requires 82 expertise [12, 13]. Alternatively, Fourier-transform infrared (FT-IR) spectroscopy may 83 provide a cost-effective approach for outbreak analysis by generating a biochemical 84 fingerprint of the bacterial composition using the full infrared spectrum (4000-400 cm −1), 85 albeit with variable cluster defining cut-offs among different bacterial species and 86 dependencies on growth conditions [13-19]. 87 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 5 The aim of this study was to evaluate whether any GAS genomic variations had emerged. To 88 this end, we compared iGAS strains isolated before, during, and after the COVID-19 89 pandemic (from January 2013 until May 2023) from patients hospitalized in nine hospitals in 90 the Regions of Zurich, Graubünden and Geneva using the current gold standard WGS. 91 Furthermore, we aimed to assess whether FT-IR spectroscopy is a reliable alternative for a 92 time-efficient detection of GAS outbreaks compared to WGS. 93 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 6

94 Study period, population, clinical and epidemiological information 95 The study encompassed 93 strains sourced from the same number of patients with GAS 96 infection. Among these, 15 strains were retrospectively identified between January 2013 and 97 December 2019, 18 strains were retrieved from the COVID-19 pandemic period between 98 February 2020 and September 2022. The remaining 60 strains were prospectively identified 99 during the latest iGAS resurgence (October 2022 to May 2023) from patients hospitalized in 100 nine Swiss hospitals: University Hospital Zurich, University Hospital of Geneva, City 101 Hospital of Zurich, Cantonal Hospital of Graubünden, Cantonal Hospital of Winterthur, 102 Regional Hospitals of Wetzikon, Uster, Limmatal, and Männedorf. These hospitals 103 collectively serve an area inhabited by approximately 2 million individuals. Only three of the 104 participating hospitals admitted pediatric patients. Two patients did not have an invasive 105 infection (GAS_036 and GAS_037), but were included in the study because they were related 106 to a patient who died from iGAS infection (GAS_035). Basic demographical, clinical, and 107 epidemiological data were collected for 74 patients from electronic health records, including 108 sex, age, diagnosis, co-infections, sampling date and site, as well as the geographical location 109 of the sampling. Outcomes included admission to the intensive care unit (ICU), intubation and 110 death related to iGAS. No clinical data were available from the Regional Hospitals of 111 Limmatal and Männedorf. 112 In order to determine the prevalence of GAS over the years, the number of all GAS strains 113 detected by culture or molecular techniques in patients admitted to the University Hospital of 114 Zurich between January 2012 and December 2023 was collected. The GAS strains originating 115 from blood, joint punctures, cerebrospinal fluid, bronchoalveolar lavage, pleura fluid and soft 116 tissue were defined as invasive. The definition of the Centre for Disease Control regarding 117 invasive infection was adopted [20]. 118 119 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 7 Ethics approval 120 This study was approved by the Ethic Committee of Nordwest- and Central Switzerland with 121 the agreement of all Swiss Ethical Committees (BASEC-ID: 2019-01291 as part of the Swiss 122 Pathogen Surveillance Platform, www.spsp.ch). Patients declining to sign a general consent or 123 any other declining statement against using data for research purposes were excluded. 124 125 Bacterial strains 126 93 GAS strains were collected from the same number of patients hospitalized in the nine 127 hospitals mentioned above and stored at -80°C. Species identification of all samples was 128 carried out using Matrix Assisted Laser-Desorption Ionization – Time of Flight (MALDI-129 TOF) (MBT Compass 4.1, Bruker Daltonics, Bremen, Germany) using the database BDAL 130 DB.13. 131 132 Sample preparation for FT-IR spectroscopy and WGS 133 Three to five macroscopically identical colonies were subcultured on 5 % sheep blood agar 134 plates (Biomérieux) for 24±2 hours at 37°C in a static incubator in the presence of 5% CO 2. 135 For FT-IR spectroscopy, a 1 μ l-loop was overloaded with bacteria cells that were then 136 resuspended in 50 μ l ethanol solution (70% v/v) in a 1.5 ml tube containing metal rods. The 137 suspension was then vortexed extensively and 50 μ l of deionized water were added, followed 138 by another round of vortexing. Fifteen microliters of the homogenized bacterial suspension 139 were placed on a 96-spot silicon plate (Bruker Daltonics, Bremen, Germany) in four technical 140 replicates per bacterial isolate. The plates were dried at 37°C for approximately 20 minutes 141 followed by infrared measurements. For WGS, a lawn was streaked from frozen stocks and 142 samples were processed as previously described [21, 22]. 143 144 Infrared measurements 145 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 8 An IR-Biotyper (Bruker Daltonics) was used for all measurements according to the 146 manufacturer’s instructions and as previously described [19]. In brief, 12 µl of test standard 1 147 and 2 (ITRS 1/ ITRS 2) were placed on the 96-spot silica plate and used as controls. The 148 samples were measured using the OPUS software v8.2.28 (Bruker Daltonics), detecting 149 carbohydrates spectra within the wavenumber range of 1200 and 900 cm -1. Data analysis was 150 performed using the same software. Isolates with three or more valid replicates in the same 151 run were included in the analysis. 152 153 Whole genome sequencing of isolates and comparative genomics 154 Whole genome sequencing of the clinical isolates was performed at the Institute for Medical 155 Microbiology, University of Zurich, Switzerland as previously described [21]. Emm-typing 156 was extrapolated from the WGS data using the emm-typing tool [23] and MLST typing was 157 inferred from the WGS data using MLST v.2.7.6 [24], which makes use of the PubMLST 158 website (https://pubmlst.org/) developed by Keith Jolley [25] and sited at the University of 159 Oxford. The development of that website was funded by the Wellcome Trust. 