Antimicrobial Susceptibility Testing in Secondary Viral Co-infections

Overview and Principles of Antimicrobial Susceptibility Testing in Secondary Viral Co-infections

The clinical and diagnostic landscape of infectious disease is profoundly complicated by the phenomenon of secondary bacterial and fungal infections that arise in the wake of a primary viral insult. This is not a niche occurrence but a central driver of morbidity and mortality in both human and veterinary medicine, as evidenced by historical pandemics and contemporary outbreaks. The 1918 influenza pandemic, the 2009 H1N1 pandemic, and most recently the COVID-19 pandemic have all underscored a critical, recurring theme: the majority of severe outcomes and fatalities are frequently attributable not to the virus alone, but to the ensuing bacterial superinfection [9, 21]. In the veterinary context, analogous dynamics are observed across species, from the Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) predisposing swine to secondary bacterial pneumonia [1] to Avian Influenza Virus facilitating secondary colibacillosis in poultry [16]. This establishes the fundamental principle that antimicrobial susceptibility testing (AST) in the context of secondary viral co-infections is not merely a routine laboratory procedure; it is a critical, time-sensitive, and mechanistically nuanced intervention that sits at the intersection of virology, bacteriology, immunology, and clinical pharmacology.

The core principle governing AST in this setting is the recognition that the host-pathogen interface is fundamentally altered by the antecedent viral infection. The virus does not simply provide a passive opportunity for bacterial invasion; it actively engineers a permissive environment. This immunopathogenesis is the bedrock upon which the rationale for AST is built. Viral infections, particularly those caused by respiratory viruses like influenza, SARS-CoV-2, and Canine Distemper Virus, induce a state of profound immune dysregulation. This includes impaired mucociliary clearance, disruption of epithelial barrier integrity, and, critically, a functional paralysis of key innate immune cells such as alveolar macrophages and neutrophils [2, 17]. For instance, influenza virus infection has been shown to decrease the phagocytic capacity of macrophages and their production of reactive oxygen species, creating a window of vulnerability for opportunistic pathogens like Streptococcus pneumoniae and Staphylococcus aureus [2, 13]. This mechanistic understanding dictates that the bacterial pathogens isolated from a post-viral infection are not random; they are often specific, highly adapted opportunists that exploit these precise immune deficits. Consequently, the AST profile of these pathogens must be interpreted within this context of an immunocompromised host, where even an organism with an intermediate susceptibility profile might be clinically resistant due to the host's diminished ability to clear the infection.

A second foundational principle is the distinction between primary co-infection (present on admission) and secondary infection (developing after 48-72 hours of hospitalization). This temporal dichotomy has profound implications for AST interpretation and antimicrobial stewardship. Primary co-infections, often seen in community-acquired pneumonia (CAP), are frequently caused by pathogens such as Streptococcus pneumoniae, Haemophilus influenzae, or Staphylococcus aureus [4, 10]. The AST profiles for these organisms are often more predictable and may be guided by local community-level antibiograms. In stark contrast, secondary infections, particularly in patients requiring intensive care and mechanical ventilation, are overwhelmingly nosocomial in origin. The causative agents shift dramatically to multidrug-resistant (MDR) Gram-negative bacilli such as Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species [6, 8, 12]. In the context of COVID-19, studies have consistently shown that while early co-infection rates are low (3-5%), secondary infection rates in ICU patients can be alarmingly high (14-50%), with a predominance of carbapenem-resistant organisms [12, 18, 20]. This shift is driven by prolonged hospitalization, invasive devices, and the intense selective pressure of widespread empiric antibiotic use [9, 23]. Therefore, the AST strategy must be dynamic: initial testing should focus on common community-acquired pathogens, while subsequent testing must be broadened to include a comprehensive panel of nosocomial MDR pathogens, with a particular emphasis on detecting carbapenemases (e.g., KPC, NDM, OXA) and ESBLs.

The methodology of AST itself must be adapted to the unique challenges of the post-viral environment. The standard of care remains phenotypic AST, using broth microdilution or disk diffusion to determine minimum inhibitory concentrations (MICs) [7, 22]. However, several factors complicate this process. First, the high rate of empiric antibiotic administration in these patients - often exceeding 70-90% in ICU cohorts [9, 11, 12] - can suppress bacterial growth in culture, leading to false-negative results. This necessitates the use of more sensitive detection methods, such as collecting samples prior to antibiotic administration or utilizing culture-independent techniques. Second, the polymicrobial nature of these infections is a major challenge. The porcine respiratory disease complex (PRDC) is a prime example, where co-detection of multiple pathogens like PRRSV-1, Glaesserella parasuis, Streptococcus suis, and Mesomycoplasma hyorhinis is the rule, not the exception [1]. In such cases, isolating a single dominant pathogen for AST may not reflect the full resistance profile of the consortium. Genotypic methods, such as multiplex PCR panels and targeted next-generation sequencing (tNGS), are increasingly vital here. They can detect the presence of resistance genes (e.g., mecA for MRSA, blaCTX-M for ESBLs, erm for macrolide resistance) directly from the clinical sample, bypassing the need for culture and providing a rapid, comprehensive resistance profile even in the face of polymicrobial growth or prior antibiotic use [4, 5, 14]. The WHO and CDC have endorsed the integration of such molecular surveillance for AMR, recognizing its superiority in detecting the true burden of resistance in complex clinical scenarios.

A critical principle that must guide the entire AST process is the imperative for antimicrobial stewardship (AMS) . The COVID-19 pandemic served as a stark lesson in the dangers of empiric antibiotic overuse. Despite low documented rates of bacterial co-infection, a vast majority of hospitalized patients received broad-spectrum antibiotics, often including azithromycin for its unproven antiviral effects [9, 23]. This practice, driven by diagnostic uncertainty and fear of missing a co-infection, has been a major driver of the post-pandemic surge in AMR [23]. The role of AST in AMS is therefore twofold. First, a negative AST result (i.e., no significant bacterial growth or negative molecular panel) is arguably as important as a positive one. It provides the objective evidence needed to justify antibiotic de-escalation or discontinuation, a cornerstone of stewardship. The use of biomarkers like procalcitonin (PCT) in conjunction with AST can further strengthen this decision-making, as a low PCT combined with a negative bacterial panel strongly argues against bacterial infection [24]. Second, a positive AST result must be used to guide targeted therapy. The goal is to move away from the shotgun approach of empiric broad-spectrum regimens (e.g., meropenem plus vancomycin) towards a focused, narrow-spectrum agent based on the specific pathogen and its MIC. For example, if AST reveals a K. pneumoniae that is susceptible to a first-generation cephalosporin, that agent should be preferred over a carbapenem, even if the patient is critically ill. This principle of "de-escalation" is the most powerful tool in the AMS arsenal and is entirely dependent on high-quality, timely AST data.

Finally, the principles of AST in veterinary secondary infections must be viewed through a One Health lens. The same MDR clones of Escherichia coli, Klebsiella pneumoniae, and Acinetobacter baumannii that plague human ICUs are also found in livestock, companion animals, and the environment [19]. The use of critically important antimicrobials in food-producing animals, often to control secondary bacterial outbreaks following viral infections like Porcine Epidemic Diarrhea Virus or Avian Influenza Virus, creates a reservoir of resistance genes that can be transmitted to human pathogens via the food chain or direct contact [16, 19]. Therefore, AST in veterinary medicine is not just about treating the individual animal; it is a critical component of global AMR surveillance. The World Organisation for Animal Health (WOAH) emphasizes the need for standardized AST methods and data sharing across the human-animal-environment interface. The principles of AMS - culture-guided therapy, de-escalation, and avoidance of prophylactic use of last-resort drugs - apply equally in veterinary practice. The detection of a multidrug-resistant Salmonella Typhimurium in a duck co-infected with Duck Hepatitis A Virus or a carbapenem-resistant Aeromonas veronii in a koi carp co-infected with Koi Herpesvirus is a sentinel event that signals a broader ecological threat [3, 15]. In this context, the veterinary clinical pathologist serves as a sentinel for public health, and the AST result is a data point in a global surveillance network.

Molecular Pathogenesis of Bacterial Superinfection Following Viral Respiratory Tract Infections

The phenomenon of bacterial superinfection following viral respiratory tract infections represents one of the most clinically consequential and mechanistically intricate paradigms in infectious disease medicine. The molecular pathogenesis underlying this sequential or concurrent microbial assault is not merely a matter of two pathogens occupying the same anatomical niche; rather, it involves a sophisticated, multi-layered cascade of host-pathogen interactions that fundamentally alter the respiratory ecosystem. Understanding these molecular mechanisms is paramount for the rational design of antimicrobial susceptibility testing protocols and for guiding therapeutic interventions in the context of secondary viral co-infections. This section delineates the core molecular pathways - from epithelial barrier disruption and immune dysregulation to altered pathogen clearance and receptor modulation - that collectively create a permissive environment for bacterial opportunists following viral insult.

Viral-Induced Epithelial Barrier Dysfunction and Receptor Upregulation

The respiratory epithelium serves as the first line of defense against microbial invasion, and viral pathogens have evolved sophisticated strategies to compromise this barrier. Infection with viruses such as Avian Influenza Virus, Swine Influenza A Virus, and Canine Influenza A Virus leads to widespread desquamation and necrosis of ciliated epithelial cells, stripping the airway of its mucociliary escalator and exposing the basement membrane. This physical disruption creates adhesion sites for bacterial pathogens, particularly Streptococcus pneumoniae and Staphylococcus aureus, which possess surface adhesins that bind to exposed extracellular matrix components such as fibronectin, laminin, and collagen [9, 17, 21]. The synergistic relationship between influenza virus and S. aureus is particularly well-documented; specific staphylococcal proteases can cleave the influenza hemagglutinin molecule, enhancing viral replication and spread, while the virus simultaneously upregulates bacterial adhesion receptors on host cells [17]. This bidirectional molecular facilitation creates a self-amplifying cycle of pathogenesis.

Beyond physical disruption, viral neuraminidase activity plays a critical role in unmasking bacterial receptors. Influenza virus neuraminidase cleaves sialic acid residues from glycoproteins on the airway surface, exposing cryptic binding sites for bacterial adhesins. This enzymatic activity is not merely a byproduct of viral replication but a direct molecular mechanism that facilitates bacterial adherence. Computational modeling of influenza-S. pneumoniae co-infection has demonstrated that antiviral treatment with oseltamivir, a neuraminidase inhibitor, can partially mitigate this effect, reducing both viral titers and secondary bacterial burdens, albeit with limited antibacterial efficacy that is dose-dependent [28]. The clinical relevance of this mechanism is underscored by the observation that bacterial superinfection rates during influenza pandemics have historically exceeded 50% of fatal cases, with S. pneumoniae and S. aureus being the predominant isolates [9, 21].

Innate Immune Dysregulation: The Double-Edged Sword

The host innate immune response to viral infection is a carefully orchestrated cascade designed to contain and eliminate the pathogen. However, the same inflammatory milieu that controls viral replication can paradoxically impair antibacterial defenses, creating a window of vulnerability for secondary bacterial invasion. This immunopathological state is characterized by profound alterations in macrophage and neutrophil function, cytokine dysregulation, and natural killer (NK) cell-mediated immunopathology.

Macrophage and Neutrophil Dysfunction: Alveolar macrophages are the sentinel phagocytes of the lower respiratory tract, responsible for the rapid clearance of inhaled bacteria. Following influenza virus infection, these cells exhibit a state of functional paralysis, characterized by reduced phagocytic capacity, impaired intracellular killing, and diminished reactive oxygen species (ROS) production [2, 17]. The molecular underpinnings of this dysfunction involve type I and type II interferon signaling, which downregulates the expression of scavenger receptors and NADPH oxidase components. Chen et al. demonstrated that targeting Toll-like receptor 7 (TLR7) signaling with the antagonist IRS661 could reverse this macrophage dysfunction in a murine model of influenza-S. aureus co-infection, restoring phagocytosis, bactericidal activity, and ROS production while simultaneously reducing the levels of pro-inflammatory cytokines such as IL-6, IL-1B, and TNF-α [2]. This finding highlights the central role of TLR7-driven inflammatory signaling in mediating macrophage suppression and identifies a potential therapeutic target for mitigating secondary bacterial infections.

Neutrophils, the first responders to bacterial infection, are similarly compromised. Viral infection induces a state of neutrophil exhaustion, characterized by reduced chemotaxis, impaired degranulation, and accelerated apoptosis. The production of anti-inflammatory cytokines, particularly IL-10, further suppresses neutrophil function. Aguilera et al. identified NK cells as a critical source of IL-10 during S. pneumoniae infection, and depletion of NK cells resulted in reduced bacterial burdens, suggesting that NK cell-derived IL-10 actively facilitates bacterial pathogenesis [13]. This finding challenges the conventional view of NK cells as purely antiviral effectors and positions them as key orchestrators of the immunosuppressive environment that permits bacterial superinfection.

