Received: 3 December 2022 Revised: 20 April 2023 Accepted: 8 May 2023 DOI: 10.1002/vetr.3056 ORIGINAL RESEARCH Risk factors associated with Streptococcus suis cases on pig farms in Spain Carlos Neila-Ibáñez1,2,3 Sebastián Napp1,2,3 Lola Pailler-García1,2,3 Lorena Franco-Martínez4 José Joaquín Cerón4 Virginia Aragon1,2,3 Jordi Casal5 1 Unitat mixta d’Investigació IRTA-UAB en Abstract Sanitat Animal, Centre de Recerca en Sanitat Animal, Universitat Autònoma de Barcelona, Background: Streptococcus suis can cause meningitis, polyarthritis and acute Bellaterra, Spain death in piglets. However, the risk factors associated with S. suis infection 2 Institut de Recerca i Tecnologia remain incompletely understood. Therefore, a longitudinal study was carried Agroalimentàries, Centre de Recerca en out, in which six batches from two Spanish pig farms with S. suis problems Sanitat Animal, Universitat Autònoma de were repeatedly examined to determine possible risk factors. Barcelona, Bellaterra, Spain Methods: A prospective case–control study was conducted, and potential risk 3 OIE Collaborating Centre for the Research factors were evaluated using mixed-effects logistic regression models. The and Control of Emerging and Re-Emerging Swine Diseases in Europe, Bellaterra, Spain explanatory variables included: (a) concomitant pathogens; (b) biomarkers 4 Interdisciplinary Laboratory of Clinical associated with stress, inflammation and oxidative status; (c) farm environ- Analysis, University of Murcia, Espinardo, mental factors; and (d) parity and S. suis presence in sows. Three models were Spain built to study the effect of these variables, including two to assess the risk 5 Department of Animal Health and factors involved in the subsequent development of disease. Anatomy, Faculty of Veterinary Medicine, Results: Risk factors for S. suis-associated disease included porcine reproduc- Universitat Autònoma de Barcelona, Bellaterra, Spain tive and respiratory syndrome virus co-infection at weaning (odds ratio [OR] = 6.69), sow parity (OR = 0.71), haptoglobin level before weaning (OR = 1.01), Correspondence relative humidity (OR = 1.11) and temperature (OR = 0.13). Jordi Casal, Department of Animal Health Limitations: Laboratory diagnosis was done at the batch level, with individ- and Anatomy, Faculty of Veterinary Medicine, Universitat Autònoma de ual diagnosis based on clinical signs only. Barcelona, Bellaterra, Spain. Conclusions: This study confirms the multifactorial nature of S. suis- Email: Jordi.casal@uab.cat associated disease, with both environmental factors and factors related to the host involved in disease development. Controlling these factors may, Funding information therefore, help prevent the appearance of disease. Program for Innovative Global Prevention of Streptococcus suis (PIGSs), from program Horizon 2020 of the European Commission. KEYWORDS Carlos Neila-Ibáñez was funded by the same animal stress, co-infection, environmental factors, PRRSV, risk factors, Streptococcus suis project, Grant/Award Number: 727966 INTRODUCTION allow strains from the microbiota of healthy ani- mals to produce clinical disease are not completely Streptococcus suis is one of the main bacterial known.5 In addition, S. suis is considered an emerging pathogens affecting the pig industry globally, caus- zoonotic pathogen, with some important outbreaks in ing economic losses due to substantial postweaning Southeastern Asian countries in the present century.6 morbidity and mortality and the costs associated S. suis is classified into 29 different serotypes with disease control.1,2 Although S. suis is a normal depending on the capsular polysaccharide, with coloniser of the porcine upper respiratory tract,3 it serotypes 1, 2, 7 and 9 being the most frequently iso- can cause disease, characterised mainly by meningi- lated from clinical cases.7 Serotypes are distributed tis, polyarthritis and acute death, in pigs, especially worldwide, and the lesions they produce are not piglets from 5 to 10 weeks old.1 Even though globally serotype dependent.8 Although many virulence fac- the proportion of pigs colonised by S. suis is close to tors have been described for S. suis, they are not always 100%,4 the level of associated disease is much lower, present in clinical isolates.4 Furthermore, there are with incidence rates below 5% and mortality below other factors (unrelated to the pathogen) that may 1% reported in nursery units in Spain, Germany and influence the development of the disease, such as the the Netherlands.2 However, the circumstances that number of piglets weaned per sow, which seemed This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2023 The Authors. Veterinary Record published by John Wiley & Sons Ltd on behalf of British Veterinary Association. Vet Rec. 2023;e3056. wileyonlinelibrary.com/journal/vetr 1 of 10 https://doi.org/10.1002/vetr.3056 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2 of 10 VETERINARY RECORD to influence mortality during a S. suis outbreak in piglets.27 In addition, biomarkers of oxidative status, suckling piglets.9 S. suis outbreaks have also been such as hydrogen peroxide (H2 O2 ) or advanced oxida- associated with some concomitant viral infections, tion protein products, can be used as pain indicators such as with porcine reproductive and respiratory and to assess oxidative stress.28 The levels of these syndrome virus (PRRSV), porcine circovirus 2 (PCV-2) biomarkers may be altered as a result of infectious pro- or swine influenza virus (SIV). These associations are cesses, as demonstrated in SIV and M. hyopneumoniae often observed under field conditions, where mixed infections.26 infections are frequent. However, experimental infec- With the aim of evaluating possible risk factors for tions to confirm these hypotheses are complex, so only S. suis-associated disease, G. parasuis, SIV, PRRSV and a few in vivo studies have been published.10 Piglets PCV-2 presence, environmental parameters, parity of born to sows infected with PRRSV during gestation the dams and biomarkers of stress, inflammation and and challenged with S. suis at 5 days of age were more oxidative status in piglets were analysed in a