Influenza A H1N1 in HIV-infected adults

Authors


  • *The data were presented in part at the 17th Conference on Retroviruses and Opportunistic Infections, San Francisco, California, 16–19 February 2010 (Abstract 802LB), XIV Congress of the Spanish Society for Infectious Diseases and Clinical Microbiology, Barcelona, 19–22 May 2010 (Abstract 118), and XIII National Congress on AIDS, Santiago de Compostela, 16–18 June 2010 (Abstract CO5.01).

Prof. Esteban Martínez, Infectious Diseases Unit, Hospital Clínic-Institut d'Investigaciones Biomèdiques August Pi i Sunyer, University of Barcelona, C/Villarroel 170, Barcelona 08036, Spain. Tel: +34 93 2275430; fax: +34 93 4514438; e-mail: esteban@fundsoriano.es

Abstract

Objectives

HIV-infected adults are considered to be at higher risk for influenza A H1N1 complications but data supporting this belief are lacking. We aimed to compare epidemiological data, clinical characteristics, and outcomes of influenza A H1N1 infection between HIV-infected and -uninfected adults.

Methods

From 26 April to 6 December 2009, each adult presenting with acute respiratory illness at the emergency department of our institution was considered for an influenza A H1N1 diagnosis by specific multiplex real-time polymerase chain reaction. For every HIV-infected adult diagnosed, three consecutive adults not known to be HIV-infected diagnosed in the same calendar week were randomly chosen as controls.

Results

Among 2106 adults tested, 623 (30%) had influenza A H1N1 infection confirmed. Fifty-six (9%) were HIV-positive and were compared with 168 HIV-negative controls. Relative to HIV-negative controls, HIV-positive patients were older, more frequently male, and more frequently smokers (P≤0.02). In the HIV-positive group, prior or current AIDS-defining events were reported for 30% of patients, 9% and 30% had CD4 counts of <200 and 200–500 cells/μL, respectively, and 95% had HIV-1 RNA <50 copies/mL. Pneumonia (9%vs. 25%, respectively, in the HIV-positive and HIV-negative groups; P=0.01) and respiratory failure (9%vs. 21%, respectively; P=0.04) were less common in the HIV-positive group. Oseltamivir (95%vs. 71% in the HIV-positive and HIV-negative groups, respectively; P=0.003) was administered more often in HIV-positive patients. Three patients (all HIV-negative) died. In the HIV-positive group, CD4 cell count and plasma HIV-1 RNA did not differ before and 4–6 weeks after influenza A H1N1 diagnosis (P>0.05).

Conclusions

HIV infection did not increase the severity of influenza A H1N1 infection, and influenza A H1N1 infection did not have a major effect on HIV infection.

Introduction

Influenza is a common cause of acute respiratory illness in HIV-infected adults [1,2]. Before the widespread use of effective combination therapy, small case series and anecdotal reports suggested that low CD4 cell counts or concomitant respiratory or cardiovascular comorbidities were associated with a higher risk for complications [3–8]. It is unclear to what extent effective antiretroviral therapy may have affected the risk for severe or complicated influenza, but HIV-infected patients are still considered to be at a higher risk [9] and for that reason they are preferentially targeted for influenza vaccination [10–12].

Human infections with a novel A H1N1 influenza virus were first identified in April 2009 [13,14] and they were increasingly reported throughout the world in the following weeks [15]. The rapid spread of the infection and the extensive reporting of associated deaths occupied the attention of the media and contributed to increased awareness in the general population [16,17]. Data from the beginning of the epidemics suggested that many influenza A H1N1 infections were not necessarily severe [18] and that HIV-infected patients were not overrepresented among those hospitalized or severely ill [14,19–21]. Nevertheless, health authorities considered that HIV-infected patients were at a higher risk for influenza A H1N1 complications, as they were for seasonal influenza, and this assumption remains unchanged [22–24].

With open access to combination antiretroviral therapy, many HIV-infected adults show sustained suppression of HIV replication in plasma, resulting in immunological and clinical benefits [25]. In Spain, health care for chronic conditions such as HIV infection and also for acute conditions and emergencies is provided free of charge by the public health care system [26]. From the beginning of the influenza A H1N1 epidemics, Spanish health authorities recommended persons considered to be at high risk of complications, including HIV-infected adults, to seek medical advice for early diagnosis and timely therapy as soon as any symptoms suggestive of influenza developed [24]. We hypothesized that HIV infection would not increase the severity of influenza A H1N1 infection, and that H1N1 influenza would not have a major impact on the control of HIV infection.

