Between March and June 2009, an out-of-season influenza outbreak, first detected in Mexico and then worldwide, led to identification of a novel virus, H1N1 influenza A virus (nH1N1 influenza) (1). This triple-reassortant influenza, containing genes from swine, avian, and human influenza lineages, is similar to the 1918 influenza pandemic virus and has been described as the “greatest pandemic threat since the emergence of Influenza A (H3N2) virus in 1968” (2). Like seasonal influenza disease, nH1N1 influenza disease seems to have increased infection rates among pediatric and young adult patients and a variable spectrum of disease severity. Although the full virulence of the novel influenza strain will not be known until a greater proportion of the population is infected, we report the first case series of critically ill children with nH1N1 influenza disease based on patients admitted to our pediatric intensive care unit (PICU) during the spring and summer of 2009. The Johns Hopkins School of Medicine Institutional Review Board reviewed and approved the study.
MATERIALS AND METHODS
Setting and Case Identification
The Johns Hopkins Hospital Children's Center is a regional Children's Hospital that serves Maryland and the surrounding area. The 26-bed PICU cares for >1300 children annually with an average length of stay of 5.7 days. Hospital Epidemiology and Infection Control monitors health care-associated infections and epidemiologically significant organisms for The Johns Hopkins Hospital Children's Center, using a system that combines administrative, microbiological, radiologic, and other data sources to produce reports (TheraDoc, Inc., Salt Lake City, UT). Because of an active respiratory virus prevention program, systematic surveillance and prevention planning includes prospective tracking of influenza. All patients admitted to The Johns Hopkins Hospital Children's Center during respiratory season (including the present pandemic) are tested for respiratory virus infection. Nasopharyngeal aspirates are tested, using direct fluorescent antibody (DFA) for influenza A, influenza B, respiratory syncytial virus, human metapneumovirus, adenovirus, and parainfluenza virus types 1-3. All residual specimens were stored at −80°C. Shell vial (R-Mix Too, Diagnostic Hybrids Inc., Athens, OH) and tube cultures (rhesus monkey kidney and A549 cells, Diagnostic Hybrids, Inc.) were inoculated in parallel with DFA. Shell vials were stained at 48 hrs for all viruses detected by DFA, except human metapneumovirus. Tube cultures were examined daily for 7 days, then weekly for an additional 2 wks until cytopathic effect was observed. All samples from inhospital patients identified as influenza A by DFA, shell vial, or tube culture were sent to the Maryland Department of Health and Mental Hygiene for confirmatory nH1N1 testing, using the U.S. Centers for Disease Control and Prevention (CDC) reagents for reverse transcription/real-time polymerase chain reaction.
Data Collection, Definitions, and Analysis
The nH1N1 influenza-infected patients in The Johns Hopkins Hospital PICU were identified from the Hospital Epidemiology and Infection Control database. Clinical information on each patient was verified and supplemented by medical record review with a standardized data abstraction form. Demographic data, past medical history, presenting symptomatology, microbiological and laboratory data, radiographic studies, and treatment, and outcome data were collected.
Patients were considered to have clinical evidence of pulmonary bacterial superinfection if they developed new radiographic findings consistent with bacterial pneumonia, such as lobar consolidation plus recurrent fever, dyspnea, tachypnea, increased or purulent secretions (3) after initial clinical improvement. Patients were considered to have evidence of bronchospasm if they clinically demonstrated wheezing or improvement with β-agonist therapy. Patients treated with oseltamivir were divided into early and late treatment groups based on timing of therapy relative to symptom onset (≤48 hrs or >48 hrs, respectively).
Data are presented as mean ± sd, where appropriate. Analysis of variance was used to determine statistical significance of differences between early and late treatment groups. A p < .05 was considered significant.