160 De novo assemblies were built with SPAdes v.3.14.1 using the –careful option [26]. 161 Annotation of the resulting assemblies was performed with Prokka v1.14.6 [27]. The created 162 GFF files were used as input to Roary v3.13.0 to construct the pangenome, for which we used 163 the command -e -mafft that generates a multi-fasta alignment of the core genes [28]. The 164 average nucleotide identity (ANI) was inferred with pyani v0.2.12 c hoosing the method 165 ANIm, which makes use of MUMmer / NUCmer to align the input sequences [29]. 166 Phylogenetic trees were inferred from the aligned core genome using IQTREE [30] under the 167 GTR+G+I model. Trees and metadata were processed and plotted using the R-packages ape 168 [31] and ggtree [32]. 169 The overlap between clusters derived from FT-IR and whole bacterial genome data 170 (specifically Average Nucleotide Identity, ANI) was quantified as outlined in Scheier et al 171 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 9 [19] . In brief FT-IR clusters were inferred by cutting the hierarchical (UPGMA) clustering 172 tree derived from the Euclidean distance of the FT-IR spectra at a given height, which was 173 varied over a broad range. ANI-based clusters were defined as the components of the network 174 obtained by connecting those pairs of nodes whose ANI was above a given threshold. We 175 used the v-measure to quantify the overlap between these two networks. 176 The genomic data with matched phenotypic information can be downloaded here ( National 177 Center for Biotechnology Information (nih.gov)). 178 179 Antibiotic susceptibility testing 180 Antibiotic susceptibility testing to penicillin, clindamycin and erythromycin was conducted on 181 pandemic and post-pandemic GAS strains as part of the routine microbiological assessment. 182 The Kirby-Bauer method was perfomed in line with the recommendations of the European 183 Committee on Antimicrobial Susceptibility Testing (EUCAST) in the ISO accredited Institute 184 of Molecular Microbiology (IMM) in Zurich. Briefly, a 0.5 McFarland standard inoculum 185 preparation was used. Mueller-Hinton agar was supplemented with 5% defibrinated horse 186 blood and 20 mg/L β -NAD (MH-F agar). After evenly inoculating the agar with the bacterial 187 suspension, antimicrobial disks were applied to the agar surface. To detect inducible 188 clindamycin resistance, erythromycin and clindamycin disks were placed 12-16 mm apart 189 (edge to edge). The plates were incubated at 35 ± 1°C in with 4-6% CO 2 for 18 ± 2 hours, 190 after which zones of inhibition were measured to determine susceptibility based on EUCAST 191 breakpoint tables (version 14.0). 192 193 Data Analysis 194 For the patients’ clinical characteristics we performed a descriptive statistical analysis. 195 Continuous variables were presented as mean with range, except for age which was presented 196 as median with range. Categorical variables were presented as frequency tables. For the 197 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 10 association of the patient-related factors with the severity of the disease we chose as 198 dependent variables the ICU admission and the all-cause in-hospital mortality and performed 199 a univariable analysis. Due to the rare event problem for the case of obesity (4/74 incidences), 200 we collapsed this variable with two other related medical conditions, diabetes mellitus and 201 cardiovascular disease (CVD), based on their clinical relevance and exhibited overlap. 202 Variables that demonstrated significant (or marginally significant) associations (p<0.07) in 203 the univariable analysis were taken to multivariable binary logistic regression analysis. The 204 significance level was set at p<0.05. The model fit was evaluated using the Hosmer-205 Lemeshow test. All analyses were performed with IBM SPSS Statistics V26 (IBM Corp., 206 Armonk, NY, USA). 207 208 Funding information 209 This study was supported by the University Of Zurich CRPP Personalized Medicine Of 210 Persisting Bacterial Infections Aiming to Optimize Treatment and Outcome to S.D.B, and 211 A.S.Z.; and by grant (INOV00121) from University Hospital Zurich to T.C.S. 212 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 11

213 Report of laboratory-confirmed GAS cases at the University Hospital Zurich between January 214 2012 and December 2023 215 The total number of iGAS strains detected with cultures or molecular techniques from January 216 2012 to December 2023 reported at the University Hospital Zurich is shown in Figure 1. GAS 217 infections typically have a seasonal pattern, showing a peak during late spring in the northern 218 hemisphere [33]. During the COVID-19 pandemic (from March 2020 until the suspension of 219 the pandemic prevention methods in Switzerland, in March 2022), the detection of GAS cases 220 declined compared to the previous years (Suppl. Figure 1) and the number of iGAS was also 221 historically low (Figure 1). In the first quarter of 2023, there was a notable surge in invasive 222 and non-invasive GAS cases, marking a threefold increase compared to the same month 223 average in pre-pandemic years (Figure 1, Suppl. Figure 1). The number of isolated GAS 224 strains remained unusually high throughout the first post-pandemic year. 225 226 Demographic and clinical characteristics of the patients with iGAS infections 227 93 patients, from whom the 93 GAS strains were isolated, were included in this study. The 228 clinical data of 74 patients hospitalized in seven Swiss hospitals were included. The clinical 229 data of the remaining 19 patients (14 from the retrospectively collected strains and five from 230 the prospectively collected ones) could not be retrieved. 