longitudi- susceptible to infection and disease than those born to nal study carried out on two Spanish commercial pig non-infected sows or those infected only with S. suis.11 farms. In two other experimental studies, piglets inoculated with PRRSV 7 days before being challenged with S. suis had higher mortality and more severe lesions than MATERIALS AND METHODS piglets challenged with only one of the pathogens.12,13 When S. suis challenge was performed in piglets 5 days Selection of the farms after PCV-2 infection, co-infected piglets exhibited more severe clinical signs and lesions than those inoc- The study was carried out on two pig farms, farms ulated with only one of the pathogens.14 Similar obser- A and B, located in Catalonia (north-eastern Spain). vations were reported for the SIV co-infection, with Both farms had a history of S. suis-associated disease, co-infected piglets showing more severe clinical signs which was confirmed by the isolation of the agent from and increased gene expression of pro-inflammatory the cerebrospinal fluid of suckling piglets (farm A) and mediators than those inoculated with only one of the weaners (farm B) with neurological clinical signs. pathogens.15 In addition, mixed infections with other Farm A was a family farm with 500 sows (White bacteria can increase the severity of the lesions caused Large × Landrace). The nursery was located 2.4 km by S. suis, as was reported in a co-infection study with from the maternity unit. Sows received a metaphy- Bordetella bronchiseptica, which was used to predis- lactic treatment with oxytetracycline in their feed pose the nasal mucosa for the S. suis inoculation.16 from 7 days pre-farrowing until piglets were weaned. In that co-infection, S. suis was found in lungs with Piglets were treated intramuscularly with amoxicillin bronchopneumonia only if B. bronchiseptica was also and gentamicin 2 days before weaning, and vacci- present, suggesting that S. suis should be considered a nated against M. hyopneumoniae and PCV-2 1 day secondary pathogen. before weaning. In the nursery, piglets with clini- Moreover, the role of the respiratory microbiota in cal signs compatible with S. suis-associated disease determining the presence and abundance of S. suis were treated intramuscularly twice with amoxicillin, deserves to be further studied, as the composition of enrofloxacin and dexamethasone, and when S. suis the nasal microbiota may predispose to disease devel- clinical disease appeared, all the animals in the batch opment by other early colonisers.17,18 Recently, Niazy were treated with amoxicillin in their drinking water et al.19 found a different composition of the tonsillar for 6 days. The status of the farm was ‘positive stable’ microbiota in S. suis-affected piglets compared with for PRRSV (i.e., PRRSV was not circulating either in the the uninfected group. One of the species found in maternity unit or in the nursery29 ), and neither PCV- different abundance was Glaesserella (Haemophilus) 2 nor influenza outbreaks were detected during the parasuis, another pig pathogen with similar clinical study. However, a porcine epidemic diarrhoea (PED) manifestations that are often misidentified.20 outbreak took place in the nursery at the time of the Environmental and management factors that irri- second batch in the study. tate the respiratory tract (e.g., high air pollution load) Farm B, with 3500 sows (Duroc × White Large × or induce stress in piglets (e.g., excessive temperature Landrace), belonged to a large commercial produc- fluctuations or overcrowding) have been previously tion company. The breeding herd and the nursery correlated with S. suis clinical disease in pigs.21–23 were located on the same farm. Animals with clinical Although animal stress can be evaluated using dif- signs compatible with S. suis were treated with amox- ferent biomarkers, cortisol is probably the most icillin and dexamethasone intramuscularly, suckling commonly used in pigs.24 The intensity of inflam- piglets were treated once and weaners twice. The farm matory processes can be measured by acute phase was ‘positive stable’ for PRRSV, and neither PCV-2 nor proteins25 such as haptoglobin, as shown in piglets influenza outbreaks were detected during sampling. co-infected with Mycoplasma hyopneumoniae and SIV On both farms, before the start of the study, swabs H1N1, which had higher levels of haptoglobin than in were collected from lesions on seven animals found non-infected animals.26 In contrast, transcription of dead or euthanased due to welfare concerns, and the haptoglobin gene was not altered in blood after these were analysed to confirm an S. suis outbreak. an S. suis challenge in caesarean-derived colostrum- S. suis presence was determined after swab plating deprived piglets when compared with non-inoculated and molecular identification of pure culture bacterial 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License VETERINARY RECORD 3 of 10 growth. S. suis identification and serotyping was per- CO2 (ppm), while the data logger placed at farm formed by PCR,30,31 confirming S. suis as the cause of B recorded only temperature and relative humid- disease and showing that the clinical isolates belonged ity (HD208.1NTCI—HP3517TC1.2, DeltaOHM, Italy). to serotype 7 on both farms. In addition, S. suis isolates Data were measured every 5 minutes for 60–90 hours obtained from lesions were analysed by enterobac- after each of the three sampling visits. terial repetitive intergenic consensus (ERIC)-PCR to determine the number of different strains involved in the outbreak, following the protocol described by Pathogen detection Versalovic et al.,32 but lowering the annealing temper- ature to 43◦ C. Nasal and vaginal swabs were resuspended in 500 µL of PBS, and blood was centrifuged at 1048 g and 4◦ C for 10 minutes to obtain serum. Both types of Sampling and data collection samples were stored at −80◦ C until they were pro- cessed. DNA and RNA were extracted using MagMAX Animal sampling was done under institutional autho- Pathogen RNA/DNA kit (Applied Biosystem), follow- risation (Ethics Commission in Animal Experimenta- ing the manufacturer’s recommendations, and then tion of the Generalitat de Catalunya protocol number stored at −80◦ C until molecular analysis was per- 11199) and followed good veterinary practices, in formed. accordance with European (Directive 2010/63/EU) The presence of S. suis, specifically serotypes 2 and Spanish (Real Decreto 53/2013) regulations. (or 1/2, which is indistinguishable from serotype On each farm, three different batches were sampled. 