Introduction

HIV infection is associated with abnormal humoral and T-cell–mediated immune responses, resulting in increased susceptibility to numerous infections, including viral respiratory infections [1]. In the spring of 2009, a pandemic influenza A (H1N1) virus (pH1N1) emerged [2]. During the summer period (June–August), influenza activity continued to occur in the Northern Hemisphere and peaked in the Southern Hemisphere [3]. In the fall (September–November), a second wave of pandemic H1N1 activity occurred in the United States, peaking in late October [4]. Seasonal influenza virus infection is a common cause of respiratory illness in HIV-infected adults [5]; descriptions of the epidemiology and clinical presentation of pH1N1 virus infections in HIV-infected adults have recently been reported. In this paper, we review the existing literature regarding pH1N1 virus susceptibility, clinical presentation, and severity in HIV-infected adults. We also describe diagnostic and treatment challenges specific to HIV-infected adults and summarize recent studies of pH1N1 viral shedding and vaccine response in HIV-infected adults.

Susceptibility to pH1N1 Virus Infection and Illness Among HIV-Infected Adults

Several studies have reported serologic evidence of pH1N1 virus infection in HIV-infected adults. The first was a multicenter study of women in the United States which reported incidence rates of pH1N1 virus infection (defined as a ≥fourfold increase in pH1N1-specific hemagglutination inhibition antibody titer in sera collected March to September 2009 compared with the same months in 2008 or 2007) of 20.6 (95% CI 6.1–69.6) and 24.6 (95% CI 7.4–81.4) per 100 person-years among HIV-infected women (n = 1,270) and similar HIV-uninfected (n = 531) women, respectively [6]. The second study among 199 HIV-infected adults in Australia found the seroprevalence (defined as pH1N1-specific hemagglutination inhibition antibody [HIA] titer ≥40 in sera collected from the June–September 2009 winter season) of 34.2% [7•]; this was reported to be similar to the post-pandemic seroprevalence rate of 28.4% reported in the general community [8]. The authors also reported a seroconversion rate (defined as a fourfold or greater increase in pH1N1-specific HIA titer between the winter sample and the sample collected in the preceding 3 to 6 months) of 14.6%, similar to the 15.6% rate reported in the general community. CD4 cell counts and HIV viral loads were not statistically different between patients who seroconverted and those who did not; however, the majority of participants in this study had a CD4 cell count of at least 350 cells/μL (74%) and an undetectable viral load (69%). Finally, two European pH1N1 vaccine studies measured pre-vaccine pH1N1-specific antibody levels in November and December 2009 (following the peak of pH1N1 activity) in HIV-infected persons compared with HIV-uninfected persons. Both studies reported higher rates of pre-vaccination serologic evidence of pH1N1 virus infection by HIA titer in HIV-infected versus HIV-uninfected persons (13% vs 3%, and 80% vs 35%, respectively) [9, 10•]. An important limitation of these studies is that serologic response may be absent in some persons with laboratory-confirmed influenza virus infection, and it is unknown if the proportion who do not develop a serologic response differs between HIV-infected and HIV-uninfected persons. It is also impossible to assess whether HIV-infected and HIV-uninfected persons in these studies had similar exposures to circulating influenza viruses.

Reports of the prevalence of HIV infection among persons with laboratory-confirmed pH1N1 virus infection provided some early data about influenza illness susceptibility among HIV-infected adults. In Brazil (0.6% estimated adult HIV prevalence [11]), 1.3% of 5,747 laboratory-confirmed pH1N1 cases reported during the first 4 months of the pandemic were among patients with HIV infection [12], suggesting that HIV-infected persons were not considerably overrepresented among persons with laboratory-confirmed pH1N1 virus infection. A recent study from France reported the incidence of influenza-like illness (ILI) during the pH1N1 pandemic period among 908 HIV-infected persons as 4.5% (95% CI, 3.7–5.4) [13]. HIV-infected persons with ILI were more likely to report at least one high-risk comorbid condition (Table 1) [14••] compared with HIV-infected persons without ILI (34% vs 12%, P < 0.0001). AIDS diagnosis and CD4 cell count were not associated with development of ILI, though the majority of patients were virologically suppressed with CD4 cell counts >350 cells/μL.