44 patients (59.5%) were male with a 231 median age of 47.5 years (range 1 to 94 years), nine (6.5%) were under 18 years of age. 232 Cardiovascular disease (35.1%), diabetes mellitus (17.6%), and solid tumors (9.5%) were the 233 most common comorbidities. The most common diagnosis of infection was skin and soft 234 tissue infection (SSTI) with bacteremia (18.9%), whereas necrotizing fasciitis was diagnosed 235 in five (6.8%) patients. The primary sampling site, representing the majority of cases (64.9%), 236 was blood, followed by tissue biopsies (13.5%). Notably, 24.3% of cases featured viral co-237 infections; six patients (8.2%) were diagnosed with influenza and three (4.1%) with SARS-238 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 12 CoV-2, three (4.1%) with a co-infection with SARS-CoV-2 and influenza and three (4.1%) 239 with respiratory syncytial virus (RSV). Two patients were diagnosed with varicella zoster 240 virus (VZV) reactivation. Fifteen patients (20.3%) suffered from toxic shock syndrome (TSS). 241 A large number required ICU admission (43.2%) and intubation (29.7%). Surgical 242 debridement was performed in 39 patients (52.7%). Combination antimicrobial therapy was 243 employed in 38 (51.4%) of cases; in 27 of them (71%) clindamycin was co-administered with 244 another antibiotic. Immunoglobulin was administered to all fifteen patients with TSS (20.3%). 245 The average hospitalization duration was 15.6 days, (range 0 - 100 days). We observed an all-246 cause mortality rate of 9.5%, with 6 out of 7 deaths specifically attributed to iGAS. The 247 demographic and clinical characteristics of the patients are summarized in Table 1. 248 The univariable logistic regression analysis revealed significant correlations between ICU 249 admission and viral coinfections (OR 3.600, 95% CI 1.172-11.057). A marginally significant 250 correlation between ICU admission, female sex (OR 0.396, 95% CI 0.152-1.027), and the 251 collapsed variable of obesity, cardiovascular disease and diabetes mellitus (OR 2.500, 95% CI 252 0.954-6.552) was detected (Table 2). All four patients with obesity were admitted to the ICU 253 (p-Fisher 0.03). No significant correlations were found between the investigated patient-254 related variables and the all-cause in-hospital mortality (Suppl. Table 1). A multivariable 255 bimodal logistic regression model for ICU admission confirmed statistically significant 256 coefficients for female sex (OR 0.256, 95% CI 0.081 - 0.804), viral coinfection (OR 4.094, 257 95%CI 1.199-13.986) and the collapsed variable (obesity, cardiovascular disease or diabetes 258 mellitus) (OR 4.595, 95% CI 1.424-14.835) (Table 2). The Nagelkerke R square was 0.256 259 indicating an adequate level of explanatory power and the significance level of the Hosmer-260 Lemeshow test was 0.393 indicating a good model fitting. 261 262 Clinical isolates characteristics 263 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 13 Clinical isolates were genetically characterized to determine their ST and emm-type (Table 264 3). The emm typing yielded successful results in 92 out of 93 strains, detecting 18 different 265 emm types. Among these, emm1-type (emm1.0, emm 1.2, emm1.3, emm1.6, emm1.7) emerged 266 as the predominant type in 63 cases (68%), followed by emm12 (emm12.0, emm12.2, 267 emm12.4) in 7 cases (7.6%). The emm1-type/ST-28 was prevalent, and associated with higher 268 mortality, since it was detected in 5/7 (71%) patients who died. 269 The Kirby-Bauer method was performed on 79 strains in order to determine phenotypical 270 antimicrobial resistance (Table 3). All tested strains were susceptible to penicillin. Only one 271 strain was found resistant to clindamycin and erythromycin. No i nducible clindamycin 272 resistance was detected. 273 274 Molecular typing of iGAS 275 A comprehensive genomic analysis utilizing WGS was performed on a cohort comprising 91 276 iGAS und 2 non-invasive GAS strains, whose sampling periods spanned through before, 277 during, and after the COVID-19 pandemic (Figure 2A and 2B). Notably, we found a 278 consistent preservation of sequence-types (STs) across all sampling times, indicating genetic 279 stability within the GAS population. Intriguingly, most strains associated with the current 280 GAS upsurge were found to originate from existing genetic lineages circulating before the 281 pandemic, suggesting a continuum in transmission dynamics rather than a pandemic-induced 282 shift. The phylogenies did not exhibit large clusters of recent and post-pandemic strains 283 clearly distinct from the older reference variants refM1uk, ref5005, and ref5448. Ten strains 284 (GAS_086, GAS_032, GAS_031, GAS_082, GAS_070, GAS_027, GAS_052, GAS_046, 285 GAS_076, GAS_048) exhibited a genetic resemblance by core genome phylogeny, possessing 286 different ST types compared to the predominant ST-28 type. These strains were not confined 287 to a specific geographic location but were dispersed throughout the country. A high level of 288 genetic relatedness was found between GAS_035 and GAS_037 strains, originating from a 289 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 14 patient that died from an iGAS infection and an individual related to the patient, respectively. 290 FT-IR spectroscopy showed a poor concordance compared to the core genome analysis 291 performed with the WGS, failing to reliably differentiate genetically distant GAS strains (Fig. 292 Suppl. 2). 293 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 15

