Influenza A/H1N1/09-10 Infections in a NICU During the 2009–2010 H1N1 Pandemic

Abstract

Few cases of the pandemic influenza A H1N1 have been reported in very low birth weight infants. We report here a small outbreak in our NICU of 3 cases of influenza A/H1N1/09-10 in very low birth weight infants during the 2009–2010 H1N1 pandemic and describe their clinical presentations and favorable outcomes despite the lack of treatment.

Annually, influenza A infections are a cause of significant morbidity and mortality in all age groups, and children and the elderly are especially vulnerable.^1,2 In June 2009 the World Health Organization declared a worldwide pandemic; the novel strain of influenza A/H1N1/09-10 (pH1N1) particularly affected children, young adults, and pregnant women.^3,4

Few cases of novel influenza pH1N1 have been reported to date in infants,^5–7 and only 1 case in a very low birth weight premature infant has been reported.⁵ Their clinical manifestations have been protean, and fever and apnea have been among the most common manifestations. We report here 3 cases of pH1N1 disease in low birth weight infants within a NICU during the 2009–2010 H1N1 pandemic.

MATERIALS AND METHODS

We reviewed the laboratory records for respiratory samples that tested positive for influenza A between June 1, 2009, and April 1, 2010, and originated from the newborn intensive care unit at Rainbow Babies and Children's Hospital in Cleveland, Ohio. This is an 82-bed level III and level II facility with all private rooms.

Influenza A testing was performed by either direct immunofluorescence assay or by an in-house real-time reverse-transcriptase polymerase chain reaction (RT-PCR) using primer sets that targeted the influenza A matrix gene and the hemagglutinin gene of pH1N1 according to Centers for Disease Control and Prevention protocol (sensitivity near 100%).^8–10 The RT-PCR was run daily during the pandemic season. Nasopharyngeal samples were transported immediately to the laboratory in M4 viral media, where they were stored at 4°C until tested.

As part of an ongoing respiratory viral study at our institution, respiratory samples were frozen at −80°C and stored for future use. The previously identified positive samples that had been stored were confirmed by additional molecular amplification at our facility or at an outside reference laboratory.¹⁰ Medical records were reviewed for demographic, clinical, and laboratory information. Chart analysis was approved by the University Hospitals Case Medical Center institutional review board.

RESULTS

Between October 1 and November 4, 2009, 5 children in the NICU tested positive for influenza A. pH1N1 activity was widespread in the state of Ohio at the time the infants were ill. One patient was transferred to another hospital shortly after diagnosis and was excluded from our analysis. One other patient, who tested positive for influenza A (not pH1N1) but tested negative when additional molecular amplification was performed at our laboratory, was also excluded. Demographics and clinical characteristics of the 3 remaining infants are shown in Table 1.

TABLE 1.

Clinical Characteristics of Patients With Influenza A in the NICU

Patient No.	Date of Positive Sample	Gestational Age, wk	Birth Weight, g	Gender	Age at Onset, d	Weight at Illness, g	Underlying Medical Problems	Symptoms	Matrix Protein Real-Time Influenza A PCR Result	H1N1 Real-Time RT- PCR Result	Retest Conventional RT-PCR
1	Oct 6, 2009	29	1160	Male	21	1260	Cocaine exposure, grade 1 IVH	Apnea, bradycardia, clinical seizures, encephalopathy, increased secretions	Positive	Positivea	NA
2	Oct 10, 2009	25	710	Female	51	1165	CLD, grade 2 IVH	Apnea, bradycardia, hypothermia, abdominal distension, increased oxygen requirement, increased secretions, rales	Positive	Positive	Positive influenza A matrix protein and H1N1/09 hemagglutinin
3	Oct 16, 2009	28	1058	Male	16	1095	CLD, grade 1 IVH	Apnea and bradycardia, abdominal distention, increased oxygen requirement, increased secretions, rales	Positive	Positive	Positive influenza A matrix protein and H1N1/09 hemagglutinin

CLD indicates chronic lung disease; IVH, intraventricular hemorrhage; NA, not applicable.