Methods

Patients

From 26 April to 6 December 2009, a specific protocol for adults (≥18 years old) presenting with any acute respiratory illness at our institution (Hospital Clinic, Barcelona, Spain) was established by the Hospital Clinic Influenza A H1N1 Committee in accordance with the recommendations of the Spanish Ministry of Health and the World Health Organization. The protocol comprised standardized clinical, chest X-ray and laboratory data collection, including oro- and nasopharyngeal swabs for influenza A H1N1. Chest X-ray was not routinely obtained in pregnant women. Respiratory, blood and urine samples were also obtained to confirm bacterial aetiology whenever bacterial infection was clinically suspected. The protocol was approved by the Ethics Committee of the Hospital Clinic and written informed consent was obtained from patients or their relatives. Because vaccination for influenza A H1N1 was not available in Spain until 16 November 2009, its impact on the results of this study can be considered negligible.

A patient was considered to have a delayed influenza A H1N1 diagnosis if he or she had undergone at least one previous medical visit because of current symptoms in which a diagnosis of influenza A H1N1 was not suspected. Pneumonia was defined as the presence of any new, not previously known lung consolidation on chest X-ray. Respiratory failure was defined as a partial pressure of oxygen (PaO2) <60 mmHg.

Whenever influenza A H1N1 infection was confirmed, specific therapy with oseltamivir was prescribed at the discretion of the attending physicians according to the recommendations of the Hospital Clinic Influenza A H1N1 Committee at the time of diagnosis, which were similar to those released by major health authorities [27–29]. In general, patients belonging to any group considered at high risk for complications (including HIV-infected patients), those presenting with more severe illness, and those diagnosed in the first months of the epidemics were more likely to receive oseltamivir. Antibacterial therapy was considered whenever a bacterial aetiology was suspected or confirmed and in patients with more severe infections. Specific complications developing during a hospital stay were identified and treated accordingly. Patients were followed during admission until discharge or death, and shortly after discharge to confirm clinical recovery.

Laboratory studies

Nucleic acids from any DNA/RNA viruses present in oro- and/or nasopharyngeal swabs were extracted from 200 μL of fresh specimen using NucliSense easyMAG (bioMérieux, Marcy l'Etoile, France) according to the manufacturer's instructions. Two specific one-step multiplex real-time polymerase chain reactions (RT-PCRs) were used for typing (A/B) and subtyping (H1/novel H1/H3/H5) of the influenza virus. The first assay consisted of primers and probes specific to the matrix (M1) gene of the influenza A and influenza B viruses for typing. The second assay employed primers and probes specific to the haemagglutinin (HA) gene of the human H1, novel human H1, human H3 and avian H5 subtypes in order to identify the most prominent subtypes capable of infecting humans (H1N1, pandemic H1N1, H3N2 and H5N1). Nontemplate controls and positive-template controls for all primer/probe sets were included in each run. An additional third assay amplified a housekeeping gene (RNase P) from host cells to check the progress of DNA extraction and to confirm the absence of PCR inhibitors as an internal control. The Centers for Disease Control and Prevention (CDC) Realtime RT-PCR Protocol for Detection and Characterization of Swine Influenza [30] supplied by the CDC (Atlanta, GA) was used to confirm positive results. The RT-PCR was carried out on Mx3000P or Mx3005P instruments (Stratagene, Agilent Technologies, Santa Clara, CA, USA). Blood cells (leucocytes, lymphocytes and platelets), chemistry [C-reactive protein (CRP), lactate dehydrogenase (LDH), creatin phosphokinase (CPK), creatinine and aspartate aminotransferase (AST)] and coagulation (Quick prothrombin time) were assessed using routine laboratory procedures at admission.