294 In summary, this multicenter study shows that WGS core genome analysis of GAS strains 295 isolated in Switzerland before, during, and after the COVID-19 pandemic did not reveal 296 genetically distinct, emerging variants. The emm1 ST28 was predominat in our cohort. FT-IR 297 spectroscopy showed poor overlap with WGS and did not effectively differentiate genetically 298 distant GAS strains, confirming WGS as the current gold standard for the characterization of 299 GAS clinical isolates. Moreover, viral coinfections, mainly influenza and SARS-CoV-2, were 300 significantly associated with a severe iGAS infection requiring an ICU admission. 301 While we found a large emm1/ST28 cluster, including the reference strains (refMIuk, ref5005 302 and ref5448), the absence of outbreak characteristics, such as clustered cases in a specific 303 geographical region and a high genetic relatedness within potential transmission chains, 304 suggests that the strains of the current resurgence were part of the diversity already circulating 305 before the COVID-19 pandemic. The emm1/ST28 genotype has been commonly reported 306 causing iGAS infections [34], consistently indicating its dominance over the years [9, 10, 35, 307 36]. Accordingly, we observed that in our cohort emm1/ST28 was the predominant variant, 308 present in most patients (5/7) who died from an iGAS infection. Interestingly, we did not find 309 any clusters originating from the three reference strains refM1uk, ref5005 and ref5448, known 310 to cause invasive infections [9, 37, 38]. Furthermore, the phenotypic susceptibility testing 311 revealed a high susceptibility to penicillin, erythromycin and clindamycin, confirming the 312 known susceptibility patterns over the years without a shift towards more resistant strains, as 313 reported worldwide [4, 6, 9]. 314 The COVID-19 pandemic highlighted the need for rapid epidemic diagnosis in controlling 315 outbreaks. To our knowledge, this study is the first to compare the time- and cost-effective 316 FT-IR spectroscopy with WGS for GAS typing and outbreak investigation. FT-IR 317 spectroscopy showed poor overlap with WGS and did not effectively differentiate genetically 318 distant GAS strains. GAS, known for its adaptive nature and highly recombining genome, 319 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 16 undergoes phenotypic changes during invasive infections, potentially impacting FT-IR 320 spectroscopy's efficacy, a technique reliant on biochemical fingerprints [39, 40]. Our findings 321 underscore FT-IR spectroscopy's limitations in characterizing GAS strains during outbreaks, 322 supporting WGS as the gold standard typing method for GAS. Furthermore, WGS is a 323 valuable tool for GAS characterization, since it can provide important information on emm 324 type, virulence factors and antibiotic resistance genes. 325 Our epidemiological records from multiple centers across Switzerland confirm the dramatic 326 increase in iGAS infections, also documented across numerous countries globally [4]. Even 327 though these infections result in about 500,000 deaths annually worldwide, imposing a 328 substantial economic burden [11, 41], data on the incidence of iGAS clinical manifestations 329 remain scarce. At the University Hospital of Zurich, a threefold increase of the number of 330 invasive and non-invasive GAS strains was reported since December 2022 compared to the 331 pre-pandemic years, which aligns with the reports of other countries [6, 9]. Within our study, 332 the most prevalent diagnosis was SSTI, in line with findings from a study conducted in New 333 Zealand [42], followed by primary bacteremia and pneumonia. In accordance with other 334 studies [43], viral coinfections, mostly influenza and SARS-CoV-2, were relatively frequent 335 in our study cohort and significantly associated with a severe disease leading to ICU 336 admission, supporting the current hypothesis linking the recent iGAS infections outbreak to a 337 high prevalence of circulating respiratory viruses [10, 35]. Furthermore, we found that female 338 sex and the metabolic risk factors (cardiovascular disease, obesity and diabetes mellitus) were 339 also significantly associated with a severe disease requiring an ICU support. The all-cause 340 mortality in our cohort was 9.5%, comparable with the data published in England [35], 341 highlighting the severity of iGAS infections. 342 The cause of the global post-pandemic increased GAS infections prevalence remains elusive. 343 Our findings point out that the recent upsurge does not arise from pathogen-related factors, 344 prompting further exploration of alternative causes, as proposed in prior studies [6, 9, 10, 44]. 345 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 17 The current hypothesis suggests that the lack of exposure to GAS and common seasonal 346 respiratory viruses during the COVID-19 pandemic, due to isolation measures, resulted in 347 decreased immunity to GAS and seasonal respiratory viruses [35, 37, 43, 45]. After lifting the 348 COVID-19 pandemic protection measures, an increase and a shift of the seasonal pattern in 349 seasonal respiratory viral and iGAS infections was observed after two years of their 350 historically low incidence. 351 This study has some limitations that affect its broader applicability. The sample origin, limited 352 to specific Swiss regions, may not represent a broader GAS strain diversity. The 353 retrospectively collected subset of data and the unavailability of clinical information for some 354 patients does not allow determining risk factors associated with the current increased iGAS 355 prevalence. Only three of the participating hospitals admitted pediatric patients, leading to an 356 underrepresentation of minors. To draw more decisive conclusions regarding the risk factors 357 of iGAS infections, future studies should include larger, diverse samples cohorts from various 358 regions, encompassing both invasive and non-invasive strains and performing genotypic as 359 well as phenotypic virulence studies. 360 To the best of our knowledge, this is the biggest data set of invasive GAS strains across 361 multiple centers in Switzerland, giving a broad view of the post-pandemic increased 362 prevalence of iGAS infections. The long time-span of the examined GAS strains allowed us to 363 explore potential genetic changes before and after the COVID-19 pandemic. The absence of 364 genetic diversity and shift to more resistant strains throughout the years suggests that non-365 pathogen related factors may lead to resurgence of iGAS, supporting the theory that the lack 366 of immunity to GAS and the common respiratory viruses due to the social distancing during 367 the COVID-19 pandemic could be associated with severe iGAS infections with many 368 fatalities. 369 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 18

