SEARCH

SEARCH BY CITATION

Keywords:

  • Bronchiolitis obliterans syndrome;
  • H1N1 subtype;
  • influenza A;
  • lung transplantation;
  • outcomes;
  • pandemic influenza

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

Influenza A H1N1 2009 led to 189 deaths during the Australian pandemic. Community-acquired respiratory viruses not only can cause prolonged allograft dysfunction in lung transplant recipients but have also been linked to bronchiolitis obliterans syndrome (BOS). We report the impact of the 2009 H1N1 pandemic on Australian lung transplant recipients. An observational study of confirmed H1N1 cases was conducted across five Australian lung transplant programs during the pandemic. An electronic database collected patient demographics, clinical presentation, management and outcomes up to a year follow-up. Twenty-four H1N1 cases (mean age 43 ± 14 years, eight females) were identified, incidence of 3%. Illness severity varied from upper respiratory tract symptoms only in 29% to lung allograft dysfunction (≥10% decline FEV1) in 75% to death in 5 (21%) cases (pre-existing BOS grade 3, n = 4). Treatment with oseltamivir occurred in all but one case confirmed after death, reduced immunosuppression, n = 1, augmented corticosteroid therapy, n = 16, and mechanical/noninvasive ventilation, n = 4. There was BOS grade decline within a year in six cases (32%). In conclusion, Australian lung transplant recipients were variably affected by the H1N1 pandemic mirroring the broader community with significant morbidity and mortality. After initial recovery, a considerable proportion of survivors have demonstrated BOS progression.


Abbreviations: 
CARV

community-acquired respiratory virus

FEV1

forced expiratory volume in 1 second

BOS

bronchiolitis obliterans syndrome

ICU

intensive care unit

ECMO

extracorporeal membrane oxygenation

ISHLT

International Society of Heart and Lung Transplantation

NPS

nasopharyngeal swab

PCR

polymerase chain reaction

RT-PCR

reverse transcriptase—polymerase chain reaction

URTI

upper respiratory tract infection

MRSA

methicillin-resistant staphylococcus aureus

ARDS

acute respiratory distress syndrome

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

The lung allograft is known to be particularly susceptible to community-acquired respiratory viral (CARV) infections leading to significant morbidity and mortality in lung transplant recipients (1–4). Extrapolated data from the 2006 Asian H5N1 influenza outbreak suggested a considerable impact could be expected from a pandemic on lung transplant programs (5).

Therefore, concern was understandable when Australia experienced its part of a global influenza pandemic during the winter months of 2009. Pandemic influenza A (H1N1) was a highly transmissible CARV associated with an initial impression of a severe infection (6). The first wave of the pandemic led to 189 deaths in the general population (7). There were unprecedented demands placed on Australian medical services. In particular, Australian intensive care units (ICUs) were overwhelmed by the number of patients that required extracorporeal membrane oxygenation (ECMO) who were generally young, pregnant or obese (8).

A recent large, predominantly North American cohort study has found that lung transplant recipients (33 out of 273 cases) were not affected preferentially more by pandemic H1N1 than other solid-organ transplant recipients yet there was considerable morbidity and mortality (9). A total of 70% of these recipients were admitted to hospital, 16% to ICU and there was a mortality rate of 4%. In another study from Israel, their single-center lung transplant program confirmed H1N1 infection in 5% of their recipient population, two cases requiring mechanical ventilation but no deaths occurred (10).

CARVs like influenza have also been linked to the onset of bronchiolitis obliterans syndrome (BOS), the main limiting factor for long-term survival after lung transplantation (11). Kumar et al. have reported a large case–control study showing BOS developing within a year in 6 of 50 (12%) lung transplant recipients with microbiologic-proven CARV infections compared with no established BOS in another 50 cases without CARV infection (12).

We report on the combined Australian experience with 2009 pandemic H1N1 focusing specifically on the immediate impact on our lung transplant recipient population and provide longer term clinical outcomes including changes in BOS grade.

Material and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

Patient identification

There are five adult lung transplant programs in Australia located at Sydney (in the state of New South Wales), Melbourne (Victoria), Brisbane (Queensland), Adelaide (South Australia) and Perth (Western Australia). Patients were included if they were lung or heart–lung transplant recipients with microbiologically confirmed H1N1 infection. Patients awaiting lung transplantation who became infected with pandemic H1N1 were excluded from this study. A single investigator at each transplant program was responsible for identifying appropriate patients for inclusion.