Design and statistical analysis

The study was designed as a prospective, observational, single-site, case series study with randomly selected controls. Participants included adults with a confirmed diagnosis of influenza A H1N1 infection irrespective of severity or any other indication for admission. For the purpose of the study, for each HIV-infected adult diagnosed with influenza A H1N1 infection, three consecutive adults not known to be HIV-infected diagnosed in the same calendar week were randomly chosen as unmatched controls. This study did not interfere with the clinical management of the patients. Epidemiological, clinical and outcome characteristics were prospectively collected and compared between the HIV-infected and HIV-uninfected groups. Because the presence and type of comorbidities were presumably different in HIV-positive and HIV-negative patients, and this could be a source of bias, we pre-planned a subanalysis considering only patients without comorbidities other than HIV infection.

For the HIV-infected group, data regarding probable route of HIV transmission, time from HIV diagnosis, CD4 cell count nadir, log10 HIV-1 RNA zenith, prior/current AIDS-defining events, hepatitis C virus coinfection, and most recent CD4, CD8 and log10 HIV-1 RNA measurements were collected. CD4 cell count, CD8 cell count and log10 HIV-1 RNA were also assessed 4–6 weeks after discharge. CD4 cell counts, CD8 cell counts and log10 HIV-1 RNA measurements prior to influenza diagnosis and 4–6 weeks after discharge were compared. Fisher's exact and Mann–Whitney U-tests were used to compare proportions and continuous variables, respectively. Risk factors for pneumonia and death were analysed by uni- and multivariate analysis (stepwise logistic regression). All statistical analyses were carried out using the spss software (version 15.0; SPSS Inc., Chicago, IL, USA).

Results

Incidence of influenza A H1N1

Among 2106 adults tested in the study period, 623 (30%) had influenza A H1N1 infection confirmed. Of these, 56 (9%) were HIV-positive. Figure 1 shows the number of patients tested for influenza A H1N1, the proportion of patients with a confirmed influenza A H1N1 diagnosis, and the number of HIV-negative and HIV-positive patients with confirmed influenza A H1N1 infection per calendar week. In both groups, there were two parallel peaks in late August and November.

Figure 1.

 (a) Number of patients tested for influenza A H1N1 per calendar week. (b) Percentage of patients positive for influenza A H1N1 per calendar week. (c) Number of HIV− and HIV+ patients positive for influenza A H1N1 per calendar week.

Characteristics of the patients

HIV-positive patients were older, more frequently male, and more frequently smokers compared with the HIV-negative controls (n=168) (Table 1a). As expected, the prevalence of comorbidities differed between HIV-positive and HIV-negative patients. Chronic lung diseases such as chronic obstructive pulmonary disease and asthma (5%vs. 26% in the HIV-positive and HIV-negative groups, respectively; P=0.0009) and pregnancy (0%vs. 11%, respectively; P=0.0232) were significantly less prevalent in the HIV-positive group than in the HIV-negative group (Table 1a).

Table 1.   Demographic data, comorbidities or concomitant conditions, and HIV-related characteristics for the patients included in the study
 HIV-positive
patients
(n=56)
HIV-negative
patients
(n=168)
P-value
(a)
Demographics
 Gender [n (%) male]44 (79)74 (44)<0.0001
 Age (years) (mean ± SD)44 ± 839 ± 150.0153
 Active smokers [n (%)]30 (54)21 (13)<0.0001
 Travel/contacts* [n (%)]4 (7)40 (24)0.0061
 Comorbidities or concomitant conditions [n (%)]8 (14)103 (61)<0.0001
Specific comorbidities or concomitant conditions
 Chronic lung disease [n (%)]3 (5)44 (26)0.0005
 Neoplasia [n (%)] (solid/blood)2 (4)12 (7)NS
(0/2)(4/8) 
 Pregnancy [n (%)]0 (0)18 (11)0.0082
 Drug-induced immunosuppression [n (%)]1 (2)12 (7)NS
 Diabetes mellitus [n (%)]1 (2)7 (4)NS
 Cirrhosis [n (%)]1 (2)2 (1)NS
 Abuse of toxic substances (drugs/alcohol) [n (%)]0 (0)2 (1)NS
 Chronic kidney disease [n (%)]0 (0)1 (1)NS
 Obesity (BMI >30 kg/m2) [n (%)]0 (0)3 (2)NS
 Ischaemic cardiovascular disease [n (%)]0 (0)2 (1)NS
  1. (a) Demographic characteristics and comorbidities in HIV-positive patients and HIV-negative controls. (b) Characteristics of HIV-positive patients related to their HIV infection.

