Avian Influenza

Influenza viruses belong to the family Orthomyxoviridae and are enveloped viruses with a segmented, negative-sense RNA genome. Four genera of influenza viruses have been identified: Influenza A, B, C, and D. Influenza A and B viruses possess eight RNA segments, whereas Influenza C and D viruses typically contain seven segments, although they frequently package eight ^4,5. Influenza A viruses (IAVs) exhibit a broad host range, infecting birds, humans, and various other mammals. In contrast, Influenza B and C viruses primarily infect humans, while Influenza D virus is predominantly found in bovine hosts ^4,6. In humans, influenza disease is caused mainly by viruses from the A and B genera, with IAVs being of particular concern due to their pandemic potential. These genera are distinguished by the antigenic properties of their internal virion proteins, while IAVs are further differentiated into subtypes by the antigenic properties of their two external proteins ^4,6.

The segmented nature of the influenza virus genome facilitates frequent reassortment events. This process occurs when a host is co-infected with two different influenza viruses; during replication, genome segments from both viruses can be packaged together into progeny virions ^4,7. Pigs, which are susceptible to multiple influenza virus subtypes, have traditionally been considered key hosts for reassortment and are often referred to as “mixing vessels” ⁷. IAVs display substantial antigenic diversity in their two principal surface glycoproteins: hemagglutinin (H) and neuraminidase (N). This variability allows for a binary classification of IAVs into serotypes, comprising 19 H subtypes and 11 N subtypes, with more than 130 distinct subtype combinations identified to date ^8,9. The primary reservoir of this genetic and antigenic diversity is avian IAVs, which have been the source of multiple influenza pandemics-often through reassortment events between avian and human influenza viruses ⁷.

Human influenza viruses bind preferentially to α-2,6-linked sialic acid receptors on the surface of epithelial cells, whereas avian influenza viruses have a higher affinity for α-2,3-linked sialic acids. These glycosidic linkages play a critical role in host specificity and cross-species transmission of influenza viruses. Consequently, both the variant and structural configuration of sialic acid receptors influence viral binding affinity to the H glycoprotein of influenza viruses ¹⁰.

Avian influenza A viruses (AIAVs) can evolve from low pathogenicity (LPAIV) to highly pathogenic forms (HPAIV), a classification based on their virulence in chickens. A defining molecular feature of HPAIV is the presence of a furin cleavage site in the H protein ³. Proteolytic cleavage of H into HA1 and HA2 subunits is a crucial step in viral replication, triggering a pH-dependent conformational change that exposes the fusion peptide, enabling fusion with endosomal membranes and subsequent release of the viral genome into the host cytoplasm. HPAIVs contain a polybasic cleavage site that allows processing by furin-like proteases, which are expressed in multiple tissues, thereby facilitating systemic viral dissemination ¹¹.

Notably, HPAIVs have emerged multiple times independently from ancestral LPAIVs, but only among viruses of the H5 and H7 subtypes ¹². Since the first emergence of HPAI H5N1 in 1996, nearly 1,000 human infections have been reported across 24 countries, with a case fatality rate of around 50%^13,14.

The influenza pandemics of the past

Over the past three centuries, ten major influenza pandemics have been documented, averaging approximately three per century (Fig. 1). These events have varied widely in intensity, morbidity, and mortality ¹.

Fig. 1 Major events in human and avian Influenza. The timeline shows the known Influenza pandemics in the past centuries (in blue) and major events in avian influenza over the last two centuries (in orange). The subtype of IAV responsible for the pandemic of the previous two centuries are shown ^{1,12,18,23,30}. 

In the 18th century, three pandemics were recorded. The first, in 1729-1730, had a global reach, with high morbidity but relatively low mortality-a pattern that was repeated in the subsequent 1732-1733 outbreak. The 1781-1782 pandemic was particularly extensive, reportedly infecting 70-80% of the population, though it also exhibited low mortality ¹.

The 19th century witnessed two notable pandemics in 1830-1831 and 1833, which differed in severity, with the latter associated with higher mortality. The 1889-1890 pandemic, commonly known as the “Russian flu,” was particularly significant because it coincided with the emergence of the germ theory of disease. This conceptual shift- from miasmatic to microbial causation- marked a turning point in public health, paving the way for modern preventive strategies such as improved hygiene and social distancing ¹.

