Pandemic 2009 Influenza A (H1N1) virus infection in cancer and hematopoietic stem cell transplant recipients; a multicenter observational study.

Background: During March 2009 a novel Influenza A virus emerged in Mexico. We describe the clinical picture of the pandemic Influenza A (H1N1) Influenza in cancer patients during the 2009 influenza season. Methods: Twelve centers participated in a multicenter retrospective observational study of cancer patients with confirmed infection with the 2009 H1N1 Influenza A virus (influenza-like illness or pneumonia plus positive PCR for the 2009 H1N1 Influenza A virus in respiratory secretions). Clinical data were obtained by retrospective chart review and analyzed. Results: From May to August 2009, data of 65 patients were collected. Median age was 51 years, 57 % of the patients were female. Most patients (47) had onco-hematological cancers and 18 had solid tumors. Cancer treatment mainly consisted of chemotherapy (46), or stem cell transplantation (SCT) (16). Only 19 of 64 patients had received the 2009 seasonal Influenza vaccine. Clinical presentation included pneumonia (43) and upper respiratory tract infection (22). Forty five of 58 ambulatory patients were admitted. Mechanical ventilation was required in 12 patients (18%). Treatment included oseltamivir monotherapy or in combination with amantadine for a median of 7 days. The global 30-day mortality rate was 18%. All 12 deaths were among the non-vaccinated patients. No deaths were observed among the 19 vaccinated patients. Oxygen saturation <96% at presentation was a predictor of mortality (OR 19.5; 95%CI: 2.28 to 165.9). Conclusions: In our cancer patient population, the pandemic 2009 Influenza A (H1N1) virus was associated with high incidence of pneumonia (66%), and 30-day mortality (18.5%). Saturation <96% was significantly associated with death. No deaths were observed among vaccinated patients.


Abstract
During Data from different studies on the impact of this new virus in the adult cancer and SCT population are somewhat contradictory. Many studies from different countries were reported 1,6-16 . In these studies, the incidence of pneumonia ranges from 20 11 to 52% 8 , while the reported mortality rate ranges from 0-10% 6,11,12,14,16 to as high as 21-31% 7-10,15 .
In this study, we examined the effects and severity of pandemic H1N1 Influenza during the 2009 influenza season, in patients with cancer and SCT in two cities of Argentina.

Methods
This is a multicenter retrospective observational study that included 12 medical centers. From May to August 2009, cancer and SCT patients older than 16 years, who presented a confirmed influenza infection by real-time PCR were included.
The following data were obtained anonymously: underlying illness, type and date of SCT, whether patients were or were not receiving immunosuppressive treatment, at the time of the influenza diseases, immunization for seasonal influenza, clinical presentation (influenza like or pneumonia), laboratory and radiology results, anti-viral treatment, and outcome. In addition, data on the time between the onset of symptoms and the initiation of antiviral therapy, need for ventilation support, and presence of co-infections were also collected. Hypoxemia was defined as an oxygen saturation value lower than 96%. We diagnosed lymphopenia when the absolute lymphocyte count was less than 1000/µL. Data on overall clinical presentation and outcome are shown in Table 1. Pneumonia and pneumonia with oxygen saturation <96% were the most common clinical presentations (43/65, 66% and 30/65, 46%, respectively). Co-infections were present in a minority of cases (9/65, 14%) and only among patients with community acquired Influenza.
Patients started treatment at a median of two days from onset of symptoms (range 0 to 45 days). Sixty eight percent (43/63) of patients started treatment within the 48h after the onset of symptoms. Some patients received combined antiviral treatment because of the potential circulation of seasonal Influenza A H1N1 known to be resistant to oseltamivir 17 .
Most patients acquired the infection in the community (58, 89%) while 7 (11%) of infections were acquired in the hospital setting despite the implementation of adequate standard precautions and

Amendments from Version 1
The data in Dataset 1 have been anonymised further as concerns were raised that the previous version included too many details ('indirect identifiers), which, once combined, might allow the identification of patients. In order to protect these individuals, Dataset 1 has also been replaced in Version 1 of this article. isolation measures during this outbreak. Detailed descriptions of the outcome of patients with community acquired (CAPIA) and nosocomially-acquired (NAPIA) pandemic Influenza A H1N1 infection patients are described in Figure 1 and Figure 2. The 30-day mortality was higher among patients with NAPIA (3/7, 43%) than among those with CAPIA (9/58, 15,5%).