Table 1 On the basis of epidemiologic studies of seasonal influenza or pH1N1, list of persons at higher risk for influenza complications [14••]
Full size table

Several studies have specifically reported the frequency of laboratory-confirmed pH1N1 virus infection in HIV-infected adults presenting with respiratory symptoms compared with HIV-uninfected adults during the pandemic. (These studies were conducted during an influenza pandemic. The relatively high proportion of patients presenting with respiratory symptoms who were confirmed to have pH1N1 virus infection may not be representative of the proportion of acute respiratory illness attributable to influenza viruses during a non-pandemic influenza season.) The first study found 11 (55%) of 20 HIV-infected emergency room patients in Mexico City with ILI had a nasopharyngeal specimen that was positive for pH1N1 virus, while only 120 (15%) of 812 HIV-uninfected patients tested positive (P < 0.001) [15]. Notably, among patients with pH1N1, HIV-infected patients were older (median age 47 vs 36 years) and more likely to smoke tobacco (63% vs 13%, P = 0.001). Conversely, in a study from Barcelona, pH1N1 virus was isolated in only 56 (3%) nasopharyngeal specimens collected from HIV-infected adults with acute respiratory illness compared with 567 (27%) specimens collected from HIV-uninfected patients with similar illness; like the first study, HIV-infected patients with pH1N1 were older (median age 44 vs 39 years, P = 0.02) and more likely to smoke tobacco (54% vs 13%, P < 0.0001) [16]. Finally, in a third study, pH1N1 virus was detected from nasopharyngeal swabs of 30 (24%) of 126 HIV-infected patients (median CD4 cell count 163 cells/μL, 50% receiving highly active antiretroviral therapy [HAART]) with respiratory symptoms seen at a tertiary center in Mexico City between April 2009 and February 2011 [17•].

In summary, there is not clear evidence that HIV-infected adults without advanced immunosuppression or other comorbid illnesses are more susceptible to pH1N1 virus infection or pH1N1-related illness compared with the general adult population. The role of CD4 count in susceptibility to influenza virus infection remains unclear, primarily due to the limited data for persons with very low CD4 counts, though one study observed a high frequency of pH1N1 virus infection among HIV-infected patients with relatively low CD4 cell counts. Other risk factors for influenza virus infection, particularly tobacco smoking, are more prevalent among HIV-infected individuals [18], and these factors may be important mediators of the relationship between HIV infection and influenza illness.

Clinical Presentation and Diagnosis of pH1N1 Illness in HIV-Infected Adults

Reports from Singapore, Spain, Chile, Mexico, and the United States demonstrate that HIIV-infected adults have similar clinical symptoms of pH1N1 as the general population, including cough (84%–91%), fever (44%–100%), myalgias (27%–80%), sore throat (27%–64%), rhinorrhea (8%–63%), dyspnea (9%–62%), gastrointestinal symptoms (9%–48%), and headache (0%–39%) [15, 16, 19, 20, 21•, 22•]. The accuracy of clinical diagnosis of influenza on the basis of symptoms alone, however, is limited because symptoms can overlap considerably with symptoms from illness caused by other pathogens [23], particularly in HIV-infected patients, for whom the differential diagnosis of fever and respiratory symptoms can include numerous infections, such as bacterial pneumonia, Pneumocystis pneumonia, and tuberculosis [24, 25••]. Influenza diagnostic testing in patients with ILI can help guide clinical decision-making and treatment decisions in HIV-infected patients. Diagnostic tests available for influenza include rapid influenza diagnostic (antigen) tests (RIDTs), immunofluorescence assays (direct [DFA] or indirect [IFA] antibody staining), reverse transcription-polymerase chain reaction (RT-PCR) assays, rapid cell culture, and conventional viral cell culture [14••, 25••, 26]; the impact of HIV infection on diagnostic yield is not known. Respiratory tract specimens should be obtained as close to illness onset as possible (ideally within 72 h). Preferred respiratory samples for influenza testing include nasopharyngeal or nasal swab, and nasal wash or aspirate [27]; the types of specimens acceptable for use (ie, throat, nasopharyngeal, or nasal aspirates, swabs, or washes) vary by which type of test is used [28]. Upper and lower respiratory tract samples (eg, endotracheal aspirate and bronchoalveolar lavage specimens) should be obtained from patients undergoing mechanical ventilation [25••].

