Influenza-Associated Pediatric Deaths in the United States, 2004–2012

Influenza-Associated Pediatric Death Surveillance

An influenza-associated pediatric death is defined as a death in a US resident <18 years of age with laboratory-confirmed influenza virus infection. The confirmation of influenza virus infection can occur before or after death by any of the following laboratory methods: rapid diagnostic test, viral culture, fluorescent antibody, enzyme immunoassay, reverse transcription polymerase chain reaction, or immunohistochemical staining of tissue. Diagnostic testing is performed according to routine clinical care and/or postmortem examination if available. State and local health departments report influenza-associated pediatric deaths to the Centers for Disease Control and Prevention by using a standardized case report form that includes information about demographic characteristics, laboratory results, clinical course, medical conditions, and influenza vaccination status. Health departments verify vaccination status by reviewing medical records, contacting the child’s health care provider, or checking the state vaccine registry, if available. Bacterial coinfection is defined as bacteria cultured from a normally sterile site (blood, cerebrospinal fluid, pleural fluid) after illness onset or from postmortem lung tissue if the specimen was collected <24 hours after death.

An influenza season was defined as October 1 through September 30 with the exception of the pH1N1 period, during which seasons were defined as October 1, 2008, through August 29, 2009, and August 30, 2009, through September 30, 2010. Rates of influenza-associated death were reported with respect to census estimates of the population <18 years of age during the same influenza season.²³ The ACIP high-risk medical conditions include the following: asthma, neurologic/neurodevelopmental disorders, chronic lung disease, heart disease including congenital heart disease, blood disorders, endocrine disorders, kidney disorders, liver disorders, metabolic disorders, immunosuppression, and pregnancy.²⁰ In this analysis, children with neurologic/neurodevelopmental disorders were considered to have a high-risk medical condition for all influenza seasons. Analyses of influenza antiviral treatment data were limited to seasons in which they were consistently collected (2010–2011 and 2011–2012).

Groups recommended for influenza vaccination by the ACIP changed during the study period. During the 2004–2005 and 2005–2006 seasons, all children 6 to 23 months of age and children <18 years of age with high-risk medical conditions were recommended for vaccination.^16,17 Neurologic disorders were added to the ACIP high-risk medical conditions in 2005.¹⁷ In the 2006–2007 and 2007–2008 seasons, vaccination recommendations were expanded to all children 6 to 59 months of age.¹⁸ This recommendation was expanded further to all children 6 months to 18 years of age in the 2008–2009 season.^19–21 All children ≥6 months of age were considered eligible for influenza vaccination in this analysis, even if they were not specifically recommended for vaccination by the ACIP on the basis of age, medical conditions, or household contacts during the season of their death, because vaccine formulations approved by the Food and Drug Administration for persons ≥6 months of age were available during the entire study period. During the 2009–2010 season, the group eligible for the pH1N1 monovalent vaccine was limited in this analysis to those who died on or after November 1, 2009, to account for timing of vaccine distribution. Children were considered fully vaccinated if they received the recommended number of doses of influenza vaccine ≥14 days before illness onset that season. Selected characteristics were compared between children with and without high-risk medical conditions.

Analysis

Data were analyzed by using SAS, version 9.3 (SAS Institute, Cary, NC). A Wilcoxon rank-sum test was used to evaluate differences between non–normally distributed data. A t test was used to evaluate differences between means, and a χ² test was used to evaluate differences between proportions. Asymptotic 95% confidence intervals (CIs) are reported for relative risks (RRs) where the natural log of the RR is assumed to be approximately normally distributed. All P values are 2-sided, and P < .05 was considered significant.

Characteristics of Influenza-Associated Deaths in Children

As of August 2, 2013, a total of 830 laboratory-confirmed influenza-associated pediatric deaths occurring from October 1, 2004, through September 30, 2012, were reported from all 50 states and Guam. The most deaths reported in 1 season, 282 (3.8 deaths per 1 million children), occurred during the 2009–2010 season, and the fewest deaths, 35 (0.5 deaths per 1 million children), occurred during the 2011–2012 season (Fig 1). Deaths for most seasons peaked in February or March (Fig 2). After the pH1N1 virus was identified in the United States in April 2009, an early wave of pediatric deaths occurred during June 2009 followed by a larger peak in October 2009.