370 In summary, our study aimed to investigate the genetic characteristics of GAS strains 371 responsible for the current resurgence of iGAS infection following the COVID-19 pandemic. 372 WGS-based molecular typing did not uncover genetic changes regarding emm-type and 373 sequence-type (ST) pre- during- and post-pandemic Moreover, antibiotic susceptibility testing 374 revealed unchanged susceptibility patterns. The FT-IR spectroscopy limitations in GAS 375 typing underscore the ongoing importance of WGS in molecular analysis. Our results suggest 376 that non-pathogen related factors, such as the lack of immunity to GAS and other common 377 respiratory viruses due to social distancing and wearing masks during the COVID-19 378 pandemic, may be responsible for the post-pandemic increased prevalence of the iGAS 379 infections. Further research is needed to improve surveillance methods of iGAS strains. 380 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 19

381 1 . Nel son, G.E. , et al ., Epi demi ol ogy of Inv a siv e Gr o up A Stre ptoc occ al In fec tio ns i n th e U ni t e d 382 S tat e s , 2005 -2012 . Clin In fe ct D is, 2016. 63 (4): p . 478 -86. DOI: 1 0.1093 /ci d/ c iw24 8 . 383 2 . Wal ker, M .J ., e t a l., Di s e a s e m ani fe st ati o ns a nd p a tho genic mec ha nis ms of G rou p A 384 S tr e pt o coc cus . Clin Micr obiol Rev, 20 14. 27 (2): p . 264-301 . D O I: 10 .1128 /cmr.001 0 1-13. 385 3 . Andreo n i, F., et al ., Cl ind amy cin A f fec t s G roup A S tr e pto coc cus Virule nc e F actor s a nd 386 I mpro ve s Clini cal Outc om e. J Inf ec t Di s, 2017. 215(2): p . 26 9-277 .D O I: 10.1093 / in f dis /jiw22 9. 387 4. World H e a lth O r gan i z a tion (15 Dece m be r 2022 ) . Di s e a s e Outbr eak N e w s; I ncrea sed inci denc e 388 o f s c arlet feve r a nd in v a s i ve Grou p A S t re p t oc occ us in fec tio n - m ulti -c ou ntry. Ava il able f r o m: 389 http s://ww w.wh o.int / em e r g enci e s/di sea se -ou tbre ak-n ew s/i tem/2022 - DO N429 . 390 5. B ar n e s , M . , et a l ., N o te s fro m th e F ield: I ncrease i n Pe dia t ric In va siv e Gr o u p A Str eptoc occ us 391 I nf ection s - Col or a d o a nd Mi n ne so t a , O c t ober -Dece m ber 2022. MMWR Morb Mor tal Wkly 392 Re p, 2023 . 72 (10 ): p. 265 -267. DO I: 10.1 5 585/mmwr.mm7210 a4. 393 6 . Abou lho s n , A., e t al. , Increa s e s in gro u p A st r ep toc occ al in fectio ns i n t h e pedi at ric popul ati on 394 i n Hou s t on, TX, 20 22. Cli n In fect Di s, 202 3. D O I: 10. 1093 /cid/ci ad197. 395 7. I ncrea se in I nva sive Gr o u p A Stre p In fe cti on s, 2022 –2023 . 2023 F eb ruary 2, 2023; Ava ilabl e 396 from: h ttp s ://ww w.cd c.gov / g r ou pa strep /iga s-in fec t io n s-inv e s tig a tion. html . 397 8. In v as iv e I nf e k t io n e n m i t G r up p e A St r e p t ok o k ke n ( iG A S ) b ei K i nd er n ; U pdat e v o m 3 1. M ä r z 398 2 023 - - S ta t e me nt der PI G S . Avail abl e fr om: 399 http s://ww w.pae dia trie schwe iz.c h/n ew s / zunahm e-i nva siver -grup pe -a -s tre pt o kok k en-400 in f e k t io n e n- i gas / . 401 9 . Alc olea - Me d ina, A ., e t al. , T he on go in g St reptoc occ us pyog e ne s ( G ro up A Str e pt ococ cus) 402 o utb r e ak in Lon do n, Uni ted Kin gd om in De cembe r 2022: a mol ec ular ep i demi ol ogy stu dy. 403 Cl inica l Mic r obi ology and In fec ti on, 202 3 .DOI: h ttp s://doi. org/10 .101 6/ j. cmi.20 23 .03.00 1 . 404 1 0. de Gie r, B. , et a l ., I nc r e ase in in v as ive gr o u p A st re ptoc occ al (S t rep toc occ us pyog e nes ) 405 i nfecti on s (i GAS) in you ng chil dr e n in t he N e t h e r l an d s, 2022. Euro Su rve ill, 20 23. 28 (1). DO I: 406 1 0.2807/1560 - 7 917 .Es .2023. 28 .1.220 09 41. 407 1 1. Brouw er, S ., e t al. , Pa t h o ge ne sis , epid e m iolog y an d control of Grou p A S t rep t o coc cus 408 in f e c t io n. N a t R ev Mic r o biol, 20 23 . 21 (7) : p. 431-447. DO I: 10.1 038 /s 41 579 -023 -00 86 5-7. 409 1 2. Be s s en, D.E. , P.R . Sme e ster s , and B. W. Be all, Mol ecul ar Epide m iol ogy , Ec ology, a nd Ev ol utio n 410 o f G r oup A Str e p tococ ci. M i crobiol Spec tr , 20 18. 6 (5) .D OI: 10 .112 8/ mi crobi ols pec . CP P3-0009 -411 2 018. 412 1 3. T eng, A.S . J. , et al . , Com pari so n o f fa s t Fo urie r tra ns for m in frared sp ectro scopy bio typ ing w it h 413 w hole gen ome s e que nc ing -ba se d ge noty ping i n co mmo n no soc omi al p a t h oge n s. Anal Bioan al 414 Ch em, 202 2. 414(24 ): p . 7179-71 89. DOI: 10.1007/ s00216-022 -04270-6 . 415 1 4. Nova is, Â. , et al . , Fo urier tra n sf or m in frare d s pectro scopy : unl oc king f und ame n tal s an d 416 p r os pec t s f or ba cteri al s tr a in typin g . Eur J Clin Micro biol In fect Di s, 2019. 38 (3 ): p . 42 7-417 4 48.DOI: 10.1007 / s10096-018 - 3 431-3 . 418 1 5. Z arnowie c, P. , et al., Fo urier Tran s form In frare d Sp ectro s c opy (FTIR) as a T ool f or t he 419 Id e nt if i c at ion a nd D if fe r e nt i at i on of P a tho g en i c Ba c t e r ia . Curr M ed Ch e m, 2015 . 22 (14): p . 420 1 710-8. DO I : 10.2174 / 092 9867322 66615 03 1115280 0. 421 1 6. Hu, Y ., et al ., E v alu at i on o f t h e IR B i ot y pe r f o r K l e bs i e l la p n eu mon i a e t yp i n g and its pot ent i a ls 422 i n ho spi tal hy gie ne ma na ge me nt . Microb Bi otechnol , 2021. 14 (4 ): p. 13 43 -1352. D OI: 423 1 0.1111/1751 - 7 915 .1370 9. 424 1 7. Azrad, M ., e t a l. , Com p ari s on of FT -IR w it h wh ole -ge no me se que n cing fo r i de nti fic at i o n of 425 ma t er na l- t o /i1ne ona t e t r a ns m is s i on of ant i b i ot i c - r es is t a n t b a c t e r i a. J M i cr obiol Me t hod s, 426 2 022. 202 : p. 106 603.D O I : 10.1016 / j.m i met.2022.1 06603 . 427 1 8. Ma r t ak , D. , et al . , Fo urier -Tran s form I nfr aRed S pec t ro scop y Can Quic kly Type Gr am - N ega t i ve 428 Ba cilli Res pon si ble f or Hos pit al Ou tbreak s. Fr ont Mic ro biol, 20 19. 10 : p . 