Cytokine Storm and Immunopathology: The excessive and dysregulated production of pro-inflammatory cytokines, often termed a "cytokine storm," is a hallmark of severe viral respiratory infections, including those caused by highly pathogenic Avian Influenza Virus, Porcine Reproductive and Respiratory Syndrome Virus, and SARS-CoV-2. While this hyperinflammatory state is intended to control viral replication, it paradoxically contributes to tissue damage, epithelial barrier disruption, and immune paralysis. Elevated levels of TNF-α, IL-6, and IL-1B directly impair macrophage and neutrophil function, while also inducing the expression of adhesion molecules on endothelial cells that facilitate bacterial translocation from the airway into the bloodstream [32, 33]. The resulting sepsis, whether viral or mixed, reflects disease severity and may contribute to functional decline and susceptibility to reinfections [18].

Adaptive Immune Suppression and Altered Lymphocyte Dynamics

The adaptive immune response is also profoundly affected by preceding viral infection, with consequences that extend well beyond the acute phase. Lymphopenia, particularly depletion of CD4+ and CD8+ T cells, is a consistent feature of severe viral respiratory infections and is associated with increased risk of secondary bacterial infections [32]. The mechanisms underlying this lymphopenia include direct viral cytopathicity, activation-induced cell death, and redistribution of lymphocytes to peripheral tissues. In a rare case of Avian Influenza Virus (H5N1) co-infection with HIV, Fox et al. documented profound CD4+ T cell depletion (admission CD4 count of 100 cells/μL) and aberrant CD8+ T cell activation (shift to a CD27+CD28- phenotype), which, while not preventing viral clearance, likely increased susceptibility to a fatal secondary bacterial infection [29]. This case illustrates the critical role of adaptive immunity in controlling bacterial superinfection, even when antiviral immunity remains intact.

The humoral immune response is also compromised. Viral infection can induce a state of B cell anergy or exhaustion, reducing the production of pathogen-specific antibodies. In the context of Porcine Reproductive and Respiratory Syndrome Virus infection, which is known to cause profound immunosuppression, co-infection with bacterial pathogens such as Glaesserella parasuis and Streptococcus suis is a common and often fatal complication [1]. The large-scale landscape study of porcine respiratory disease complex (PRDC) by Chacón-Pérez et al. revealed that secondary opportunistic bacteria were predominantly detected in post-weaning piglets, a period coinciding with waning maternal immunity and heightened susceptibility to viral infections [1]. This epidemiological pattern underscores the interplay between adaptive immune status and viral-bacterial synergy.

Altered Bacterial Clearance and the Role of the Microbiome

The respiratory microbiome plays a critical role in maintaining immune homeostasis and providing colonization resistance against pathogens. Viral infection disrupts this delicate microbial ecosystem, leading to dysbiosis that favors the outgrowth of opportunistic bacteria. The mechanisms include direct viral-mediated killing of commensal bacteria, alterations in the local nutrient environment, and immune-mediated selection pressures [21, 25]. The resulting loss of microbial diversity and depletion of protective commensals, such as Corynebacterium and Dolosigranulum species, creates a niche that can be exploited by pathogens.

Furthermore, viral infection can directly impair the mechanical clearance of bacteria from the airway. The mucociliary escalator, which relies on coordinated ciliary beating and mucus production, is severely compromised by viral-induced epithelial damage. Mucus hypersecretion, driven by goblet cell hyperplasia and increased mucin gene expression, can paradoxically impair bacterial clearance by creating a stagnant, nutrient-rich environment that promotes bacterial growth [18]. In patients with chronic obstructive pulmonary disease (COPD), pre-existing mucus plugging is exacerbated by viral infection, creating a particularly permissive environment for secondary bacterial infections [18].

Specific Viral-Bacterial Synergy: Molecular Examples

The molecular pathogenesis of superinfection is not uniform across all viral-bacterial pairs; rather, specific combinations exhibit unique synergistic interactions. The influenza-S. pneumoniae paradigm remains the most extensively studied. Influenza virus neuraminidase cleaves sialic acid residues from the pneumococcal capsule, exposing bacterial adhesins and facilitating adherence to host cells. Conversely, pneumococcal neuraminidase can activate latent influenza virus, enhancing viral replication [21]. This bidirectional enzymatic cooperation creates a positive feedback loop that amplifies both infections.

The interaction between Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) and Mesomycoplasma hyorhinis provides another compelling example. PRRSV infection induces a state of profound immunosuppression, characterized by reduced alveolar macrophage function and impaired interferon production. This creates a permissive environment for M. hyorhinis, a commensal that can become pathogenic under conditions of immune compromise. The co-occurrence network analysis by Chacón-Pérez et al. revealed that M. hyorhinis showed strong associations with nine of the evaluated pathogens, including G. parasuis (OR = 4.05), S. suis (OR = 1.57), and PRRSV-1 (OR = 2.55), highlighting its role as a pivotal pathogen in the PRDC [1].

In the context of Canine Influenza A Virus (H3N2) infection, Zhou et al. documented a novel co-infection with Enterococcus faecalis in dogs, resulting in severe pneumonia and the formation of cervical cysts [27]. The authors hypothesized that the viral infection promoted the secondary invasion of this opportunistic bacterium, leading to more severe and complicated clinical outcomes. This case illustrates that the molecular mechanisms of superinfection extend beyond the well-characterized influenza-pneumococcus paradigm and can involve a diverse array of bacterial pathogens.

Implications for Antimicrobial Susceptibility Testing

The molecular pathogenesis of bacterial superinfection has direct implications for antimicrobial susceptibility testing (AST). The altered host environment created by viral infection can influence the pharmacokinetics and pharmacodynamics of antimicrobial agents, potentially affecting their efficacy. For example, the inflammatory milieu can alter drug penetration into tissues, while the presence of biofilm-forming bacteria, such as Klebsiella pneumoniae and Pseudomonas aeruginosa, can further complicate treatment [12, 30]. The high prevalence of multidrug-resistant (MDR) organisms in secondary infections, particularly in the intensive care unit (ICU) setting, underscores the need for rapid and comprehensive AST to guide targeted therapy [12, 20, 26].

The emergence of carbapenem-resistant Klebsiella pneumoniae (CRKP) and Acinetobacter baumannii as predominant pathogens in secondary infections following COVID-19 and influenza highlights the critical role of horizontal gene transfer in the dissemination of resistance determinants [12, 31]. The co-colonization of patients with distinct CRKP clones, as documented by Fang et al., can facilitate the transfer of resistance plasmids, enabling hypervirulent strains to acquire carbapenem resistance and cause secondary bloodstream infections [31]. This dynamic underscores the need for genomic surveillance and molecular characterization of resistance mechanisms in the context of viral-bacterial co-infections.

In conclusion, the molecular pathogenesis of bacterial superinfection following viral respiratory tract infections is a complex, multi-factorial process involving epithelial barrier disruption, innate and adaptive immune dysregulation, altered bacterial clearance, and specific viral-bacterial synergistic interactions. Understanding these mechanisms is essential for the rational design of diagnostic and therapeutic strategies, including the appropriate use of antimicrobial susceptibility testing to guide targeted therapy and mitigate the emergence of antimicrobial resistance.

Methodological Approaches and Standardized Protocols for Susceptibility Testing in Co-infected Clinical Samples

The diagnostic interrogation of clinical specimens derived from hosts experiencing concurrent viral and bacterial infections presents a unique and formidable challenge to the clinical microbiology laboratory. Unlike monomicrobial infections, where a single etiological agent is expected and standard interpretive criteria are well-established, co-infected samples are characterized by a polymicrobial landscape where the viral pathogen has initiated a cascade of immunopathological and physiological changes that dictate the very nature of the secondary bacterial community. This section delineates the specialized methodological approaches and standardized protocols required for accurate antimicrobial susceptibility testing (AST) in this complex context, emphasizing the critical need to move beyond simplistic culture-based paradigms towards integrated, multi-modal diagnostic algorithms. The astute clinical pathologist must recognize that the viral infection is not merely a bystander but an active modulator of both the bacterial phenotype and the host's response to therapy, necessitating a nuanced interpretation of AST results.

Foundational Challenges in Polymicrobial Diagnostic Stewardship

The primary obstacle in performing AST on co-infected samples is the inherent difficulty in distinguishing clinically significant pathogens from colonizers or contaminants, a dilemma magnified by the altered microenvironment created by the viral infection. Viral-induced tissue damage, such as the desquamation of respiratory epithelium seen with Avian Influenza Virus or Infectious Bronchitis Virus, provides adherence sites and nutrients for opportunistic bacteria like Escherichia coli or Staphylococcus aureus [17, 21]. Similarly, the immunosuppression characteristic of infections with Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) or Canine Distemper Virus can allow for the overgrowth of commensal organisms that would otherwise be held in check [1, 41]. Consequently, a positive culture for a potential pathogen from a respiratory sample in a patient with SARS-CoV-2 (or the veterinary equivalent) does not automatically denote a true co-infection requiring targeted therapy; it may represent viral shedding or irrelevant colonization. The landmark study by Cox et al. [9] highlighted that during the COVID-19 pandemic, up to 74.5% of ICU patients received antibiotics despite documented bacterial co-infection rates being far lower, a stark illustration of the diagnostic uncertainty that plagues this field. Therefore, any protocol for AST in co-infection must be underpinned by rigorous clinical correlation and, ideally, quantitative or semi-quantitative culture methods to differentiate significant growth from background flora.

Pre-analytical Considerations: Specimen Selection and Transport

The adage "garbage in, garbage out" is nowhere more pertinent than in AST for co-infected samples. The choice of specimen type is paramount and must be guided by the anatomical site of the viral insult. For respiratory viral co-infections, bronchoalveolar lavage (BAL) fluid or protected brush specimens offer superior diagnostic yield over swabs or expectorated sputum, as they sample the lower airways where secondary pneumonia manifests [10, 33]. In the context of enteric co-infections, such as those involving Porcine Epidemic Diarrhea Virus (PEDV) or Canine Parvovirus, fecal samples must be processed promptly to prevent overgrowth of fastidious organisms and to minimize the loss of obligate anaerobes that may be relevant [36, 39]. Regardless of the source, the sample must be transported in appropriate media that preserves both viral integrity (if molecular testing is pending) and bacterial viability, while also suppressing the growth of commensal microbiota that could outcompete the target pathogen.

A critical and often overlooked pre-analytical step is the implementation of a "viral-first" screening algorithm. Before dedicating resources to bacterial culture and AST, the clinical laboratory should ideally confirm the presence of the suspected viral agent using sensitive molecular methods like RT-qPCR or multiplex respiratory panels. This confirmation not only validates the clinical suspicion of co-infection but also informs the selection of culture media and incubation conditions. For instance, a sample from a pig with respiratory disease confirmed to be positive for Swine Influenza A Virus should be specifically cultured for Glaesserella parasuis and Streptococcus suis, pathogens known to synergize with the virus [1]. Conversely, in a case of koi carp mortality where Koi Herpesvirus is detected, the focus of bacterial culture should shift towards opportunistic aquatic pathogens like Aeromonas veronii or Shewanella spp., as demonstrated by Tasci et al. [3] in a genome-resolved co-infection study.

Culture-Based Strategies for Isolation in the Presence of Virus

The core challenge in culture-based AST for co-infections is obtaining a pure isolate of the bacterial pathogen. The presence of viral cytopathic effect (CPE) in cell culture is not a concern for standard bacterial agar plates, but the altered host microbiota can be. For respiratory samples, the standard use of selective and differential media is non-negotiable. MacConkey agar (for Gram-negative bacilli), Mannitol Salt Agar (for staphylococci), and Chocolate agar (for fastidious organisms like Haemophilus spp. or Histophilus somni) should be inoculated alongside non-selective blood agar [22, 34]. However, the veterinary clinical pathologist must be prepared to use more specialized media.

• For swine respiratory co-infections involving PRRSV or Porcine Circovirus 2, the isolation of Mycoplasma hyopneumoniae or M. hyorhinis requires specific Friis or Hayflick media, which must be inoculated even if the sample is heavily contaminated with other bacteria. Chacón-Pérez et al. [1] found M. hyorhinis to be a pivotal pathogen in co-detection networks, yet its fastidious nature means it is frequently missed if only standard media are used.
• For avian co-infections with Infectious Bursal Disease Virus or Newcastle Disease Virus, the isolation of E. coli (avian pathogenic E. coli, APEC) is routine, but the presence of Salmonella spp. requires enrichment broths (e.g., Selenite F or Rappaport-Vassiliadis) followed by selective plating on XLD or Hektoen enteric agar [15, 16].
• For aquatic co-infections, the isolation of bacteria from fish with Channel Catfish Virus or Tilapia Lake Virus presents unique challenges. Samples must be cultured on Lowenstein-Jensen or specific media for Streptococcus agalactiae and Francisella noatunensis, and incubation at lower temperatures (25-28 degrees C) is required [3].