^aSample confirmed at an outside reference laboratory.

Patients were premature (mean gestational age: 27$\raisebox{1ex}{$3$}\!\left/ \!\raisebox{-1ex}{$7$}\right.$ weeks), very low birth weight (VLBW) (mean birth weight: 976 g) infants with a mean weight of 1173 g at the time of diagnosis. The time interval between positive samples was 4 days between the first and second case and 6 days between the second and third case. The first and third patients were located in the same area of the unit, each one in a private room.

Predominant respiratory findings included increased secretions from the endotracheal tube (n = 3), increase in oxygen requirements (n = 3), apnea (n = 3), and rales (n = 2). Two of the 3 patients were on continuous positive airway pressure (CPAP) at the time of diagnosis, and 1 patient was not ventilated. One patient required subsequent intubation secondary to respiratory distress, whereas 2 others required an increase in CPAP settings. All of them required supplemental oxygen (mean fraction of inspired oxygen: 53%).

None of our patients were febrile. Other findings included an increase in abdominal girth that required discontinuation of feeds in 2 patients. One patient developed clinical seizures. Neurologic workup of this patient was performed, including cerebrospinal fluid (CSF) analysis; the results were normal. Results of a herpes simplex virus PCR from CSF were negative; however, the CSF was not sent for pH1N1 PCR. MRI was performed and revealed no intracranial abnormalities. An electroencephalogram was also unremarkable.

The mean white blood cell count was 13 400/μL with no predominant shifts. The mean C-reactive protein level at the time of illness was 0.6 mg/dL. Two patients were started on antibiotics for the possibility of sepsis. None were bacteremic. All patients had a respiratory viral antigen panel performed on nasopharyngeal aspirates by direct immunofluorescence assay, which included pH1N1, and the results were negative. All 3 patients tested positive for the influenza A matrix protein by PCR of nasopharyngeal secretions and for pH1N1 hemagglutinin by real-time RT-PCR.

Two stored frozen specimens were available for retesting by conventional PCR for this study. Samples revealed the same results as originally found, testing positive for influenza A by matrix protein and positive for pH1N1 hemagglutinin.

None of the patients were treated with antiviral agents active against influenza; however, all patients improved after a mean of 8.3 days (range: 5–15 days). Two patients had repeated PCRs for influenza from nasopharyngeal samples at a mean of 7 days after the initial positive sample and tested negative. The infants were discharged at a mean of 46 days after the pH1N1 infection (range: 35–53). None required supplemental oxygen or neurologic follow-up at discharge.

Infection-control investigations failed to show any direct contact between the affected patients and known pH1N1- or influenza A–infected family or hospital staff. None of the mothers had influenza-like symptoms at the time of delivery. However, 20 nurses and other medical staff called out sick with respiratory illnesses during the periods of positive samples; none were confirmed positive for pH1N1. None of these nurses took care of the infants in the 4 days before the infants became ill.

DISCUSSION

Influenza A infection is common among the pediatric population and affects 10% to 40% of healthy children each year.¹¹ During the 2009 pandemic, children were particularly affected; mortality rates exceeded 4 times the average of previous influenza seasons, and 18% were younger than 2 years.^3,12

More reports have emerged regarding the impact of the pH1N1 virus on neonates. Barak et al⁵ described the case of a premature VLBW infant who tested positive for pH1N1 at 50 days of life. His initial presenting symptoms included recurrent vomiting and deep apneas accompanied by cyanosis and bradycardia that required positive-pressure ventilation. The infant was started on 5 days of oseltamivir and supportive care, which resulted in complete resolution in his symptoms. Zenciroglu et al⁷ described a series of 10 infants (who ranged in age from 5 to 70 days) in Turkey who tested positive for pH1N1. In their series, only 3 of the infants were premature (30–34 weeks). The most common symptoms were fever and cough. Apnea was the initial symptom in 3 of the patients. The majority of them had rales on auscultation, infiltrates on chest radiographs, cough, and close contact with others with influenza-like symptoms. Despite respiratory symptoms predominating, only 2 of them required oxygen support. From Japan, Morioka et al⁶ reported apneas, fever, and poor sucking as prominent symptoms. They reported the case of an infant who presented with severe pneumonia and was treated with oseltamivir for 5 days.⁶