  2. BMI, body mass index; IQR, interquartile range; NS, not significant; SD, standard deviation.

  3. *Known exposure to persons suffering a respiratory infection in the previous two weeks.

(b)
Route of transmission [n (%)]
 Men who have sex with men26 (46)
 Injecting drug use18 (32)
 Heterosexual contact12 (21)
Time from HIV-1 diagnosis (years) [median (IQR)]14 (5–19)
Nadir CD4 count (cells/μL) [median (IQR)]222 (134–379)
Nadir CD4 percentage [median (IQR)]18 (14–22)
Log10 HIV-1 RNA zenith (copies/mL) [median (IQR)]5.2 (4.7–5.5)
Prior/current AIDS-defining events [n (%)]16 (29)
Hepatitis C [n (%)]21 (38)
Most recent CD4, CD8 and HIV-1 RNA measurements at influenza A H1N1 diagnosis
CD4 count (cells/μL) [median (IQR)]583 (370–715)
CD4 percentage [median (IQR)]26 (23–33)
CD8 count (cells/μL) [median (IQR)]995 (828–1485)
CD8 percentage [median (IQR)]50 (41–60)
Log10 HIV-1 RNA (copies/mL) [median (IQR)]1.7 (1.7–1.7)
Patients with >50 copies/mL (n, log10 HIV-1 RNA copies/mL)3 (2.7, 3.5 and 4.3)
CD4 count categories [n (%)]
 <50 cells/μL1 (2)
 50–100 cells/μL1 (2)
 101–200 cells/μL3 (5)
 201–500 cells/μL17 (30)
 >500 cells/μL34 (61)

In the HIV-positive group, 16 patients (29%) experienced prior (n=15) or current (n=1; toxoplasma encephalitis under acute therapy) AIDS-defining events (Table 1b). Twenty-two HIV-positive patients (39%) had a CD4 count of either <200 cells/μL (n=5) or between 200 and 500 cells/μL (n=17), but 53 (95%) showed virological suppression in plasma within a period of 4 months preceding the diagnosis of influenza A H1N1 infection (Table 1b).

Clinical presentation

Among several symptoms assessed using the protocol, dysthermia, cough, arthromyalgias and fatigue were the most common, each being present in >50% of both the HIV-positive and HIV-negative patients (Table 2a). There were no significant differences between the groups in the symptoms assessed other than gastrointestinal symptoms, which included nausea, vomiting, abdominal discomfort and diarrhoea. Interestingly, gastrointestinal symptoms were significantly more common in HIV-positive patients (38%) than in HIV-negative patients (19%) (P=0.0035).

Table 2.   Clinical (a) and laboratory (b) data
 HIV-positive patients (n=56)HIV-negative patients (n=168)P-value
  • (a) Clinical characteristics of HIV-positive patients and HIV-negative controls presenting with influenza A H1N1 infection. (b) Blood cell and chemistry data and concomitant bacterial infections detected in HIV-positive patients and HIV-negative controls presenting with influenza A H1N1 infection.

  • Pneumonia was defined as a new, not previously detected X-ray lung infiltrate.

  • Respiratory failure was defined as partial pressure of oxygen (PaO2) <60 mmHg.

  • AST, aspartate aminotransferase; CPK, creatin phosphokinase; LDH, lactate dehydrogenase; SD, standard deviation.

  • *

    X-ray was not performed in 15 pregnant HIV-negative women.

  • Bacteria were detected in blood cultures and/or urine antigens and/or valid respiratory samples.

