The clinical course of 2009 H1N1 influenza in Allo-SCT patients is unknown. Data were collected in the UK from October 2009 to April 2010 on laboratory-confirmed cases of H1N1 influenza in Allo-SCT recipients. H1N1 infection was diagnosed in 60 patients, median age 42 years, at a median of 10 months post-SCT. Twenty-one patients (35%) developed pneumonia and nine (15%) required admission to intensive care units. Actuarial mortality was 7% at 28 days and 19% 4 months post-diagnosis of 2009 H1N1 influenza. Increasing age and pre-existing lung disease were risk factors for pneumonia (P=0.006 and 0.037, respectively); older age was a risk factor for death (P=0.012). Morbidity and mortality from 2009 H1N1 influenza in SCT patients exceeds that of immunocompetent patients, but parallels that in other critically ill hospitalised cohorts; the elderly and those with chronic pulmonary disease are at greatest risk.
2009 H1N1 influenza, a novel triple reassortant influenza virus, containing genes derived from swine, human and avian influenza A viruses, first emerged as an international public health problem in April 2009. The first human infections caused by the virus were reported in Mexico and rapidly spread to the southern United States and then globally, resulting in the World Health Organisation (WHO) raising the level of influenza pandemic alert from phase 5 to phase 6 on 11 June 2009.1 As of 1 August 2010, there have been at least 18 449 deaths worldwide.2 Many respiratory viral infections that are relatively innocuous in the immunocompetent can follow a more severe course in immunocompromised patients, resulting in progressive pneumonia and death.3, 4, 5 There are few reports of 2009 H1N1 influenza infection in immunocompromised patients or patients with haematological malignancies.6, 7, 8, 9 The objectives of this study were to describe the clinical features, treatment and outcome of 2009 H1N1 influenza infection in patients after Allo-SCT.
Design and methods
A retrospective multicentre cohort study of adult and paediatric Allo-SCT recipients infected with 2009 H1N1 influenza was undertaken between October 2009 and April 2010. Participating centres were enrolled through the British Society of Blood and Marrow Transplantation (BSBMT) Registry. The diagnosis of H1N1 infection was made locally by the transplant centre and defined as detection of 2009 H1N1 influenza by PCR in a respiratory sample (including throat swab, nasopharyngeal aspirate, sputum or bronchoalveolar lavage).10 Individual transplant centres reported all Allo-SCT recipients with laboratory-confirmed 2009 H1N1 influenza directly to the BSBMT. Data were collected locally by review of medical records in participating centres. A study-specific spreadsheet was used and investigators at each site were asked to submit demographic and clinical information on all adult and paediatric microbiologically confirmed cases of H1N1 infection. Spreadsheets were collected and analysed by the BSBMT Registry. Data were sought on patient demographics (including gender, age, underlying diagnosis, co-morbidities), transplant-specific information (conditioning, type of donor, stem cell source, current immunosuppressive therapy) and the clinical features, treatment and complications of 2009 H1N1 influenza infection.
We analysed potential risk factors associated with progression to pneumonia (diagnosed clinically or radiologically), intensive care unit (ICU) admission and mortality secondary to 2009 H1N1 influenza infection. Stata statistical software (StataCorp. 2009: Release 11) (StataCorp LP, College Station, TX, USA) was used. For univariate analyses, Fisher's exact test was used for binomial variables, and log-rank test was used for survival. For multivariate analyses, logistic regression was used for pneumonia and ICU admission, and Cox's regression was used for overall survival; age was analysed by two methods, in two cohorts (greater than or less than the median) and as a continuous variable.
Sixty cases of 2009 H1N1 influenza in Allo-SCT recipients were reported from 16 UK transplant centres. Reporting centres represented 63% of adult and 77% of paediatric Allo-SCT activity in the United Kingdom (Table 1). The reported frequency of 2009 H1N1 influenza infection in adult SCT recipients was higher than that observed in paediatric patients.
Table 2 shows the demographic data, including underlying diagnosis, type of transplant, immunosuppression and co-morbidities in addition to the clinical manifestations and complications of H1N1 infection. The median age at diagnosis of H1N1 influenza was 42 years (range 10 months to 68 years). Most patients (88%) were aged 18–65 years. AML and myelodysplasia were the most common underlying diagnoses (33%). Twenty-six patients (43%) had a matched sibling and 32 (53%) an unrelated donor (including three (5%) mismatched donors). Forty-seven (79%) had received a T-cell-depleted graft. Twenty (33%) had received myeloablative conditioning. The median time from transplant to diagnosis of H1N1 was 10 months; 27 patients (45%) were more than 1 year post transplant. Twenty-two patients (37%) had active GVHD at the time of H1N1 influenza diagnosis; grade II–IV acute GVHD was observed in 17 (28%) patients and extensive chronic GVHD in five (9%) patients. Thirty-four patients (57%) were receiving immunosuppressive agents at the time of H1N1 diagnosis; many were on multiple agents. Relevant co-morbidities included chronic lung disease in nine patients (15%), diabetes in four (7%) and cardiac disease in one (2%). Hypogammaglobulinaemia was documented in eight patients (21%). The median lymphocyte count at H1N1 diagnosis was 0.71 × 109/L; 34 patients (59%) were lymphopenic (lymphocytes less than 1 × 109/L) at diagnosis.