Our infants were younger and had lower birth weights than those described in the literature. Similar to the previous cases reported, respiratory symptoms predominated in our infants. Other manifestations seen in our patients were increase in abdominal girth, which prompted the discontinuation of feeds in 2 infants, and 1 patient had clinical seizures. Seizures have been described with pH1N1 infections in children along with other neurologic manifestations, including encephalopathy.^13–15 Although one could anticipate that this population would have severe disease with pH1N1 infection, all of our patients did well; their symptoms improved within 7 days, on average. This result was similar to that reported in 1980 by Glezen,¹ who described subclinical infection or afebrile upper respiratory illnesses in infants with influenza.

Low infection rates, particularly in neonates, may be related to underrecognized disease or to the minimal exposure that neonates have to sick contacts. This is especially true in NICUs, in which additional barriers to infection exist. Yet, influenza infections do occur in neonates, and outbreaks in NICUs have been described.^16–19 To our knowledge, ours is the first report in the medical literature of a pH1N1 outbreak in a US NICU. In our cases, there was no documentation of illness in any of the parents or hospital caregivers, although influenza-like illnesses did occur among nurses in our NICU.

None of our patients received treatment because of their moderate clinical disease and lack of an adequate dose of oseltamivir for preterm infants. No prophylaxis was given to other infants in the NICU. During the pandemic, the US Food and Drug Administration issued an emergency-use authorization for use of oseltamivir to treat pH1N1 in children younger than 1 year. However, no data existed on the dosing of the drug in premature infants until Acosta et al²⁰ published their findings in August 2010. They found that neonates achieved higher levels of the oseltamivir metabolite, oseltamivir carboxylate, even when using a lower dosage per kilogram, which was likely related to diminished renal clearance of the drug secondary to renal immaturity. They concluded that 1 mg/kg per dose twice daily in infants younger than 38 weeks would be equivalent to the 3 mg/kg per dose twice daily used in term neonates and young infants.

Establishing the diagnosis of influenza infections during the pH1N1 pandemic was problematic because of the lack of sensitivity of the immunoassays; therefore, RT-PCR became the preferred method for diagnosis.³ In our patients, results of all the direct immunofluorescence assays were negative. Yang et al²¹ found that low viral loads in samples from patients with pH1N1 infections affected the performance of antigen-based tests. To et al²² compared pH1N1 to historical influenza A samples and found that the viral loads in patients at days 0 to 3 of illness were lower in the pH1N1 samples. Although we did not perform viral load tests on our specimens, one could speculate that lower viral loads would elicit less inflammatory responses and, hence, less severe disease in these VLBW infants. It is interesting to note that the majority of severe cases of pH1N1 in children occurred in those older than 5 years.^3,12 This age trend was also observed in reports of neurologic complications of this pandemic strain.^13,15 Younger age has been associated with prolonged shedding of pH1N1²³; however, there have been no studies, to our knowledge, that have examined pH1N1 viral loads in neonates.

Two of our patients had repeat samples collected 7 days after the initial positive sample, and they tested negative. This result is consistent with the literature, in which the median length of positive RT-PCR results was 6 days.²⁴

CONCLUSIONS

Influenza infections caused by pH1N1 in VLBW infants can present with mild-to-moderate symptoms or a sepsis-like syndrome. All of our patients had a favorable outcome despite the lack of therapy. Although a high level of suspicion is required to establish the diagnosis, the benefit of treatment for this infection in premature VLBW infants is unclear, and further studies are needed.