2 by PCR), 7 and 9 (the most common virulent In each batch, 30 piglets were selected from 10 dif- serotypes in European farms), in the nasal and vaginal ferent sows. All piglets with clinical signs compatible samples was evaluated. Meanwhile, the presence of with S. suis were chosen, and the group was completed G. parasuis, both virulent and non-virulent strains, by randomly selecting healthy piglets until a group was evaluated only in the nasal samples. Screening for size of 30 was reached. Selected animals were ear- these pathogens was performed using conventional tagged the week before weaning. Animal ages ranged PCR assays, with the primers and conditions described between 17 and 22 days. Clinical signs were checked by Ishida et al.30 for the presence of S. suis, Okura daily by the veterinarian in charge of the farm accord- et al.31 for serotypes 1/2–2, 7 and 9, and Galofré- ing to the protocols of the company. Animals were Milà et al.33 for G. parasuis. Nasal samples were also classified as diseased if they presented clinical signs tested for influenza viral RNA using a quantitative compatible with S. suis, such as neurological signs or reverse transcription-PCR (RT-qPCR) assay based on lameness, before the sample collection by the veteri- the amplification of the conserved segment of the narian in charge of the study. Nasal swabs and blood matrix gene, as described by López-Valiñas et al.34 samples were taken from the selected piglets, and The presence of PRRSV and PCV-2 in the serum was nasal and vaginal swabs from their dams. Piglets were determined by real-time qPCR assay with commercial sampled again approximately 2 weeks after weaning, kits (VetMAX PRRSV EU and NA 2.0 Kit, Life Tech- when they were between 31 and 36 days of age. If nologies, and VetMAX Porcine PCV2 Quant Kit, Life any of the animals not sampled initially subsequently Technologies, respectively). developed lameness or neurological signs, they were For S. suis detection in lesions, a sterile cotton swab also ear-tagged and sampled (up to 10 more). Two was moistened in the lesion or with the fluid in the weeks later, when piglets were between 45 and 54 days case of the cerebrospinal fluid, plated into a choco- of age, the clinical status of the animals in relation to S. late agar plate (Biomerieux), and incubated at 37◦ C suis-associated disease was also recorded. As new indi- and 5% CO2 overnight. The pure culture compatible viduals were included in the cases group during the with S. suis was recovered and saved in PBS. DNA was follow-ups, the study can be classified as a prospective extracted using a chelex-based instagene matrix (Bio- case–control study. Rad Laboratories, Hercules, CA, USA) following the In total, 117 piglets were sampled between October manufacturer’s instructions. S. suis was confirmed by and December 2019 on farm A, and 90 piglets were PCR, using the protocol described by Ishida et al.30 sampled between March and May 2021 on farm B. The number of piglets sampled at each visit varied between 30 and 40, whereas the number of sows sampled was Analyses of cortisol, haptoglobin and 10 for each batch. Information about the ages and hydrogen peroxide sex distribution of the sampled animals is included in Table S1. Cortisol concentration was measured with a solid- During the first two visits, an environmental data phase, competitive chemiluminescent enzyme logger was placed in the rooms where the piglets immunoassay that uses a polyclonal rabbit anti- were located, at a height of approximately 30 cm. cortisol antibody (Immulite/Immulite 1000 cortisol, The data logger located at farm A (MHD21ABE17, Siemens Medical Solutions Diagnostics) previously DeltaOHM, Italy) recorded temperature (◦ C), relative validated for porcine saliva samples.35 Haptoglobin humidity (%), temperature-humidity index (THI) and concentrations were measured using a commercial 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4 of 10 VETERINARY RECORD quantitative turbidimetric test (Spinreact, S.A.U, ing dependent variables as fixed effects. Specifically, Spain) in an automated analyser (Olympus AU600), only those variables with a p ≤ 0.25 in the bivari- which was previously validated by Kaiser et al.36 ate analysis were further evaluated in the multivariate H2 O2 was assessed based on the method of Rhee analysis.39 The final model selection was performed et al.37 in an automated analyser (Olympus AU600), as via manual backward selection to better avoid multi- previously validated.38 collinearity, based on the Akaike information criterion, including only those variables with a p ≤ 0.05 and excluding those with a variance inflation factor greater Statistical analysis than 5. When building the regression model, the serotypes To evaluate the effect of the different variables on of S. suis analysed were not considered. As CO2 and S. suis-associated disease, three different mixed- THI were only recorded on farm A, the statistical effects logistic regression models were chosen, analysis was repeated to include these two variables depending on the age of the animals and the use associated with the ventilation only for this farm. of retrospective data: Statistical analyses were conducted with R (v. 4.0.2),40 using the packages lme4 (v. 1.1-23)41 and rsq ∙ Model 1: General risk factor model, considering the (v. 2.2).42 presentation of S. suis-associated disease at any of the visits, and using all explanatory variables; ∙ Model 2: Weaning risk factor model, considering the RESULTS presentation of the disease at the second and third visits (i.e., during the post-weaning period), and S. suis isolation and disease prevalence using the data collected previously in the farrowing unit; and During the study, S. suis was isolated from lesions of ∙ Model 