A critical methodological point is the use of enrichment culture for samples with low suspected bacterial burden. In the early stages of a secondary infection, or when the patient has already been treated empirically, the number of viable bacteria may be low. Inoculating a sample into a non-selective broth (e.g., Brain Heart Infusion) for 4-6 hours before plating can resuscitate stressed cells and increase the yield of positive cultures, enabling subsequent AST [35].

The Problem of Polymicrobial Growth on Primary Plates

A frequent occurrence in co-infected samples is the growth of multiple bacterial colony morphotypes on the primary isolation plate. The protocol must define a clear decision tree for selecting which colony to test. The general rule is to prioritize the most abundant morphotype in quantitative cultures or, in semi-quantitative cultures, colonies growing in the second or third streak quadrant. However, in co-infections, less abundant but highly virulent pathogens may be of greater clinical significance. For example, in a bovine respiratory disease complex case involving Bovine Herpesvirus 1 and Mannheimia haemolytica, the latter, even if present in lower numbers than commensal Pasteurella multocida, should be the target for AST due to its known leukotoxin-producing capability. The use of MALDI-TOF MS allows for rapid identification of individual colonies from mixed cultures, guiding the selection of isolates for susceptibility testing [3]. If MALDI-TOF is unavailable, a standardized algorithm based on colony morphology, Gram stain, and rapid biochemical tests (e.g., oxidase, catalase) must be followed.

Phenotypic Antimicrobial Susceptibility Testing Protocols for Co-infection Isolates

Once a pure bacterial isolate is obtained, the performance of AST follows established standards (CLSI or EUCAST), but with specific modifications for the co-infection context. The standard disk diffusion (Kirby-Bauer) method or broth microdilution (MIC) is acceptable, but the panel of antibiotics tested must be expanded to reflect the clinical reality of the co-infected patient.

1. Antibiotic Panel Selection: The panel must be tailored to the pathogen spectrum typical of the specific viral co-infection. For instance, for secondary infections following Canine Influenza A Virus, where Staphylococcus pseudintermedius and E. coli are common, the panel should include amoxicillin-clavulanate, cephalexin, enrofloxacin, and doxycycline to cover both the Gram-positive cocci and Gram-negative rods that may be present [27]. For COVID-19 associated secondary infections, the data clearly show a high prevalence of multidrug-resistant (MDR) Acinetobacter baumannii and Klebsiella pneumoniae [6, 8, 12]. Therefore, the AST panel must include carbapenems (e.g., meropenem, imipenem), colistin, and tigecycline to identify the narrow therapeutic window available for extensively drug-resistant (XDR) pathogens [6, 26]. Standard panels that only test first-line agents (e.g., tetracycline, penicillin) are wholly inadequate in this setting.

2. Interpretation of Results in the Context of Host Status: The standard MIC breakpoints are derived from pharmacokinetic/pharmacodynamic (PK/PD) data in immunocompetent hosts. In a co-infected patient, the host's immune system is compromised by the viral infection, which may impair the clearance of bacteria. For example, a minimum inhibitory concentration (MIC) that is considered "susceptible" for Streptococcus pneumoniae in a healthy individual might be clinically insufficient in a patient co-infected with Avian Influenza Virus, where viral neuraminidase cleaves sialic acid residues and impairs phagocytosis [13, 28]. While clinical microbiology laboratories cannot routinely adjust breakpoints for individual patients, the reporting pathologist must be aware of this synergy. This is where the concept of antibiotic efficacy in the co-infected niche becomes critical. The protocol should include a comment in the report if the isolated pathogen is a known protagonist in viral-bacterial synergy, alerting the clinician to the potential need for higher doses or combination therapy.

3. Broth Microdilution for Fastidious Organisms: For pathogens like Mycoplasma hyopneumoniae or Haemophilus parasuis, which are common in viral co-infections, agar dilution or broth microdilution using specialized, supplemented media (e.g., PPLO broth for mycoplasmas, or Haemophilus Test Medium) is mandatory. Disk diffusion is unreliable for these organisms. The protocol must specify the use of cation-adjusted Mueller-Hinton broth with appropriate growth factors (e.g., NAD for Histophilus somni or Glaesserella parasuis) and a standardized inoculum (0.5 McFarland standard) [22].

4. Screening for Resistance Mechanisms: Given the high rate of MDR in secondary infections, phenotypic screening for key resistance mechanisms should be integrated into the standard workflow. This includes:

   • ESBL confirmation: Using the combination disk method (e.g., cefotaxime vs. cefotaxime/clavulanate) for any E. coli or Klebsiella spp. isolate with elevated MICs to third-generation cephalosporins [30, 39].
   • Carbapenemase detection: The modified carbapenem inactivation method (mCIM) or EDTA-modified CIM (eCIM) is essential for any Gram-negative isolate showing resistance to meropenem or imipenem, given the high prevalence of carbapenem-resistant organisms in co-infected ICU patients [26, 31].
   • Methicillin resistance (MRSA/MRS): Cefoxitin disk screening is a standard surrogate for mecA gene detection in staphylococci. This is particularly critical in veterinary medicine for S. pseudintermedius in canine and feline co-infections [40, 42].

Molecular and Genotypic Approaches as Adjuncts to Phenotypic AST

Phenotypic culture and AST remain the gold standard, but their drawbacks - long turnaround time and failure to detect non-cultivable organisms - are amplified in the co-infection scenario. Genotypic methods are rapidly becoming indispensable.

Targeted Next-Generation Sequencing (tNGS) and Metagenomics: Direct sequencing of clinical samples (e.g., sputum, BAL) can circumvent the need for culture, providing a comprehensive snapshot of the entire polymicrobial community, including the virus and all bacteria. Beloussov et al. [5] demonstrated that sputum tNGS detected substantially more microbial diversity and a greater number of antimicrobial resistance genes (ARGs) than culture from K. pneumoniae CAP cases. This is because tNGS can detect bacteria that are present in low numbers, are non-viable, or are difficult to culture. However, it is critical to acknowledge that tNGS does not provide phenotypic MIC values. It only identifies the presence of resistance genes. The correlation between genotype and phenotype is not perfect; a gene may be present but not expressed, or a novel mutation may confer resistance not captured by the database [5, 38].

The Integrated Diagnostic Algorithm: The most robust protocol for AST in co-infected samples is one that integrates both phenotypic and genotypic data. The workflow should be:

1. Direct Molecular Screening: Use a multiplex PCR panel (e.g., FilmArray Respiratory Panel) on the primary sample to rapidly identify the viral agent and common bacterial co-pathogens [10, 43]. This provides actionable information within 1-2 hours.
2. Standard Culture and AST: Inoculate the sample onto appropriate agar and incubate. If a pathogen is isolated, perform standard AST.
3. Targeted NGS for Discordant Cases: If the patient is not responding to therapy guided by phenotypic AST, or if the culture is negative but molecular screening suggests a pathogen, perform tNGS or whole genome sequencing (WGS) on the primary sample. This can reveal hidden pathogens or resistance mechanisms not detected by routine AST [3, 26].
4. Genomic Resistome Surveillance: For epidemiological purposes, the resistome (the collection of ARGs) from co-infected samples should be characterized using metagenomic approaches. This data can inform hospital infection control and antimicrobial stewardship programs, identifying emerging threats like the blaKPC-2 plasmid in carbapenem-resistant K. pneumoniae [31] or the tet(S/M) mosaic gene in S. pneumoniae [37].

Standardization and Quality Control in Veterinary Diagnostic Laboratories

For the veterinary clinical pathologist, adherence to internationally recognized standards (e.g., CLSI VET01, VET02, VET03 for aquatic species) is paramount. Key quality control measures specific to co-infection testing include:

• Use of validated selective media: Each batch of selective agar must be QC tested with known positive and negative control organisms to ensure it supports growth of the target pathogen while suppressing contaminants.
• Inoculum standardization: For broth microdilution, the use of a spectrophotometer to standardize the inoculum to approximately 5 x 10^5 CFU/mL is preferred over visual turbidity standards, especially for mucoid or fastidious organisms.
• External Proficiency Testing: Laboratories must participate in external quality assessment schemes that include challenging polymicrobial specimens, not just pure cultures.
• Reporting of Mixed Cultures: Protocols must include clear criteria for when to report a mixed culture as "normal respiratory flora" versus when to proceed with identification and AST of individual components. A general principle is to proceed with AST if a single morphotype constitutes >75% of the growth on the primary plate, or if a known primary pathogen (e.g., A. baumannii in a COVID-19 patient) is present in any amount [8, 11].

In conclusion, the methodological approach to susceptibility testing in co-infected clinical samples is far from routine. It requires a deliberate, stepwise strategy that begins with rigorous pre-analytical sample selection, employs specialized culture conditions tailored to the specific viral-bacterial synergy, expands the AST panel to cover MDR profiles, and, increasingly, integrates molecular methods to overcome the limitations of culture. The veterinary clinical pathologist must act as a diagnostic integrator, synthesizing virology, bacteriology, and clinical history to ensure that the AST results guide effective therapy and contribute to the global fight against antimicrobial resistance.

Clinical Application and Performance of Susceptibility Assays in Swine and Human Respiratory Co-infection Models

The clinical utility of antimicrobial susceptibility testing (AST) in the context of secondary viral co-infections extends far beyond simple determination of minimum inhibitory concentrations (MICs). In both swine and human medicine, the performance characteristics of these assays - their sensitivity, specificity, turnaround time, and predictive value for clinical outcomes - must be rigorously evaluated within the specific ecological and immunological framework of co-infection. The respiratory tract, during active viral infection, represents a dynamically altered microenvironment where standard AST parameters may not fully capture the in vivo efficacy of antimicrobial agents. This section critically examines how susceptibility assays perform when applied to clinical specimens derived from swine and human patients with documented viral-bacterial co-infections, drawing on large-scale surveillance data, prospective diagnostic comparisons, and mechanistic co-infection models.

Performance of Molecular and Phenotypic Assays in Human Co-infection Diagnostics

The diagnostic landscape for bacterial co-infection in human viral respiratory disease has been profoundly reshaped by the introduction of multiplex molecular panels. A prospective study of 111 hospitalized pneumonia patients in Jordan, comparing conventional culture with multiplex real-time PCR (FTD Respiratory Pathogens 33), demonstrated that PCR detected pathogens in 74.8% of patients versus 57.7% by culture (p<0.001), with a sensitivity of 96.9% compared to 86.3% for culture [4]. However, this enhanced sensitivity came with a higher false-positive rate, and critically, the molecular approach identified bacterial-viral co-infections in 36.9% of cases, revealing a complex etiological landscape that conventional methods would have missed [4]. The implication for AST is profound: rapid molecular detection of a bacterial pathogen does not provide a phenotypic susceptibility profile. Clinicians are thus forced to rely on local antibiograms or empiric guidelines, a strategy that proved problematic during the COVID-19 pandemic when widespread empiric antibiotic use - despite low documented co-infection rates - fueled antimicrobial resistance [9, 23]. The disconnect between molecular pathogen detection and phenotypic AST remains a central challenge; while syndromic panels like the Allplex Respiratory Panels detect 16 viruses and seven bacteria, they "do not yield viable organisms" and therefore cannot provide AST results, forcing a reliance on conventional culture for resistance profiling [10].

The comparative performance of genotypic versus phenotypic resistance prediction has been explored in targeted next-generation sequencing (tNGS) studies. In a pilot analysis of Klebsiella pneumoniae community-acquired pneumonia, sputum tNGS revealed more than sixfold unique antimicrobial resistance genes (ARGs) compared to matched pure cultures (38 vs. 7), including clinically relevant determinants that were absent from culture isolates [5]. This suggests that culture-based AST may systematically underestimate the resistome present in the native polymicrobial community, particularly when the dominant pathogen is outcompeted in vitro or when resistance genes reside in co-colonizing organisms. However, concordance between sputum tNGS and culture tNGS was low (k = 0.279), and major ESBL and carbapenemase genes were not detected by sequencing in isolates that exhibited phenotypic multidrug resistance, indicating that "alternative resistance mechanisms" or gene expression differences may confound purely genotypic approaches [5]. Thus, the optimal diagnostic strategy appears to be an integrated one: tNGS or multiplex PCR for broad pathogen and resistome detection, followed by targeted phenotypic AST on confirmed isolates to validate resistance phenotypes and guide therapy.