A 17-yr-old female with known sickle cell anemia, hypertension secondary to sickle cell nephropathy, and asthma presented to our emergency department with fever, hypoxia, dyspnea, and chest pain after a routine outpatient transfusion. No abnormalities were noted on chest radiography. Respiratory and blood cultures were obtained. She was admitted for pain management and antimicrobials and was discharged home 48 hrs later after clinical improvement. The patient returned to the emergency department several days after discharge with recurrent fever, cough, dyspnea, lethargy, vomiting, and diarrhea. Hemoglobin was noted to be 4.7 g/dL and a repeat chest radiograph demonstrated a left upper-lobe consolidation consistent with acute chest syndrome. The patient was admitted to the PICU. The DFA for influenza was negative, and she improved clinically after double-volume exchange transfusion and initiation of bilevel positive airway pressure ventilation and antibacterial therapy. However, ongoing low-grade hemolysis was suspected due to decreasing hemoglobin levels and mild hematuria. On day 3, she developed severe dyspnea and was intubated for respiratory failure. Chest radiography showed diffuse alveolar infiltrates consistent with acute respiratory distress syndrome. Intravenous corticosteroids were added to bronchodilator therapy and antibacterial coverage was broadened. The patient subsequently developed oliguric renal failure and hypotension requiring inotropic therapy, and she continued to have unexplained high fever and bronchospasm. On day 10 of illness and day 5 after obtaining the nasopharyngeal aspirates on admission, influenza A (nH1N1 influenza) was isolated from tube culture and oseltamivir therapy was begun via nasogastric tube. Her course was notable for paroxysms of severe cough with hypoxia requiring neuromuscular blockade in addition to deep sedation/anesthesia while intubated. Renal function ultimately recovered to baseline without renal replacement therapy. She remained intubated for a total of 14 days and remained in the intensive care unit for a total of 26 days. On hospital day 32, the patient was transferred to a rehabilitation hospital and was subsequently discharged. She continues outpatient transfusion therapy and is neurologically intact.
A 21-yr-old female with a history of congenital human immunodeficiency virus infection complicated by encephalopathy (with paraplegia), lymphocytic interstitial pneumonitis, and poor compliance with medical therapy presented with 2 days of fever, tachypnea, pharyngitis, and cough. Absolute CD4 count on admission was 470/mm3. She was admitted to the PICU for management of hypoxia, dyspnea, and hypotension requiring transient vasopressor therapy. Antimicrobial therapy, including oseltamivir, was initiated despite negative DFA results because chest radiography demonstrated increased bilateral perihilar markings consistent with viral infection compared to a previous study. Within 24 hrs, the patient had returned to her baseline respiratory condition and was downgraded to a noncritical hospital bed. Shell vial culture was negative, but tube culture grew influenza A (nH1N1 influenza) at 3 days; the patient completed a 5-day course of oseltamivir.
A 14-yr-old female with congenital cytomegalovirus infection and associated cerebral palsy, mental retardation, obstructive sleep apnea (requiring overnight continuous positive airway pressure), and gastrostomy tube dependence was admitted with increased frequency of emesis. After admission, she developed progressive respiratory distress requiring intubation and transfer to the PICU for mechanical ventilation. A nasopharyngeal aspirate was sent due to the decline in respiratory status and the DFA was positive for influenza A (nH1N1 influenza). Oseltamivir therapy was initiated (approximately 24 hrs after onset of initial symptoms). The patient required invasive mechanical ventilation for 7 days and remained in the PICU for 16 days. Because airway instrumentation occurred before the suspicion of viral illness, the patient was not appropriately isolated and multiple staff members on the inpatient floor, the rapid response team, and the ICU team were exposed to influenza.
Summary Results for the Entire Series
Between June 1, 2009 and August 7, 2009, we diagnosed 140 patients <264 months (22 yrs) of age with nH1N1 infection at The Johns Hopkins Hospital; 13 (9.3%) required admission to the PICU (three patients, aged 20-21 yrs, were admitted to adult ICUs). All patients were exposed to nH1N1 in the community. The median age was 114 months (range, 5-263 months), with a male predominance (62% male) (Table 1). Almost all patients had underlying comorbid illness and risk factors for complications of influenza, of which asthma was most prevalent (85%) followed by neuromuscular disease (38%). Based on the Pediatric Risk of Mortality III score (4.5 ± 4.36), the severity of illness at admission to the PICU was mild to moderate (4). DFA proved highly insensitive with a 62% false-negative rate. DFA detected only 38% of cases; an additional 38% of cases were detected at 48 hrs by shell vial. Twenty-three percent of cases were detected only by tube culture. The mean times to diagnosis of all samples requiring shell vial or tube culture were 5.3 (± 2.8) days and 2.4 (±1.2) days from symptom onset and PICU admission, respectively. The mean times to diagnosis of samples requiring tube culture were 7.3 (±3.1) days and 3.3 (±1.5) days from symptom onset and PICU admission, respectively.