3: Nursery risk factor model, considering the animals with clinical signs, confirming S. suis as the presentation of the disease late in the nursery, and most likely cause of the disease. In total, five differ- using the data collected in the first nursery visit. ent S. suis isolates were recovered, one from a tarsal joint and three from cerebrospinal fluid on farm A The mean and range values of each environmental (serotypes 7, 9 and other serotypes than 1/2, 2, 7 or parameter and stress marker were treated as continu- 9), while one S. suis isolate was recovered from fibrin ous variables. Age (in days) was classified as a discrete located in the thoracic cavity on farm B (a serotype value, while animal status for the different bacterial different to 1/2, 2, 7 and 9). On both farms, strains and viral pathogens was dichotomised as negative or were recovered from weaned animals. S. suis identifi- positive. cation was confirmed by recN PCR. Isolates from farm First, a bivariate analysis to test the associations A showed different fingerprinting by ERIC-PCR, and between the dependent variables (S. suis-associated they were also different to the isolate from farm B, indi- clinical signs) and the explanatory variables was car- cating that several S. suis strains were causing disease ried out. Considering the hierarchical structure of the on farm A. data, due to the potential effects of sow, batch and The prevalence of S. suis-associated disease differed farm, a mixed-effect logistic regression was used to between farms and batches (Table 1). Overall, farm model the binary independent outcome, including A had more diseased animals (from 5.5% to 17.6% sow, batch and farm as random effects and the remain- in early nursery) than farm B (less than 0.3% for TA B L E 1 Prevalence of Streptococcus suis-associated disease and mortality in the six batches at each sampling point. Farm A Farm B Batch 1 Batch 2 Batch 3 Batch 1 Batch 2 Batch 3 S. suis- Suckling piglets 0% 0.3% 0% 0.2% 0.1% 0.1% associated (0/300) (1/335) (0/363) (3/1786) (2/1906) (2/1805) disease Early nursery 7.0% 17.6% 5.0% 0.0% 0.1% 0.0% (21/300) (59/335) (18/363) (0/1786) (1/1906) (0/1805) Late nursery 9.1% 13.0% 4.7% 0.1% 0.3% 0.1% (27/298) (43/330) (17/360) (1/1785) (5/1901) (2/1795) Mortality Early nursery 0.7% 1.5% 0.8% 0.1% 0.3% 0.5% (2/300) (5/335) (3/363) (1/1786) (5/1906) (10/1805) Late nursery 7.4% 14.2% 8.3% 0.7% 0.8% 0.5% (22/298) (47/330) (30/360) (13/1785) (15/1901) (9/1795) Nursery 8.0% 15.5% 9.1% 0.8% 1.0% 1.1% (24/300) (52/335) (33/363) (14/1786) (20/1906) (19/1805) 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License VETERINARY RECORD 5 of 10 TA B L E 2 Prevalence of the pathogens obtained in sampled piglets and sows. Farm A Farm B Batch 1 Batch 2 Batch 3 Batch 1 Batch 2 Batch 3 Piglets Maternity Number of samples 30 30 30 30 30 30 unit S. suis-associated disease 0% 3% 0% 3% 7% 7% Nasal S. suis 93% 83% 97% 100% 100% 80% Nasal S. suis serotype 2 (1/2) 3% 0% 0% 13% 100% 3% Nasal S. suis serotype 9 20% 77% 57% 0% 10% 0% Nasal G. parasuis virulent strains 93% 93% 60% 50% 10% 93% Nasal G. parasuis non-virulent strains 100% 97% 70% 100% 100% 100% Nasal SIV 20% 0% 0% 7% 10% 0% Blood PRRSV 0% 10% 17% 7% 0% 37% Blood PCV-2 0% 0% 10% 0% 10% 3% Early Number of samples 40 40 37 30 30 30 nursery S. suis-associated disease 38% 30% 24% 0% 3% 0% Nasal S. suis 100% 100% 100% 100% 100% 57% Nasal S. suis serotype 2 (1/2) 13% 0% 0% 3% 90% 37% Nasal S. suis serotype 9 90% 100% 100% 10% 93% 0% Nasal G. parasuis virulent strains 87% 90% 92% 100% 100% 100% Nasal G. parasuis non-virulent strains 87% 93% 84% 97% 100% 77% Nasal SIV 5% 0% 3% 0% 0% 0% Blood PRRSV 36% 75% 41% 17% 70% 47% Blood PCV-2 0% 10% 5% 0% 143% 0% Late Number of samples 40 40 37 30 30 30 nursery S. suis-associated disease 55% 30% 16% 3% 0% 0% Sows Nasal Number of samples 10 10 10 10 10 10 S. suis 90% 70% 50% 90% 100% 60% S. suis serotype 2 (1/2) 30% 0% 0% 60% 40% 80% S. suis serotype 9 20% 40% 70% 20% 20% 0% G. parasuis virulent strains 50% 50% 30% 50% 50% 80% G. parasuis non-virulent strains 80% 60% 30% 100% 100% 50% Vaginal S. suis 90% 10% 50% 20% 0% 0% S. suis serotype 2 (1/2) 30% 0% 0% 0% 0% 0% S. suis serotype 9 10% 10% 30% 0% 0% 0% Abbreviations: G. parasuis, Glaesserella parasuis; PCV-2, porcine circovirus type 2; PRRSV, porcine reproductive and respiratory syndrome virus; S. suis, Streptococcus suis; SIV, swine influenza virus; Vir, virulent. the different phases and batches). Weaners were usu- farm A belonged to serotype 7, the presence of this ally more affected than suckling piglets, especially on serotype was analysed in samples from this farm, and farm A. it was detected in 23 out of 117 animals. Although G. parasuis was detected in all piglets throughout the study, the prevalence of virulent Prevalence of concomitant pathogens strains varied according to age, being especially high in early nursery pigs (90% and 100% on farms A and B, The prevalence of the pathogens in both piglets and respectively). sows is shown in Table 2. S. suis was detected in all In general, the prevalence of PCV-2 and SIV was batches and in a high proportion of nasal samples low. In contrast, the prevalence of PRRSV on farms A (93.0% for piglets and 76.7% for sows), but it was less and B was relatively high, particularly in nursery pigs common in the vagina of sows (56.7%). S. suis was (Table 2). not detected (i.e., absent in both farrowing and nurs- ery units) only in five animals, all from the third batch of farm B. Data on serotypes 2 (or 1/2) and 9 were Biomarker determination included in the analyses because they are the most common in Europe. Serotype 2 (or 1/2) was more The levels of cortisol, haptoglobin and H2 O2 varied prevalent on farm B, contrary to farm A where the most significantly between the various ages and batches prevalent was serotype 9. As one clinical isolate from (Table S2). In general, piglets with S. suis-associated 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 6 of 10 VETERINARY RECORD disease had higher levels of cortisol, H2 O2 and hap- at weaning (Table 3), namely, the level of hap- toglobin than healthy piglets of the same age and toglobin in the blood (which