Swine Respiratory Co-infection Models: Network-Based Approaches to AST Interpretation

In swine production, the porcine respiratory disease complex (PRDC) serves as a natural laboratory for evaluating AST performance in co-infection contexts. A landmark retrospective analysis of 6,017 diagnostic submissions from Spanish swine farms between 2020 and 2024 identified over 800 distinct co-detection patterns among 12 key respiratory pathogens [1]. Glaesserella parasuis (56.7%), Streptococcus suis (56.4%), and Mesomycoplasma hyorhinis (47.8%) were the most frequently detected agents, with co-occurrence network analysis revealing that M. hyorhinis was a pivotal hub, showing robust associations with nine other evaluated pathogens, including odds ratios of 4.05 with G. parasuis and 2.55 with Porcine Reproductive and Respiratory Syndrome Virus (PRRSV-1) [1]. These data underscore a critical performance consideration for AST in swine co-infection models: the susceptibility profile of a single bacterial isolate may not predict the response of co-infecting pathogens, particularly when viral immunosuppression - such as that induced by PRRSV-1 or Swine Influenza A Virus - alters bacterial colonization dynamics and antibiotic pharmacokinetics at the tissue level.

The clinical application of AST in swine PRDC must therefore account for the temporal and hierarchical structure of co-infection. The Spanish surveillance data demonstrated that secondary opportunistic bacteria were disproportionately detected in post-weaning piglets, and their frequency rose steadily until 2022 before declining in parallel with decreases in primary pathogens like PCV2 and M. hyopneumoniae [1]. This suggests that AST-guided therapy should be dynamic, targeting the dominant opportunist at the appropriate stage of disease progression. Furthermore, the detection of M. hyorhinis in nearly half of submissions, coupled with its strong linkage to G. parasuis and S. suis, indicates that empiric AST panels in swine must include agents effective against these fastidious or slow-growing organisms, which are often overlooked in standard veterinary AST protocols [1].

Mechanistic Insights from Co-infection Models: Implications for AST Interpretation

The performance of susceptibility assays cannot be divorced from the immunopathophysiology of co-infection. In murine models of influenza-associated Staphylococcus aureus co-infection, targeting toll-like receptor 7 (TLR7) with the antagonist IRS661 significantly improved survival and reduced pulmonary damage by enhancing macrophage phagocytosis and bactericidal activity [2]. This finding raises a crucial question for AST interpretation: does an in vitro susceptibility result remain predictive when the host immune response is pharmacologically altered? The study demonstrated that TLR7 antagonism upregulated reactive oxygen species production and modulated JNK phosphorylation, effectively lowering the bacterial burden even in the absence of antibiotics [2]. In such a scenario, a bacterial isolate classified as "intermediate" or even "resistant" by standard MIC testing in vitro might be cleared in vivo if the host response is appropriately augmented. Conversely, the profound immunosuppression induced by influenza virus - including impaired neutrophil function and reduced antimicrobial peptide production [17] - may render an in vitro "susceptible" result clinically irrelevant, as the host cannot sustain effective drug action at the infection site.

This disconnect is particularly well-documented in influenza-Streptococcus pneumoniae co-infection models, where numerical simulations of oseltamivir treatment strategies revealed that the standard 75 mg regimen yielded only 47% antiviral efficacy and a mere 9% antibacterial efficacy, while a 150 mg dose increased antibacterial efficacy to just 16% [28]. The computational model indicated that the presence of bacterial co-infection fundamentally altered the pharmacodynamics of the antiviral, limiting its ability to reduce viral replication and secondarily impacting bacterial clearance [28]. For AST to be clinically meaningful in co-infection, susceptibility breakpoints may need to be adjusted to account for the altered tissue penetration, immune cell recruitment, and microbial community dynamics that characterize the co-infected lung.

Comparative Performance of Culture-Based vs. Molecular Susceptibility in COVID-19 Cohorts

The COVID-19 pandemic provided an unprecedented opportunity to evaluate AST performance at scale. A multicenter Welsh cohort study comparing 299 ICU patients with COVID-19 pneumonia to 173 patients with no viral pneumonia and 48 with influenza A/B demonstrated that the incidence of respiratory co-infection was 40.5% in COVID-19 patients, with Staphylococcus aureus predominating within the first 48 hours and Gram-negative Enterobacterales emerging later [44]. The antimicrobial susceptibility patterns differed markedly between early and late infections, with late-onset Gram-negative isolates showing higher carbapenem resistance [44]. This temporal shift has direct implications for AST strategy: a single admission culture may not predict the susceptibility profile of superinfecting organisms acquired during prolonged ICU stay. Studies from Iran and Lebanon reported that Klebsiella pneumoniae and Acinetobacter baumannii were the most common secondary pathogens, with carbapenem resistance exceeding 80% in many centers, and that 84.8% of isolates displayed high resistance patterns, with all S. aureus being methicillin-resistant [12, 26]. In Lebanon, whole genome sequencing of 13 Gram-negative isolates revealed a "plethora of AMR determinants" and demonstrated that efflux pump inhibition with PAβN could restore levofloxacin activity against P. aeruginosa isolates that were phenotypically resistant [26].

These findings highlight a critical performance gap: conventional broth microdilution AST, even when performed according to CLSI standards, may fail to detect resistance mediated by efflux pumps or biofilm formation in the co-infected lung. In COVID-19 patients, biofilm-forming, ESBL-producing K. pneumoniae were isolated from sputum, with 75% harboring the blaCTX-M gene [30]. The presence of biofilm in vivo can increase MICs by 100- to 1000-fold compared to planktonic AST, rendering standard breakpoints non-predictive of therapeutic success. Similarly, in a study of Bordetella bronchiseptica isolates from swine, extension of incubation time from 16-20 hours to 24 hours resulted in 1-2 dilution steps higher MIC50 values for several antimicrobial agents, demonstrating that AST protocols optimized for fast-growing pathogens may underestimate resistance in slower-growing or stress-adapted bacteria [22].

One Health Implications and the Need for Harmonized Assay Performance Standards

The translational value of AST in co-infection models is amplified when considering zoonotic and reverse-zoonotic transmission. Genomic characterization of Aeromonas veronii and Shewanella sp. co-infection in koi carp revealed convergent resistance mechanisms, including OXA-436, tet(A), and aadA3, with the Shewanella isolate showing reduced fluoroquinolone and phenicol susceptibility due to rsmA [3]. The authors emphasized that "genotype-phenotype agreement was high" for multidrug resistance, supporting the use of genomics to "guide dual-coverage, mechanism-aware therapy" in ornamental fish medicine while simultaneously informing zoonotic risk mitigation for aquarists [3]. In poultry, co-infection of Novel Duck Orthoreovirus with multidrug-resistant Salmonella Typhimurium resulted in higher mortality (up to 26.67%) and earlier viral replication peaks, with the Salmonella isolate confirmed as multidrug-resistant by AST, complicating therapeutic interventions [15]. These examples illustrate that susceptibility assays must be validated not only for the target pathogen but for the specific co-infection context, as viral-induced immunosuppression (such as that caused by duck orthoreovirus) facilitates bacterial dissemination and may reduce the effective concentration of antibiotics at the site of infection [15].

The World Organisation for Animal Health (WOAH) and the Clinical and Laboratory Standards Institute (CLSI) have established standardized AST methods for veterinary pathogens, but these protocols are typically optimized for single-agent infections. The data reviewed here argue strongly for the development of co-infection-specific AST guidelines that account for altered pharmacokinetics, biofilm formation, and the potential for horizontal transfer of resistance plasmids between co-colonizing organisms, as demonstrated by the emergence of carbapenem-resistant hypervirulent Klebsiella pneumoniae following in-host plasmid transfer from classical to hypervirulent clones [31]. Without such harmonized performance standards, the predictive value of AST in co-infection will remain suboptimal, perpetuating the empiric antibiotic use that drives the AMR crisis.

Interpretation of Antimicrobial Susceptibility Data in the Context of Viral Pathogen Co-detection and Immune Modulation

The interpretation of antimicrobial susceptibility testing (AST) data in patients with confirmed or suspected viral co-detection represents one of the most complex and clinically consequential challenges in contemporary veterinary and human infectious disease medicine. Standard interpretive criteria - whether derived from Clinical and Laboratory Standards Institute (CLSI) or European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines - are predicated on pharmacokinetic/pharmacodynamic (PK/PD) thresholds established in immunocompetent hosts under steady-state conditions. These frameworks assume that the host immune system will cooperate with antimicrobial therapy to eradicate the pathogen. However, when a viral pathogen is concurrently present, the host immune landscape is fundamentally remodeled, often in ways that render standard AST interpretations inadequate or even dangerously misleading. This section provides a comprehensive, mechanistic analysis of how viral co-detection and viral-induced immune modulation must be integrated into the interpretation of AST data, drawing on experimental, clinical, and epidemiological evidence across multiple host species and viral families.

The Immunological Paradigm Shift: How Viruses Alter the Therapeutic Context

Viral infections, particularly those affecting the respiratory, enteric, or lymphoid systems, induce profound alterations in innate and adaptive immunity that directly impact the efficacy of antimicrobial therapy. The classic example is the synergistic interaction between influenza viruses and bacterial pathogens such as Streptococcus pneumoniae and Staphylococcus aureus. Influenza A virus infection disrupts the respiratory epithelial barrier, impairs mucociliary clearance, and downregulates antimicrobial peptide production, thereby reducing the bacterial inoculum required to establish infection [17, 21]. More critically, influenza-induced type I interferon signaling suppresses neutrophil recruitment and macrophage phagocytic function, creating a permissive environment for bacterial proliferation that may persist even when the bacterial pathogen is classified as "susceptible" by in vitro AST [2, 13]. This immunological vulnerability means that a minimum inhibitory concentration (MIC) that would be clinically achievable and effective in a naive host may be insufficient to clear infection in the context of viral co-detection.

The mechanistic underpinnings extend beyond influenza. Infection with Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) in swine induces a well-characterized state of immune dysregulation, including suppression of alveolar macrophage function, altered cytokine profiles, and impaired pulmonary clearance mechanisms. In a large-scale retrospective study of porcine respiratory disease complex (PRDC) in Spanish swine, PRRSV-1 was detected in 52.4% of submissions and showed robust co-occurrence with opportunistic bacteria such as Mesomycoplasma hyorhinis (OR = 2.55) and Glaesserella parasuis (OR = 4.05) [1]. This network of associations underscores that the presence of a viral pathogen like PRRSV fundamentally changes the host-pathogen dynamic, rendering standard AST breakpoints - established in immunocompetent animals - potentially non-predictive of clinical outcome. Similarly, Infectious Bursal Disease Virus in poultry induces B-cell depletion and immunosuppression that predisposes to secondary bacterial infections, including colibacillosis caused by avian pathogenic Escherichia coli (APEC). In a study of broiler colibacillosis outbreaks, APEC serotypes O78, O2, and O1 showed multidrug resistance to ampicillin, tetracycline, and sulfamethoxazole-trimethoprim but remained sensitive to ciprofloxacin [16]. However, the clinical response to ciprofloxacin in immunosuppressed birds may be diminished compared to immunocompetent flocks, a factor rarely captured in AST interpretive guidelines.

The Inadequacy of Standard Breakpoints in the Co-detection Setting

Standard AST interpretive criteria are derived from population distributions of MICs and PK/PD targets that achieve bacterial stasis or killing in animal models or human clinical trials conducted in patients without concurrent viral illness. The implicit assumption is that the host contributes a relatively constant "background" level of antimicrobial activity. Viral co-detection invalidates this assumption. The consequence is that a bacterial isolate categorized as "susceptible" by CLSI or EUCAST breakpoints may be associated with clinical failure when the host is simultaneously infected with a virus that impairs immune effector functions.

This phenomenon has been observed across multiple viral-bacterial combinations. In the context of Canine Influenza A Virus (H3N2) infection in dogs, co-infection with Enterococcus faecalis was documented, and although antimicrobial susceptibility testing guided the selection of enrofloxacin, the authors emphasized that the viral infection likely facilitated bacterial colonization and complicated clinical outcomes [27]. Similarly, experimental co-infection of ducklings with novel duck orthoreovirus (N-DRV) and Salmonella Typhimurium demonstrated that viral-induced immunosuppression - characterized by suppressed humoral responses and elevated viral loads in immune organs (spleen, thymus, bursa of Fabricius) - facilitated bacterial dissemination and systemic infection [15]. Importantly, the S. Typhimurium isolate was confirmed to be multidrug-resistant by in vitro AST, but even a fully susceptible isolate would have faced an uphill battle against the backdrop of viral-induced immune compromise. These observations indicate that the "effective MIC" in vivo is higher than the in vitro MIC when viral co-detection is present, a concept that is not captured by current interpretive frameworks.