(a)
 Clinical symptoms assessed using the protocol
  Dysthermia [n (%)]56 (100)159 (95)0.1165
  Cough [n (%)]48 (86)145 (86)1.0000
  Arthromyalgias [n (%)]44 (79)128 (76)0.8553
  Fatigue [n (%)]38 (68)128 (76)0.2225
  Headache [n (%)]22 (39)78 (46)0.4379
  Sore throat [n (%)]21 (38)46 (27)0.1781
  Gastrointestinal symptoms [n (%)]21 (38)31 (19)0.0057
  Rhinorrhoea [n (%)]18 (32)57 (34)0.8710
  Expectoration [n (%)]23 (41)57 (34)0.3390
  Dyspnoea [n (%)]10 (18)36 (21)0.7030
 Time from onset (days) (mean ± SD)2.8 ± 1.63.2 ± 2.00.1359
 Axillar temperature (°C) (mean ± SD)37.9 ± 0.937.7 ± 1.00.1685
 Delayed influenza A H1N1 diagnosis [n (%)]4 (7)21 (13)0.3342
 Pneumonia [n (%)]5 (9)42 (27)*0.0045
 Respiratory failure [n (%)]5 (9)36 (21)0.0450
(b)
Leucocytes (cells/μL) (mean ± SD)6367 ± 23417602 ± 51450.0839
Lymphocytes (cells/μL) (mean ± SD)1511 ± 7191033 ± 4860.0001
Platelets (cells/μL) (mean ± SD)174 768 ± 61 494194 207 ± 71 2740.0691
C-reactive protein (mg/dL) (mean ± SD)3.4 ± 3.55.3 ± 5.90.0283
LDH (U/L) (mean ± SD)337 ± 109 (n=17)537 ± 370 (n=27)0.0364
CPK (U/L) (mean ± SD)153 ± 86 (n=17)268 ± 521 (n=15)0.3778
Creatinine (mg/dL) (mean ± SD)1.0 ± 0.31.0 ± 0.50.4478
AST (U/L) (mean ± SD)33 ± 21 (n=20)48 ± 39 (n=36)0.1253
Quick prothrombin time (%) (mean ± SD)89.3 ± 12.2 (n=27)84.4 ± 16.7 (n=90)0.1623
Concomitant bacterial infection detected [n (%)]4 (7)13 (8)1.0000
 Streptococcus pneumoniae39 
 Staphylococcus aureus04 
 Capnocytophaga spp.10 

HIV-infected patients had a shorter period from the onset of symptoms to hospital admission, but this difference was not significant. There were no significant differences in the proportion of patients with a delayed influenza A H1N1 diagnosis or in axillar temperature at admission. Interestingly, HIV-positive patients presented with pneumonia (9%vs. 27% for HIV-positive and HIV-negative patients, respectively; P=0.0045) and respiratory failure (9%vs. 21%, respectively; P=0.0450) less often than did HIV-negative patients (Table 2a).

Laboratory data

HIV-positive patients had higher lymphocyte counts and lower concentrations of plasma C-reactive protein than HIV-negative patients (Table 2b). There was also a trend towards lower leucocyte and platelet counts in HIV-positive patients relative to HIV-negative patients. The other laboratory parameters were assessed only in a fraction of patients. LDH was lower in HIV-positive patients, but the other laboratory parameters, namely CPK, creatinine, AST and Quick prothrombin time, did not differ significantly between the groups.

Roughly similar proportions of HIV-positive (7%) and HIV-negative (8%) patients had bacteria detected in valid respiratory samples and/or blood cultures and/or urine antigens at admission (Table 2b). Streptococcus pneumoniae was the most common bacterium, accounting for 12 (71%) of the 17 bacteria detected.

Prognosis

As expected, a substantial proportion of HIV-infected patients (95%; n=53) were treated with oseltamivir. This proportion was higher than that in HIV-negative patients (71%; n=119) (P=0.0003) (Table 3). However, roughly similar proportions of HIV-positive (52%; n=20) and HIV-negative (49%; n=82) patients received antibacterial therapy (P=0.6997). There was a trend towards a shorter duration of hospital stay (mean±standard deviation) in HIV-positive patients (1.1±2.3) than in HIV-negative patients (2.0±3.4) (P=0.0812), and fewer HIV-positive patients (n=15; 27%) were admitted for 1 day or longer compared with HIV-negative patients (n=70; 42%) (P=0.0564). Concordantly, a higher proportion of HIV-positive patients (77%; n=43) than HIV-negative patients (56%; n=94) showed clinical recovery in less than 1 week (P=0.0068). None (0%) of the HIV-positive patients died, but three (2%) of the HIV-negative patients died. Causes of death in each patient were multifactorial.