Laboratory data were consistent with viral infections. The average admission total leukocyte count was in the normal range (8724 ± 4449/mm3); some patients presented with leukopenia or leukocytosis (Table 2). Lymphopenia was common with the mean lymphocyte count of 1473 (±1110)/mm3 at admission. Hepatic transaminase enzymes and renal function tests were normal or mildly elevated during the PICU stay.
Twenty-three percent (3 of 13 patients) of the study population presented with normal chest radiography and never developed any abnormality on chest radiography. Some patients with radiographic abnormalities on admission had multiple types of abnormalities. Of patients with abnormal admission chest radiographs, 60% (6 of 10) demonstrated increased interstitial markings consistent with viral disease, whereas 30% (3 of 10) revealed diffuse alveolar infiltrates and 40% (4 of 10) had findings consistent with hyperinflation. Thirty-one percent (4 of 13) of all patients already had evidence of lobar consolidation on chest radiography at PICU admission and one patient (8%) had a pleural effusion at admission. Peak radiographic findings occurred at 4.6 (±3.4) days after admission and consisted of diffuse alveolar infiltrates in 70% (7 of 10) of patients with abnormal chest radiographs.
Eighty-five percent (11 of 13 patients) of patients with nH1N1 received antiviral therapy (Table 3); all treated patients received standard-dose oseltamivir. Among those treated, 45% (5 patients) were treated within 48 hrs of initial symptom onset (early treatment group) and 55% (6 patients) were treated after >48 hrs of symptom onset (late treatment group). Forty-six percent (6 of 13 patients) of the entire study population received mechanical ventilation, four patients via endotracheal tube and two patients by noninvasive methods (e.g., bilevel positive airway pressure). Consistent with the radiographic prevalence of increased interstitial markings and hyperinflation, a majority of patients (77%) also had symptoms of bronchospasm that were alleviated with β-agonist therapy. More than half of patients received adjuvant therapies for bronchoconstriction, including therapeutic magnesium and/or systemic corticosteroids. Three patients (23%) required inotropic therapy for hypotension; echocardiography for two of these patients demonstrated normal left ventricular function. The third patient did not undergo echocardiography due to limited duration of vasopressor therapy. No patient required renal replacement therapy or extracorporeal support.
Twenty-three percent (3 of 13 patients) of patients in this case series were treated for a suspected bacterial superinfection on the basis of clinical and radiographic findings (Table 3). None of the bacterial cultures obtained from patients grew.
Patient outcomes were divided into subgroups based on early treatment or late treatment (Table 4). ICU length of stay was 4.2 (±6.6; median, 1.0) days and 6.8 (±8.8; median, 3.5) days for the early and late treatment groups, respectively. Duration of mechanical ventilation for the early treatment group was 4.0 ± 4.2 days compared with the late treatment group duration of 7.8 ± 6.5 days. The differences in ICU length of stay and duration of mechanical ventilation were not statistically significant between those who were part of the early or late treatment groups. All patients survived through hospital discharge.
The nH1N1 disease has primarily impacted young people, including children. The major findings from this series are: 1) underlying chronic illness (especially respiratory illness) seems associated with critical nH1N1 influenza disease in children. 2) The respiratory illness is highly variable from patient to patient and within a single patient involving bronchoconstriction and alveolar consolidation. Respiratory support and sedation therapies must be individualized and rapidly adjusted. 3) Diagnosis and treatment were often delayed, reflecting initial inexperience with nH1N1 disease during the start of the pandemic. 4) The duration of critical illness was not different between early and late treatment groups although our patient numbers are small. 5) Bacterial superinfection occurred in one quarter of patients, more commonly than previously reported. 6) Moderate nH1N1 influenza disease, including respiratory failure and hypotension, had 100% survival in our series.
Association With Underlying Chronic Illness
The CDC has described underlying conditions that, for seasonal influenza disease, place patients at increased risk for complications (5). This series confirms the significant vulnerability of patients with comorbidities for nH1N1 influenza infection. Ninety-two percent of patients admitted to our PICU with nH1N1 influenza disease were previously known to have one of these high-risk conditions (Table 1), whereas only one patient was admitted with no significant past history of disease. As this pandemic continues, additional data will likely emerge to better define high-risk groups for nH1N1 influenza; however, our data support the use of high-risk groups as defined in seasonal influenza outbreaks as preliminary risk factors for critical illness in nH1N1 disease.