The problem is compounded by the fact that viral infections themselves can alter the expression of bacterial resistance determinants. For instance, influenza A virus infection has been shown to induce oxidative stress and inflammation that can upregulate efflux pump expression in co-infecting bacteria, transiently increasing the MIC to fluoroquinolones and tetracyclines [2]. Furthermore, the inflammatory milieu created by viral infection - characterized by elevated cytokines such as IL-6, IL-1β, and TNF-α - can select for bacterial subpopulations with pre-existing resistance mechanisms, a phenomenon that may be missed by standard AST performed on single colonies [26, 45]. In COVID-19 patients, studies have consistently shown that secondary bacterial infections are associated with extensively drug-resistant and multidrug-resistant organisms, including carbapenem-resistant Klebsiella pneumoniae and Acinetobacter baumannii [6, 8, 12, 30]. While some of this resistance reflects nosocomial transmission, the selective pressure exerted by the viral-induced inflammatory and immunological environment cannot be discounted.

Viral Determinants of Resistance Phenotype Expression and Clinical Outcome

The specific identity of the co-detected virus profoundly influences how AST data should be interpreted. Viruses that directly infect and impair immune cells - such as Feline Immunodeficiency Virus (FIV), Aleutian Disease Virus, or Bovine Viral Diarrhea Virus - induce a state of secondary immunodeficiency that broadly compromises the host's ability to clear bacterial infections, regardless of in vitro susceptibility [41]. In these cases, even bacterial isolates with low MICs to bactericidal antibiotics may require prolonged therapy, higher doses, or combination regimens to achieve clinical cure. The risk of clinical failure is disproportionately high when the infecting bacterial species is an obligate or facultative intracellular pathogen, such as Salmonella spp., Mycobacterium spp., or Brucella spp., where bacterial clearance depends heavily on T-cell-mediated immunity.

Respiratory viruses that primarily damage epithelial barriers - such as Avian Influenza Virus, Bovine Respiratory Syncytial Virus, and Infectious Laryngotracheitis Virus - create a different set of interpretive challenges. The loss of epithelial integrity and mucociliary function allows bacteria to access the lower respiratory tract in higher numbers and to form biofilms on damaged mucosa. In this setting, the bactericidal activity of β-lactams and aminoglycosides may be compromised by the altered microenvironment, including reduced oxygen tension and pH changes that can affect antibiotic activity [49]. Moreover, the viral infection can induce a shift in the pulmonary microbiome, promoting the overgrowth of potentially pathogenic bacteria that may not be detected by standard culture but are identified through molecular methods such as multiplex PCR or targeted next-generation sequencing [4, 10]. These "occult" pathogens may have AST profiles that differ from the dominant cultured isolate, leading to discordance between AST results and clinical outcome.

The temporal dynamics of viral and bacterial co-detection also require careful interpretive consideration. In the PRDC study, secondary opportunistic bacteria were predominantly detected in post-weaning piglets, a period coinciding with waning maternal immunity and increased susceptibility to Porcine Circovirus 2 and PRRSV [1]. Similarly, in COVID-19 patients, early co-infections (within 48 hours of admission) were more frequently caused by S. aureus, while late-onset secondary infections were dominated by Gram-negative Enterobacterales, reflecting the different immunological and environmental pressures at each stage [44]. These temporal patterns should inform the interpretation of AST data: an isolate obtained early in the course of viral illness may reflect a community-acquired strain with predictable susceptibility, whereas isolates obtained after prolonged hospitalization and antimicrobial exposure are more likely to harbor acquired resistance mechanisms and require broader therapeutic coverage.

Clinical Case Studies in Co-detection: Lessons from the Pandemic Era

The COVID-19 pandemic provided an unprecedented, large-scale natural experiment in the interpretation of AST data in the context of viral co-detection. Multiple studies consistently demonstrated that bacterial co-infection rates at admission were low (3-6%), yet empiric antibiotic use was exceedingly high (70-90% of hospitalized patients) [9, 23, 33]. This discrepancy reflects the clinical uncertainty generated by overlapping clinical signs of viral and bacterial pneumonia and the lack of rapid, reliable diagnostic tools to differentiate the two. More importantly, when secondary bacterial infections did occur in COVID-19 patients, they were disproportionately caused by multidrug-resistant organisms, and the mortality rate among co-infected patients was markedly elevated (51-83% in ICU cohorts) [6, 12, 20, 46].

These findings have direct implications for AST interpretation. In a study of hospitalized COVID-19 patients in Iran, 100% of patients with secondary bacterial infections had received empiric antibiotics before the first positive culture, predominantly meropenem (86.2%) and levofloxacin (73.8%) [12]. The most common causative agents were K. pneumoniae and A. baumannii, both of which showed extensive drug resistance, with carbapenem resistance rates exceeding 60%. Standard AST would classify these isolates as "resistant" to most β-lactams, but the clinical relevance of this resistance was amplified by the underlying viral pneumonia and the immunological compromise induced by SARS-CoV-2. The virus itself infects endothelial cells and induces a hyperinflammatory state that can impair antibiotic penetration into lung tissue and alter the host response to infection [48]. Similar patterns were observed in Lebanon, where carbapenem and colistin-resistant Gram-negative bacilli were identified in COVID-19 ICU patients, and efflux pump inhibitors were required to restore levofloxacin activity in some isolates [26]. These observations highlight that in vitro resistance may be a surrogate for immunological failure, not just pharmacological failure.

The veterinary literature provides analogous examples. In the context of Newcastle Disease Virus infection in pigeons, bacterial co-infections with E. coli, Salmonella spp., and S. aureus were common, and the bacterial isolates showed high resistance to ampicillin, tetracycline, ceftriaxone, and sulfamethoxazole-trimethoprim [47]. Similarly, in ducks co-infected with novel duck orthoreovirus and S. Typhimurium, the bacterial isolate was multidrug-resistant, and co-infected ducklings exhibited higher mortality rates (up to 26.67%), earlier peaks in viral replication, and suppressed humoral responses compared to single infections [15]. These case studies underscore that AST data must be interpreted in the context of the specific viral pathogen, the host immune status, and the clinical stage of infection.

Toward an Integrated Interpretive Framework

Given the complexity of viral-bacterial interactions, it is clear that AST data cannot be interpreted in isolation. An integrated framework is required that incorporates the viral pathogen, the host immune status, and the local antimicrobial resistance epidemiology. Several principles should guide this interpretation:

First, the likelihood of clinical success for a given antimicrobial regimen is reduced in the presence of viral co-detection, and this risk is proportional to the degree of immunosuppression or epithelial damage induced by the virus. For viruses known to cause profound immune dysregulation - such as Feline Leukemia Virus, Canine Distemper Virus, or PRRSV - even fully susceptible bacterial isolates may require combination therapy, extended treatment duration, or higher-than-standard dosing to achieve clinical cure. In such cases, the AST result should be considered a necessary but not sufficient predictor of outcome.

Second, the specific bacterial species and its intrinsic pathogenicity must be considered. Opportunistic pathogens such as E. coli, K. pneumoniae, and A. baumannii are common causes of secondary infection in virally compromised hosts and are frequently multidrug-resistant [8, 46]. In contrast, primary pathogens such as S. pneumoniae or G. parasuis may cause severe disease even when the MIC falls within the susceptible range, particularly if the immune response is blunted by viral infection [17]. The interpretive weight of the AST result should therefore be adjusted according to the pathogen's intrinsic virulence and the host's immunological reserve.

Third, the use of rapid molecular diagnostics - including multiplex PCR panels and targeted metagenomics - can provide critical context for AST interpretation. These methods can detect viral and bacterial co-pathogens simultaneously, identify antimicrobial resistance genes directly from clinical specimens, and reveal polymicrobial infections that are missed by culture [4, 10, 14]. In the setting of viral co-detection, a molecular result showing the presence of a resistance gene such as mecA, blaKPC, or mcr-1 should be interpreted with heightened concern, as the immunological window for effective therapy is narrower [14]. Conversely, the absence of molecular resistance markers can provide confidence that standard therapy may succeed, even when the host is immunocompromised.

Fourth, biomarker-guided strategies - including procalcitonin (PCT) and C-reactive protein (CRP) monitoring - can assist in integrating AST data with the host response. In patients with viral pneumonia, low PCT levels (<0.25 ng/mL) are associated with a low probability of bacterial co-infection and can support decisions to withhold or discontinue antibiotics [24, 43]. When bacterial co-infection is confirmed by culture and AST, a persistently elevated PCT may indicate that the chosen antimicrobial regimen is ineffective, either due to in vitro resistance or due to inadequate host response [50]. Serial biomarker measurement should thus be considered a complementary tool to AST interpretation in the co-detection setting.

Finally, antimicrobial stewardship programs must be adapted to the unique challenges of viral co-detection. During the COVID-19 pandemic, targeted interventions - including prospective audit and feedback, rapid molecular testing, and biomarker-guided algorithms - demonstrated the feasibility of reducing unnecessary antibiotic use even in crisis conditions [23]. These principles are directly transferable to veterinary practice, where the empiric use of broad-spectrum antibiotics in virally infected animals is common and often inappropriate. A stewardship framework that integrates AST data with viral diagnostic results and host-response biomarkers can optimize antimicrobial selection, reduce the selection pressure for resistance, and improve clinical outcomes.

In summary, the interpretation of antimicrobial susceptibility data in the context of viral pathogen co-detection and immune modulation requires a fundamental departure from traditional paradigms. Standard breakpoints must be applied with caution, and the viral status of the host should be considered a critical variable in the therapeutic equation. Only through an integrated, multi-dimensional approach - incorporating viral identity, host immunocompetence, pathogen virulence, molecular resistance profiling, and biomarker monitoring - can the clinician hope to achieve reliable predictions of clinical outcome and deliver effective, personalized antimicrobial therapy.

Epidemiological Co-detection Patterns and Their Implications for Susceptibility Testing Priorities in Porcine Respiratory Disease Complex

The porcine respiratory disease complex (PRDC) epitomizes the multifactorial aetiology that challenges both clinical diagnosis and antimicrobial stewardship in modern swine production. Unlike single-agent respiratory infections, PRDC arises from intricate, often synergistic interactions among primary viral initiators - such as Porcine Reproductive and Respiratory Syndrome Virus (PRRSV), Porcine Circovirus 2 (PCV2), and Swine Influenza A Virus - and a diverse array of secondary bacterial opportunists. Large-scale epidemiological surveillance is therefore indispensable for delineating which co-detection patterns are clinically relevant and, critically, for informing which bacterial pathogens must be prioritised in antimicrobial susceptibility testing (AST) panels. A landmark retrospective analysis of 6,017 routine diagnostic submissions from Spanish commercial swine farms (2020-2024) has provided the most comprehensive landscape of PRDC-associated co-detections to date [1]. This study, which examined 12 key viral and bacterial agents, revealed that over 800 distinct co-detection patterns exist, underscoring the extraordinary complexity facing diagnosticians and clinicians.

The Co-detection Landscape and Central Bacterial Hubs

The Spanish surveillance data identified Glaesserella (Haemophilus) parasuis (56.7%), Streptococcus suis (56.4%), PRRSV-1 (52.4%), and Mesomycoplasma hyorhinis (47.8%) as the most frequently detected agents [1]. However, it is the co-occurrence network analysis that provides the most actionable insights for AST prioritisation. M. hyorhinis emerged as a pivotal hub, demonstrating strong pairwise associations with nine of the other evaluated pathogens. Notably, the odds ratio (OR) for co-detection of M. hyorhinis with G. parasuis was 4.05 (95% CI high significance), with S. suis OR = 1.57, and with PRRSV-1 OR = 2.55 [1]. This suggests that when PRRSV infection is present, the probability of isolating M. hyorhinis more than doubles, and when M. hyorhinis is found, the likelihood of co-isolating G. parasuis is quadrupled. Such robust associations are not merely statistical curiosities; they have profound implications for laboratory workflow. For example, in a piglet with PRRSV-positive lung tissue, the clinician should explicitly request AST for M. hyorhinis and G. parasuis even if these bacteria are not immediately apparent on primary culture, because their co-detection is highly probable and their antimicrobial resistance profiles may diverge significantly from those of the primary viral insult.

Secondary opportunistic bacteria - including G. parasuis, S. suis, and M. hyorhinis - were significantly more frequent in post-weaning piglets (p < 0.05), a period coinciding with waning maternal immunity and peak PRRSV/PCV2 circulation [1]. This age-stratified co-detection pattern dictates that AST panels for weaner-age pigs must be broadened to include these fastidious organisms, whereas finishing pigs may require more emphasis on Bordetella bronchiseptica and Pasteurella multocida - pathogens that become more prominent in older animals with chronic viral carriage. The study also noted a temporal decline in PCV2 and M. hyopneumoniae detection after 2022, mirroring a subsequent decrease in secondary opportunists [1]. This temporal coupling reinforces the concept that primary viral pathogens create the ecological niche for secondary bacteria; thus, when a farm experiences an outbreak of PRRSV or swine influenza, the laboratory must anticipate a surge in specific bacterial co-infections and adjust its AST menu accordingly.