RT-PCR is the most accurate and sensitive test for detecting influenza viruses, including the pH1N1 virus [14••]. Immunofluorescence is often available at hospital laboratories and has moderately high sensitivity and high specificity, but requires an adequate specimen with endothelial cells present, a fluorescence microscope, and a trained clinical laboratory scientist [29]. Commercial RIDTs can detect influenza virus antigens within 15 min of testing and are widely available in outpatient settings [30]. RIDTs have high specificity (>90%) but low to moderate sensitivity (20%–70%) for pH1N1 virus compared with other assays [31••, 32, 33•, 34]. Because of the limited sensitivity, a negative RIDT result should not be used for decision-making regarding influenza treatment. Among HIV-infected patients with RT-PCR–confirmed pH1N1 virus infection, only 47% had a positive RIDT result, which was comparable to the 52% yield observed in non–HIV-infected adults [21•]. A case report of an HIV-infected woman (CD4 = 166 cells/μL) with RT-PCR–confirmed pH1N1 virus infection illustrated the limitations of RIDTs [35•]. This patient was suspected to have pH1N1, but an initial RIDT, immunofluorescence test, and viral culture were all negative, and antiviral treatment was delayed for several days. Influenza diagnosis was later confirmed by RT-PCR, but the patient died. Although a positive RIDT result is reliable when community influenza activity is high, a negative RIDT does not rule out influenza virus infection. If a definitive diagnosis is needed, providers should consider confirming negative test results with more sensitive and specific influenza testing. Updated information concerning influenza diagnostic testing is available at http://www.cdc.gov/flu/professionals/diagnosis/.

HIV Testing in Persons with Influenza Illness

In the United States, 21% of HIV-infected persons are unaware of their diagnosis (0.09% of the population) [36]; thus, the Centers for Disease Control and Prevention (CDC) recommends routine HIV screening for all patients aged 13–64 years in health care settings [37••]. Hospital admission for any illness, including influenza, presents an important opportunity for HIV testing. Given the high prevalence of HIV infection among hospitalized patients with pH1N1, HIV screening should be performed in patients aged 13–64 years admitted with influenza or any other condition who have not had a previous HIV test or who have new risk factors for HIV infection since their last test. (In all health care settings, screening for HIV infection should be performed routinely for all patients aged 13–64 years. Health care providers should initiate screening unless a prevalence of undiagnosed HIV infection in their patients has been documented to be <0.1%. In the absence of existing data for HIV prevalence, health care providers should initiate voluntary HIV screening until they establish that the diagnostic yield is <1 per 1,000 patients screened, at which point such screening is no longer warranted [37••].)

Acute HIV infection often presents with “flu-like” symptoms, and case reports have documented patients with acute HIV infection who were misdiagnosed with pH1N1 [38] or had dual infection with acute HIV and pH1N1 virus [39]. An estimated 40%–90% of patients with acute HIV infection will experience symptoms [40] and a substantial number will seek medical care [41]. However, acute HIV infection often is not recognized by clinicians because the symptoms (including fever, fatigue, pharyngitis, lymphadenopathy, and rash) resemble those of influenza, infectious mononucleosis, and other viral illnesses [41]. Acute HIV infection should be considered in patients with “flu-like” symptoms and can be diagnosed by detecting HIV RNA in plasma from persons with a negative or indeterminate HIV antibody test [37••].