indicates inflamma- batch. For example, in the nursery of farm A, healthy tion) (OR = 1.01), the average relative humidity in animals had a median of 77.80 mg/dL of haptoglobin farrowing units (OR = 1.11) and the sow parity compared to 243.37 mg/dL in animals with clinical (OR = 0.71). signs (Figure S1). Model 3: Nursery risk factor model Environmental data The effect of the variables was also studied for the Mean room temperatures ranged between 25.8◦ C and nursery period with Model 3. The influence of PRRSV 28.4◦ C in the farrowing unit and between 26.3◦ C and co-infection (OR = 6.69) and the average temperature 28.4◦ C in the nursery, while the mean relative humid- (OR = 0.13) at the beginning of the nursery period ity ranged between 37.3% and 58.6% in the farrowing were significant, in addition to the parity of the sow unit and between 24.0% and 49.3% in the nursery (OR = 0.55). Sow parity had the same protective effect (Table S3). Relative humidity could not be recorded on as observed in Models 1 and 2. one of the visits to farm A due to a device malfunc- tion. CO2 concentration in the nursery of farm A was almost twice that in the farrowing unit (mean of 2857 Models for CO2 and THI vs. 1484 ppm, respectively), while THI values were sim- ilar in both units (63.9% in farrowing rooms and 61.0% CO2 and THI were measured only on farm A. On this in the nursery) (Table S3). farm, only CO2 range was statistically significant in Models 1 and 2, whereas in Model 3 mean CO2 was also significant (Table S5). Compared with the models built Bivariate analysis for the two farms, co-infection with PRRSV and sow parity remained as significant factors with similar val- The results of the bivariate analysis in the two farms, ues, while haptoglobin was significant in nursery pigs including the p-values and odds ratios of the variables instead of in suckling piglets. with p ≤ 0.25, are shown in Figure 1. While the odds ratios fluctuated between models, the effects of the different factors (identified as either a risk or a pro- DISCUSSION tective factor) were consistent throughout all of them. p-Values and odds ratios for all variables and models S. suis-associated disease is one of the most sig- are shown in Table S4. nificant diseases in the pig industry, particularly in intensive pig production systems.6 Despite the fact Multivariable model that the bacterium is highly prevalent on pig farms due to its role as a natural inhabitant of the micro- Model 1: General risk factor model biota of the porcine upper respiratory tract,4,43,44 the proportion of animals that are clinically affected is rel- The results from Model 1 (i.e., S. suis-associated dis- atively low.2 This is in accordance with the result of ease at any time) are presented in Table 3. The this study, in which S. suis was detected in the nasal results indicate that animals that were sampled 1 day cavity of all but five of the sampled piglets (97.6% for older (corresponding to the age at which piglets were both farms), but only a small percentage of the ani- weaned) present an almost five-fold increase in the mals developed associated disease (0.12%, 1.52%, and odds of developing S. suis clinical disease. The ani- 1.47% of the total piglets for each of the sampling mals with higher levels of cortisol and haptoglobin at points, considering both farms). the first sampling were also more prone to develop S. suis-associated disease has been associated with disease. The presence of PRRSV in nursery pigs was co-infection with other agents, especially with viruses. also statistically significant in this model, becoming Rieckmann et al.45 reported that even low virulence the most influential factor (OR = 6.40). Regarding S. suis strains resulted in the development of S. suis environmental factors, higher mean relative humidity disease in PRRSV-positive herds. In the present study, in farrowing rooms increased the odds of S. suis- PRRSV-infected animals presented a higher risk of associated disease (OR = 1.10). The only sow factor developing S. suis clinical signs following weaning. The that was retained in the model was the sow parity, with absence of statistical significance for the other two younger sows being more likely to have piglets with porcine viruses studied, PCV-2 and SIV, does not imply S. suis problems (OR = 0.69). a lack of effect, as their prevalence was low. An impor- tant question that arises in co-infection scenarios is which is the primary pathogen, as the simultaneous Model 2: Weaning risk factor model detection of several pathogens in diseased animals does not allow us to establish the order of infection. Some farrowing variables had a significant impact However, by looking at the risk factors in the previ- on the development of S. suis-associated disease ous stages (Models 2 and 3), we were able to relate the 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License VETERINARY RECORD 7 of 10 F I G U R E 1 Results of the bivariate analysis for all factors with p ≤ 0.25 on both farms. Model 1: General risk factor model; Model 2: Weaning risk factor model; Model 3: Nursery risk factor model. G. p. No Vir: G. parasuis non-virulent strain; G. p. Vir: Glasserella parasuis virulent strain; H2 O2 : hydrogen peroxide; PRRSV: porcine reproductive and respiratory syndrome virus; RH: relative humidity; SIV: swine influenza virus; S. suis: Streptococcus suis PRRSV infection at 5 weeks of age with the subsequent Piglets were intentionally sampled before and after development of clinical signs consistent with S. suis at weaning, which represents a period of high stress due 7 weeks of age. to various factors such as an abrupt separation from Despite both G. parasuis and S. suis being early the sow, change from milk to solid feed, movement to colonisers of the porcine upper respiratory tract and weaner pens or the creation of new hierarchical groups affecting young pigs, a direct relationship was not by commingling litters; and thus, the frequency of identified in the studied farms. The use of different disease was expected to be higher.47 The risk of devel- host cell receptors may explain the absence of inter- oping clinical disease also depends on the duration