REVISED
Most (45/58; 78%) of outpatients required hospital admission. Reasons for patient admission included mainly oxygen desaturation, but, in many cases, patients were admitted because of their severe state of immune suppression and the lack of information about this emergent virus, especially when the patient's social environment prevented him/her from easy access to medical care.
Outpatients who presented with upper respiratory tract symptoms (URTI), had the most benign course since the majority (11/19, 52%) resolved their infections with antiviral therapy in the outpatient setting, and, among the 8 (42%) who were admitted, none of them required ICU admission or developed signs of pneumonia. The 30-day mortality among CAPIA URTI was 0.
Patients, who developed NAPIA, belonged to 3 different centers and started having symptoms at median of 20 days after admission (range 2 to 33). This group had the poorest prognosis since the 30-day mortality rate was 43% (3/7).
One of three (33%) NAPIA URTI progressed to pneumonia, while none of the 19 patients with CAPIA URTI did. Therefore, the overall progression from URTI to pneumonia was of 4.5% (1/22).
The 30-day mortality according to the clinical presentation and setting is best described in Table 2 for comparison. It is shown that having pneumonia at presentation and developing of the infection in the hospital setting tended to be associated with a higher 30-day mortality without achieving statistical significance.
No deaths were observed among patients who had been vaccinated against seasonal influenza in the same year. We could not collect data on the date of vaccination. However, we do know that the first case was detected on May 12 th , while the seasonal influenza vaccine was available since March. Therefore, there is a high probability that at least 14 days might have passed between vaccination and the onset of symptoms.
The presence of any co-infection (bacterial or viral) at onset of symptoms and the delay in treatment were not associated to death or mechanical ventilation. By univariate analysis lack of history of vaccination, and the following baseline characteristics: pneumonia, oxygen saturation <96%, and lymphocyte count <800 cells/µL, were associated to 30-day mortality and mechanical ventilation. By multivariate analysis only lack of history of vaccination (OR did not apply because none died in the vaccinated group) and baseline oxygen saturation <96% (OR 19.5; 95% CI 2.28-165.9; P=0,007) were associated to mechanical ventilation and death. There might be a bias regarding the apparent benefit of vaccination because in cancer patients, immunization is usually advised when the period of major immunosuppression has finished.

Discussion
Our study shows the clinical course of the infection by the 2009 pandemic Influenza A H1N1 virus in 65 cancer patients from 12 institutions located in two cities of Argentina. Eleven percent of these infections were nosocomially acquired. Overall we found a high rate of pneumonia (66%) and mortality (18%). The clinical course was less severe in those who presented with an URTI in the outpatient setting in contrast to those who presented with pneumonia and desaturation especially in the hospital setting. We also found that the best predictors of death were oxygen desaturation at presentation and lack of vaccination against seasonal Influenza.
The 30-day mortality rate we show in this study is more than three times the one observed in Argentina in the same patient population when looking at infections by other respiratory viruses such as Adenovirus, Influenza, Parainfluenza, and RSV (5%) 3 . However, it is similar to the mortality rate reported in hematological patients with pneumonia by Influenza, Parainfluenza, Picornavirus and RSV at a US institution (15%) 24 .
It is well known that seasonal influenza-induced pneumonia is independently associated with mortality after HCT (adjusted HR 2.6; 95% CI 1.40-4.86) 20 . The pandemic Influenza A H1N1 virus is an independent risk factor for progression to LRTI (OR 5.64; 95% CI 1.3-25) and hypoxemia (OR 5.91; 95% 1.4-24) compared with seasonal influenza virus in HCT recipients 1 . In addition, immunosuppression was a main risk factor for early mortality among In contrast, patients with LRTI required hospitalization with a high number of them requiring admission to ICU for ventilation support. The dismal outcome seen in these patients despite treatment with oseltamivir probably indicates that this high-risk group needs to be treated differently from patients with isolated URTI. Some authors have suggested an initial treatment with high dose of oseltamivir and/or combination therapy approaches in the case of respiratory failure 22 . Higher doses could also be considered in a setting of profound lymphocytopenia 30 . All our patients received standard dose of oseltamivir (75 mg po twice a day) for a minimum of 10 days based on data on slower viral clearance 30 .
Nosocomial outbreaks of seasonal 32 and pandemic 2009 Influenza A H1N1 infection 33 can develop even in the setting of appropriate infection control measures. Seven (11%) of our patients had hospitalacquired influenza. Three of them (43%) died. This mortality rate is higher than the previously reported 13-27%, however, the number of patients in our report is too small to make any conclusion.
Seasonal influenza vaccination is recommended yearly for all patients with cancer and HSCT recipients 34 . In our study all deaths occurred among the non-vaccinated patients, while there were no deaths among the vaccinated patients. Individuals vaccinated against seasonal Influenza A or with previous seasonal influenza infection may benefit from preexisting cross-reactive memory CD4 + T cells and CD8 + T cells reducing their susceptibility to Influenza A H1N1 infection or explaining, at least in part, the unexpected mild illness in the community [35][36][37][38][39] . Whether the trivalent seasonal Influenza vaccine is protective against the pandemic Influenza A H1N1 virus in cancer patients is still a matter of debate 38 .
In conclusion, we report a series of cancer patients with the pandemic Influenza A H1N1 infection with a high incidence of hospitalization, severe pneumonia, ICU admission, mechanical ventilation, and 30-day mortality. In our series hypoxemia and lack of vaccination with seasonal trivalent influenza vaccine were predictors of mechanical ventilation requirement and death. A larger study is needed to evaluate the possibility of cross protection with the seasonal influenza vaccination. 337 Argentinean patients admitted to ICU with influenza like illness and respiratory failure that required mechanical ventilation. Mortality was highly associated with refractory hypoxemia 25 . These data explain our high mortality rate observed among the 17 patients who were admitted to ICU (11/17; 65%) or among the 12 patients who developed respiratory failure (11/12; 92%). These values are comparable to those reported in the same type of population 1,8,12,22 , but are higher than those reported in other populations, which ranged from 0-24% 21,26-29 . Indeed, the overall mortality rate observed among Argentinean patients admitted to ICU and requiring mechanical ventilation was 46% 25 . To further support the high mortality of patients with pandemic 2009 Influenza pneumonia, we identified hypoxemia at onset of symptoms as an independent predictor of mortality.