Severity of pH1N1 Illness in HIV-Infected Adults

Previous studies of seasonal influenza have noted higher hospitalization rates [42, 43] and increased mortality [44] among persons living with HIV; however, these studies were limited mostly to persons with low CD4 cell counts or AIDS. Several recent studies have reported the frequency of HIV infection among hospitalized adults with pH1N1 (Table 2). In the United States, three separate case series [21•] of hospitalized patients with pH1N1 were reported from California [45], New York City [46], and nationally (excluding California and New York City) [47]. These studies detected a higher prevalence (3.4%) of HIV infection among hospitalized patients with pH1N1 compared with the prevalence of HIV in the general US adult population (0.45% [37••]). Similarly, in a study evaluating critically ill patients with pH1N1 in Canada, 2 (2%) of 118 adults had HIV infection [48], although the HIV prevalence in Canada is estimated to be 0.4% [11]. A study from 13 hospitals in Spain identified HIV infection in 59 (10%) of 589 adults hospitalized with confirmed pH1N1 virus infection [22•], although the HIV prevalence in Spain is estimated to be 0.4% [11]. Conversely, a French study did not detect a higher prevalence of HIV infection among patients with severe pH1N1; in France, where HIV prevalence is 0.4%, only 2 (0.8%) of 244 adults with laboratory-confirmed pH1N1 virus infection requiring intensive care unit (ICU) admission had HIV infection [49].

Table 2 Reported rates of hospitalization, ICU admission, mechanical ventilation, and death among HIV-infected persons with pH1N1 virus infection compared with HIV-uninfected persons
Full size table

Recent studies have also compared the severity of laboratory-confirmed pH1N1 virus infection among HIV-infected persons with HIV-uninfected persons and have demonstrated similar clinical outcomes between these groups (Table 2). In the previously mentioned US case series, while a substantial proportion of HIV-infected patients hospitalized with pH1N1 were admitted to the ICU (29%), required mechanical ventilation (21%), and died (13%), these clinical outcomes were comparable to the experience of HIV-uninfected hospitalized adults. A report from Mexico City evaluated patients who presented to an emergency room with laboratory-confirmed pH1N1 virus infection and observed similar rates of hospitalization (27% vs 25%, P = 0.81), severe disease (9% vs 7%, P = 0.68), and death (9% vs 5%, P = 0.86) in HIV-infected persons compared with HIV-uninfected persons [15]. Of note, 73% of patients had a CD4 cell count >350 cells/μL, and 82% were receiving HAART. In the above-described study from Barcelona, although HIV-infected patients were over-represented among patients seeking care with confirmed pH1N1 virus infection, they were also not more likely to be hospitalized (23% vs 24%, P = 1.0), have longer hospitalizations (mean 1.1 vs 2.0 days, P = 0.08), have complications (13% vs 11%, P = 0.71), or die (0% vs 2%, P = 0.74) compared with HIV-uninfected patients. In this study, 95% of HIV-infected patients were virologically suppressed (HIV viral load <50 copies/mL) and 91% of HIV-infected patients had a CD4 cell count >200 cells/μL [16]. In another study of hospitalized Spanish adults with confirmed pH1N1 virus infection, the 26 HIV-infected patients did not differ from the 559 HIV-uninfected patients in terms of frequency of ICU admission (12% for both), rate of mechanical ventilation (9% vs 8%), or duration of hospital stay (5.9 vs 6.9 days) [22•]. Finally, a report from Singapore also demonstrated that clinical outcomes were similar among 11 HIV-infected patients with confirmed pH1N1 virus infection (64% with CD4 cell counts >200 cells/μL, 82% on HAART) and 66 HIV-uninfected controls matched on age and hospitalization status [19].