The 20th and 21st centuries saw four influenza pandemics, all caused by IAVs. The deadliest of these was the 1918-1919 “Spanish flu,” caused by an H1N1 virus. It infected an estimated 500 million people-about 30% of the global population-and resulted in 50 to 100 million deaths over three devastating waves. The impact was magnified by the worldwide disruption and troop movements associated with World War I. Later pandemics included the 1957 “Asian flu” (H2N2) and the 1968 “Hong Kong flu” (H3N2), which were less severe but still responsible for an estimated 1-3 million and 1-4 million deaths, respectively, despite widespread transmission ¹.

The only influenza pandemic of the 21st century to date occurred in 2009, caused by a novel H1N1 virus that originated in North America. Although initially feared to rival the severity of the 1918 pandemic, it ultimately proved relatively mild, with an estimated 125,000 to 400,000 deaths worldwide ¹.

The panzootic H5N1 clade 2.3.4.4b

In addition to classification by subtype, IAV genome sequences are further categorized based on phylogenetic relationships, including common ancestry and descendant lineages. These classifications include clades (e.g., 1.1, 2.2, 2.3), subclades (e.g., 2.3.2.1c, 2.3.3.4b), lineages (such as Eurasian and American), and genotypes (e.g., A1, B1, B3.13) ^3,15,16. Clades and subclades are defined primarily through phylogenetic analysis of the hemagglutinin (H) gene, reflecting evolutionary divergence. In contrast, genotypes are determined by the constellation of gene segments present in each strain, offering insight into reassortment events and genomic composition ¹⁶.

The subclade 2.3.4.4b emerged in 2016 (Fig. 1) in an H5N8 virus in China and subsequently spread across Asia and into Europe. This virus caused an unexpected epidemic peak in 2020/2021 ¹⁷. The 2.3.4.4b H5N8 rearranged with a LPAI H5N1 around 2020, leading to the 2.3.4.4b H5N1. This virus caused the wave of 2021/2022, the current panzootic, with several human cases, and was the most devastating avian influenza epidemic ^17,18. After this subclade was introduced into North America, several reassortment events with circulating LPAIVs led to the emergence of new genotypes. At least three genotypes have circulated among infected marine birds in South America since 2022, and these genotypes have also shown infectivity and virulence toward marine mammals ^19,20.

Both avian- and human-type receptors are present in swine respiratory epithelia. Swine are generally the intermediate hosts involved in the reassortment and adaptation of AIAVs to mammals before spillover to humans. However, new evolutionary pathways may be considered for the emergence of a pandemic, given the frequent infection of many mammalian species (both marine and terrestrial), particularly cattle in the USA^20,21.

Cattle are typically infected with the Influenza D virus but not with IAVs. However, during the current panzootic, the first documented spillover of Influenza A into cows involved an H5N1 virus of genotype B3.13- originally circulating in wild animals (Fig. 1). This genotype is a reassortant between B3.6 and an LPAIV endemic in the United States. Phylogenetic evidence suggests that a single spillover event was responsible for the emergence of this panzootic clade in cattle ²². In cattle, the primary site of AIAV replication appears to be the udder, as mammary tissue expresses avian-type α-2,3-linked sialic acid receptors. This raises the possibility of virus transmission through unpasteurized milk. The initial spillover event is believed to have occurred in Texas, USA, with subsequent spread to cattle in other states likely facilitated by contaminated milking equipment ¹⁸. Since February 2024, at least two additional spillover events involving a different genotype, D1.1, have been reported in U.S. cattle ^23,24. Alarmingly, the first fatal human case associated with this panzootic in the United States was linked to infection with this genotype ²³.

Most human cases of H5N1 infection during the current panzootic have been mild so far. However, this apparent low pathogenicity may be partly attributed to host-related factors, including the young age of many patients and the presence of pre-existing immunity to N1 and other conserved cellular immunity epitopes ^25,26. In contrast, experimental infection of ferrets with the H5N1 genotype B3.13-currently circulating in U.S. cattle-resulted in high lethality ²⁷. Another study reported reduced mortality in mice infected with a clade 2.3.4.4b H5N1 virus, despite the strain exhibiting high pathogenicity in chickens ²⁸. Notably, differences in pathogenicity among genotypes within the panzootic clade have also been observed in both teals and poultry ²⁹.

