Influenza viruses enter host cells by binding of HA to sialic acid residues on glycoproteins and glycolipids. Binding affinity for specific sialic acid residues is an important determinant of host range, tissue tropism, and the potential for cross-species transmission (recently reviewed in ref [39]). In short, human influenza viruses preferentially bind to sialic acids with an α2,6 linkage to galactose (SAα2,6Gal) whereas avian influenza viruses preferentially bind to α2,3 linkages (SAα2,3Gal) [40]. In people, SAα2,6Gal is found predominantly in the upper airway and SAα2,3Gal mainly in the lower airway [41]. In contrast, birds express SAα2,3Gal predominantly in the upper airway and intestinal tract [42]. Thus, mutations that allow avian influenza viruses to bind to SAα2,6Gal in the upper respiratory tract of people are likely required for efficient human-to-human transmission via respiratory droplets and aerosols.

Influenza viruses replicate in the nucleus of respiratory and intestinal epithelial cells. Replication peaks at 48 h after infection and virus is shed for approximately 6–8 days [43]. Severity of disease is associated with viral replication in the lower respiratory tract [25]. While symptoms result from a combination of virus-mediated damage to epithelial cells and host immune responses (immunopathology), it is generally accepted that immunopathology plays the largest role in tissue damage [44]. Infected epithelial cells and innate immune cells release pro-inflammatory cytokines and chemokines that are important to control infection, but also lead to bystander damage to epithelial and endothelial cells. Excess neutrophil recruitment [45,46] and inflammatory cytokines such as IL-6, IL-8, IL-10, TNFα, CXCL10, IL-2R, GCSF, MCP1, and MIP1α are associated with disease severity and poor outcomes [35,47,48,49].

Because lung specimens are usually collected during autopsy, pathology is only documented in fatal cases of human IAV infection. Characteristic findings of viral pneumonia are similar between pandemic and non-pandemic years [25,32,34,50,51,52,53,54] so are described together. Grossly, the trachea and bronchi are hemorrhagic and often filled with blood-tinged, foamy fluid. The lungs are dark red and edematous, reflecting the underlying hemorrhagic bronchopneumonia that is frequently complicated by secondary bacterial infection [32,34,53]. Histologically, the trachea and bronchi have epithelial necrosis and desquamation in addition to submucosal edema, congestion, and hemorrhage. In the lower airways, there is necrotizing bronchiolitis and alveolitis, interstitial mononuclear inflammation, interstitial and alveolar edema, thrombi, hyaline membranes, and type II pneumocyte hyperplasia. These changes are consistent with the exudative phase of diffuse alveolar damage (DAD), which is the histologic hallmark of acute respiratory distress syndrome (ARDS) [55]. With time, epithelial regeneration, interstitial fibrosis, and bronchiolitis obliterans may develop. Secondary bacterial pneumonia consisting of overwhelming neutrophilic inflammation, extensive necrosis, and hemorrhage can obscure the underlying viral effects [25]. Hemophagocytosis is a prominent feature of some cases of fatal disease and is thought to be mediated by hypercytokinemia [33,51]. Hemophagocytosis was present in the lymph nodes of 18 of 36 (53%) pediatric patients dying of non-pandemic influenza [52] and in 25 of 41 (61%) patients dying of 2009 pH1N1 [33]. The clinical significance of hemophagocytosis in these infections is unclear.

While overall histologic findings are similar in fatal cases from pandemic and non-pandemic years (Table 1), antigen distribution appears to differ. In a study of 47 pediatric patients dying of seasonal influenza pneumonia from 2003–2004, antigen was detected primarily in the bronchial epithelial cells and mucous glands of the trachea, bronchi, and large bronchioles [52]. In contrast, antigen was mainly present in type I and II alveolar pneumocytes in patients dying of 2009 pandemic H1N1 (pH1N1) [33]. These findings likely reflect differences in tissue tropism based on receptor location and the variable course of disease at the time of sampling.

As the case numbers are significantly lower compared with seasonal and pandemic H1N1, there are fewer reports on the pathology of fatal avian H5N1 infection despite a case fatality rate of 56% [56]. Diffuse alveolar damage is the main histologic feature and most cases also display mild lymphocytic interstitial pneumonia, alveolar histiocytosis, hemorrhage, and type II pneumocyte hyperplasia [51,57,58,59,60,61,62,63,64]. Depending on the time course of disease, DAD may be exudative or organizing and fibrotic [62,63]. Extrapulmonary lesions including lymphoid depletion, hepatic necrosis, and acute tubular necrosis, are frequently reported [51,58,60,61,65]. Viral RNA can be detected outside of the respiratory tract in the spleen, liver, intestines [62,63,64], and cerebrospinal fluid [66] although extrapulmonary antigen is only rarely documented [61], suggesting that systemic spread may not be the culprit of multiorgan failure. In the respiratory tract, viral antigen and RNA are found most commonly in type I and II pneumocytes although they can also be present in macrophages, sloughed epithelial cells, non-ciliated and ciliated bronchiolar epithelium, and tracheal epithelium [57,58,60,61,62,63]. The predominant viral distribution in pneumocytes is consistent with the lower airway distribution of SAα2,3Gal, the receptor for avian influenza viruses [41]. Hemophagocytosis is a frequent finding in the lungs, lymph nodes, and spleen [51,57,58,60,61,63,64]. In severe cases, reactive hemophagocytic syndrome, consisting of pancytopenia, abnormal clotting times, and reduced liver function, is the most prominent finding [51,60]. A combination of high viral loads and an intense cytokine response appear central to the pathogenicity of H5N1 in people [57,67,68,69].