Two studies provide evidence that advanced HIV infection may result in increased pH1N1 severity as has been described with seasonal influenza. In the previously mentioned study from a tertiary referral center in Mexico City, high proportions of hospitalization (53%), mechanical ventilation (23%), and death (20%) were reported among the 30 HIV-infected patients with laboratory-confirmed pH1N1 virus infection (median CD4 cell count 163 cells/μL, 50% receiving HAART, 40% with concomitant opportunistic infections [OIs]) [17•]. Not being on HAART was associated with longer duration of hospitalization (P = 0.0006) and use of mechanical ventilation (P = 0.001). Among the six deaths in this study, the median CD4 cell count was 60 cells/μL, five were co-infected with Pneumocystis jirovecii pneumonia, and only one was receiving HAART. Patients with pH1N1 virus infection and OIs were more likely to be hospitalized (P < 0.0001), need mechanical ventilation (P = 0.0086), and die (P = 0.03) compared with patients without OIs, suggesting that influenza severity was associated with immunosuppression from advanced HIV disease and was exacerbated by concomitant OIs. However, patients with OIs also had a longer time of illness before hospitalization (P = 0.0003) and initiation of oseltamivir (P = 0.004), which may have influenced disease severity. In the second study, 24 HIV-infected patients in France who were classified has having severe influenza illness had lower nadir and current CD4 cell counts than all HIV-infected persons with ILI [13].

There are mixed reports about whether fatal pH1N1 occurs with higher frequency in the context of HIV infection. Reports from the United States and South Africa have described the distribution of HIV infection among fatal cases of confirmed pH1N1 virus infection. In a study from New York City, where HIV prevalence is approximately 1.4% [50], 7 (14.9%) of 47 confirmed fatal pH1N1 virus infections occurred in HIV-infected persons [51]. However, another study of fatal pH1N1 virus infections in the United States (excluding data from New York City) observed that only 7 (2.2%) of 324 confirmed fatal pH1N1 virus infections occurred in HIV-infected persons (Fowlkes, unpublished data) [52]. A report of 91 patients who died of confirmed pH1N1 virus infection from South Africa (where HIV prevalence is approximately 18% [11]) observed that 17 (53%) of 32 patients with a known HIV status were HIV-infected [53•]. Tuberculosis co-infection was also reported in 10% of patients who died of influenza, and nearly two thirds of HIV-infected patients who died were also pregnant. Data on CD4 cell counts and the use of HAART were not provided for these HIV-infected patients, and the relative contribution of pregnancy, tuberculosis, and immunosuppression from HIV infection to these deaths is difficult to discern. The aforementioned studies from South Africa and Mexico raise concern that HIV infection increases influenza severity and risk of death, and that during influenza pandemics, countries with a high HIV prevalence (especially if antiretroviral treatment coverage is low) could experience substantial excess influenza-related mortality.

In summary, several studies suggest that HIV-infected persons with pH1N1 virus infection may be at increased risk of hospitalization, although the potential bias to test and hospitalize HIV-infected persons with symptoms suggestive of pH1N1 virus infection may have increased their representation in these studies. Among HIV-infected persons hospitalized, symptoms, severity, and clinical outcomes appear to be similar to HIV-uninfected persons, especially in settings where a majority of patients are receiving antiretroviral therapy and are not severely immunosuppressed. Studies including a high frequency of patients with advanced HIV disease, however, suggest that influenza severity may be increased in these settings. Finally, the contribution of high-risk comorbidities (which are becoming more frequent in the HAART era) to influenza severity in HIV-infected persons has not been well defined.

Influenza Treatment in HIV-Infected Adults

Empiric influenza antiviral treatment for HIV-infected patients with suspected influenza is recommended [14••]. Based on antiviral resistance profiles for circulating influenza A and B viruses in the United States, as of April 2011, oseltamivir and zanamivir are recommended; updates on antiviral resistance among circulating influenza viruses in the United States are available at http://www.cdc.gov/flu/weekly. There is no evidence that oseltamivir or zanamivir interact with HIV antiretroviral medications, though data are limited.

Sporadic oseltamivir-resistant pH1N1 virus infections have been identified, including rare episodes of transmission [54], but the public health impact has been limited to date; infection with antiviral-resistant virus should be considered in patients who do not respond to treatment. Oseltamivir resistance has been reported within 10 days of treatment initiation in an HIV-infected patient; this patient did not have severe illness [55]. In the United States, 76% of 37 patients with oseltamivir-resistant pH1N1 virus infection had severe immunosuppression (including AIDS) compared with 11% of 65 matched patients with oseltamivir-susceptible pH1N1 virus infection [56]. Most patients with oseltamivir-resistant pH1N1 virus infection, however, had hematologic malignancies requiring chemo- or immunosuppressive therapy; only one of these 28 patients was HIV-infected (diagnosed with AIDS), and this patient also had lymphoma. Although the impact of antiviral-resistant influenza viruses has been limited to date, infection control measures, including reminding oseltamivir-treated persons to continue hand hygiene during treatment, are critical to prevent the further transmission of oseltamivir-resistant viruses.

