During the pandemic (H1N1) 2009 outbreak in the United Kingdom, the Influenza Clinical Information Network (FLU-CIN) collected clinical and epidemiologic data for patients with virologically confirmed pandemic (H1N1) 2009 virus infection admitted to hospitals (27). Data included demography, symptoms, medical history, influenza vaccination history, relevant timelines, investigations and results, treatment (e.g., antiviral and antibacterial drugs), outcome, and cause of death when available. Trained health care workers abstracted data from case notes. During May 11, 2009–January 31, 2010, data were accrued from 75 National Health Service hospitals in 31 cities or towns in England, Scotland, Wales, and Northern Ireland.

From this source cohort, we defined patients with nosocomial pandemic (H1N1) 2009 as those admitted to a hospital for a reason other than acute respiratory infection in whom respiratory symptoms developed >72 hours (3 days) after admission. In addition, we included infants who had not left the hospital since birth in whom pandemic (H1N1) 2009 had developed. We included transfers from other hospitals when a transfer was for a reason other than influenza and when the history of influenza clearly indicated that it had been acquired at another hospital. FLU-CIN procedures were reviewed and approved by the Ethics and Confidentiality Committee of the National Information Governance Board for Health and Social Care in England for collection, storage, and use of personal data for surveillance purposes.

Of 1,520 patients in the FLU-CIN cohort, illnesses in 30 (2.0%) (15 children) met the criteria for nosocomial influenza (Tables 1, 2 [adults] and Tables 3, 4 [children]). Patient ages ranged from 41 days to 76 years at onset of influenza symptoms (median age 44 years for adults and 1 year for children).

Although pandemic (H1N1) 2009 produced a generally mild illness, in the United Kingdom, as elsewhere, severe illness developed in a small proportion of relatively young patients who required hospitalization (28). Although nosocomial outbreaks of pandemic (H1N1) 2009 have been described (24–26), sporadic nosocomial cases of pandemic (H1N1) 2009 identified during surveillance activities have not been described. The present case series has the advantage of being derived from a larger cohort of hospital inpatients in whom confirmation of pandemic (H1N1) 2009 was obtained by using nationally standardized PCR criteria, from settings where clinical management and infection control precautions were driven by national guidelines (29,30), and with data abstracted by trained nurses (27).

We based our definition of nosocomial influenza on a recent study of health care–associated influenza in children (31). A recent systematic review of incubation periods of acute respiratory viral infections found that the median incubation period for influenza was 1.4 days for influenza A, and symptoms developed in 95% of patients in <2.8 days (32). These findings suggest that our cutoff point, 3 days after admission, make inclusion of community cases unlikely. In addition, in no patients did onset of respiratory illness occur <4 days after hospital admission; median length of hospitalization before symptom onset was 11 days for adults and 13 days for infants. Therefore, inadvertent inclusion of community-acquired cases is highly unlikely.

On the basis of information obtained in the study, we cannot determine where and from whom patients acquired influenza. However, 3 routes are possible. First, infection could have been acquired from other patients; 1 patient shared a bay with a patient who was presymptomatic at the time but for whom influenza was diagnosed 1 day later. Second, transmission from visitors of patients cannot be ruled out. Although national guidelines strongly discourage persons with influenza-like symptoms from visiting patients (29), this recommendation may have been difficult to implement, particularly for parents of sick children who often provide most hands-on care in a hospital. Third, transmission may have occurred from an infectious health care worker (because staff continue to work when infected with influenza [33]) or from contaminated hands of a health care worker. Transmission from asymptomatic persons might occur in all 3 instances (34).

Nosocomial cases in this study occurred equally in adults and children. Consistent with previous findings (3), most patients had >1 serious underlying illnesses, notably hematologic malignancies, congenital disorders, or prematurity. Staff and caregivers of patients with hematologic malignancy and prematurity are often particularly vigilant for symptoms suggestive of infectious disease. Although we detected nosocomial influenza in patients admitted to nonmedical areas (for emergency or elective surgery), many cases of nosocomial infection in other patient groups probably have been overlooked, particularly because influenza in these groups is likely to have been milder. Additionally, some patients are likely to have been discharged from a hospital during the incubation period of nosocomially acquired pandemic influenza. Thus, in this case series, detecting such patients would not have been possible.

More than half of patients required level 2 or level 3 care, which is higher than that required by the source cohort (12%) (27). Approximately one fifth of children and one third of adults died. Although the deaths of 2 patients were expected because of the stage of their underlying malignancy, this case-fatality rate is far higher than that for patients with pandemic (H1N1) 2009 and concurrent conditions in the source cohort (5%) (27). The combined factors of increased host susceptibility (18–20), prolonged virus shedding in immunocompromised children (35), and increased likelihood of development of drug resistance (36) raise questions about the need for enhanced infection control procedures in special-care–infant units, pediatric wards, and hemopoietic transplant units and a requirement that staff working in these areas be vaccinated (37–39). Precautions should include restricting unnecessary movement of patients to units with particularly vulnerable patients and postponement of semi-elective (nonurgent) procedures for hematology patients during peak pandemic activity.

Vaccine against pandemic (H1N1) 2009 became available at the end of October 2009. Assuming a 2-week period for vaccine administration, case-patients in groups at risk with influenza onset dates after November 30, 2009, could have been vaccinated and would have had time to seroconvert (14 days). Using these criteria, we determined that 4 cases (in 3 adults and 1 child) (13%) were potentially preventable by vaccination; 2 of these patients required escalated care, and 1 patient died. Although 72% of patients received antiviral therapy, similar to 75% in the source cohort (27), we observed avoidable delays between recording of respiratory symptoms and start of specific antiviral therapy in some adults and all children. Although under ordinary circumstances the complex clinical picture of such patients might result in delayed or incidental finding of influenza, in a pandemic situation or during a seasonal epidemic clinicians should be alert to the possibility of influenza. Delays encountered in this series most likely reflect a failure to consider such a diagnosis early. Other reasons are caution or uncertainty in using oseltamivir in patients younger than the drug licensing permits (12 months) in nonpandemic situations, reluctance to empirically instigate antiviral treatment in advance of a confirmed diagnosis of pandemic (H1N1) 2009, lack of confidence about absorption of oseltamivir by nasogastric tube insertion in patients already receiving mechanical ventilation or concerns about potential gastrointestinal side effects.

Nosocomial infections with pandemic (H1N1) 2009 in this case series were associated with high rates of illness and death. This finding highlights the need for adherence to infection control guidelines for staff and visitors (including the need to urge visitors not to visit when they are ill, particularly when providing hands-on care for vulnerable children), staff vaccination, maintenance of clinical suspicion for influenza in areas of high risk, prompt (empirical) antiviral treatment for vulnerable patients in whom influenza is possible or likely, and consideration of postponing nonurgent procedures for hematology patients during periods of known high influenza activity. This report demonstrates that nosocomial transmission is a recurrent problem when the prevalence of influenza is high and the total effect of nosocomial influenza is underestimated by outbreak reports alone.