1440.D OI : 429 1 0.3389/fmic b .2019.01 440. 430 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 20 1 9. S chei er, T. C. , e t al ., Fou r i e r- t r ans fo r m i nfr a r e d sp e ct r os c o py f o r t y p i ng of v anc o myc i n -431 resi sta nt E nteroco ccu s fae ciu m: per form a nce ana lysis and o u t b r e ak i nv e st i ga tion . Mic r o biol 432 Sp e c t r , 2 02 3. 11 (5) : p. e0098 423. DOI: 10 .1128/ spec trum.009 84 -23. 433 20 . Ca s e defi niti on s f or in fe ctiou s co n ditio n s under pu blic he alt h s urvei lla nc e. Cen te rs for Di s e a s e 434 C on t r o l a nd P r e ven t io n. MMWR R ec om m Rep, 1997. 46 (Rr-10 ) : p. 1 -55. 435 2 1. Rä z , A . K ., e t al. , Lim i t e d A da pt atio n o f St aphyl ococ cus aur eu s duri ng Tra ns i t i o n fro m 436 Co loniz atio n t o Inv a sive I nfecti on . Micro biol Spec t r , 20 23. 11 (4) : p. e0259 021 .D OI: 437 1 0.1128/ spec trum.0259 0 -21. 438 2 2. Ko lesnik -G oldmann, N ., e t al. , C om pari s o n of Disk Dif fu s i on, E -Te st, a nd Br o t h Mic rodilu tio n 439 M et hods fo r Te s t i n g I n V i t r o A c t i vi t y o f Cef id e ro c o l i n A ci n e to ba ct er b au m an n ii. A ntibiotic s 440 (Ba s el ), 2023. 12(7). D O I: 10 .3390 /an tibio tic s 12 071212. 441 2 3. Ka patai, G ., et al ., Wh ol e ge no me s equ en c ing o f gr oup A Strep toc occ us : de ve lo pm ent a nd 442 ev a l uat ion of a n aut oma t ed p i pel i n e f or em m g en e t y p i ng. Pe erJ , 2017 . 5 : p. e 322 6.DO I: 443 1 0.7717/pe erj.32 26. 444 2 4. S eeman n, T. , ml s t Gi th ub . 445 2 5. Jol ley, K .A. and M .C . M a ide n, BI GSd b : Scalable an aly s is of ba cterial g en ome v aria tion a t t he 446 po pu l at i o n l ev el . BM C Bioin fo rmatic s , 20 10. 11 : p. 59 5. D OI : 10 .1186/1471 -21 05 -1 1-595. 447 2 6. Ba nkev ich, A . , et al ., SPA de s: a n e w ge no me a sse mbly alg orith m a n d it s a pplic ati o ns to 448 sin gl e -cel l s e q uenc in g. J Co mput Biol, 20 12. 19 (5): p . 455-77 . DOI: 10 .1 089/cmb.2 0 12.0021. 449 2 7. S eeman n, T. , Prok ka : rapi d prokaryo tic geno me an no tati on . B ioin forma tic s, 2 014. 30 (1 4 ) : p . 450 2 068-9. DO I : 10.1093 /bioi n for ma t i cs / b tu 153. 451 2 8. P age, A.J ., e t al., R o ar y : rapi d l ar g e- s c ale p r ok aryote p an g eno me a naly s i s. Bi o inf o r ma tic s, 452 2 015. 31 (2 2): p . 3691 -3 .DOI: 1 0.1093 /bio informa tic s /b tv421 . 453 29 . Pr i t cha r d , L ., e t a l. , G e nom ic s an d taxo no m y in d iagn o stics for f oo d sec uri t y : so ft-r otti ng 454 e nter obacteri al pl ant pa tho g en s. Analyti c al Method s , 20 16. 8 (1 ): p. 12 -24. DO I : 455 1 0.1039/C5 AY02 550H. 456 3 0. Nguy en, L .T., e t a l. , IQ -TREE: a fa s t an d e ffec t i ve s tocha s tic algori thm for e sti ma ti ng 457 ma x i m u m - li k e l iho od p hy l o ge ni es . M ol Bi ol Evol, 2015. 32 (1): p . 268 -74. DOI: 458 1 0.1093/molbe v / msu300 . 459 3 1. P ara dis , E. a nd K. S chlie p , ape 5 .0: a n e nv ironme nt for mo dern phy loge n etic s an d 460 e volu tio n ar y anal yse s i n R. Bioi n fo r ma t i c s, 2019. 35 (3): p . 52 6-528. D O I : 461 1 0.1093/bioin form atic s /b ty633 . 462 3 2. Y u, G., e t al . , gg tree: an r pack ag e for v i sua liz a t io n an d a n no ta t i on o f p hy logen eti c trees w it h 463 the ir cov ari ate s a nd o ther as s o cia ted dat a. Me thod s in Ec ology and E volution, 20 17. 8 (1) : p. 464 2 8-36. DO I : http s:/ /d oi.org /10.1 111 /2041 -210X. 12628 . 465 3 3. Ke nnis , M., e t al . , Sea s on al vari atio n s an d r i s k f actor s of Stre pt ococ cus py ogen e s i n fecti on: a 466 m u l t i c e n te r re s e a r c h n e t w o rk s t u d y . The r Adv Infec t Di s, 202 2. 9 : p . 467 2 0499361 2211321 01.DO I : 10.1177/2049 9361221 132101. 468 3 4. E kelund , K., et a l., V a r ia t i ons i n em m t y pe a m on g gr oup A s t r e pt o co c ca l i s o l at es ca us in g 469 i nva sive or noni nv as i ve i n fection s i n a na tionw id e s tudy. J Clin Mic r o b iol, 200 5. 43 (7): p . 470 3 101-9. DO I : 10.1128 / jc m.43.7 .3 101-310 9. 2005. 471 3 5. Guy , R., e t a l. , I n cr eas e i n in vas i ve gr oup A s t re p to co c ca l i n f ec t i on not if i c at i o ns , Engl a nd , 472 2 022. Euro S urveill , 2023. 28 (1). D O I: 10.2 807/1560 -7917.E s .2023 .28.1 .220 0942. 473 3 6. Gher ardi, G ., L. A. Vital i , and R. Cre t i , Pr ev al ent e mm Types amo n g I nva s i v e GAS in Europe a n d 474 Nort h Americ a s i nce Year 200 0. Fron t Pu bli c Heal t h, 2018. 6 : p. 59. DO I: 475 1 0.3389/f pubh.20 18 .00059 . 476 3 7. P ee te rman s, M. , et al . , Clini cal a nd m ole cular e pi demi o logic al fea ture s o f c r itic ally ill p atie nt s 477 w ith inv a s i ve gro up A Stre pt ococ cus in f e c tion s: a Be lgi an m ulticen t e r case - s e r i e s . A n n 478 In t e n s iv e Car e, 2 024. 14 (1 ): p . 19.D OI : 10.11 86/s1361 3-024 -01249 - 7 . 479 3 8. Wilk ening , R. V. a nd M. J . Fed erl e, Evol u t i onary Con s tr aint s Sh apin g Strep toco ccu s py ogene s-480 Hos t In teracti on s. T r e n d s Mic robiol , 20 17. 25 (7) : p . 562-572 .DOI: 10 . 1016/j. tim.2 017. 01.007. 481 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 21 3 9. F eil, E.J ., e t al. , R ec ombi nati on wit hin n atu r al po pul a t io n s o f pa tho geni c bac t eria: sh ort-ter m 482 em p i r i c al es t i ma t es an d l ong- t er m ph y l o ge n et i c c o n s e que n c e s . Proc N atl Acad Sci U S A, 483 2 001. 98 (1 ) : p . 182-7 .D OI: 10 .1 073/pna s . 98.1.182 . 484 4 0. Ba o, Y. -J ., e t a l., Phe no typic differ enti ati o n of Stre pt oco ccu s py ogene s p op ula tion s is ind u ced 485 b y reco mbina tion -dri v en g e ne - s p ecific sw ee p s . S cie n t i f i c R epo rt s, 2016. 6 (1 ): p. 36 644.DOI: 486 1 0.1038/ sr e p3 6644. 487 4 1. Ca r a peti s, J.R ., e t al. , The glo bal b ur d e n o f gro up A s tre pt oco cca l disea se s. L an ce t In f e ct D is , 488 2 005. 5 (11 ) : p . 685-94 .DOI: 10 . 