The benefits of influenza antiviral therapy are greatest when initiated early, but treatment should still be started for patients who present >48 h after the onset of symptoms and require hospitalization, are at high risk for influenza complications, or present with debilitating illness [57]. Randomized, controlled efficacy trials of oseltamivir or zanamivir, conducted primarily among outpatients with mild illness, demonstrated a reduction of illness duration by approximately 1 day when administered within 48 h of illness onset compared with placebo [23, 58–61]. Observational studies have also suggested that oseltamivir reduces severe clinical outcomes in patients hospitalized with influenza [62]. Additionally, earlier neuraminidase inhibitor treatment was associated with decreased severity and improved survival in observational studies of patients hospitalized with pH1N1 [47, 51, 63]. Few HIV-infected patients were represented in the randomized clinical trials and observational studies, but among HIV-infected patients with pH1N1 in Mexico City, delayed administration of oseltamivir in patients with concomitant OIs was significantly associated with mortality (P = 0.002) [17•].

Since severe [35•] and persistent [55] pH1N1 virus infections have been reported in persons with HIV infection, in certain clinical situations (ie, severe influenza illness), treatment regimens may need to be altered. Extending treatment regimens longer than 5 days for patients whose illness is prolonged, and doubling the dose of oseltamivir (eg, 150 mg twice daily in adults) for patients with severe illness, has been advocated as an appropriate strategy in certain situations [14••, 25••, 57].

Influenza Viral Shedding in HIV-Infected Adults

Knowledge about viral shedding in HIV-infected patients compared with HIV-uninfected patients may inform infection control measures, interpretation of diagnostic tests, and influenza prevention and treatment guidelines; however, comprehensive data on shedding of circulating influenza viruses, either seasonal or pH1N1, among HIV-infected adults are not available. In healthy adults, the median time to cessation of viral shedding has been reported to be 4.5 days [64]; early antiviral therapy may decrease the duration of shedding [65]. Case reports in persons with hematologic malignancies have described prolonged pH1N1 virus shedding for weeks to months, often in association with antiviral resistance [66, 67], which has potential implications for nosocomial transmission and emergence and spread of antiviral resistance.

One study predating the pH1N1 pandemic described shedding after administration of live attenuated influenza vaccine and included 57 HIV-infected (all with CD4 cell counts >200 cells/μL and HIV viral loads <10,000 copies/mL) and 54 HIV-uninfected adults [68]. Culturable virus (an influenza B strain) was recovered from one adult, an HIV-infected participant, 5 days after vaccination. Among the cases described in one of the Mexico City series of pH1N1 virus infections (median CD4 cell count 163 cells/μL), shedding (ie, RT-PCR positive nasopharyngeal swab) was detected among two (20%) of 10 HIV-infected adults tested 6 and 7 days after beginning oseltamivir treatment [17•]. In another report, an HIV-infected patient (CD4 cell count 168 cells/μL) with pH1N1 was found to have influenza viral shedding 9 days after initial presentation after 6 days of oseltamivir treatment; a neuraminidase inhibitor resistance mutation was detected, but subsequent testing did not demonstrate viral shedding, and the patient’s clinical symptoms resolved [55]. Finally, in a pH1N1 outbreak among 15 HIV-infected children in Germany (all with CD4 cell count >350 cells/μL), viral shedding lasted a median of 9 (range 5–14) and 6 (range 3–11) days as measured by RT-PCR and culture, respectively, similar to the general population; oseltamivir treatment was significantly associated with reduced duration of shedding [69•].