Of children who died, 53% were male (Table 1). The median age of children was 7 years. Children who died during the 2005–2006 season had the youngest median age of 3 years, and children who died during the 2009–2010 season had the oldest median age of 8.5 years (Table 2). Overall, 649 (78%) children with influenza-associated death had influenza A virus infection, 165 (20%) had influenza B virus infection, and 1 (<1%) was coinfected with influenza A and B viruses; 15 (2%) children had influenza virus infection for which the type was not distinguished. Except for the 2009–2010 season during which only 1% of viruses identified were influenza B, influenza B virus infection was identified in 23% to 38% of deaths each season (Table 2).

Of 794 children with influenza-associated death and a reported medical history, 453 (57%) had ≥1 high-risk medical condition. Neurologic disorders were the most commonly reported high-risk medical condition (33%). Overall, 26% had a pulmonary disorder, including 16% with asthma; 12% had a chromosome or genetic abnormality associated with a high-risk condition; and 11% had congenital heart disease or other cardiac disease (Table 1). Approximately one-third or more of children who died each season had no known high-risk medical conditions, and during the 2006–2007 season the majority (62%) of children had no high-risk medical conditions (Table 2).

Among children with influenza-associated death, the proportion without high-risk medical conditions varied by the type and subtype of influenza virus identified (P < .001). Among 323 deaths associated with pH1N1 virus, 115 (36%) occurred in children with no high-risk medical conditions. For deaths associated with seasonal influenza A (H1N1), influenza A (H3N2), and influenza B viruses, the proportions occurring in children without high-risk medical conditions were 20 of 32 (63%), 41 of 57 (72%), and 72 of 150 (48%) children, respectively.

More than one-third of children died before hospital admission, including 18% who died in the emergency department and 16% who died outside the hospital (Table 1). The median duration of illness was 5 days (interquartile range: 3–12 days). Of children who had a bacterial culture obtained from a normally sterile site, 154 (40%) had ≥1 bacterial coinfection. Almost half of those with bacterial coinfections had Staphylococcus aureus infection, and 22 (14%) had Streptococcus pneumoniae infection. Other bacterial coinfections identified included Streptococcus species other than S pneumoniae (n = 38), of which the most common was Streptococcus pyogenes (n = 20), Enterococcus species (n = 5), and Gram-negative organisms such as Pseudomonas aeruginosa (n = 6), Enterobacter cloacae (n = 4), and Klebsiella pneumoniae (n = 3). Among 769 children with information on complications, the most common complication was radiographically confirmed pneumonia (51%); other complications included seizures (11%), encephalopathy/encephalitis (9%), and coinfection with noninfluenza viruses (5%) (Table 1).

Among 126 children who died during the 2010–2011 and 2011–2012 seasons whose treatment status was known, 56 (44%) received influenza antiviral treatment. When limiting those deaths to the 74 that occurred after hospital admission, 46 (62%) received antiviral treatment; and when limiting further to the 65 children ≥1 year of age, 43 (66%) received antiviral treatment. Of 511 children ≥6 months of age whose vaccination status was known, 84 (16%) had been fully vaccinated with seasonal influenza vaccine. During the 2009–2010 season, of 66 children ≥6 months of age whose pH1N1 vaccination status was known and who died on or after November 1, 2009, 2 (3%) had been fully vaccinated with pH1N1 monovalent vaccine.

Comparison of Influenza-Associated Deaths in Children Without and With High-Risk Medical Conditions

There were 341 (43%) children without high-risk medical conditions and 453 (57%) children with ≥1 high-risk medical condition who died; medical history was unknown for 36 children. Children without high-risk medical conditions were more likely to be <5 years old than those with high-risk medical conditions (RR: 1.3; 95% CI: 1.1–1.6) (Table 3). Those without high-risk medical conditions were more likely to die outside the hospital or in the emergency department (RR: 1.9; 95% CI: 1.6–2.4) and ≤3 days after illness onset (RR: 1.6; 95% CI: 1.3–2.0). Among children who were tested, bacterial coinfection was more common among children without high-risk medical conditions (RR: 2.0; 95% CI: 1.5–2.5).

Among those who ≥6 months of age whose vaccination status was known, 19 of 203 (9%) children without high-risk medical conditions and 64 of 290 (22%) children with high-risk medical conditions had been fully vaccinated with seasonal influenza vaccine. During the 2010–2011 and 2011–2012 seasons, children without high-risk medical conditions were less likely to receive influenza antiviral treatment than those with high-risk medical conditions, although the difference was not statistically significant (35% vs 54%; RR: 0.7; 95% CI: 0.4–1.0).