Data collection and statistical analysis

Data were collected from investigators via an electronic database that was set up in May 2009 (the onset of the H1N1 pandemic in Australia). The database contained patient demographic and public health information, clinical presentation, investigation and management of lung transplant recipients with pandemic H1N1. Clinical outcomes for these patients were updated at 1-year follow-up. This data were obtained by medical record review and subsequently transferred to The Prince Charles Hospital (Brisbane, Queensland, Australia) for collating and analysis. Our data analysis used descriptive statistics reported as means, standard deviations, ranges and counts. This observational study was a quality assurance project approved by each participating center. Patient confidentiality was maintained, as the database was deidentified. No funding was required for this study.

Definitions of clinical parameters

Lung allograft dysfunction in this study was defined as a fall in FEV1 equal or greater than 10% from the lung transplant recipient's baseline. Prolonged allograft dysfunction was defined as delayed recovery to baseline lung function over more than 2 weeks. Sore throat and rhinorrhea were considered typical symptoms of an upper respiratory tract infection (URTI). Fever was defined as a body temperature ≥37.6°C. BOS was defined by ISHLT criteria (13).

Nasopharyngeal swabs (NPSs)

NPSs were performed by medical and nursing staff across Australia in both the primary care and hospital settings. This technique involved a cotton swab being moistened in physiologic saline before being inserted into each nostril to access the posterior nasopharynx, then gently rotated to absorb secretions. A second swab was used for the tonsils and posterior oropharynx. Each swab was placed into 2 mL of viral transport medium (14).

Microbiologic analysis

Influenza A was identified by real-time RT-PCR of infected exfoliated nasopharyngeal cells and cell culture at the viral microbiologic reference laboratory located in each Australian state. Samples that tested positive in this assay were confirmed as positive or negative for the H1N1 subtype in a second validated real-time PCR assay incorporating primers and probe specific for the hemagglutinin gene of that virus (Applied Biosystems Pty Ltd., Australia) (15). Any positive samples were referred to the World Health Organization Collaborating Centre for Influenza Reference and Research where an attempt to culture an isolate was made.

Antiviral, immunosuppression and antibiotic therapy

Lung transplant recipients with suspected influenza during the pandemic were invariably commenced on oral oseltamivir (Tamiflu, Roche, Australia) at 75 mg twice daily until NPS results were known. Continuation of oseltamivir beyond 5 days, hospital admission, changes in immunosuppression and treatment with antibiotics were decided by the treating transplant physician. When augmentation of corticosteroid therapy was considered appropriate, prednisolone (or equivalent) was commenced at 1 mg/kg, to a maximum dose of 60 mg/day and weaned by 5 mg every second day to the previous maintenance level.

National pandemic plan

The response to the pandemic in Australia was implemented according to phases outlined in the 2008 Australian Health Management Plan for Pandemic Influenza (AHMPPI) (16). However, given the geographic and climatic differences across Australia, each individual state health service was responsible for ensuring the appropriate public health measures outlined in this document were carried out rather than a uniform approach. Additional measures were established at the discretion of the individual lung transplant program such as service reduction with cancellation of clinics and nonessential bronchoscopic procedures.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

There were 24 lung transplant recipients with microbiologically confirmed pandemic H1N1 during the study period. Table 1 shows patient demographics, transplant parameters and the baseline immunosuppression. The overall incidence of infection was calculated to be 3% based on a total of 826 lung transplant recipients across Australia. Figure 1 shows the geographical distribution of our cohort compared with mortality data from pandemic H1N1 in the general Australian population.

Table 1.  Baseline characteristics of Australian Lung transplant recipients with pandemic influenza A H1N1 2009
 Findings (n = 24)
Mean (range) age ±SD43 (19–66) years ±14
Male gender16 (67%)
Mean (range) time from transplant5 (0.4–12) years
Type of transplant
 Bilateral lung18 (75%)
 Heart-lung4 (17%)
 Single lung1 (4%)
 Redo bilateral lung1 (4%)
Native disease
 Cystic fibrosis9 (38%)
 COPD +/− alpha1-antitrypsin deficiency8 (33%)
 Congenital heart disease2 (8%)
 Interstitial lung disease2 (8%)
 Other3 (13%)
Pre-existing BOS grade
 011 (46%)
 0-p2 (8%)
 14 (17%)
 22 (8%)
 35 (21%)
Baseline irnmunosuppression
 Corticosteroids/Mycophenolate/Tacrolirnus9 (38%)
 Corticosteroids/Mycophenolate/Cyclosporine A7 (28%)
 Corticosteroids/Everolimus3 (13%)
 Corticosteroids/Azathioprine/Tacrolimus;3 (13%)
 Cortticosteroids/Azathioprine/Cyclosporine A1 (4%)
 Corticosteroids/Tacrolimus1 (4%)
image

Figure 1. Geographical distribution of pandemic H1N1 cases.