Viral-Bacterial Synergy and Its Impact on Susceptibility Testing Priorities

The immunopathogenic mechanisms underlying viral-bacterial synergy have been extensively characterised in human influenza and COVID-19, and similar principles apply to PRDC. For instance, PRRSV infection impairs alveolar macrophage function and disrupts mucociliary clearance, creating a permissive environment for S. suis and G. parasuis adhesion and invasion. In a mouse model of influenza-Staphylococcus aureus co-infection, TLR7 antagonism enhanced macrophage phagocytosis and bactericidal activity, highlighting how the host immune response to the virus fundamentally alters bacterial clearance [2]. Translating this to PRDC means that the choice of antimicrobial agent must consider not only the bacterial isolate's MIC but also the host's virus-induced immune state. For example, during acute PRRSV viraemia, β-lactam monotherapy for S. suis may be less effective because virus-triggered inflammation impairs neutrophil recruitment, necessitating a combination with a drug that has intracellular activity, such as florfenicol or tiamulin. Susceptibility testing should therefore be interpreted in the context of the co-infecting virus; a "susceptible" result in vitro may not guarantee clinical success if viral immunosuppression limits drug penetration or bacterial killing.

The Spanish co-detection network [1] also revealed strong links between PRRSV-1 and G. parasuis (OR not explicitly stated but implied by the M. hyorhinis hub). G. parasuis is a fastidious, NAD-dependent bacterium that is notoriously difficult to culture, and many laboratories do not routinely perform AST on it. Yet, given its high co-occurrence with PRRSV, its omission from susceptibility panels represents a critical gap. Similarly, S. suis is a major zoonotic pathogen with increasing reports of resistance to tetracyclines, macrolides, and even penicillin in some regions [1]. The co-detection of S. suis with multiple other pathogens (OR = 1.57 with M. hyorhinis) suggests that empirical therapy targeting S. suis alone may fail if co-infecting organisms with different resistance profiles are not also covered.

Implications for Designing Surveillance-Driven AST Panels

The existence of over 800 discrete co-detection patterns [1] precludes a one-size-fits-all AST panel. Instead, a risk-stratified approach is required, based on the most frequent and clinically impactful co-occurrences. The following priorities emerge from the epidemiological evidence:

1. Universal inclusion of M. hyorhinis and G. parasuis in all respiratory AST panels from post-weaning pigs, given their central position in the co-detection network and their fastidious growth requirements. Many commercial AST systems (e.g., broth microdilution panels) do not include these species, forcing reliance on off-label breakpoints. Laboratories should validate custom panels for these organisms, ideally using the CLSI or EUCAST veterinary-specific guidelines where available.

2. Virus-triggered expansion of the AST menu. When a farm is confirmed to be PRRSV-positive, the laboratory should automatically reflex to include AST for S. suis, G. parasuis, and M. hyorhinis on all respiratory specimens, even if the primary culture does not yield these pathogens. Similarly, during an Swine Influenza A Virus outbreak, the addition of B. bronchiseptica and P. multocida AST is warranted, as influenza virus-induced epithelial damage preferentially predisposes to these colonisers [22].

3. Temporal and geographic adaptation. The Spanish study noted marked regional differences, with PRRSV-1 positivity significantly higher in northeastern Spain [1]. Farms in such high-prevalence regions should employ a PRRSV-focused AST algorithm that prioritises the bacteria most strongly associated with PRRSV. Conversely, in areas where PCV2 remains endemic, the panel should emphasise M. hyopneumoniae and B. bronchiseptica, as these are classic PCV2-associated co-pathogens. Seasonal influenza peaks may further shift the bacterial profile toward S. aureus and S. pneumoniae-like organisms in swine, mirroring human patterns [21, 33].

4. Integration of genotypic resistance surveillance. The observation that many PRDC-associated bacteria are multidrug-resistant (MDR) [1] underscores the need for complementary molecular AST methods. Targeted metagenomics or multiplex PCR panels that detect resistance genes (e.g., mecA, ermB, tet(M), bla TEM) can provide rapid, culture-independent resistance profiling, especially when fastidious organisms fail to grow. In human respiratory co-infections, next-generation sequencing-based AMR surveillance has already demonstrated the ability to detect resistance markers not captured by phenotypic methods [14]. For PRDC, similar approaches could be adapted to identify the erm and tet genes commonly harboured by S. suis and M. hyorhinis, enabling early adjustment of therapy.

5. Animal- and production-stage-specific breakpoints. The current veterinary AST breakpoints are often extrapolated from human data or from a limited number of animal isolates. The high co-detection frequencies observed in post-weaning piglets [1] demand that clinical breakpoints for G. parasuis and M. hyorhinis be specifically validated in that age group, taking into account the altered pharmacokinetics and pharmacodynamics caused by immature renal function and concurrent viral infection. Without such refinement, a "susceptible" result may mislead therapy.

Evidence from Human Parallels and the Call for Integrated Stewardship

The human pandemic experience reinforces the lessons from PRDC. During the 2009 H1N1 influenza pandemic, secondary bacterial pneumonia - most often caused by Streptococcus pneumoniae, S. aureus, and Haemophilus influenzae - was responsible for a substantial proportion of deaths [21]. Similarly, in COVID-19 patients admitted to ICUs, bacterial co-infections were identified in 40.5% of cases, with S. aureus predominating early and Gram-negative rods (e.g., Klebsiella pneumoniae, Acinetobacter baumannii) emerging later [44]. These patterns mirror the PRDC temporal shift from early, primary viral-driven opportunistic bacteria (e.g., S. suis, G. parasuis) to late, nosocomial-type Gram-negative infections in prolonged hospitalised pigs. The concurrent detection of AMR genes in human respiratory samples - such as erm, tet, and mecA [14] - highlights the universal challenge: susceptibility testing must be dynamic, reflecting the evolving co-infection landscape.

In summary, the epidemiological co-detection patterns of PRDC, as revealed by the largest study of its kind [1], provide an evidence-based framework for prioritising which bacterial pathogens to include in routine AST panels. The central roles of M. hyorhinis, G. parasuis, and S. suis, their strong associations with PRRSV and PCV2, and their age- and region-specific distributions mandate a shift from generic AST to targeted, virus-aware testing algorithms. Such an approach not only improves therapeutic success but also supports antimicrobial stewardship by reducing empirical overuse of broad-spectrum agents. Future efforts should focus on developing rapid, culture-independent AST methods for fastidious PRDC pathogens and on establishing veterinary-specific clinical breakpoints that account for the complex immunological milieu of viral-bacterial co-infection.

Emerging Technologies and Rapid Diagnostics for Antimicrobial Susceptibility Assessment in Viral-Bacterial Co-infections

The conventional diagnostic paradigm for antimicrobial susceptibility testing (AST) is fundamentally challenged by the complex pathophysiology of viral-bacterial co-infections. Standard culture-based phenotypic AST, while remaining the reference standard for definitive resistance profiling, suffers from turnaround times of 48-72 hours, a limitation that is often clinically untenable in critically ill patients with rapidly progressive pneumonia or sepsis secondary to a viral insult. In the context of co-infections, where viral-induced immunopathology and epithelial barrier disruption create a permissive niche for opportunistic bacteria, the clinical imperative for rapidly actionable susceptibility data is amplified. The experience from the COVID-19 pandemic, where widespread empirical antibiotic use occurred despite documented low bacterial co-infection rates [9, 23], has galvanized the development and clinical integration of a suite of emerging technologies designed to bridge the gap between rapid pathogen detection and comprehensive resistance profiling. These platforms - spanning syndromic multiplex PCR, targeted and untargeted next-generation sequencing (NGS), matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, and biomarker-guided stewardship algorithms - represent a paradigm shift from a purely culture-centric model to a multi-modal, genomically-informed diagnostic ecosystem.

Syndromic Multiplex Molecular Testing: From Pathogen Detection to Early Stewardship Cues

The most widely implemented rapid diagnostic for respiratory co-infections is the syndromic multiplex polymerase chain reaction (mPCR) panel. These panels, targeting 16-33 viral and bacterial pathogens simultaneously from a single respiratory specimen (e.g., nasopharyngeal swab, tracheal aspirate, bronchoalveolar lavage), have demonstrated significantly higher sensitivity than conventional culture for pathogen detection in community-acquired and hospital-acquired pneumonia. In a prospective study of 111 hospitalized pneumonia patients, mPCR detected pathogens in 74.8% of cases versus 57.7% by culture (p<0.001), revealing a 36.9% rate of bacterial-viral co-infections [4]. Similarly, the Allplex Respiratory Panels, when compared to conventional diagnostics in 110 mechanically ventilated ICU patients, showed high utility in community-acquired cases but exhibited lower yield for hospital-acquired and ventilator-associated pneumonia, largely due to the panels' limited bacterial targets and inability to perform phenotypic AST [10]. The critical insight from these studies is that syndromic mPCR excels as a screening tool for common pathogens - detecting Avian Influenza Virus, respiratory syncytial virus, rhinovirus, Haemophilus influenzae, and Streptococcus pneumoniae - but it does not provide direct AST data. The absence of an antibiogram means that a positive bacterial detection by mPCR cannot, by itself, differentiate a susceptible from a resistant strain.

However, the clinical value of mPCR in co-infection management lies in its ability to guide antimicrobial stewardship through a "rule-out" mechanism. For patients presenting with respiratory symptoms and a positive viral-only result on a rapid mPCR, the probability of a concomitant bacterial co-infection is substantially reduced. Studies from emergency department settings have demonstrated that the deployment of rapid viral mPCR resulted in appropriate antibiotic management (defined as no antibiotics prescribed in patients without risk factors for bacterial co-infection) in 82.4% of cases [43]. In that cohort, 65.5% of patients were not prescribed any empiric antibiotics following a positive viral result. The presence of purulent sputum or unilobar infiltrate were independent risk factors for inappropriate antibiotic continuation, highlighting that radiological and clinical features must be integrated with molecular results [43]. In the veterinary context, the application of such panels is expanding. For instance, in the investigation of the Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) and Porcine Circovirus 2 complex, syndromic panels targeting both viruses and primary/secondary bacterial opportunists (Glaesserella parasuis, Streptococcus suis, Mesomycoplasma hyorhinis) have revealed over 800 distinct co-detection patterns, with M. hyorhinis emerging as a pivotal node in co-occurrence networks, showing strong associations with nine of twelve evaluated pathogens [1]. The rapid identification of such polymicrobial signatures allows for more nuanced empirical therapy, but the lack of concomitant AST remains a significant limitation.

Targeted and Metagenomic Next-Generation Sequencing: Unlocking the Resistome Directly from Clinical Specimens

To address the AST gap inherent in syndromic panels, targeted next-generation sequencing (tNGS) and metagenomic NGS (mNGS) have emerged as transformative tools. The central advantage of these platforms is their ability to interrogate both pathogen identity and antimicrobial resistance gene (ARG) carriage directly from primary clinical material, bypassing the need for culture and thus reducing turnaround time to 24-48 hours. The Respiratory Pathogen ID/AMR Panel (RPIP, Illumina), a target-capture-based tNGS assay targeting 282 respiratory pathogens and 2,097 AMR markers, has been deployed at scale for community surveillance. In a cohort of 4,390 nasal swabs collected during the COVID-19 pandemic, RPIP detected at least one pathogen in 85.1% of samples, at least one AMR gene in 41%, and co-detection of both in 38.4% [14]. The most prevalent resistance determinant was Erm 23S ribosomal RNA methyltransferase (65.5%), co-detected with S. aureus and S. pneumoniae, while the mecA gene, conferring methicillin resistance, was found in 5.3% of S. aureus-positive specimens. This technology, when adapted for veterinary use, could revolutionize surveillance for resistance in pathogens associated with viral-bacterial co-infections in livestock and companion animals, such as Avian Influenza Virus and Pasteurella multocida in poultry, or Canine Influenza A Virus and Enterococcus faecalis in dogs [27].