action, as was observed in previous in vitro studies.46 of the stressful situation, which may result in acute Regarding the environmental conditions, relative or chronic stress, and is also influenced by the age at humidity and CO2 can be taken as indirect measures which this stress occurs.48 It is also important to point of ventilation and air renewal. Our results showed that out that stress evaluated in the context of an infec- piglets located in farrowing units with a higher rela- tious disease can be considered either a consequence tive humidity or in nursery units with a higher CO2 of the disease or a possible cause.49 The longitudinal concentration were more prone to develop S. suis- study design allowed us to evaluate whether changes associated disease, which reflects the importance of in the assessed biomarkers preceded the S. suis- keeping animals in well-ventilated spaces. associated disease. Prolonged stress stimuli have been 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 8 of 10 VETERINARY RECORD TA B L E 3 Variables included in the three mixed-effect logistic regression models built for both farms. Variable OR 95% CI Beta coefficient p-Value Model 1: General risk factor Age of suckling piglets at sampling 4.95 1.87–13.12 1.59 0.001 Cortisol in suckling piglets 1.88 1.32–2.69 0.63 <0.001 Haptoglobin in suckling piglets 1.01 1.01–1.02 0.01 <0.001 Mean of relative humidity in the farrowing unit 1.10 1.02–1.18 0.09 0.013 PRRSV presence in nursery pigs 6.40 1.74–23.53 1.85 0.005 Sow parity 0.69 0.52–0.93 −0.36 0.016 Model 2: Weaning risk factor Haptoglobin in suckling piglets 1.01 1.00–1.02 0.01 0.028 Mean of relative humidity in the farrowing unit 1.11 1.05–1.17 0.10 <0.001 Sow parity 0.71 0.52–0.97 −0.34 0.031 Model 3: Nursery risk factor PRRSV presence in early nursery pigs 6.69 1.55–28.85 1.90 0.011 Mean of temperature in early nursery 0.13 0.05–0.37 −2.04 <0.001 Sow parity 0.55 0.37–0.83 −0.59 0.004 Abbreviations: CI, confident interval; OR, odds ratio; PRRSV, porcine reproductive and respiratory syndrome virus. associated with elevated levels of cortisol50 ; however, biota, as has been observed in sows vaccinated against in our study, the association of high cortisol levels and G. parasuis.52 S. suis-related disease was observed only in the gen- Our findings suggest that control of infections with eral model. Haptoglobin has been shown to increase other pathogens, improvement of the environmental in feed-deprived piglets,51 which can occur when an conditions and reduction of stress may help reduce animal does not have access to feed due to mobil- the incidence of S. suis disease, with the consequent ity problems caused by the arthritis typical of S. suis decrement in the cost for the producers and the posi- infection, as shown in Model 1. However, in Model 2 tive effect on public health due to both a reduction in we also found that high values of haptoglobin in suck- the use of antimicrobials and a reduction of the risk for ling piglets were correlated with the appearance of the human infections.53 disease 2 weeks later, at the beginning of postwean- However, one of the limitations of the study was that ing. Despite being a statistically significant variable in it was not possible to obtain a laboratory confirma- our models, the magnitude of this influence turned tion of the diagnosis for each case of S. suis-associated out to be low, increasing the possibility of developing disease, as that would have implied the sacrifice of S. suis-associated disease by 1.01 for each mg/dL of the animals. Because of that, laboratory diagnosis was haptoglobin in serum (observed values of haptoglobin carried out only at the farm level, while diagnosis of ranged from 8 to 322.9 mg/dL). individuals was based on the presence of clinical signs As sampling was done just before weaning, animals compatible with S. suis-associated disease. Further- that would be weaned a few days older were more more, given the logistical difficulties associated with a likely to develop S. suis-associated disease. Although longitudinal study, the evaluation of risk factors could the explanation for this relationship seems unclear, it only be carried out on two farms. Therefore, the con- might be associated with differences between the two sistency of the risk factors identified would need to be farms, and also with the presence of different strains corroborated with further studies using a larger num- that would disperse and colonise at different rates. ber of farms, as well as studies in other countries, According to Gebhardt et al.,43 the sex of the animal where other potential factors may influence the risk of does not influence postweaning mortality, nor did it S. suis. influence the appearance of S. suis-associated disease in our study. Hopkins et al.9 observed that piglets from sows CONCLUSION whose previous litters presented with S. suis problems were less likely to develop the disease. In our study, Our study highlights the multifactorial nature of S. we showed that piglets born from older sows were less suis-associated disease, for which both environmen- likely to present problems, and that result was very tal factors and factors related to the host seem to consistent across all the models. The reasons for this be involved in the development of the disease. The finding could be the higher immunological protection information presented in this study can help farmers conferred by the colostrum intake, or changes in the in preventing S. suis outbreaks, enabling them to iden- sow vaginal or nasal microbiota, which consequently tify and potentially control some of the risk factors may influence the development of the piglet micro- involved in the appearance of disease. 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License VETERINARY RECORD 9 of 10 AUTHOR CONTRIBUTIONS 7. Segura M, Aragon V, Brockmeier SL, Gebhart C, Greeff A, Conceptualisation: Carlos Neila-Ibáñez, Sebastián Kerdsin A, et al. Update on Streptococcus suis research and pre- vention in the era of antimicrobial restriction: 4th International Napp, Virginia Aragon and Jordi Casal. Methodol- Workshop on S. suis. Pathogens. 