Postexposure Chemoprophylaxis and pH1N1 Vaccine Among HIV-Infected Adults

Oseltamivir and zanamivir postexposure chemoprophylaxis are approximately 70%–90% effective at preventing influenza illness among household and close contacts of persons with laboratory-confirmed influenza in randomized trials among healthy adults [65, 70–72]. Postexposure antiviral chemoprophylaxis can be considered in HIV-infected patients who have not been vaccinated against influenza and who have a recent close contact with a person with suspected or confirmed influenza, depending on the severity of immunosuppression and the type and duration of contact. Chemoprophylaxis should only be used if antivirals can be started within 48 h of influenza exposure [14••]. Patients receiving chemoprophylaxis should seek medical advice if they develop a febrile respiratory illness to evaluate for infection with an antiviral-resistant influenza virus [66]. As an alternative to chemoprophylaxis, HIV-infected patients who are close contacts of persons with influenza may instead be counseled about the early symptoms of influenza and advised to contact their health care provider for possible early antiviral treatment if they develop these symptoms [14••].

Yearly influenza vaccination for adults infected with HIV is recommended by the CDC and the Infectious Disease Society of America and should be provided at the time of chemoprophylaxis [73]. Furthermore, HIV-infected persons hospitalized for any reason should be assessed regarding their influenza vaccination status and offered influenza vaccine if indicated prior to discharge. A recent meta-analysis reviewed the clinical efficacy of seasonal influenza vaccine in HIV-infected persons and found the pooled relative risk reduction in prevention of influenza by vaccination from the three studies that met the criteria for inclusion was 65% (95% CI 36%–82%); however, the single randomized controlled trial included in this analysis demonstrated a lower relative risk reduction of 41% (95% CI 2%–55%) [74••]. A subsequent randomized, double-blind, placebo-controlled trial of trivalent, inactivated influenza vaccine in HIV-infected adults (median CD4 cell count >350 cells/μL in vaccine and placebo groups) conducted in Johannesburg, South Africa in 2008 demonstrated an efficacy of 75.5% (95% CI 9.2%–95.6%) against confirmed influenza illness [75•].

Concerns were raised in the past about the potential for influenza vaccine to increase HIV viral load and worsen HIV disease progression. Two recent vaccine studies (including one with pH1N1 vaccine) did not demonstrate a vaccine effect on viral load, virologic suppression, or CD4 cell count [75•, 76•]. Questions have also arisen regarding the ability HIV-infected persons to mount an adequate antibody response to influenza vaccine. Several studies have shown lower antibody responses in HIV-infected persons compared with HIV-uninfected persons after seasonal influenza vaccination, with some suggesting that persons with low CD4 cell counts or AIDS have the most limited antibody response to influenza vaccination [77–81]. However, a few HAART-era studies conducted before the pH1N1 pandemic suggest that immune response may be related to HIV viral load rather than CD4 cell count [82–84]. Studies examining protection from clinical influenza post-vaccination, rather than antibody response, suggest that influenza vaccination of HIV-infected adults may be effective despite variable antibody responses [74••, 85], but data on the clinical effectiveness of influenza vaccination in HIV-infected persons are limited.

There are conflicting reports of the immunogenicity and efficacy of the pH1N1 vaccine among HIV-infected persons, though most studies suggest adequate immune response in patients without advanced disease. One prospective, observational study found seroconversion (defined as HIA titers ≥40 or ≥4 fold higher after vaccination) in 69% of 160 HIV-infected persons (mean CD4 cell count 523 cells/μL, 90% receiving HAART) after one dose of AS03-adjuvanted influenza A/California/7/2009 (H1N1) vaccine, which was lower than the 79%–98% seroconversion rate reported for the general population [86]. In this study, seroconversion was associated with younger age and higher CD4 cell count. After a second vaccine dose, seroconversion increased to 92% in the 135 HIV-infected patients who received both doses; seroconversion after two doses was associated with a higher nadir CD4 cell count and shorter duration of HIV but was not affected by current CD4 cell count, HIV viral load, or HAART status [87]. Similarly, another prospective, observational study of a single dose of non-adjuvanted monovalent influenza A/California/7/2009 (H1N1) vaccine found that, despite a median CD4 cell count of 581 cells/μL and 82% receiving HAART, HIV-infected persons were significantly less likely to achieve a protective antibody titer (HIA titers ≥40) compared with HIV-uninfected persons (56% vs 80%, P = 0.003); CD4 cell count, HIV viral load, and receipt of HAART, however, were not associated with vaccine response [76•]. Adjuvanted vaccines have demonstrated improved immunogenicity in HIV-infected persons compared with non-adjuvanted vaccines [88] but are not available in the United States.