A critical refinement in tNGS methodology is the comparison of ARG repertoires between matched culture isolates and primary clinical samples (e.g., sputum). A pilot study involving Klebsiella pneumoniae in community-acquired pneumonia demonstrated that sputum tNGS revealed more than sixfold unique ARGs compared to pure culture tNGS (38 vs. 7 genes), including clinically relevant determinants that were absent from the corresponding cultured isolates [5]. This suggests that culture-based AST may underestimate the true resistome by selecting for a dominant clone or by losing subpopulations of resistant strains during in vitro growth. Concordance between MALDI-TOF species identification and culture tNGS was high (k = 0.712), but concordance between sputum tNGS and culture tNGS was low (k = 0.279), underscoring the complementary nature of the methods [5]. For viral-bacterial co-infections, this disparity is particularly salient: the viral infection creates a heterogeneous microbial landscape within the lung, where multiple bacterial strains - some resistant, some susceptible - may coexist. tNGS provides a snapshot of this collective resistome, enabling the clinician to consider the "worst-case" resistance scenario when selecting empiric therapy.

Untargeted mNGS, which sequences all nucleic acids in a specimen without prior target enrichment, offers the broadest unbiased detection capability. This approach is particularly valuable for identifying novel or unexpected co-infecting pathogens, including fungi and fastidious organisms, and for detecting resistance genes that may be present on mobile genetic elements. The application of mNGS to COVID-19 ICU cohorts has revealed the presence of carbapenem-resistant Acinetobacter baumannii and Klebsiella pneumoniae clones that were missed by routine culture [9, 26]. In a genomic investigation of a fatal Aeromonas veronii and Shewanella sp. co-infection in koi carp, mNGS-level genome sequencing uncovered a virulome of 194 loci in A. veronii and 152 in Shewanella sp., annotated with adhesion, secretion, iron uptake, and immune-evasion functions, alongside resistomes encoding OXA-436, tet(A), and aadA3 - resistances that were phenotypically confirmed by conventional AST [3]. The convergence of virulence and resistance in a single co-infection event highlights the power of genomic diagnostics to delineate the full pathogenic potential of a polymicrobial infection, supporting mechanism-aware, dual-coverage therapy.

Whole-Genome Sequencing for Outbreak Investigation and Resistance Mechanism Discovery

Whole-genome sequencing (WGS) of bacterial isolates from co-infection settings has become indispensable for high-resolution epidemiological tracing and the discovery of novel resistance mechanisms. The detection of the mosaic tetracycline resistance gene tet(S/M) in a multidrug-resistant Streptococcus pneumoniae CC230 lineage that underwent capsular switching in South Africa exemplifies this utility. WGS of 12,254 pneumococcal isolates across 29 countries identified tet(S/M) in 131 isolates, all of which were phenotypically tetracycline-resistant. Birth-death modeling of CC230 revealed an unrecognized outbreak in South Africa between 2000-2004 (effective reproductive number R ≈ 2.5) that was driven by this resistance determinant [37]. Critically, the WGS analysis revealed capsular switching from vaccine serotype 14 to non-vaccine serotype 23A within the sublineage, allowing the resistant clone to evade pneumococcal conjugate vaccine pressure. In the context of viral-bacterial co-infection, such genomic surveillance is essential: the serotype and resistance profile of S. pneumoniae that secondarily invades an Avian Influenza Virus-infected host may differ from circulating strains in the absence of viral co-infection, as influenza-induced inflammation can alter the ecological niche for pneumococcal colonization and selection.

Similarly, WGS of carbapenem-resistant K. pneumoniae (CRKP) from a single patient co-infected with distinct ST268 (hypervirulent, serotype KL20) and ST4496 (classical, serotype KL47) clones demonstrated that in-host co-colonization facilitated the horizontal transfer of a blaKPC-2-encoding plasmid from the classical to the hypervirulent background, enabling the evolution of CR-hvKP and subsequent secondary bloodstream infection [31]. Such plasmid-mediated resistance escalation has profound implications for AST: a blood culture isolate may show carbapenem susceptibility at the time of initial testing, but the presence of a resistant subpopulation within the respiratory tract (as revealed by WGS of colonization isolates) warrants consideration of carbapenem-sparing or combination therapy from the outset. The integration of WGS into real-time clinical decision-making remains aspirational in most veterinary settings due to cost and bioinformatics expertise, but its value for reference laboratories and national surveillance networks (e.g., WOAH, national veterinary institutes) is undeniable.

MALDI-TOF Mass Spectrometry and its Integration with AST

MALDI-TOF MS has become a cornerstone of rapid bacterial identification in clinical microbiology, offering species-level identification within minutes from a single colony. Its role in co-infection diagnostics is primarily as a high-throughput identification tool to confirm culture results. In studies of Bacillus spp. co-infections in pediatric viral respiratory cases, MALDI-TOF was used to identify 20 isolates among 166 bacterial co-infections, with a predominance in respiratory syncytial virus-positive patients, highlighting that even environmental opportunists can be clinically relevant in the context of viral injury [35]. In the koi carp co-infection study, MALDI-TOF MS was the initial identification step prior to genomic characterization [3].

The critical limitation of MALDI-TOF is that it cannot, in its standard diagnostic application, provide quantitative AST data. However, emerging research is exploring the use of MALDI-TOF for direct detection of resistance mechanisms - such as β-lactamase hydrolysis assays or detection of carbapenemase activity within 1-2 hours of colony growth. These "phenotypic-to-proteomic" approaches, if validated for veterinary pathogens, could fill the void between rapid identification and full AST. For now, the most practical implementation is to use MALDI-TOF to accelerate the identification of cultured isolates, thereby shortening the time to definitive report of phenotypic AST results (e.g., from 48 to 24 hours), which remains a tangible improvement for guiding therapy in critically ill animals with suspected secondary bacterial infections.

Biomarker-Guided Stewardship and Host-Directed Therapeutics

Beyond pathogen-focused diagnostics, host-response biomarkers have emerged as powerful adjuncts to determine the likelihood of bacterial co-infection and to guide antibiotic discontinuation. Serum procalcitonin (PCT) has been extensively studied in COVID-19 and influenza cohorts. Despite the widespread use of PCT, a retrospective analysis of 294 patients admitted with lower respiratory tract infections found that the unrestricted ordering of PCT and respiratory viral panels, without formal antimicrobial stewardship intervention, did not reduce total antibiotic days of therapy (median 7 days in both groups), though inappropriately managed patients had higher 30-day readmission rates [24]. This highlights that the diagnostic technology alone is insufficient; it must be embedded within a structured stewardship framework. In the COVID-19 setting, biomarker-guided antibiotic discontinuation algorithms have been proposed as a key strategy to mitigate antimicrobial resistance (AMR), with evidence suggesting that rapid molecular testing combined with PCT can safely reduce antibiotic exposure in patients without confirmed bacterial co-infection [23].

A more innovative frontier is the development of host-directed immunomodulatory therapies that reduce susceptibility to secondary bacterial infection. Targeting the Toll-like receptor 7 (TLR7) pathway has shown promise in experimental influenza-Staphylococcus aureus co-infection models. Both TLR7 deficiency and the TLR7 antagonist IRS661 significantly improved survival in co-infected mice, enhanced macrophage phagocytosis and bactericidal activity, and reduced pro-inflammatory cytokine storm while preserving IFN-γ responses [2]. These findings suggest that rapid assessment of host immune dysregulation - rather than solely the pathogen - could inform the need for adjunctive immunomodulation alongside antimicrobial therapy. Transcriptomic host-response classifiers are being developed to distinguish viral from bacterial infection and to predict the risk of secondary infection, with the potential to be deployed as point-of-care assays.

Integration into a Diagnostic Algorithm for Viral-Bacterial Co-Infections

The optimal diagnostic strategy for antimicrobial susceptibility assessment in viral-bacterial co-infections is not a single technology but rather a tiered, integrated algorithm. At the first level, a syndromic mPCR panel (including both viral and common bacterial targets) should be performed on the lower respiratory tract specimen (e.g., bronchoalveolar lavage or endotracheal aspirate) at the time of clinical suspicion. A positive viral result without bacterial targets, in a patient without clinical or radiographic risk factors for bacterial infection (e.g., no purulent sputum, no lobar consolidation, normal PCT), can safely support withholding or early discontinuation of antibiotics [24, 43]. For patients with suspected bacterial co-infection - either by initial mPCR detection of a bacterial pathogen or by clinical severity - specimens should be reflexed to both conventional culture (for definitive phenotypic AST and isolate archiving) and tNGS or mNGS (for comprehensive resistome profiling). The tNGS result, available within 48 hours, provides a genotypic AST prediction that can guide definitive therapy while the culture-based AST is pending. This combined approach has been shown to improve pathogen detection, identify resistance determinants missed by culture, and guide targeted antimicrobial therapy in K. pneumoniae CAP [5].

For ICU patients at highest risk of MDR secondary infection - such as those with prolonged mechanical ventilation following viral pneumonia, or those infected with Porcine Reproductive and Respiratory Syndrome Virus or Avian Influenza Virus - serial sampling with mNGS may be justified to monitor for the emergence of resistance or superinfection. The economic and logistical barriers to routine mNGS in veterinary medicine remain substantial, but the decreasing cost of sequencing and the development of portable, rapid benchtop sequencers (e.g., Oxford Nanopore) are making real-time genomic AST increasingly feasible. Ultimately, the goal is to transition from a reactive, one-size-fits-all empirical antibiotic approach to a precision-based, dynamically adaptive strategy that integrates rapid pathogen identification, comprehensive resistome mapping, host immune status, and pharmacokinetic/pharmacodynamic principles to optimize outcomes for individual animals while preserving the antimicrobial arsenal for future generations.

References

1. Chacón-Pérez G, Pastor-Calonge A, Del Caso-Yagüe S, Martínez-Martínez S, Arnal-Bernal J, Gutiérrez-Martín C, et al.. Large-scale landscape of porcine respiratory disease complex-associated pathogens in Spanish swine production. The Veterinary Journal. 2026. DOI: https://doi.org/10.1016/j.tvjl.2026.106664

2. Chen T, Liu J, Yu R, Wu Y, Luo Q, Xu H, et al.. Targeting TLR7 reprograms macrophage function to attenuate influenza-associated Staphylococcus aureus coinfection . Microbial Pathogenesis. 2026. DOI: https://doi.org/10.1016/j.micpath.2025.108208

3. Tasci G, Ajmi N, Altun S, Satıcıoğlu IB, Duman M. Genome-Resolved Co-Infection by Aeromonas veronii and Shewanella sp. in Koi Carp: A Zoonotic Risk for Aquarists. Microorganisms. 2025. DOI: https://doi.org/10.3390/microorganisms14010036

4. Alsayed A, Zihlif M, Abuata O, Permana A, Zihlif M. Molecular Diagnostics, Antimicrobial Resistance Patterns, and Clinical Outcomes in Hospitalized Pneumonia Patients: A Prospective Study from Jordan and a Call for National Guideline Integration. Infection and Drug Resistance. 2026. DOI: https://doi.org/10.2147/IDR.S585095

5. Beloussov V, Strochkov V, Sandybayev N, Lavrinenko A, Solomadin M. A Pilot Study of Klebsiella pneumoniae in Community-Acquired Pneumonia: Comparative Insights from Culture and Targeted Next-Generation Sequencing. Diagnostics. 2026. DOI: https://doi.org/10.3390/diagnostics16010154

6. Touny A, Rageh F, Riad E, Sakr M, Abdelhady SA, Elgamal R, et al.. Incidence of Co-infection and its Impact on COVID-19 Patients admitted in the Intensive Care Unit. Egyptian Journal of Anaesthesia. 2023. DOI: https://doi.org/10.1080/11101849.2023.2175404

7. Bwanali AN, Lubanga A, Kondowe S, Nzima E, Mwale A, Kamanga W, et al.. Trends and patterns of antimicrobial resistance among common pathogens isolated from adult bloodstream and urinary tract infections in public health facilities in Malawi, 2020-2024. BMC Infectious Diseases. 2025. DOI: https://doi.org/10.1186/s12879-025-11335-1

8. Sharifipour E, Shams S, Esmkhani M, Khodadadi J, Fotouhi-Ardakani R, Koohpaei A, et al.. Evaluation of bacterial co-infections of the respiratory tract in COVID-19 patients admitted to ICU. BMC Infectious Diseases. 2020. DOI: https://doi.org/10.1186/s12879-020-05374-z

9. Cox M, Loman N, Bogaert D, O'Grady J. Co-infections: potentially lethal and unexplored in COVID-19. The Lancet Microbe. 2020. DOI: https://doi.org/10.1016/S2666-5247(20)30009-4

10. Uršič T, Erzar K, Seme K, Keše D, Jereb M, Strle F, et al.. Syndromic molecular testing in mechanically ventilated patients with severe pneumonia: a supportive diagnostic approach. Microbiology spectrum. 2025. DOI: https://doi.org/10.1128/spectrum.02155-25

11. Sharma B, Sreenivasan P, Biswal M, Mahajan V, Suri V, Sehgal IS, et al.. Bacterial coinfections and secondary infections in COVID-19 patients from a tertiary care hospital of northern India: Time to adhere to culture-based practices. Qatar medical journal. 2021. DOI: https://doi.org/10.5339/qmj.2021.62