Download figure to PowerPoint

The mean onset of symptoms prior to commencing treatment was 5 days (range 1–15). Eighteen cases (75%) had lung allograft dysfunction with abnormal pulmonary infiltrates on chest radiograph and/or CT present in eight of these cases. The mean FEV1 decline at initial presentation was 16% (range 10–23%). Isolated URTI symptoms were experienced in six cases (29%). Other common symptoms included increased dyspnoea in 16/20 (80%), productive cough in 17/24 (71%), myalgias in 11/19 (58%), fever in 11/20 (55%) and hypoxia in 6/24 (25%).

All cases were diagnosed with PCR on NPS except two on bronchial washing. Sixteen of NPS were performed in the hospital setting and seven in primary care. Additional pathogens were isolated from bronchial washings or spontaneous sputum cultures in 12 (60%) of 20 cases. These were five cases with Pseudomonas aeruginosa, three cases with MRSA, two with Aspergillus spp. and one each with Scedosporium prolificans and Rhizopus spp. Transbronchial biopsies were performed in two cases, with one showing no abnormality but the other interestingly showing resolving bronchiolitis from pandemic H1N1.

Fifteen (63%) patients from our cohort were admitted to hospital with four (17%) of these requiring ICU. The mean hospitalization duration was 11 days (range 2–29). Treatment with oseltamivir was administered in 23/24 confirmed cases. One case was confirmed after death and no oseltamivir was administered ante-mortem. Table 2 outlines details of oseltamivir administration and other treatment that was required including immunosuppression changes. Recovery from pandemic H1N1 (with resolution of symptoms and/or return to baseline lung function) occurred within 2 weeks in 10 (42%) of 24 cases. However, prolonged allograft dysfunction occurred in nine (39%) cases and death was directly attributable to pandemic H1N1 in five (21%) cases. Table 3 shows the specific details of these deaths individually.

Table 2.  Management of pandemic influenza A H1N1 2009 in Australian lung transplant recipients
 Findings (n=24)
  1. *Necessitated by rhizopus spp. coinfection; mycophenolate ceased and tacrolimus/prednisolone reduced.

  2. **Further two cases required treatment for fungal pneumonia after H1N1 recovery.

Oseltamivir23 (96%)
 75 mg twice daily for 5 days10 (42%)
 75 mg twice daily for 10 days10 (42%)
 75 mg twice daily for 10 days followedby 150 mg twice daily for 5 days2 (8%)
 75 mg twice daily for 15 days1 (4%)
Immunosuppression
 Augmented corticosteroid therapy16 (67%)
 Unchanged only7 (29%)
 Reduced* only1 (4%)
Other management interventions
 Antibiotics administered concurrently19/20 (95%)
 Antifungal therapy administered concurrently**3 (13%)
 Noninvasive ventilation3 (13%)
 Mechanical ventilation1 (4%)
 Anti-CMV therapy administered concurrently1 (4%)
Table 3.  Lung transplant recipient deaths in Australia related to pandemic influenza A H1N1 2009
Patient no.Case detailsImaging findingsTime to death (days)BOS gradeState*
  1. *NSW (New South Wales) is Australia's most populous state (estimated 7 million).

  2. **SA (South Australia) is 5th most populous state (estimated 1.6 million).