Variability in Respiratory Disease
A surprising aspect of this series was the variability of symptomatology among the most severely affected children. Both clinically and radiographically, our patients displayed features of alveolar consolidation/acute respiratory distress syndrome and reversible obstructive lung disease (bronchoconstriction). This combination proved a particular challenge in critical care management.
In our case series, the severe bronchospasm was treated with intravenous magnesium sulfate, systemic corticosteroids, and short-acting β-agonist therapy in 54%, 54%, and 77% of patients, respectively. The most severely affected patients also experienced intense coughing spells during invasive mechanical ventilation requiring both anesthetic doses of sedative hypnotic agents and (in three patients) persistent neuromuscular blockade. Given that management strategies for bronchospasm and acute respiratory distress syndrome are not well aligned, mechanical ventilation and management of critically ill children with nH1N1 infection proved to be challenging. When the composite picture included alveolar consolidation, hypoxemia, and bronchospasm, we utilized a strategy of high positive end-expiratory pressure, low tidal volume, and low ventilator breath frequency to maintain protective low lung volumes at the same time prolonging the expiratory phase to allow alveolar emptying and prevent air trapping. However, the management of these patients should be individualized and regularly reassessed and rapidly adjusted to reflect the heterogeneous and labile nature of the lung involvement.
Delayed Diagnosis and Prevention of Disease
Sixty-two percent (8 of 13) of critically ill influenza patients in this series had an initial negative DFA result. The low sensitivity of DFA and the subsequent reliance on culture for detection resulted in a delay in definitive diagnosis for these patients. As demonstrated in the case of Patient 1, viral tube culture is very sensitive but virus growth may require ≥5 days. Nucleic acid amplification tests, such as the CDC's real-time polymerase chain reaction assay, combine high sensitivity with faster time-to-result and, therefore, may play a more central role as the pandemic evolves (6, 7).
The importance of detailed hospital disaster plans and staff preparedness cannot be underestimated. As demonstrated by patient 3, delayed diagnosis and failure to anticipate the possibility of an influenza-related etiology for symptoms will lead to unnecessary staff exposure. We subsequently determined that our rapid response team would carry personal protective equipment to all off-unit events where intubation or other high-risk procedures may take place. Note that this practice is also not without risk, as it involves a delay in care as personal protective equipment is applied and decreased efficiency of communication when members of the team are subjected to the noise and isolation of a Powered Air-Purifying Respiratory device. As noted by Aziz, it is essential for hospitals to prepare for pandemic influenza with written (and rehearsed) disaster policies (8).
There are few data regarding preventive chemoprophylaxis in the ICU; however, Oliveira et al recommended routine chemoprophylaxis with neuraminidase inhibitors for all adult ICU patients on a daily basis from the start of an outbreak (confirmed case in an ICU patient) until approximately 1 wk after the end of the outbreak (9). Current CDC recommendations suggest antiviral chemoprophylaxis with neuraminidase inhibitors should be reserved for those patients with likely nH1N1 exposure who are at high risk for complications (5). In accordance with this, we have adopted a strategy of heightened surveillance to identify symptomatic patients, early isolation, and treatment of patients with unexplained fever or other symptoms (before diagnostic testing results). However, we have not used chemoprophylaxis routinely for asymptomatic, unexposed ICU patients during either seasonal influenza outbreaks or the present pandemic. This practice is based on our prior experience with seasonal influenza disease. Of note, both the CDC recommendations and our hospital and PICU policies regarding the diagnosis, isolation, and management of patients infected with nH1N1 influenza have evolved throughout the pandemic and are expected to continue to change as new data become apparent.