These findings underscore the potential for this lineage’s pathogenicity to evolve further through reassortment and mutation. Comprehensive studies are urgently needed to assess the virulence and transmissibility of these viruses in humans. Regardless of the current severity, it is imperative to implement preparedness and response plans now to mitigate the risk of a future influenza pandemic.

Pandemic risk?

The current panzootic caused by Influenza A-H5N1 has been marked by frequent spillover events into a wide range of terrestrial and marine mammalian species, with around 50 species affected to date ^3,30,31. The United States currently reports the highest number of human cases associated with this outbreak, with 70 confirmed infections between 2024 and the end of April 2025 ¹³. However, this number remains below the highest annual record of human H5N1 cases, which occurred in 2015, when 145 cases were reported globally, 136 of them in Egypt alone ¹³.

A prerequisite for the development of a pandemic is that the panzootic influenza virus acquires the ability not only to infect humans, but also to sustain efficient human- to-human transmission-something that has not occurred to date. Efficient human-to-human transmission of avian influenza viruses likely requires the cumulative acquisition of several key mutations (Fig. 2) ^30-33.

Fig. 2 Major mutations in four of the AIAV associated with efficient mammalian transmission and replication. Mutations in four viral proteins, potentially related to human adaptation, are shown ^22,30,32-36. 

H mutations enhancing affinity for human receptors: Adaptation to the human α-2,6-linked sialic acid receptor is critical. Notably, only four mutations were sufficient to render an Influenza A-H5N1 virus transmissible, via respiratory droplets, in ferrets³⁴. More recently, a single mutation, Q226L, in the H of bovine H5N1 (genotype B3.13) was shown to switch viral specificity toward mammalian receptors ³⁵. Three additional mutations have been identified as key contributors to increased human receptor affinity ³².

PB2 mutations enabling polymerase adaptation to mammalian ANP32 proteins: ANP32A and ANP32B, which serve as host cofactors for the viral polymerase, differ between avian and mammalian species. Several mutations have been characterized that allow avian influenza polymerases to utilize mammalian ANP32 proteins ³⁶ efficiently. Some of these mutations have already been detected at low frequency in viruses infecting cattle ²².

Increased viral stability at low pH: Avian influenza viruses generally exhibit reduced stability in acidic environments such as the human upper respiratory tract, posing a barrier to transmission ³³.

Host immune history: Pre-existing immunity to seasonal human influenza viruses, whether from prior infection or vaccination, may influence susceptibility and clinical outcomes following exposure to H5N1 viruses ³⁷.

The eventual emergence of a pandemic influenza virus can result not only from the gradual accumulation of key mutations but also, frequently, from reassortment events^20,21. Traditional “mixing vessels” for such reassortment have been pigs; however, emerging evidence highlights other hosts with high reassortment potential, including humans, minks, ferrets, seals, dogs, cats, and various bird species-particularly turkeys, chickens, quails, and ducks ^20,21. Interestingly, current data suggest that pigs may be less susceptible to the current panzootic IAV than some other mammals ^38,39.

Although several key mutations associated with mammalian adaptation have been detected in viruses isolated from infected mammals during this panzootic ²¹, none of the sequenced viral isolates to date possess the whole constellation of mutations required for efficient human-to-human transmission. Nevertheless, the extensive geographic spread of the panzootic clade and its ability to infect a broad range of mammalian hosts raise significant concerns about the potential emergence of a pandemic strain ⁴⁰.

Preparedness

According to the World Health Organization ⁴¹, influenza virus activity is classified into six distinct phases, progressing from circulation limited to animals to widespread human transmission at the global level (Table 1). This framework is divided into two overarching stages: Phases 1-3, which focus on preparedness, and Phases 4-6, which emphasize response and mitigation.