Table 3.   Prognosis in HIV-positive patients and HIV-negative controls presenting with influenza A H1N1 infection
 HIV-positive
patients
(n=56)
HIV-negative
patients
(n=168)
P-value
Duration of hospital stay (days) (mean ± SD)1.1 ± 2.32.0 ± 3.40.0812
≥1 day at hospital [n (%)]15 (27)70 (42)0.0564
Complications after admission [n (%)]7 (13)18 (11)0.8066
Anti-influenza therapy (oseltamivir) [n (%)]53 (95)119 (71)<0.0001
Antibiotic therapy [n (%)]29 (52)82 (49)0.6997
Clinical recovery <1 week [n (%)]43 (77)94 (56)0.0068
Evolution to death [n (%)]0 (0)3 (2)0.5750
Specific complications
diagnosed after admission
HIV-positive
patients
(n=56)
HIV-negative
patients
(n=168)
  1. SD, standard deviation.

Bacterial infection (n)44
Cardiovascular disease (n)21
Acute hepatitis (n)12
Hepatic encephalopathy (n)01
Deep venous thrombosis (n)01
Acute kidney failure (n)01
Diabetes ketoacidosis (n)02
Respiratory failure (n)03
Shock (n)01
Pleural effusion (n)02

Table 3 shows a list of specific complications in HIV-positive and HIV-negative patients identified after admission. Similar proportions of HIV-positive (13%; n=7) and HIV-negative (11%; n=18) patients developed intrahospital complications (P=0.8066). Interestingly, there were three patients (two HIV-positive and one HIV-negative) who developed myocarditis and/or ischaemic cardiovascular episodes, one of whom had no previous history of cardiovascular disease. There were also three patients with acute hepatitis (one HIV-positive and two HIV-negative); in two of these patients this was attributed to oseltamivir.

Clinical presentation and prognosis in patients without comorbidities

There were more HIV-positive (48 of 56; 86%) than HIV-negative (65 of 168; 39%) patients without comorbidities. When the two groups were compared, therapy with oseltamivir was found to be significantly more common, and there was a trend towards more frequent antibacterial therapy, in HIV-positive patients than in HIV-negative patients (Table 4). There were no significant differences between the groups in the proportion of patients with a delayed influenza A H1N1 diagnosis, pneumonia or respiratory failure. There were no differences either in the duration of hospital stay, clinical recovery, intrahospital complications and evolution to death. Nevertheless, all three patients who died belonged to the HIV-negative group without comorbidities.

Table 4.   Clinical presentation and prognosis in HIV-positive patients and HIV-negative controls without comorbidities other than HIV infection presenting with influenza A H1N1 infection
 HIV-positive patients (n=48)HIV-negative patients (n=65)P-value
  1. SD, standard deviation.

Delayed diagnosis [n (%)]4 (8)10 (15)0.3876
Pneumonia [n (%)]5 (10)14 (22)0.1349
Respiratory failure [n (%)]3 (6)6 (9)0.7308
Duration of hospital stay (days) (mean ± SD)1.0 ± 2.21.2 ± 3.40.7395
At least 1 day at hospital [n (%)]11 (23)16 (24)1.0000
Complications after admission [n (%)]7 (15)7 (11)0.5745
Anti-influenza therapy (oseltamivir) [n (%)]45 (94)26 (40)<0.0001
Antibiotic therapy [n (%)]24 (50)21 (32)0.0799
Clinical recovery <1 week [n (%)]37 (77)46 (71)0.5216
Evolution to death [n (%)]0 (0)3 (5)0.2604

Factors associated with pneumonia

Forty-seven patients presented with pneumonia; the proportion of patients with pneumonia in the HIV-positive group (9%; n=5) was lower than that in the HIV-negative group (28%; n=42) (P=0.0045). Factors associated with a higher risk of pneumonia at admission in the univariate analysis, other than being HIV-negative, were: older age (mean 45 years in those with pneumonia vs. 39 years in those without; P=0.0066), headache (31%vs. 13%, respectively; P=0.0009), tiredness (27%vs. 5%, respectively; P=0.0006), dyspnoea (35%vs. 17%, P=0.0099), longer time from the onset of symptoms to hospital admission (mean 5 vs. 2.6 days, respectively; P=0.0001), and delayed influenza A H1N1 diagnosis (56%vs. 17%, respectively; P=0.0001). In the multivariate analysis, being HIV-positive was not an independent risk factor for pneumonia at admission. We identified time from the onset of symptoms to hospital admission [odds ratio (OR) 1.82 per extra day; 95% confidence interval (CI) 1.50–2.22; P<0.0001] and tiredness (OR 4.40; 95% CI 1.19–16.23; P=0.0260) as independent factors associated with pneumonia at admission.