Effect of Early Antiviral Treatment
Aoki et al demonstrated that early treatment with neuraminidase inhibitor therapy has the greatest effect on symptom duration and severity in a healthy population. Although 48 hrs is the most frequently recommended interval for therapy, the authors demonstrated that treatment as early as 12 hrs after symptom onset had greater effects than treatment at 48 hrs (10). Eighty-five percent of our patients were treated; all those treated received the neuraminidase inhibitor oseltamivir, which has demonstrated efficacy in decreasing viral shedding and the duration of mild symptoms in uncomplicated seasonal influenza (11). Reasons for failure to treat included significant delay in diagnosis with spontaneous resolution of symptoms. However, patient 1 illustrates the risks of delayed treatment, especially in patients with significant comorbidities. Patient outcomes (Table 4) are divided into subgroups of early treatment (oseltamivir administered within 48 hrs of symptom onset) and late treatment (oseltamivir administered after 48 hrs of symptoms). The reasons for delayed treatment include delayed presentation and delayed diagnosis. Although not statistically significant, we are limited in our conclusions because we had a small number of patients. Still, we did note trends toward shorter ICU stays, shorter duration of mechanical ventilation, and shorter interval between oseltamivir and extubation (as a surrogate for clinical improvement) in the early treatment group. We report these findings because there is a paucity of data about the use of antiviral agents for severe influenza disease in critically ill patients, and there are even fewer data in children.
Our experience supports the CDC advisory of immediate, presumptive treatment of critically ill children with a history suggestive of nH1N1 influenza disease (regardless of duration of symptoms) because the risk/benefit ratio for these patients is likely in favor of therapy. DFA has a high false-negative rate and definitive culture diagnosis takes several days. Although widespread resistance to neuraminidase inhibitors among seasonal influenza A has occurred (including rare reports of nH1N1 isolates) (12), the nH1N1 influenza virus seems to still be largely sensitive to oseltamivir (13). Whitley et al suggested that oral dosing to children >12 months results in cost-effective reduction of disease burden and viral shedding (14). Oseltamivir is only approved for use in patients ≥12 months of age, but during this epidemic the U.S. Food and Drug Administration approved an Emergency Use Authorization for treatment of influenza infection in patients <12 months old. Oseltamivir is not available in a parenteral formulation; we were able to treat four of our patients (including patients 1 and 3) via nasogastric or gastrostomy tube. This practice is supported by a pharmacokinetic study of nasogastric administration of oseltamivir to adults demonstrating good absorption and extensive metabolism to the active metabolite, oseltamivir carboxylate, even in critically ill patients with severe influenza (15). The Food and Drug Administration recently approved intravenous peramivir for compassionate use, although there are very limited data for its use in pediatric patients.
The CDC has recommended peramivir in unresponsive or deteriorating patients or in patients unable or unlikely to absorb oseltamivir. Intravenous peramivir can be obtained through the CDC at http://emergency.cdc.gov≤1N1antivirals.
A key component of severe influenza management involves surveillance for and early treatment of bacterial superinfection. Reed et al found that, among children hospitalized with seasonal influenza disease, 15% had culture-proven bacterial infection (16). The most common organism among those cases was Staphylococcus aureus (including methicillin-resistance in >40% of S. aureus isolates). Patients with bacterial infection were also more likely to require critical care and had a higher patient fatality than other children with influenza. In our series, 23% of our nH1N1 influenza-infected patients developed clinical and radiographic evidence of bacterial superinfection, although none of our patients had positive bacterial respiratory cultures (3). Enhanced morbidity and mortality with influenza-related bacterial infection may be due to increased viral replication and pathogenicity caused by proteolytic activation of hemagglutinin (17). Seasonal influenza A has also been reported to down-regulate neutrophil function (18), which would theoretically provide a more hospitable milieu for bacterial infection. Critically ill children with nH1N1 influenza disease deserve extreme vigilance for bacterial superinfection.
All 13 patients in this case series survived until discharge. This finding contrasts with a report by Li et al revealing a mortality rate of 18.5% in a cohort of more severely compromised critically ill adult patients (Acute Physiology and Chronic Health Evaluation III scores of 82 ± 20 in nonsurvivors) with seasonal influenza A infection (19). The comparison between our data and those by Li et al should be interpreted cautiously because of the younger age group, the early stage of a new pandemic, and the less severe illness in our cohort. None of our patients required renal replacement therapy or extracorporeal support. Regardless, we report that, in this series of 13 children in our PICU, 12 of whom had known high-risk conditions for severe influenza disease, mortality was 0%.