140-year old female; 1-year postbilateral lung transplant for combined emphysema and pulmonary fibrosis; no growth in sputumBronchopneumonia63NSW*
256-year old male; 1.9 years postbilateral lung transplant for COPD; no growth in sputumPeribronchial ground-glass opacities73NSW*
342-year old male; 8.3 years postbilateral lung transplant for CF; MRSA in sputumBilateral bronchopneumonia260NSW*
419-year old male;2.7 years postheart/lung transplant for IPAH; Aspergillus in sputumBilateral bronchopneumonia243NSW*
560-year old female; 2 years postbilateral lung transplant for COPD; no growth in sputumBilateral bronchopneumonia33SA**

Public health information was obtained in 19 patients. There was one lung transplant recipient who had nosocomial-acquired H1N1 after she had initially been admitted for tracheo-bronchitis due to Rhizopus spp. Five patients reported household contacts with clinical features of influenza prior to their infection. Nine (38%) were in full-time employment which possibly increased their risk of pandemic H1N1 exposure.

Within a year, one further lung transplant recipient has died. This patient had made a complete recovery from 2009 pandemic H1N1 infection although there had been prolonged allograft dysfunction of 40 days. His death was related to renal failure rather than BOS. On the other hand, progression to BOS or worsening in BOS grade has been demonstrated in one-third of Australian lung transplant recipients with microbiologic-proven pandemic H1N1 (6 out of 18 cases). Table 4 provides further details of these BOS grade changes.

Table 4.  Details of BOS grade changes in Australian lung transplant recipients 1 year after pandemic influenza A H1N1 2009
 Findings (n=18)
Remained BOS 08
Decline BOS 0 to BOS 0-p1
Decline BOS 0-p to BOS 11
Decline BOS 0 to BOS 21
Remained BOS 12
Decline BOS 1 to BOS 21
Decline BOS 1 to BOS 31
Decline BOS 2 to BOS 31
Remained BOS 32

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

Influenza A H1N1 2009 has provided the first opportunity for modern lung transplant programs across the world to systematically study the impact of a pandemic on our recipients (9,10,17). Based on incidence and mortality estimates, this pandemic had a relatively benign effect on the Australian general population (37 127 confirmed cases with an overall incidence of 1.7%) when compared to the Spanish influenza pandemic of 1918 (6,7). Our study demonstrates that the incidence of pandemic H1N1 in Australian lung transplant recipients was slightly higher (3%) but not dramatically as one would expect given immunosuppression and lung allograft susceptibility to CARVs (5, 18). This difference is even less significant if we take into account that confirmation testing was abandoned for all suspected patients during the pandemic as it became apparent that Australian viral testing laboratories were being overwhelmed (6). Therefore, the incidence data for the general population are likely to be significantly underestimated.

Our overall incidence was less than that found by a single-center lung transplant program experience in Israel (10). A possible explanation for this difference was that the impact of pandemic H1N1 on lung transplant recipients varied according to geography and to the different timing of the peaks of the pandemic that occurred across Australia. New South Wales and Victoria have the two largest Australian lung transplant programs with outcomes consistent with international registries, yet there were significantly different experiences with H1N1 despite a similar timing of their pandemic peaks. More lung transplant recipients in New South Wales were affected and more severely as evidenced by deaths in a significant proportion. This data mirrors the mortality found in the general population in these respective states suggesting that Victoria was clearly not as adversely affected. By contrast, the general population in Queensland was severely affected by pandemic H1N1 with 41 deaths in a smaller population. Although the Queensland lung transplant recipient population is only half the size of New South Wales or Victoria, we were surprised that no H1N1-related deaths occurred. An explanation for this was that the peak of the pandemic occurred a month later and the Queensland lung transplant program had been able to prepare for the pandemic to some extent. This program replaced face-to-face clinics with phone reviews for recipients, nonessential services such as elective bronchoscopic procedures were cancelled and intensive patient education for early presentation was provided. This geographic variation in pandemic H1N1 incidence was not apparent in population-dense countries such as the United Kingdom (19).

Lung transplant recipients were affected variably along a spectrum from a mild self-limited URTI to prolonged allograft dysfunction to death. The causes for the prolonged allograft dysfunction included pneumonia, secondary bacterial infections and classical viral-induced bronchiolitis (as demonstrated on trans-bronchial biopsies in one case). Our mortality rate was significantly greater than previously published (21% vs. 4%) (9). This can be explained in that 80% of the deaths in our cohort had pre-existing BOS grade 3. This severe bronchiolar epithelial damage results in limited defense mechanisms to influenza (20). Bronchopneumonia on imaging was the finding in four of these deaths but peribronchial ground-glass opacities on CT were found in one other which generally preceded fatal H1N1 ARDS (21).