Among HIV-positive patients, those with pneumonia at admission were more commonly active smokers (100%vs. 49% for those with and without pneumonia, respectively; P=0.0545) and former/current injecting drug users (100%vs. 31%, respectively; P=0.0053), and more frequently had dyspnoea (60%vs. 14%, respectively; P=0.0351), respiratory failure (60%vs. 4%, respectively; P=0.0034), and concomitant bacterial infections (60%vs. 2%, respectively; P=0.0014) compared with those without pneumonia. Among HIV-positive patients, presenting with pneumonia was not associated with gender, comorbidities, travel/contacts, age, time from HIV diagnosis, CD4 cell count nadir, log10 HIV-1 RNA zenith, prior/current C events, delayed influenza A H1N1 diagnosis, time between the onset of symptoms and hospital admission, temperature at admission, or laboratory parameters, including most recent CD4 cell count, CD8 cell count and HIV-1 RNA measurement. Because of the low number of HIV-positive patients with pneumonia, multivariate analyses assessing independent risk factors could not be performed.

Changes in CD4 and CD8 cell counts and HIV-1 RNA

Most recent CD4 and CD8 cell ounts and HIV-1 RNA measurement prior to influenza A H1N1 diagnosis were available for all patients (n=56) within 4 months preceding influenza A H1N1 diagnosis (median 7 weeks; interquartile range 2–13 weeks). CD4 and CD8 cell counts and HIV-1 RNA were determined 4–6 weeks after discharge in 51 patients. Compared with values obtained before diagnosis, there were slight decreases in CD4 count (median −15 cells/μL; interquartile range −44 to 39 cells/μL), CD4 percentage (median −0.4%; interquartile range −0.8 to 2.3%), CD8 count (median −14 cells/μL; interquartile range −122 to 77 cells/μL) and CD8 percentage (median −0.7%; interquartile range −2.8 to 1.5%), but none of these changes was statistically significant (P>0.05 for all comparisons). Plasma HIV-1 RNA and the number of patients with plasma HIV-1 RNA below the detection limit remained unchanged.

Discussion

In this prospective study performed at a single referral hospital in Catalonia, we found that 9% of adults with confirmed influenza A H1N1 infection during the 2009 pandemics were HIV-positive. As the prevalence of HIV infection in adults in Catalonia is 0.6% [31], HIV-positive patients were overrepresented among those with confirmed influenza A H1N1 infection. The increased rate of diagnosis of influenza A H1N1 infection in HIV-positive adults relative to that in HIV-negative individuals might suggest that HIV-positive patients are more vulnerable to influenza A H1N1 infection than the general adult population, but the overall findings of our study, indicating that influenza A H1N1 infection in HIV-positive adults had a similar or even more benign presentation and prognosis than that in the general adult population, argue against that conclusion. Alternatively, this increased rate of diagnosis might have been a consequence of a higher proportion of HIV-positive patients relative to HIV-negative controls having a diagnosis of influenza A H1N1 infection confirmed. Because the health care of HIV-positive patients is already linked to the hospital, they are more likely than HIV-negative patients to go to hospital whenever they feel unwell, and this may be especially true for those without any underlying comorbidity or those with comorbidities not cared for at the hospital. This reasoning would explain not only the higher-than-expected representation of HIV-positive patients among those adults with confirmed influenza A H1N1 infection, but also the shorter time interval between the onset of symptoms and the diagnosis of influenza A H1N1 infection in HIV-positive patients.

Because the highest influenza A H1N1 rates have been reported in children and younger adults [13], we should have expected younger HIV-infected adults to be the individuals mainly affected. However, HIV-positive adults with confirmed influenza A H1N1 infection had representative features of the HIV-infected adult population receiving care at our institution, suggesting that influenza A H1N1 does not preferentially target a specific age group of HIV-infected adults. The clinical presentation was similar in HIV-positive and HIV-negative patients, except for gastrointestinal symptoms, which were more common in HIV-positive patients. It has been suggested that gastrointestinal symptoms occur more frequently in influenza A H1N1 infection than in seasonal influenza infection, especially in adults [32]. Gastrointestinal symptoms are a common problem in HIV-positive persons [33], and this might have contributed to the higher frequency of digestive symptoms seen in HIV-positive individuals with influenza A H1N1 infection.