12. Pourajam S, Kalantari E, Talebzadeh H, Mellali H, Sami R, Soltaninejad F, et al.. Secondary Bacterial Infection and Clinical Characteristics in Patients With COVID-19 Admitted to Two Intensive Care Units of an Academic Hospital in Iran During the First Wave of the Pandemic. Frontiers in Cellular and Infection Microbiology. 2022. DOI: https://doi.org/10.3389/fcimb.2022.784130

13. Aguilera ER, Clark SE, Lenz L. The role of NK cells during bacterial and viral/bacterial co-infection. Journal of Immunology. 2020. DOI: https://doi.org/10.4049/jimmunol.204.supp.148.30

14. Houdeshell H, Vest A, Arnold E, Thacker X, Harrison I, Clark C, et al.. B-149 Next-generation Sequencing Via Targeted Metagenomics for Broad Antimicrobial Resistance Surveillance. Clinical Chemistry. 2023. DOI: https://doi.org/10.1093/clinchem/hvad097.482

15. Li B, Mao M, Li H, Man X, Wu M, Lu C, et al.. Synergistic pathogenicity of novel duck Orthoreovirus and salmonella typhimurium in ducks. Poultry Science. 2025. DOI: https://doi.org/10.1016/j.psj.2025.104929

16. Barka S, Lawal J, Durkwa H, Balami A, Oladele O, Ogbe AO, et al.. Outbreak Investigation and Antimicrobial Resistance in Broiler Colibacillosis: A Case Report. Sahel Journal of Life Sciences FUDMA. 2025. DOI: https://doi.org/10.33003/sajols-2025-0303-48

17. Beloborodov V, Kovalev I, Sapronov GV. Current Issues of Empirical Therapy of Severe Bacterial Community-Acquired Pneumonia During the Season of Respiratory Viral Infections. Antibiotics and Chemotherapy. 2020. DOI: https://doi.org/10.37489/0235-2990-2020-65-9-10-64-70

18. Lucaciu F, Wellmann N, Mihai A, Sima A, Roșca O, Suba M, et al.. Post-COVID Respiratory Sequelae in COPD: Mucus Plugging, Infectious Complications, and Risk-Stratified Follow-Up. Journal of Clinical Medicine. 2026. DOI: https://doi.org/10.3390/jcm15082890

19. Hasona IF, Awad A, Younis G, Mohamed W. A One Health Perspective on Proteus mirabilis: The Interaction of Virulence and Antimicrobial Resistance Across Human and Animal Reservoirs. Microorganisms. 2026. DOI: https://doi.org/10.3390/microorganisms14020444

20. Khurana S, Singh P, Sharad N, Kiro V, Rastogi N, Lathwal A, et al.. Profile of co-infections & secondary infections in COVID-19 patients at a dedicated COVID-19 facility of a tertiary care Indian hospital: Implication on antimicrobial resistance. Indian Journal of Medical Microbiology. 2020. DOI: https://doi.org/10.1016/j.ijmmb.2020.10.014

21. Manna S, Weinberger D, Satzke C. Editorial: Thematic issue on bacterial-viral co-infections. FEMS Microbes. 2023. DOI: https://doi.org/10.1093/femsmc/xtac031

22. Prüller S, Rensch U, Meemken D, Kaspar H, Kopp P, Klein G, et al.. Antimicrobial Susceptibility of Bordetella bronchiseptica Isolates from Swine and Companion Animals and Detection of Resistance Genes. PLoS ONE. 2015. DOI: https://doi.org/10.1371/journal.pone.0135703

23. Kim T. Lessons From the Coronavirus Disease 2019 Pandemic: Implications for Antimicrobial Stewardship for COVID-19 Management. Journal of Korean medical science. 2026. DOI: https://doi.org/10.3346/jkms.2026.41.e72

24. Bilinskaya A, Kuti J, Linder KE. 112. Impact of Respiratory Viral PCR Panels (RVP) and Serum Procalcitonin (PCT) on Antibiotic Days of Therapy (DOT) in Patients Admitted with Lower Respiratory Tract Infections (LRTI). Open Forum Infectious Diseases. 2020. DOI: https://doi.org/10.1093/ofid/ofaa439.157

25. Kumar M, Kumar S, Kumar R, Singh R, Bhoi B, Kumar G, et al.. Chapter 39 Implementing microbiome analysis in clinical practice: overcoming challenges . The Gut Microbiome in Precision Medicine. 2026. DOI: https://doi.org/10.1016/B978-0-443-40483-2.00044-3

26. Sleiman A, Abdelkhalek P, Doumat G, Atallah F, Hamadeh L, Moussa P, et al.. The under investigated facet of the COVID-19 pandemic: Molecular analysis of secondary bacterial infections at a COVID dedicated intensive care unit within a tertiary care center in Lebanon. Frontiers in Medicine. 2023. DOI: https://doi.org/10.3389/fmed.2023.1001476

27. Zhou L, Sun H, Song S, Liu J, Xia Z, Sun Y, et al.. H3N2 canine influenza virus and Enterococcus faecalis coinfection in dogs in China. BMC Veterinary Research. 2019. DOI: https://doi.org/10.1186/s12917-019-1832-x

28. Boianelli A, Sharma-Chawla N, Bruder D, Hernández-Vargas E. Oseltamivir PK/PD Modeling and Simulation to Evaluate Treatment Strategies against Influenza-Pneumococcus Coinfection. Frontiers in Cellular and Infection Microbiology. 2016. DOI: https://doi.org/10.3389/fcimb.2016.00060

29. Fox A, Horby P, Ha NH, Hoa LNM, Lam N, Simmons C, et al.. Influenza A H5N1 and HIV co-infection: case report. BMC Infectious Diseases. 2010. DOI: https://doi.org/10.1186/1471-2334-10-167

30. Sharma S, Regmi S, Aryal S, Chalise BS, Gurung K, Thapa J, et al.. Extended Spectrum β-Lactamases Producing and Biofilm Forming Multidrug Resistant Klebsiella Pneumoniae from Bacterial Coinfection and Secondary Infection in Patients with Coronavirus Disease At A Tertiary Care Hospital, Kathmandu, Nepal. Journal of Institute of Science and Technology. 2025. DOI: https://doi.org/10.3126/jist.v30i2.82555

31. Fang H, Chen Y, Chen Y, Qi Y, Yan R, Guo F. In-host co-colonization and bloodstream infection by distinct classical and hypervirulent CRKP clones harboring a homologous blaKPC-2-harboring plasmid. Frontiers in Cellular and Infection Microbiology. 2025. DOI: https://doi.org/10.3389/fcimb.2025.1683743

32. Zheng Y, Li Y, Zeyneloğlu C, Tian W, Babayev H, D'Avino P, et al.. Risk and Protective Factors for Infection, Severe Disease, and Mortality in Epidemic Respiratory Viruses. Allergy. European Journal of Allergy and Clinical Immunology. 2026. DOI: https://doi.org/10.1111/all.70314

33. Asis A, Rodríguez A, Reyes LF, Díaz E, Nseir S, Martín-Loeches I. The double threat: bacterial and fungal co-/superinfection in viral pneumonia.. Expert Review of Respiratory Medicine. 2026. DOI: https://doi.org/10.1080/17476348.2026.2629003

34. Shomer N, Holcombe H, Villano J, Hayward A. Chapter 7 Biology and Diseases of Guinea Pigs, Gerbils, and Hamsters . Laboratory Animal Medicine. 2026. DOI: https://doi.org/10.1016/B978-0-443-13721-1.00009-9

35. Chagas AdLd, Araújo JCdS, Vieira J, Avelino M, Carneiro L. Isolation, identification, and sensitivity profile of Bacillus spp. in co-infection with respiratory viruses. Frontiers in Pediatrics. 2026. DOI: https://doi.org/10.3389/fped.2025.1718515

36. Gyem K, Pelden S, Tshering D, Penjor K, Wangchuk R, Dorji S, et al.. Aetiological, seasonal and antibiotic susceptibility patterns of diarrhoeal diseases in Bhutan (2016-2022): a retrospective study of surveillance data. BMJ Open. 2025. DOI: https://doi.org/10.1136/bmjopen-2024-086332

37. Lo S, Gladstone R, Tonder AJv, Plessis Md, Cornick J, Hawkins P, et al.. A mosaic tetracycline resistance gene tet(S/M) detected in an MDR pneumococcal CC230 lineage that underwent capsular switching in South Africa. Journal of Antimicrobial Chemotherapy. 2019. DOI: https://doi.org/10.1093/jac/dkz477

38. Ocheretina O, Shen L, Escuyer V, Mabou M, Royal-Mardi G, Collins SE, et al.. Whole Genome Sequencing Investigation of a Tuberculosis Outbreak in Port-au-Prince, Haiti Caused by a Strain with a "Low-Level" rpoB Mutation L511P - Insights into a Mechanism of Resistance Escalation. PLoS ONE. 2015. DOI: https://doi.org/10.1371/journal.pone.0129207

39. Nooh WA, Mustafa KK. Bacterial Co-Infections and Antibiotic Resistance in Urinary Tract Infection of Covid-19 Patients in Erbil City. Zanco Journal of Pure and Applied Sciences. 2024. DOI: https://doi.org/10.21271/zjpas.36.5.1

40. Rizvi A, Saeed M, Nadeem A, Yaqoob A, Rabaan A, Bakhrebah M, et al.. Evaluation of Bi-Lateral Co-Infections and Antibiotic Resistance Rates among COVID-19 Patients in Lahore, Pakistan. Medicina. 2022. DOI: https://doi.org/10.3390/medicina58070904

41. Youngs J, Arnold A. Infections in the immune-compromised host: secondary immunodeficiency. Medicine. 2021. DOI: https://doi.org/10.1016/j.mpmed.2021.07.003

42. Singh V, Upadhyay P, Reddy J, Granger J. SARS-CoV-2 respiratory co-infections: Incidence of viral and bacterial co-pathogens. International Journal of Infectious Diseases. 2021. DOI: https://doi.org/10.1016/j.ijid.2021.02.087

43. Falk CL, Dumkow LE, Geyer AC, Egwuatu N, Langholz JM, Schmidt KJ, et al.. 138. Impact of Emergency Department Rapid Viral Respiratory Multiplex PCR Testing on Antimicrobial Prescribing. Open Forum Infectious Diseases. 2023. DOI: https://doi.org/10.1093/ofid/ofad500.211

44. Pandey M, May A, Tan L, Hughes H, Jones J, Harrison W, et al.. Comparative incidence of early and late bloodstream and respiratory tract co-infection in patients admitted to ICU with COVID-19 pneumonia versus Influenza A or B pneumonia versus no viral pneumonia: wales multicentre ICU cohort study. Critical Care. 2022. DOI: https://doi.org/10.1186/s13054-022-04026-9

45. Mollasharifi T, Zamani M, Gharib A. Evaluation of Secondary Bacterial Infections and Determination of Antibiogram Susceptibility Testing in Hospitalized Patients with COVID-19. Iranian Journal of Pathology. 2024. DOI: https://doi.org/10.30699/IJP.2024.2006587.3141

46. Attal R, Deotale V. Bacterial co-infections and secondary infections and their antimicrobial resistance in Covid-19 patients during the second pandemic wave. GMS Hygiene and Infection Control. 2024. DOI: https://doi.org/10.3205/dgkh000465

47. Badr H, Shosha EA, Roshdy H, Mohammed AAE, Saad N, Aboelenin SM, et al.. Investigation of many bacterial and viral infections circulating in pigeons showing nervous symptoms. Saudi Journal of Biological Sciences. 2022. DOI: https://doi.org/10.1016/j.sjbs.2022.01.023

48. Wagner J, Bojkova D, Shumliakivska M, Luxán G, Nicin L, Aslan GS, et al.. Increased susceptibility of human endothelial cells to infections by SARS-CoV-2 variants. Basic Research in Cardiology. 2021. DOI: https://doi.org/10.1007/s00395-021-00882-8

49. Anaya BJ, Osorio-Vargas E, Monterrosa-Moreno S, Tirado DF, González-Burgos E, Serrano DR. Pulmonary Drug Delivery for Infectious Diseases: Cutting-Edge Formulations and Manufacturing Technologies. Pharmaceutics. 2026. DOI: https://doi.org/10.3390/pharmaceutics18020242

50. Patel H, Benny B, Nahar P, Landis A. The Impact of Antimicrobial Stewardship in Treating Patients with Escherichia coli Bacteremia in a Small Single Center Community Hospital. INNOVATIONS in Pharmacy. 2023. DOI: https://doi.org/10.24926/iip.v14i4.5526