Conversely, two Italian studies have documented adequate antibody responses to a single dose of MF59-adjuvanted influenza A/California/7/2009 (H1N1) vaccine. The first study reported 83% of HIV-infected persons achieved seroconversion, which was not associated with CD4 cell count, HIV viral load, or receipt of HAART [89]. The second study reported that seroconversion rates were lower in HIV-infected persons compared with HIV-uninfected persons, but this was largely due to high pre-vaccination antibody titers; post-vaccination antibody titers were similar in the two groups with seroprotection rates (HIA titers ≥40) over 97% [9]. Furthermore, two other studies demonstrated similar seroprotection rates in HIV-infected (CD4 cell count >200 cells/μL and >500 cells/μL, respectively) and HIV-uninfected adults (80% vs 70% and 88% vs 93%, respectively) after a single dose of the same vaccine; a second vaccine dose only moderately improved immunogenicity [10•, 90]. Additional studies have demonstrated improved seroconversion rates after two doses of adjuvanted pH1N1 vaccine as compared with one dose, with improved seroconversion rates with double vaccine dose in patients with lower nadir CD4 cell counts [91, 92]. None of these studies have evaluated vaccine efficacy against clinical pH1N1 disease, but these data suggest that future work is needed to determine the appropriate dose and schedule of influenza vaccine in HIV-infected persons, particularly in those with advanced disease who may be at risk for inadequate immune response to standard vaccination.

Conclusions and Future Directions

In summary, pH1N1 virus infection susceptibility and severity may not be increased in HIV-infected adults without advanced HIV disease or high-risk comorbid conditions. While some studies suggest that HIV-infected persons with pH1N1 virus infection may be at increased risk of hospitalization, among HIV-infected persons hospitalized, clinical severity appears to be similar to HIV-uninfected persons in settings where most patients are not severely immunosuppressed. Studies including a high frequency of patients with advanced HIV disease or comorbid conditions, however, suggest that influenza severity may be increased in these settings.

Respiratory symptoms in an HIV-infected patient during the influenza season or a pandemic, particularly in the setting of advanced HIV disease, can present diagnostic challenges for clinicians. RT-PCR is the most accurate and sensitive test for detecting the pH1N1 virus, although results may not be quickly available for clinical decision-making. RIDTs are also available, but a negative result does not rule out a diagnosis of influenza infection. Thus, empiric influenza antiviral treatment is recommended for HIV-infected patients with suspected influenza. Influenza antiviral therapy has the most benefit when started early, but treatment should still be started for hospitalized patients who present more than 48 h after the onset of symptoms. Hospital admission for influenza also presents an opportunity for HIV testing, and acute HIV infection should be considered in patients with “flu-like” symptoms in the appropriate clinical context. Hospital admission for HIV-infected persons also represents an opportunity to assess influenza vaccination status.

Prospective studies that evaluate hospitalization rates (not just outcomes after hospitalization) among HIV-infected persons are needed to assess the relative contribution of immunosuppression and high-risk comorbid conditions to influenza risk and severity in this population. Prevention and treatment of underlying high-risk comorbid conditions, including tobacco cessation counseling, in addition to annual influenza vaccination, should continue to be part of routine HIV clinical care to help prevent influenza illness and complications. Finally, additional information about pH1N1 vaccine immunogenicity and efficacy in HIV-infected patients, particularly in elderly patients and those with advanced HIV, would be useful in determining optimal vaccine formulations for HIV-infected persons.