2020;9:374. ogy: Carlos Neila-Ibáñez. Biomarker determination: 8. Reams RY, Glickman LT, Harrington DD, Thacker HL, Lorena Franco-Martínez and José Joaquín Cerón. Data Bowersock TL. Streptococcus suis infection in swine: a ret- analysis: Carlos Neila-Ibáñez and Lola Pailler-García. rospective study of 256 cases. Part II. Clinical signs, gross and Writing the original draft: Carlos Neila-Ibáñez. Writing microscopic lesions, and coexisting microorganisms. J Vet Diagnostic Investig. 1994;6:326–34. the review and editing: Carlos Neila-Ibáñez, Sebastián 9. Hopkins D, Poljak Z, Farzan A, Friendship R. Factors contribut- Napp, Lola Pailler-García, Virginia Aragon and Jordi ing to mortality during a Streptoccocus suis outbreak in nursery Casal. Funding acquisition: Sebastián Napp, Virginia pigs. Can Vet J. 2018;59:623–30. Aragon and Jordi Casal. All authors read and approved 10. Obradovic MR, Segura M, Segalés J, Gottschalk M. Review the final manuscript. of the speculative role of co-infections in Streptococcus suis- associated diseases in pigs. Vet Res. 2021;52:49. 11. Feng W, Laster SM, Tompkins M, Brown T, Xu J-S, Altier C, et al. ACKNOWLEDGEMENTS In utero infection by porcine reproductive and respiratory syn- The authors would like to thank the farm animal drome virus is sufficient to increase susceptibility of piglets to caretakers and veterinarians for their willingness to challenge by Streptococcus suis type II. J Virol. 2001;75:4889–95. collaborate in the study. Diego Pérez and Patricia 12. Thanawongnuwech R, Brown GB, Halbur PG, Roth JA, Royer Pleguezuelos from IRTA-CReSA are thanked for their RL, Thacker BJ. Pathogenesis of porcine reproductive and res- piratory syndrome virus-induced increase in susceptibility to help in sampling. Núria Galofré-Milà from IRTA- Streptococcus suis infection. Vet Pathol. 2000;37:143–52. CReSA is thanked for her assistance in the analysis 13. Xu M, Wang S, Li L, Lei L, Liu Y, Shi W, et al. Secondary infec- of samples. Carlos Neila-Ibáñez, Sebastián Napp, Lola tion with Streptococcus suis serotype 7 increases the virulence Pailler-García, Virginia Aragon and Jordi Casal are also of highly pathogenic porcine reproductive and respiratory grateful to the CERCA programme (Centres de Recerca syndrome virus in pigs. Virol J. 2010;7:184. 14. Wang Q, Zhou H, Hao Q, Li M, Liu J, Fan H. Co-infection de Catalunya Programme) from the Generalitat de with porcine circovirus type 2 and Streptococcus suis serotype Catalunya. 2 enhances pathogenicity by dysregulation of the immune responses in piglets. Vet Microbiol. 2020;243:108653. C O N F L I C T O F I N T E R E S T S TAT E M E N T 15. Lin X, Huang C, Shi J, Wang R, Sun X, Liu X, et al. Investigation of The authors declare they have no conflicts of interest. pathogenesis of H1N1 influenza virus and swine Streptococcus suis serotype 2 co-infection in pigs by microarray analysis. PLoS D ATA AVA I L A B I L I T Y S TAT E M E N T One. 2015;10:e0124086. 16. Vecht U, Wisselink HJ, Van Dijk JE, Smith HE. Virulence of The data are available from the corresponding author Streptococcus suis type 2 strains in newborn germ-free pigs upon request. depends on phenotype. Infect Immun. 1992;60:550–6. 17. Correa-Fiz F, Fraile L, Aragon V. Piglet nasal microbiota at E T H I C S S TAT E M E N T weaning may influence the development of Glässer’s disease Sampling of piglets was done under institutional during the rearing period. BMC Genomics. 2016;17:404. 18. Blanco-Fuertes M, Correa-Fiz F, Fraile L, Sibila M, Aragon V. authorisation from IRTA-CReSA (Protocol number Altered nasal microbiota composition associated with devel- 11199) and followed good veterinary practices in opment of polyserositis by Mycoplasma hyorhinis. Pathogens. accordance with European (Directive 2010/63/EU) 2021;10:603. and Spanish (Real Decreto 53/2013) regulations and 19. Niazy M, Hill S, Nadeem K, Ricker N, Farzan A. Compositional with the approval of the Ethics Commission. analysis of the tonsil microbiota in relationship to Streptococ- cus suis disease in nursery pigs in Ontario. Anim Microbiome. 2022;4:10. REFERENCES 20. Aragon V, Segales J, Tucker AW. Glässer’s disease. In: Zimmer- 1. Gottschalk M, Segura M. Streptococci. In: Zimmerman JJ, Kar- man JJ, Karriker LA, Ramirez A, Schwartz KJ, Stevenson GW, riker LA, Ramirez A, Schwartz KJ, Stevenson GW, Zhang J, Zhang J, editors. Diseases of swine. 11th ed. Hoboken, USA: editors. Diseases of swine. 11th ed. Hoboken, USA: John Wiley John Wiley & Sons, Inc.; 2019. p. 844–53. & Sons, Inc.; 2019. p. 934–50. 21. Vötsch D, Willenborg M, Weldearegay YB, Valentin-Weigand 2. Neila-Ibáñez C, Casal J, Hennig-Pauka I, Stockhofe-Zurwieden P. Streptococcus suis - the “two faces” of a pathobiont in the N, Gottschalk M, Migura-García L, et al. Stochastic assessment porcine respiratory tract. Front Microbiol. 2018;9:480. of the economic impact of Streptococcus suis-associated dis- 22. Dee SA, Carlson AR, Winkelman NL, Corey MM. Effect of man- ease in German, Dutch and Spanish swine farms. Front Vet Sci. agement practices on the Streptococcus suis carrier rate in 2021;8:676002. nursery swine. J Am Vet Med Assoc. 1993;203:295–9. 3. Baele M, Chiers K, Devriese LA, Smith HE, Wisselink HJ, 23. Staats JJ, Feder I, Okwumabua O, Chengappa MM. Streptococ- Vaneechoutte M, et al. The Gram-positive tonsillar and nasal cus suis: past and present. Vet Res Commun. 1997;21:381–407. flora of piglets before and after weaning. J Appl Microbiol. 24. Cerón JJ, Contreras-Aguilar MD, Escribano D, Martínez-Miró 2001;91:997–1003. S, López-Martínez MJ, Ortín-Bustillo A, et al. Basics for 4. Segura M, Fittipaldi N, Calzas C, Gottschalk M. Critical Strepto- the potential use of saliva to evaluate stress, inflammation, coccus suis virulence factors: are they all really critical? Trends immune system, and redox homeostasis in pigs. BMC Vet Res. Microbiol. 2017;25:585–99. 2022;18:81. 