1016/ s 14 73-3099(05 )70 267-x. 489 4 2. Ca nnon , J. W., et a l ., The ec on omic and h e alth bur den s o f d i s ea s e s c au s e d by gr o u p A 490 S tr e pt o coc cus i n New Zeal an d. I nt J In fe c t Di s , 2021. 103 : p . 17 6-181. DOI : 491 1 0.1016/j.iji d.2020.1 1 .193. 492 4 3. de Gie r, B. , et a l ., A ssoc iati on s be twe en c ommo n re spira tory v ir u se s an d inv asi ve grou p A 493 stre pt ococ cal infec t i on : A ti me- seri es ana ly s i s . I nf lue n z a O t he r R e s p ir V i ru ses , 2 0 19 . 13 (5) : p. 494 4 53-458. DOI : 10.111 1/ i rv .12658. 495 4 4. Ra mos A m ador, J.T. , A. B e r zo s a Sá nchez , and M. Illá n Ramo s, Grou p A Strep toco ccu s i n va sive 496 i nfecti on in c hildre n: E pi demi ologic cha n ges and i m plica t i on s . R ev Esp Quimio ter , 20 23. 36 497 Suppl 1 (Supp l 1): p. 3 3-36 .D O I : 10.3720 1 / req / s 0 1.09 .2023 . 498 4 5. Nyg aard, U. , et a l., I nv a s i v e gro up A s tre p toco cca l infec tio ns i n c hildren and adol esc ents i n 499 Den mark duri ng 2022-23 c om pare d with 2016-1 7 to 202 1-22: a n at i onwide , mul ti c entre, 500 po pu l at i o n- bas e d c o ho r t s t udy . L ance t C hild Adole s c H e al th, 202 4. 8 (2) : p. 112-1 21.DO I: 501 1 0.1016/ s2352-4642 (2 3)00295 -x . 502 503 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 22 Figures 504 Figure 1 - Number of detected invasive GAS strains at the University Hospital Zurich 505 between January 2012 and December 2023 506 Invasive GAS (iGAS) strains detected either by culturing or via molecular methods are 507 included in this graph. The blue color indicates the number of new cases monthly and the red 508 color the moving average every three months. Q1: months from January to March, Q2: 509 months from April to June, Q3: months from July to September, Q4: months from October to 510 December. 511 512 Figure 2 - Phylogenetic tree of the 93 GAS clinical isolates and three reference strains 513 (refM1uk, ref5005 and ref5448) 514 Midpoint rooted maximum likelihood phylogenetic trees based on the alignment of the 515 clinical isolates and reference strains core genes. The scale bar corresponds to the number of 516 nucleotide substitutions per position. A. Phylogenic tree of all GAS clinical isolates and the 517 three reference strains (refM1uk, ref5005 and ref5448). B . Magnified representation of the 518 large emm1/ST28 cluster (including refM1uk, ref5005 and ref5448) from panel A. 519 520 Table 1. Demographic and clinical characteristics of the patients with invasive GAS 521 infections 522 523 Table 2 . Results of the univariable and multivariable logistic regression analysis of the 524 association of patient-related factors with ICU admission (n=32) upon iGAS infection. 525 526 Table 3. Molecular and microbiological characteristics of the 93 GAS clinical isolates. 527 528 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 23 Supplementary Figure 1 - Number of detected GAS strains (invasive and non-invasive) 529 at the University Hospital Zurich between January 2012 and December 2023 530 Invasive and non-invasive GAS strains detected either by culturing or via molecular methods 531 are included in this graph. The blue color indicates the number of new cases monthly and the 532 red color the moving average every three months. Q1: months from January to March, Q2: 533 months from April to June, Q3: months from July to September, Q4: months from October to 534 December. 535 536 Supplementary Figure 2 - Assessment of average nucleotide identity (ANI) networks 537 with FT-IR spectroscopy 538 Overlap (quantified as the v-measure) between FT-IR and ANI-derived clusters for different 539 cutoffs used for FT-IR (X-axis) and ANI (color) clustering. 540 541 Supplementary Table 1. Results of the univariable and multivariable logistic regression 542 analysis of the association of patient-related factors with all-cause in-hospital mortality (n=7) 543 upon iGAS infection. 544 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 0 0 0 2 1 0 1 0 0 2 1 1 1 0 0 1 3 1 0 0 0 0 2 2 2 2 1 1 2 4 0 1 0 1 0 0 0 0 0 1 0 2 1 3 9 2 0 2 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 Number of iGAS strains detection Number of iGAS strains detection Moving Average (size=3) . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint Characteristics (N=74) Sex, n (%) Male 44 (59.5) Female 30 (40.5) Median age (range), years 47.5 (1-94) Patients < 18 years old, n (%) 9 (6.5) Hospital, n (%) University Hospital of Zurich 20 (27) University Hospital of Geneva 17 (23) City Hospital Zurich 16 (21.6) Wetzikon Hospital 9 (12.2) Cantonal Hospital Chur 7 (9.5) Uster Hospital 4 (5.4) Cantonal Hospital Winterthur 1 (1.4) Coexisting conditions, n (%) Cardiovascular disease 26 (35.1) Diabetes mellitus 13 (17.6) Renal disease 8 (10.8) Solid tumortumour 7 (9.5) Alcohol abuse 6 (8.1) Intravenous drug use 5 (6.8) Immunosuppression 4 (5.4) Adipositas 4 (5.4) Chronic obstructive lung disease 3 (4.1) Autoimmune disease 2 (2.7) Viral hepatitis 2 (2.7) HIV 1 (1.4) HematologicalHaematological malignancy 0 (0) No underlying disease, n (%) 30 (40.5) ≥ 1 underlying disease, n (%) 44 (59.5) Homeless 2 (2.7) Sampling site, n (%) Blood 48 (64.9) . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint Tissue biopsy 10 (13.5) Swab 5 (6.8) Joint fluid 3 (4.1) Pleural fluid 3 (4.1) Abscess 3 (4.1) Bone biopsy 1 (1.4) Ascites 1 (1.4) Diagnosis, n (%) SSTI with bacteraemia 14 (18.9) Pneumonia 12 (16.2) Isolated bacteraemia 10 (13.5) SSTI without bacteraemia 5 (6.8) Necrotizing fasciitis 5 (6.8) Septic arthritis 5 (6.8) Non pregnancy related endometritis 3 (4.1) Osteomyelitis 3 (4.1) Tonsilitis with abscess 3 (4.1) Septic bursitis 3 (4.1) Otitis media/mastoiditis without bacteraemia 2 (2.7) Otitis media/mastoiditis with bacteraemia 2 (2.7) Tonsilitis without abscess 2 (2.7) Pharyngitis with bacteraemia 1 (1.4) Meningitis 1 (1.4) Pregnancy related endometritis 1 (1.4) Mediastinitis 1 (1.4) Endocarditis 1 (1.4) Toxic shock syndrome, n (%) 15 (20.3) Septic shock, n (%) 32 (43.5) Viral co-infection, n (%) 18 (24.3) Influenza (A or B) 6 (8.2) SARS-CoV-2 3 (4.1) SARS-CoV-2 and Influenza (A or B) 3 (4.1) RSV 3 (4.1) . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint VZV 2 (2.7) Rhinovirus 1 (1.4) ICU admission, n (%) 32 (43.2) Intubation, n (%) 22 (29.7) Surgery, n (%) 39 (52.7) Amputation, n (%) 4 (5.4) Antimicrobial therapy, n (%) Co-Amoxicillin-Clavulanate 39 (52.7) Clindamycin 36 (48.6) CeftriaxonCeftriaxone 26 (35.1) Penicillin 21(28.4) Piperacillin-Tazobactam 15 (20.3) Vancomycin 9 (12.2) Meropenem 4 (5.4) Clarithromycin 3 (4.1) Cefepime 3 (4.1) Gentamycin 2 (2.7) Combination therapy, n (%) 38 (51.4) Immunoglobulin, n (%) 15 (20.3) Duration of hospitalization, mean (range), days 15.6 (0-100) Duration of ICU hospitalization, mean (range), days 3.2 (0-21) All-cause mortality, n (%) 7 (9.5) GAS-related death, n (%) 6 (8.1) HIV: Human Immunodeficiency Virus, SSTI: Skin and soft tissue infections, ICU: Intensive Care Unit, GAS: Group A Streptococci, SARS-CoV-2: Severe acute respiratory syndrome coronavirus type 2, RSV: Respiratory Syncytial Virus, VZV: Varicella zoster virus . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint Odds Ratio 95% Confidence Intervall p value Univariable analysis Female sex 0.396 0.152, 1.027 0.057 Cardiovascular disease 1.944 0.739, 5.116 0.178 COPD 0.645 0.056, 7.445 0.725 Diabetes mellitus 1.680 0.504, 5.600 0.398 Solid tumour 0.194 0.022, 1.696 0.138 Immunosuppression 0.419 0.042, 4.232 0.461 Alcohol abuse 1.345 0.253, 7.150 0.728 Renal disease 0.766 0.169, 3.471 0.729 Viral coinfection 3.600 1.172, 11.057 0.062 Collapsed metabolic Variable 2.500 0.954, 6.552 Multivariable analysis Sex 0.256 0.081, 0.804 0.020 Collapsed Metabolic Variable 4.595 1.424, 14.835 0.011 Viral coinfection 4.094 1.199, 13.986 0.025 Constant 0.651 0.336 COPD: chronic obstructive pulmonary disease, HIV: human immunodeficiency virus, IBD: inflammatory bowel disease, Collapsed Metabolic Variable: cardiovascular disease, diabetes mellitus and obesity . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint Inhibition zone (mm) Nr. Strain ID Lab ID Isolation year ST emm type PEN ERM CLI 1 GAS_001 23021066 2022 28 1.0 27 26 23 2 GAS_002 23021067 2023 39 4.0 22 22 19 3 GAS_003 23021068 2023 28 1.3 29 27 23 4 GAS_004 23021069 2023 28 1.0 26 24 18 5 GAS_005 23021070 2023 28 1.3 30 28 23 6 GAS_006 23021071 2023 28 1.3 34 31 26 7 GAS_007 CI1224 2017 28 NA NA NA NA 8 GAS_008 CI1271 2017 28 1.0 NA NA NA 9 GAS_009 CI1296 2017 28 1.0 NA NA NA 10 GAS_010 CI1316 2017 28 1.0 NA NA NA 11 GAS_011 CI1453 2018 28 1.0 NA NA NA 12 GAS_012 CI1800 2019 1138 1.0 32 29 23 13 GAS_013 CI2261 2019 101 1.0 NA NA NA 14 GAS_014 CI348 2013 28 1.0 NA NA NA 15 GAS_015 CI350 2013 28 1.0 NA NA NA 16 GAS_016 CI407 2013 52 28 NA NA NA 17 GAS_017 CI4711 2020 63 77.0 30 12 25 18 GAS_018 CI4839 2020 28 1.0 33 31 25 19 GAS_019 CI5359 2020 59 8.0 37 32 27 20 GAS_020 CI543 2013 28 1.0 NA NA NA 21 GAS_021 CI6416 2020 267 1.0 24 22 20 22 GAS_022 CI655 2014 NA 28 NA NA NA 23 GAS_023 CI729 2015 167 118 NA NA NA 24 GAS_024 CI7790 2021 403 11.1 31 6 9 25 GAS_025 CI780 2015 167 118 NA NA NA 26 GAS_026 CI781 2015 101 89 NA NA NA 27 GAS_027 CI8226 2023 242 12.4 30 28 22 28 GAS_028 CI8227 2023 28 1.3 32 29 23 29 GAS_029 CI8236 2023 28 1.0 25 24 19 30 GAS_030 CI8237 2023 28 1.0 32 28 22 31 GAS_031 CI8239 2023 242 12.4 33 30 24 32 GAS_032 CI8240 2023 242 1.3 30 28 22 33 GAS_033 CI8241 2023 28 1.3 31 28 23 34 GAS_034 CI8242 2023 28 1.0 32 29 23 35 GAS_035 CI8243 2023 28 1.0 33 30 24 36 GAS_036 CI8244 2023 28 1.0 33 29 24 37 GAS_037 CI8245 2023 28 1.0 31 28 23 38 GAS_038 CI8248 2023 28 1.0 31 29 23 39 GAS_039 CI8249 2023 28 1.0 27 26 22 40 GAS_040 CI8252 2022 28 1.0 30 29 23 41 GAS_041 CI8253 2022 28 1.0 31 29 23 42 GAS_042 CI8254 2022 28 1.0 31 29 23 43 GAS_043 CI8255 2023 28 1.0 32 29 23 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 44 GAS_044 CI8256 2023 28 1.0 33 28 22 45 GAS_045 CI8257 2023 28 1.0 31 29 23 46 GAS_046 CI8258 2023 36 12 31 28 23 47 GAS_047 CI8259 2023 28 1.0 32 28 23 48 GAS_048 CI8260 2023 36 12 32 28 23 49 GAS_049 CI8261 2023 62 1.0 34 30 24 50 GAS_050 CI8262 2023 28 1.0 29 28 22 51 GAS_051 CI8263 2023 28 1.0 32 29 23 52 GAS_052 CI8265 2023 36 12 32 28 23 53 GAS_053 CI8266 2023 46 22 35 14 31 54 GAS_054 CI8267 2022 39 1.0 26 26 21 55 GAS_055 CI8268 2022 28 1.0 24 26 22 56 GAS_056 CI8269 2023 28 1.0 24 22 18 57 GAS_057 CI8270 2023 28 1.0 26 22 19 58 GAS_058 CI8273 2023 148 1.3 29 26 21 59 GAS_059 CI8274 2023 28 1.3 33 28 23 60 GAS_060 CI8277 2023 28 1.0 28 24 19 61 GAS_061 CI8278 2023 28 1.0 22 23 18 62 GAS_062 CI8279 2022 28 1.0 31 30 19 63 GAS_063 CI8280 2023 314 1.0 25 12 17 64 GAS_064 CI8453 2022 28 1.0 32 29 23 65 GAS_065 CI8466 2022 101 1.0 27 25 19 66 GAS_066 CI8467 2022 28 1.0 33 28 22 67 GAS_067 CI9366 2023 28 1.4 29 25 22 68 GAS_068 CI9367 2023 101 89.4 27 28 22 69 GAS_069 CI9368 2023 62 87.9 26 23 21 70 GAS_070 CI9370 2023 242 12.4 24 22 22 71 GAS_071 CI9375 2022 28 1.7 27 24 19 72 GAS_072 CI9376 2022 28 1.4 27 24 21 73 GAS_073 CI9377 2022 28 1.7 34 32 30 74 GAS_074 CI9378 2022 178 44.0 24 24 20 75 GAS_075 CI9379 2022 190 49.3 28 28 22 76 GAS_076 CI9380 2022 36 12.2 19 22 18 77 GAS_077 CI9381 2022 NA 73.0 26 26 23 78 GAS_078 CI9382 2022 NA 80.0 30 30 27 79 GAS_079 CI9384 2022 137 104.0 33 30 25 80 GAS_080 CI9385 2022 190 49.3 27 26 21 81 GAS_081 CI9387 2022 28 1.4 31 31 24 82 GAS_082 CI9389 2022 242 252.0 26 24 22 83 GAS_083 CI9390 2022 150 212.0 26 26 22 84 GAS_084 FF09716211 2021 772 89.0 35 26 25 85 GAS_085 FS42596242 2022 11 53.0 30 26 25 86 GAS_086 FS49301899 2022 242 1.0 28 23 21 87 GAS_087 FS49818997 2022 28 1.0 32 33 24 88 GAS_088 FS55545334 2022 969 1.0 and 86.2 22 28 20 89 GAS_089 FS69217489 2023 28 1.0 27 24 22 . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint 90 GAS_090 FS72033123 2023 28 1.0 28 29 22 91 GAS_091 FS72033369 2023 28 1.0 30 24 22 92 GAS_092 FS72034424 2023 52 1.0 27 24 21 93 GAS_093 FS72034435 2023 28 1.0 29 24 20 NA: not available . CC-BY-NC-ND 4.0 International licenseIt 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 April 3, 2024. ; https://doi.org/10.1101/2024.04.03.24305261doi: medRxiv preprint