Influenza A H5N1 and HIV co-infection: case report

Influenza A/H5N1 infection is characterized by high viral loads, overproduction of pro-inflammatory cytokines and chemokines, direct lung tissue destruction, pulmonary oedema and extensive inflammatory infiltration [1‐3]. The prevailing view is that alveolar damage is the primary pathology leading to acute respiratory distress, multiple organ dysfunction syndrome and death [3]. Likewise, 2009 H1N1 infection can cause acute respiratory distress syndrome and death in previously healthy young adults very similar to the clinical syndrome seen in H5N1 [4].

It remains unclear whether lung pathology in severe influenza is a direct consequence of high viral loads and/or of ensuing inflammatory responses. The involvement of innate versus adaptive immunity in inflammation or controlling viremia is also poorly defined. Further understanding of the pathological processes is necessary to develop interventions that prevent severe lung disease. The occurrence of H5N1 infection in a patient with HIV infection offered a unique opportunity to study the pathological and immunological process when adaptive immunity is impaired.

To our knowledge only one other patient with documented HIV and H5N1 co-infection has been reported and details of the clinical course in this patient have not been published [6].

Patients with immune compromise including those with HIV are at higher risk of complications associated with seasonal influenza [7]. This may not translate to highly pathogenic H5N1 which has been fatal in the majority of previously healthy patients. Although early clinical and laboratory findings were similar to those of other fatal H5N1 patients [2, 6, 8, 9], the patient showed a transient clinical improvement and a relatively delayed time to death, which was accompanied by signs of secondary pneumonia. In addition, viral loads were relatively low with no virus detection in plasma, unlike some fatal cases described previously [2, 6, 8, 9] suggesting that H5N1 viral clearance was not compromised by HIV co-infection. H5N1 can be cleared without antivirals and there is little benefit of Oseltamivir after day 6 of illness [9, 10]. However early control of viral load appears to be important because survival rates are highest in patients treated earliest [10]. Thus, the contribution of Oseltamivir may have been negligible in this patient because viral loads were relatively low prior to administration. Lymphopenia is common in H5N1 patients [2, 6, 9] and low CD3 counts and CD4:CD8 ratios have been reported [2]. We can not determine if immune compromise preceded influenza A/H5N1 infection in this patient. Investigation of AIDS defining illnesses was not exhaustive because the patient was reported to have been healthy and the presenting symptoms and rapid onset of illness were consistent with the confirmed diagnosis of H5N1. We can not exclude the possibility of P. jirovecci or mycobacterial co-infection. CD4 counts and CD4:CD8 ratios were lower than we have seen in other H5N1 patients (unpublished findings). This suggests that CD4 T cells may not be required for viral clearance consistent with findings that CD8 T cells are more important for survival from highly virulent influenza infection in mice [11]. The patient maintained substantial numbers of peripheral CD8 T cells and a large fraction became activated, but the concomitant shift to a CD27+CD28- phenotype raises doubts about their antiviral function. This phenotype is associated with CD8 T cells that proliferate but lack cytotoxic function and accumulate in progressive HIV infection [12]. Chemokine and cytokine levels were high as in other fatal H5N1 cases and may account for the severity of the early pneumonia. The prevailing view is that excessive cytokine and chemokine release are secondary to high viral loads [2], whereas viral loads were low in this patient. Although HIV infection is also associated with increased expression of proinflammatory cytokines [13, 14], levels tend to be lower than found here [15], even with co-presentation of pneumocystis pneumonia, bacterial pneumonia or mycobacteriosis, and IFN-γ expression is often decreased [13]. Levels of IL-8, MCP-1, and IL-6 also decreased with H5N1 viral load indicting that they are primarily induced by H5N1 infection. There has been no conclusive report of secondary infection accompanying H5N1 whereas secondary pneumonia may have caused a considerable fraction of the deaths from the highly pathogenic 1918 H1N1 strain and the current H1N1 2009 strain [16]. Findings in this patient including recrudescence of fever, rising CRP levels and neutrophilia were consistent with a secondary infection despite being treated with broad-spectrum antibiotics. A. fumigatus was detected in a tracheal aspirate suggestive of invasive pulmonary aspergillosis. Corticosteroid adminsitration may have contributed to the presumed development of invasive pulmonary aspergillosis [17], but there have been numerous reports of invasive pulmonary aspergillosis developing after influenza A infection in which T cell lymphopenia occurred [18, 19]. None these patients had HIV and few received corticosteroids indicating the influenza A associated immune suppression may be sufficient for the development of invasive pulmonary aspergillosis. In conclusion the course of H5N1 infection in this case was not notably different in the presence of HIV co-infection but it is possible that HIV co-infection and profound CD4 T cell depletion increase susceptibility to secondary infection. The findings suggest that CD4 T cells may not be required for H5N1 virus clearance.