5. Segura M, Calzas C, Grenier D, Gottschalk M. Initial steps 25. Murata H. Stress and acute phase protein response: an incon- of the pathogenesis of the infection caused by Streptococ- spicuous but essential linkage. Vet J. 2007;173:473–4. cus suis: fighting against nonspecific defenses. FEBS Lett. 26. Deblanc C, Robert F, Pinard T, Gorin S, Quéguiner S, Gautier- 2016;590:3772–99. Bouchardon AV, et al. Pre-infection of pigs with Mycoplasma 6. Gottschalk M, Xu J, Calzas C, Segura M. Streptococcus suis: a hyopneumoniae induces oxidative stress that influences out- new emerging or an old neglected zoonotic pathogen? Future comes of a subsequent infection with a swine influenza virus Microbiol. 2010;5:371–91. of H1N1 subtype. Vet Microbiol. 2013;162:643–51. 20427670, 2023, 5, Downloaded from https://bvajournals.onlinelibrary.wiley.com/doi/10.1002/vetr.3056 by Readcube (Labtiva Inc.), Wiley Online Library on [27/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 10 of 10 VETERINARY RECORD 27. Neila-Ibáñez C, Brogaard L, Pailler-García L, Martínez J, Segalés cial swine production. I: review of non-infectious contributing J, Segura M, et al. Piglet innate immune response to Streptococ- factors. Transl Anim Sci. 2020;4:462–84. cus suis colonization is modulated by the virulence of the strain. 44. Goyette-Desjardins G, Auger JP, Xu J, Segura M, Gottschalk M. Vet Res. 2021;52:145. Streptococcus suis, an important pig pathogen and emerging 28. Contreras-Aguilar MD, Escribano D, Martínez-Miró S, López- zoonotic agent-an update on the worldwide distribution based Arjona M, Rubio CP, Martínez-Subiela S, et al. Application of on serotyping and sequence typing. Emerg Microbes Infect. a score for evaluation of pain, distress and discomfort in pigs 2014;3:e45. with lameness and prolapses: correlation with saliva biomark- 45. Rieckmann K, Pendzialek SM, Vahlenkamp T, Baums CG. A ers and severity of the disease. Res Vet Sci. 2019;126:155–63. critical review speculating on the protective efficacies of auto- 29. Holtkamp D, Torremorell M, Corzo CA, Linhares D, Nunes M, genous Streptococcus suis bacterins as used in Europe. Porc Yeske P, et al. Proposed modifications to porcine reproduc- Heal Manag. 2020;6:12. tive and respiratory syndrome virus herd classification. J Swine 46. Mathieu-Denoncourt A, Letendre C, Auger JP, Segura M, Health Prod. 2021;29:261–70. Aragon V, Lacouture S, et al. Limited interactions between 30. Ishida S, Tien LHT, Osawa R, Tohya M, Nomoto R, Kawamura Streptococcus suis and Haemophilus parasuis in in vitro co- Y, et al. Development of an appropriate PCR system for the infection studies. Pathogens. 2018;7:7. reclassification of Streptococcus suis. J Microbiol Methods. 47. Campbell JM, Crenshaw JD, Polo J. The biological stress of early 2014;107:66–70. weaned piglets. J Anim Sci Biotechnol. 2013;4:19. 31. Okura M, Lachance C, Osaki M, Sekizaki T, Maruyama F, 48. Proudfoot KL, Weary DM, von Keyserlingk MAG. Linking the Nozawa T, et al. Development of a two-step multiplex PCR social environment to illness in farm animals. Appl Anim Behav assay for typing of capsular polysaccharide synthesis gene Sci. 2012;138:203–15. clusters of Streptococcus suis. J Clin Microbiol. 2014;52:1714–9. 49. Martínez-Miró S, Tecles F, Ramón M, Escribano D, Hernández 32. Versalovic J, Koeuth T, Lupski R. Distribution of repetitive DNA F, Madrid J, et al. Causes, consequences and biomarkers of sequences in eubacteria and application to fingerprinting of stress in swine: an update. BMC Vet Res. 2016;12:171. bacterial genomes. Nucleic Acids Res. 1991;19:6823–31. 50. Dhabhar FS. Enhancing versus suppressive effects of stress 33. Galofré-Milà N, Correa-Fiz F, Lacouture S, Gottschalk M, on immune function: implications for immunoprotection and Strutzberg-Minder K, Bensaid A, et al. A robust PCR for the immunopathology. Neuroimmunomodulation. 2009;16:300– differentiation of potential virulent strains of Haemophilus 17. parasuis. BMC Vet Res. 2017;13:124. 51. Ott S, Soler L, Moons CPH, Kashiha MA, Bahr C, Vandermeulen 34. López-Valiñas Á, Sisteré-Oró M, López-Serrano S, Baioni L, J, et al. Different stressors elicit different responses in the sali- Darji A, Chiapponi C, et al. Identification and characterization vary biomarkers cortisol, haptoglobin, and chromogranin A in of swine influenza virus H1N1 variants generated in vaccinated pigs. Res Vet Sci. 2014;97:124–8. and nonvaccinated, challenged pigs. Viruses. 2021;13:2087. 52. Blanco-Fuertes M, Correa-Fiz F, López-Serrano S, Sibila M, 35. Escribano D, Fuentes-Rubio M, Cerón JJ. Validation of an auto- Aragon V. Sow vaccination against virulent Glaesserella para- mated chemiluminescent immunoassay for salivary cortisol suis shapes the nasal microbiota of their offspring. Sci Rep. measurements in pigs. J Vet Diagnostic Investig. 2012;24:918– 2022;12:3357. 23. 53. Rayanakorn A, Goh BH, Lee L-H, Khan TM, Saokaew S. Risk 36. Kaiser M, Jacobson M, Andersen PH, Bækbo P, Cerón JJ, Dahl factors for Streptococcus suis infection: a systematic review and J, et al. Inflammatory markers before and after farrowing in meta-analysis. Sci Rep. 2018;8(1):13358. healthy sows and in sows affected with postpartum dysgalactia syndrome. BMC Vet Res. 2018;14:83. 37. Rhee SG, Chang TS, Jeong W, Kang D. Methods for detection and measurement of hydrogen peroxide inside and outside of S U P P O RT I N G I N F O R M AT I O N cells. Mol Cells. 2010;29:539–49. Additional supporting information can be found 38. Rubio CP, Mainau E, Cerón JJ, Contreras-Aguilar MD, online in the Supporting Information section at the Martínez-Subiela S, Navarro E, et al. Biomarkers of oxida- end of this article. tive stress in saliva in pigs: analytical validation and changes in lactation. BMC Vet Res. 2019;15:144. 39. Dohoo IR, Martin SW, Stryhn H. Veterinary epidemiologic research. 1st ed. Charlottetown, Canada: AVC Inc; 2003. 40. R Core Team R: a language and environment for statistical How to cite this article: Neila-Ibáñez C, Napp S, computing. Vienna: R Foundation for Statistical Computing; Pailler-García L, Franco-Martínez L, Cerón JJ, 2020. Aragon V, et al. Risk factors associated with 41. Bates D, Mächler M, Bolker BM, Walker SC. Fitting linear Streptococcus suis cases on pig farms in Spain. mixed-effects models using lme4. J Stat Softw. 2015;67:1–48. 42. Zhang D. CRAN - Package rsq. 2021. Vet Rec. 2023;e3056. 43. Gebhardt JT, Tokach MD, Dritz SS, DeRouchey JM, Woodworth https://doi.org/10.1002/vetr.3056 JC, Goodband RD, et al. Postweaning mortality in commer-