In Phase 1, no influenza viruses circulating among animals have been reported to cause infections in humans. Phase 2 is characterized by an animal influenza virus-circulating among domesticated or wild animals - that has caused human illness and is therefore considered a potential pandemic threat. Phase 3 occurs when an animal or human-animal reassortant influenza virus causes sporadic cases or small clusters of disease in people but does not lead to sustained human-to-human transmission or community-level outbreaks. In Phase 4, human-to-human transmission results in localized community outbreaks, indicating a significant increase in pandemic risk. Phase 5 is marked by community-level outbreaks in at least two countries within one WHO region. In contrast, Phase 6 -defined by community outbreaks in multiple WHO regions-signifies the onset of a global pandemic ⁴¹.

Although the global spread of the A (H5N1) panzootic remains a serious concern, no sustained human-to-human transmission has been reported as of May 2025. Consequently, the human influenza epidemic remains classified as Phase 3, highlighting the urgent need to bolster preparedness efforts and prevent escalation to a pandemic. To mitigate this risk, countries must develop and implement coordinated response strategies at both national and international levels, emphasizing proactive preventive measures. Since 2021, the World Health Organization (WHO) has been working with Member States to draft a global treaty on pandemic prevention, preparedness, and response. This treaty was formally adopted by consensus at the 78th World Health Assembly in May 2025. It establishes guidelines for the timely sharing of epidemiological data, genomic analysis of emerging pathogens, and critical information related to potential vaccines and treatments ⁴².

National preparedness plans should incorporate robust communication strategies to promote public adherence to non-pharmaceutical interventions (NPIs). For influenza, these include individual-level protective behaviors such as staying home when symptomatic, covering coughs and sneezes, and practicing frequent hand hygiene. At the community level, recommended measures may include temporary school closures, suspending childcare services, and canceling mass gatherings during periods of heightened transmission. When a novel pandemic influenza virus emerges, the combined implementation of NPIs and antiviral therapies can significantly reduce transmission rates, particularly in the early stages before a vaccine becomes widely available ^41-44.

Pandemic preparedness for influenza depends heavily on early detection of novel strains, particularly those that originate in animal reservoirs through spillover events. Over the past two decades, repeated zoonotic influenza outbreaks have underscored the vital importance of sustained surveillance in both human and animal populations, especially among birds and swine. As noted earlier in this paper, the most pressing current threat is the ongoing panzootic of avian influenza A (H5N1). However, other avian influenza subtypes continue to circulate in more localized contexts. Genetic reassortment during coinfections plays a critical role in the emergence of pandemic-capable strains. While the surface glycoproteins H and N are traditionally viewed as central to influenza infectivity and host specificity, increasing evidence indicates that internal viral genes -particularly those encoding the polymerase complex- also significantly contribute to virulence and disease severity ⁴⁴. Since 1952, the World Health Organization has led a global initiative to monitor circulating influenza viruses, enabling a coordinated international response to emerging influenza threats ⁴⁵.

Given that a future pandemic influenza virus is likely to emerge from an avian influenza strain affecting poultry, control strategies must integrate biosecurity measures, epidemiological surveillance, targeted culling, and vaccination with strain-specific immunogens ^46-48. Culling should be prioritized in outbreaks involving HPAI strains, mainly when rapid transmission occurs within poultry populations. Protective measures for farm workers are also essential to minimize the risk of zoonotic transmission. A similar level of urgency applies to controlling H5N1 infections in other livestock species, such as cattle ⁴⁹.

During the preparedness phase, early interventions should focus on securing vaccine and antiviral stockpiles ^48-51 and on expanding the healthcare system capacity to ensure effective clinical management during a pandemic. Notably, at least three experimental vaccines targeting early H5N1 influenza virus strains have demonstrated the ability to elicit cross-reactive binding and cross-neutralizing antibodies against the HPAI clade 2.3.4.4b in humans, suggesting they could serve as interim protective tools while updated vaccines are developed⁵². Furthermore, a novel H5-based mRNA vaccine has recently been shown to generate a robust adaptive immune response in cattle⁵³. Nevertheless, the final formulation of a pandemic influenza vaccine must be tailored to the specific viral strain in circulation at the time of emergence to ensure optimal efficacy.