In agreement with current expectations for HIV-positive adults on effective antiretroviral therapy [25], most HIV-positive patients with confirmed influenza A H1N1 infection in our cohort showed good virological control. However, 30% had CD4 counts between 200 and 500 cells/μL, and a further 9% had CD4 counts <200 cells/μL. Many of these patients with lower CD4 cell counts had experienced prior AIDS-defining events because of late diagnosis of HIV infection [34], and one of them was even receiving acute therapy for toxoplasma encephalitis at the time of the influenza A H1N1 diagnosis. Although some HIV-positive patients were profoundly immune-suppressed because of recent late presentation or incomplete CD4 cell recovery [35–38], we did not find that influenza A H1N1 caused illness preferentially in these patients; nor did these patients show more severe infection. Active smoking and former/current injecting drug use, but not the degree of immunosuppression, as indicated by CD4 cell counts, were risk factors for pneumonia in HIV-positive patients with influenza A H1N1 infection. Injecting drug use [39,40] and tobacco smoking [41] are widely recognized risk factors for bacterial pneumonia in HIV-infected patients. Of note, three (60%) of the five HIV-positive patients with confirmed influenza A H1N1 infection presenting with pneumonia had concomitant infection with Streptococcus pneumoniae detected.

We found that a longer time from the onset of symptoms to hospital admission was independently associated with a more severe presentation, such as pneumonia, in this group of patients as a whole. A later diagnosis may lead to a delay in the initiation of specific anti-influenza therapy. Investigators from Mexico City reported that, in a series of severely immune-suppressed HIV-infected adults who were receiving care for underlying opportunistic respiratory infections, including Pneumocystis pneumonia and tuberculosis, a concomitant diagnosis of influenza A H1N1 infection was not initially suspected [42]. Of 27 HIV-positive patients with influenza A H1N1 infection seen during the initial months of the Mexico City epidemic, 14 required hospitalization and six died. In addition, there are anecdotal reports of HIV-infected patients with severe or fatal influenza A H1N1 infection [43,44]. These cases indicate that influenza A H1N1 infection can be severe in already severely ill HIV-positive patients presenting with concomitant opportunistic infections; in these patients, the missed diagnosis of influenza H1N1 and the subsequent delay in the provision of specific anti-influenza therapy may be fatal. However, our study and others [45,46] suggest that HIV-positive adults are generally no more likely to experience severe complications of H1N1 influenza virus infection than adults not infected with HIV. Similar to seasonal influenza [47], we confirmed that influenza A H1N1 infection did not adversely affect surrogate markers such as CD4 and CD8 cell counts and HIV-1 RNA in plasma or HIV disease progression.

Our study had several limitations. HIV-negative controls were assumed to be HIV-uninfected, but we did not confirm this. Comorbidities in HIV-positive and HIV-negative patients were not similar either in prevalence or in spectrum, and this may have had an impact on clinical presentation and outcome. However, a subanalysis considering exclusively those patients without comorbidities other than HIV infection showed that being HIV-infected was not associated with a more severe presentation. As a result of specific recommendations, almost all HIV-positive patients received oseltamivir therapy compared with 71% of HIV-negative controls. This may have had an important effect on outcome in HIV-positive patients, but certainly not on the presentation of influenza A H1N1.

In summary, in a setting of universal access to antiretroviral therapy, which allowed successful control of HIV infection, and also to emergency health care, which allowed diagnosis of influenza A H1N1 and early initiation of anti-influenza therapy, HIV infection did not increase the severity of influenza A H1N1 infection and influenza A H1N1 infection did not have a major impact on HIV infection control. Because the immunogenicity reported to date for H1N1 vaccines in HIV-infected adults is poor [48,49], the findings of this study may be of value in the management of influenza A H1N1 infection in HIV-positive adults in settings similar to that described in this study.

Acknowledgements

Financial support was received from Red Temática Cooperativa de Investigación en SIDA (RIS G03/173), Ministerio de Ciencia e Innovación (Spain).

Ancillary