The most common presenting features of 2009 H1N1 influenza were fever in 39 patients (65%), dyspnoea in 27 (47%), headache in 17 (31%) and myalgia in 18 (34%). Less common manifestations included cough in eight patients (13%), diarrhoea in five (8%) and vomiting, wheeze and lethargy in two cases (3%) each. Fifty-three patients (91%) were treated with a neuraminidase inhibitor; all received oseltamivir, including 15 (25%) who received both oseltamivir and zanamivir. The median duration of treatment with oseltamivir was 5 days, and 6 days for zanamivir. Oseltamivir resistance was reported in two cases from one centre. Only two patients (4%) had received the H1N1 vaccine; one patient had two doses, the other a single vaccine dose, both within 2 weeks of H1N1 diagnosis and over 12 months from transplant.
Pneumonia complicating H1N1 influenza was diagnosed in 21 patients (35%): by clinical signs in five (8%) and with radiological findings in 16 (27%). Actuarial mortality in patients with pneumonia was 24% at 4 months. Nine of the 60 patients (15%) were admitted to ICU, for a median duration of 7 days. Of these patients, seven required mechanical ventilation and one inotropes. Actuarial mortality in patients admitted to ICU was 70% at 4 months. Extrapulmonary complications occurred in three patients (5%): hepatitis, hallucinations and acute renal failure in one patient each. Acute GVHD flared in two patients. Concurrent bacterial infections were reported in 11 patients (18%), viral infections in 11 (18%) and fungal infections in 12 (20%). Bacterial infections included two cases of Escherichia coli, two of Staphylococcus aureus, two of Pneumocystis carinii, one case each of Haemophilus influenzae, Pseudomonas and Clostridium difficile and four unspecified bacterial infections. Viral infections included one patient with adenovirus and two cases each of CMV, respiratory syncytial virus, norovirus, Epstein–Barr virus, rhinovirus and an unspecified viral infection.
Median follow-up from diagnosis of H1N1 infection was 6 months (range 2–10 months). Actuarial overall mortality was 7% at 28 days and 19% at 4 months post-diagnosis of 2009 H1N1 influenza; mortality was 0% at 4 months in five paediatric patients (age <18 years), 18% in patients aged 18–65 years and both patients aged over 65 years died. Eight deaths were documented as secondary to 2009 H1N1 influenza; the attributable mortality rate was 14% at 6 months. The six remaining deaths were due to relapse of acute leukaemia in four patients and multi-organ failure with concurrent fungal infection and overwhelming sepsis in one patient each.
Univariate analyses of factors associated with progression to lower respiratory tract involvement, admission to ICU and death were performed (Table 3). On univariate analysis, three factors were associated with the development of pneumonia: older age, pre-existing lung disease and concurrent infections (P=0.006, 0.054 and 0.043, respectively). No significant risk factors were identified for admission to ICU. Older age and active GVHD were associated with a significantly increased risk of death at 4 months (P=0.002 and 0.006, respectively). No other transplant-specific factors predisposed to severe complications or death secondary to H1N1 influenza infection. Multivariate analysis showed increasing age and lung disease as significant risk factors for pneumonia among patients with H1N1 infection after Allo-SCT (hazard ratio 7.1 and 6.8, P=0.006 and 0.037, respectively) (Table 4). Increasing age was the only significant independent risk factor for mortality (hazard ratio 6.8, P=0.012).