This series illustrates the presentation of critical nH1N1 influenza disease in the PICU. The pediatric intensivist must be prepared for hemodynamic instability and highly individualized ventilator management, depending on whether bronchospasm or alveolar consolidation predominates. Significant sedation and even neuromuscular blockade may be necessary to manage paroxysms of severe coughing. All patients including those presenting with mild-moderate cardiovascular, respiratory, and hematologic dysfunction as well as bacterial superinfection survived. The duration of critical illness was not different between early and late treatment groups. Whether this is reflective of sample size or indicative of the importance of therapeutic intervention at any time early during infection in critically ill patients is unclear. The effect of early antiviral therapy on critically ill pediatric patients with influenza infection (nH1N1 or seasonal) is an area for future larger studies. In the meantime, practitioners are advised to have a high index of suspicion and to consider presumptive treatment of all critically ill patients, regardless of duration of symptoms, until definitive diagnostic testing is available. Additional epidemiologic studies in children are desperately needed to guide surveillance practices and infection prevention and control strategies for emerging pathogens, such as nH1N1 influenza
We thank Alicia Budd in Hospital Epidemiology and Infection Control at The Johns Hopkins Hospital whose efforts tracking the patients of nH1N1 during the emerging pandemic were invaluable in the identification of patients for this manuscript.
1. Fraser C, Donnelly CA, Cauchemez S, et al: Pandemic potential of a strain of influenza A (H1N1): Early findings. Science
2. Novel Swine-Origin Influenza A (H1N1) Virus Investigation Team, Dawood FS, et al: Emergence of a novel swine-origin influenza A (H1N1) virus in humans. N Engl J Med
3. Langley JM, Bradley JS: Defining pneumonia in critically ill infants and children. Pediatr Crit Care Med
2005; 6(3 Suppl):S9–S13
4. Pollack MM, Patel KM, Ruttimann UE: PRISM III: An updated Pediatric Risk of Mortality score. Crit Care Med
5. Centers for Disease Control and Prevention (CDC): Updated interim recommendations for the use of antiviral medications in the treatment and prevention of influenza for the 2009-2010 season. Available at http://www.cdc.gov/h1n1flu/recommendations.htm
. Accessed October 21, 2009
7. Centers for Disease Control and Prevention (CDC): Evaluation of rapid influenza diagnostic tests for detection of novel influenza A (H1N1) virus—United States, 2009. MMWR Morb Mortal Wkly Rep
8. Aziz AM: Pandemic influenza: Implications for infection control in hospital. Br J Nurs
9. Oliveira EC, Lee B, Colice GL: Influenza in the intensive care unit. J Intensive Care Med
10. Aoki FY, Macleod MD, Paggiaro P, et al: Early administration of oral oseltamivir increases the benefits of influenza treatment. J Antimicrob Chemother
11. Hayden FG, Treanor JJ, Fritz RS, et al: Use of the oral neuraminidase inhibitor oseltamivir in experimental human influenza: randomized controlled trials for prevention and treatment. JAMA
12. Hayden F: Developing new antiviral agents for influenza treatment: What does the future hold? Clin Infect Dis
2009; 48(Suppl 1):S3–S13
13. Centers for Disease Control and Prevention (CDC): Update: Drug susceptibility of swine-origin influenza A (H1N1) viruses. MMWR Morb Mortal Wkly Rep
14. Whitley RJ: The role of oseltamivir in the treatment and prevention of influenza in children. Expert Opin Drug Metab Toxicol
15. Taylor WR, Thinh BN, Anh GT, et al: Oseltamivir is adequately absorbed following nasogastric administration to adult patients with severe H5N1 influenza. PLoS One
16. Reed C, Kallen AJ, Patton M, et al: Infection with community-onset Staphylococcus aureus and influenza virus in hospitalized children. Pediatr Infect Dis J
17. Scheiblauer H, Reinacher M, Tashiro M, et al: Interactions between bacteria and influenza A virus in the development of influenza pneumonia. J Infect Dis
18. Pang G, Clancy R, Cong M, et al: Influenza virus inhibits lysozyme secretion by sputum neutrophils in subjects with chronic bronchial sepsis. Am J Respir Crit Care Med
19. Li G, Yilmaz M, Kojicic M, et al: Outcome of critically ill patients with influenza virus infection. J Clin Virol