Data availability
Lymphocytopenia has been described as a risk factor for progression from upper to lower viral respiratory tract infection in cancer patients 24,30 , and profound lymphopenia (<100 cell/µL) was reported as a significant risk factor for requirement of mechanical ventilation and death in HCT recipients infected with seasonal influenza virus 30 . In our study, having fewer than 800 lymphocytes/µL at presentation was a predictor for the need for mechanical ventilation and death in a univariate but not in a multivariate analysis. We did not analyze a lower value such as <100 of lymphocytes due to the small number of patients included with this value.
It is noteworthy that co-infections or bacterial complications developed in less than 15% of patients.
Neuraminidase inhibitor therapy appears to be effective in preventing progression to LRTI 2,30 and hypoxemia 30 when instituted early after onset of symptoms. It was reported that delaying therapy in cancer patients with the pandemic Influenza A H1N1 virus infection was significantly associated with death 16 . Early initiation of antiviral therapy in these patients may attenuate the severity of disease 21,27 . In our series, antiviral therapy was started early after a median of two days after the onset of symptoms, with a range from 0-45 days. We did not find any correlation between days from onset of symptoms to therapy or diagnosis to therapy by univariate or multivariate analysis.
It is known that patients with URTI can be treated as outpatients and can recover completely from their infection 31 . In our series half of the outpatients with URTI remained as such, while the other half was admitted but did not require ICU. There is a possibility that most of the admitted patients could have been managed as outpatients as well.
The global progression from URTI to pneumonia in our study was of 4.5% (1/22). This single patient had a nosocomial infection and died 21 days later with sepsis, respiratory failure and neutropenia. This is according to reports of progression to LRTI that may occur even after one week of symptoms 20 .
1. This is an interesting article that shows the impact of pandemic 2009 influenza A (H1N1) in a highly vulnerable population for influenza complications such as immunocompromised oncohaematological patients.

Open Peer Review
The paper is well written and I would like to make some comments or questions to the authors: Did you observe any difference in time elapsed since the initiation of symptoms between those with community acquired pandemic 2009 influenza A (H1N1) who presented with upper respiratory tract infections and those who presented with pneumonia?
Previous vaccination against seasonal influenza showed that some degree of preexisting immunity to pandemic 2009 influenza A (H1N1) strains exists, especially among adults aged >60 years.
Have you found any difference in progression from UTRI to pneumonia or to death between those older or younger than 69 years of age?
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
No competing interests were disclosed. There was a trend towards a longer period with symptoms in patients who presented with pneumonia.
Question 2: Patients older than 60 years old who had been vaccinated against seasonal Influenza (N=10) had