We report a multicentre cohort study describing the clinical course and outcome of 2009 H1N1 influenza infection in 60 Allo-SCT recipients in the United Kingdom. The morbidity and mortality observed in our cohort of allogeneic transplant patients were significant with progression to pneumonia in 35%, ICU admission in 15% and 14% attributable mortality. Indeed, the spectrum of clinical features, frequency of severe complications and death rate are similar to those observed elsewhere with seasonal influenza A infection in SCT recipients.11, 12, 13, 14, 15 Ljungman5 reported a prospective study looking at the incidence of community acquired respiratory viral infections in 793 recipients of SCT and the risk factors associated with severe complications; 30 cases of seasonal influenza A were recorded with 13% mortality. Many single centres also report experience in treating respiratory viral infections after SCT, with a frequency of pneumonia of up to 80%, and 0–20% mortality associated with seasonal influenza A infection.11, 12, 13, 14, 15
There are only limited reports of 2009 H1N1 influenza in SCT recipients and indeed our study is the largest cohort of Allo-SCT recipients infected with 2009 H1N1 influenza. Redelman-Sidi et al.16 report lower airways involvement in 27% of 12 allograft recipients included in a single-centre series of 45 cancer patients with 2009 H1N1 infection. However, the disease course was mild in this cohort and all patients received antiviral therapy, none required mechanical ventilation and there were no mortalities. Higher complications and mortality were observed in a two centre study of 27 cases of H1N1 infection in SCT recipients; pneumonia was reported in 52% and day 30 mortality was 22% overall, increasing to 43% in those with pneumonia.17 Delayed antiviral therapy and chronic steroid use predisposed to pneumonia while steroid was a risk factor for death. In a smaller report of nosocomial H1N1 infection in eight patients in a haematology unit, three cases were identified in allograft recipients; all presented with fever and upper respiratory tract infection symptoms, two (66%) of the SCT patients were admitted to ICU with pneumonia and both died.18
The clinical presenting features described in our study parallel those seen with 2009 H1N1 influenza in immunocompetent patients.19, 20, 21 However, life-threatening complications and mortality occurred at a higher frequency than in the general population, where around 9% were hospitalised20 and the case fatality rate was 26 per 100 000.22 Despite national epidemic planning and prompt antiviral therapy in the majority of patients, we describe a similar frequency of complications to that reported in recent series of critically ill hospitalised 2009 H1N1 influenza patients. In such series, progression to pneumonia was observed in 40–66%, admission to ICU in 25–31% and death reported in 7–14%, increasing to 20% in patients aged over 50 years.6, 23, 24
Older age, underlying co-morbidities25 and immune suppression7 are recognised as risk factors for severe disease in H1N1 infection. The high frequency of T-cell depletion observed in our cohort and associated protracted immune suppression may account for the cases of H1N1 infection seen many months post-SCT. Increasing age and pre-existing lung disease were significant risk factors for progression to pneumonia and older age was also associated with mortality secondary to H1N1 infection post-transplant, as reported in the immunocompetent patients.22 Although we could not calculate the incidence of H1N1 infection for the whole UK SCT population, we did assess incidence by age for participating centres. Our reporting population represented a significantly greater proportion of paediatric relative to adult patients, encompassing 77 vs 63% of UK paediatric and adult transplant activity, respectively (P=0.001). Despite this, the reported incidence of 2009 H1N1 influenza in our series was relatively higher in adult compared with paediatric SCT recipients. Moreover, we observed a significantly greater frequency of complications and death in older adults compared with paediatric patients. This contrasts with the general population, where the incidence of H1N1 infection in adults was significantly lower than that in children, although the case fatality rate was higher in the elderly.26 It has been postulated that the lower incidence of infections in the older population may be explained by cross-protective immunity from past influenza vaccines and infections.27 The absence of such immunological memory post-Allo-SCT may explain the higher relative incidence of 2009 H1N1 infection observed in our adult cohort.
We recognise our study has some limitations. Despite being the largest series of Allo-SCT patients presenting with 2009 H1N1 influenza infection to date, we accept that the numbers are relatively small, limiting our analysis of clinical manifestations and outcomes in this group. Moreover, our multivariate analysis of risk factors contributing to severe outcomes is limited by the small sample size (n=60). Centres submitted data on all laboratory-confirmed cases of H1N1 infection. However, the study design may have resulted in the omission of cases of subclinical infection in patients who were asymptomatic or had milder symptoms and did not present to their transplant centre; such patients would not have been diagnosed as cases of H1N1 infection. Indeed, evidence exists supporting the occurrence of asymptomatic viral respiratory infections in recipients of haematopoietic stem cell transplants.28 The potential omission of less severe or asymptomatic cases may have resulted in us overestimating the rate of complications and mortality seen with H1N1 infection in Allo-SCT patients. However, our findings are consistent with a recent large multicentre cohort study of 2009 H1N1 influenza infection in the recipients of solid organ transplants in terms of presenting features, frequency of use of neuraminidase inhibitors (94%), rate of progression to pneumonia (32%) and ICU admission (16%).29 One-third of the solid organ transplant recipients had an infected household contact and delayed initiation of antiviral therapy predisposed to severe disease.29
In conclusion, morbidity and mortality from 2009 H1N1 influenza in allograft recipients is similar to that seen in SCT patients during seasonal influenza A epidemics, but higher than that observed in the general population. SCT recipients remain at risk of severe complications secondary to H1N1 infection months to years after SCT; we speculate that this may be related to immune suppression secondary to persistent lymphopenia related to T-cell depletion. This report identifies subgroups of SCT recipients, older patients and those with pre-existing lung disease, at risk of severe complications and death secondary to 2009 H1N1 influenza. On the basis of our observations and the findings of other series, we recommend that these patients and their household contacts receive vaccination and that symptomatic SCT recipients receive early antiviral therapy during future H1N1 epidemics. The efficacy of these measures should be assessed by a prospective study.
We thank the data managers and transplant physicians at participating centres for providing data.