NPS was the first-line diagnostic sample used in Australia, chosen by public health officials because it could be obtained easily in primary care so that general practitioners could take the burden off the public hospital system. This is reflected in our data with 30% of NPS performed in primary care. However, there has been some concern about the accuracy of RT-PCR using NPS and the possibility of false-positives and negatives. The specific problem was NPS sampling error (inadequate epithelial cells obtained or suboptimal specimen processing) rather than RT-PCR accuracy with a recent study demonstrating no false-positive and one false-negative RT-PCR result obtained from 85 appropriately collected NPS samples (22, 23). The degree of sampling error is difficult to ascertain but there was no case identified in the Australian lung transplant recipient population who was RT-PCR H1N1 positive on bronchial washings and negative on a NPS sample. Although NPS screening of the entire Australian lung transplant recipient population would have answered this question, this would not have been a sensible approach as medical services were already stretched and would have increased exposure risk to our patients unnecessarily.

The dose and duration of antiviral treatment in each individual case was decided by the treating transplant physicians based on clinical response rather than according to American Society of Transplantation guidelines which were published after Australia's first pandemic wave (24). Oseltamivir was the predominant antiviral used in Australia. The frequent use of extended courses of oseltamivir in our cohort beyond 5 days was based on an awareness that immunosuppression generally led to prolonged viral shedding and previous Australian experience demonstrating prolonged lung allograft dysfunction with other CARVs have responded to longer courses of antiviral therapy (3,4,25). Some centers have suggested treating until NPS was negative on RT-PCR technique was the optimal approach but this was not adopted by Australian programs as the detection of a nucleic acid sequence does not always correspond to actively replicating virus and may persist beyond resolution of symptoms (26).

Augmented corticosteroid therapy was also employed in the majority of cases. Although controversial, corticosteroids have previously been shown to be effective in attenuating the inflammatory response to CARVs in vitro and in clinical studies as an adjunct to antiviral therapy (3,4,27). This inflammatory response associated with a secondary hemophagocytic syndrome has been implicated in the deaths occurring in the general population on postmortem studies (28). ECMO was not employed in any case of our cohort in dramatic contrast to the general population (n = 68) (8). This reflects the infrequency of H1N1 ARDS in Australian lung transplant recipients.

In the last decade, it has become apparent that CARVs are a significant risk factor for BOS after lung transplantation (11,12). Obliterative bronchiolitis has been described shortly after influenza pneumonia in both pediatric and adult lung transplant recipients (29,30). Early diagnosis and prompt treatment of CARVs has influenced this outcome on short-term follow-up (3,4). In our cohort, all pandemic H1N1 survivors were either at or had returned to their baseline lung function when BOS progression occurred. Therefore, if pandemic H1N1 were associated with BOS, the possible mechanism would be a virus induced upregulation of alloimmune cell reactivation. Our study suggests a significant association between pandemic H1N1 and BOS but is limited by the lack of a control group for comparison.

Our study has other limitations. Firstly, no data were obtained on those lung transplant recipients who were given prophylactic doses of oseltamivir for exposure to high-risk contacts. This included a large cohort attending the World Transplant Games in Australia during the pandemic. NPS was not required for prescription of prophylaxis. Prophylaxis was generally initiated in primary care and these prescriptions were not tracked on an accessible database. Secondly, although H1N1 vaccination was adopted by Australian lung transplant programs after the peak of the pandemic had occurred, its efficacy is still unknown and this issue was not addressed by this study.

In conclusion, to our knowledge this is the first collaborative national study to assess both the immediate and longer term impact of pandemic H1N1 on lung transplant recipients. Initial impression has been that the pandemic was benign and this would be true based on the incidence alone. However, we have demonstrated significant morbidity, with prolonged allograft dysfunction and long hospitalization times, and mortality, higher than previously published. Deaths occurred in two regions and the majority of these patients had pre-existing BOS grade 3. We have also demonstrated that the impact on lung transplant recipients varied according to geography and this mirrored where the broader community was severely affected. Outcomes appeared to be better in the lung transplant program that was able to plan for the pandemic peak by implementing a reduction in services and patient education. Our data have reiterated the possible link between influenza and BOS. Lessons learnt from the 2009 influenza A H1N1 pandemic have been invaluable for lung transplant services across Australia.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

The authors of this manuscript have received no funding for this project.

Disclosures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation. The manuscript was not prepared or funded by a commercial organization.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. Disclosures
  9. References