Influenza in Infants Born to Women Vaccinated During Pregnancy

Discussion

This study reports infant influenza outcomes in a large cohort of infants born to women self-reporting influenza immunization during pregnancy before and after the H1N1 pandemic. The overall proportion of pregnant women reporting influenza immunization was low but increased significantly after the H1N1 pandemic. Infants born to pregnant women who reported receiving influenza vaccine during pregnancy were significantly less likely to have documented health care encounters associated with ILI (64% reduction), and they had a 70% reduction in laboratory-confirmed influenza and an 81% reduction in hospitalization with laboratory-confirmed influenza. Furthermore, in this mother–infant cohort, 97% of all laboratory-confirmed influenza cases occurred in infants born to women who did not report receiving influenza vaccine during pregnancy.

Only 10% of women in this cohort spanning 9 influenza seasons reported influenza immunization, but the proportion of pregnant women reporting influenza immunization increased significantly after the H1N1 pandemic, with a more pronounced recent increase, reaching a high of 52% in the 2013 to 2014 influenza season. Nationally, prepandemic H1N1 National Health Interview Survey seasonal influenza coverage estimates for 2008 to 2009 were only 11%.²⁰ During the 2009 to 2010 flu season, reported seasonal influenza immunization coverage among pregnant women increased to 32.1% as measured by the National 2009 H1N1 Flu Survey and 47.1% as measured by the Pregnancy Risk Assessment Monitoring System.^16,17 National vaccination coverage of pregnant women has remained relatively stable since the 2010 to 2011 season, with reported immunization coverage during pregnancy of 34.6% during the 2013 to 2014 influenza season.^18,19 We observed a similar pattern. The higher rates in recent seasons may be associated with efforts by ACOG and CDC to improve influenza coverage in pregnant women. However maternal influenza immunization rates during pregnancy remain suboptimal.

In our cohort we also found that demographic characteristics differed between pregnant women who did and did not report receiving the influenza vaccine. Uninsured pregnant women and those with government insurance were less likely to report receiving influenza immunization during pregnancy. A recent Massachusetts Pregnancy Risk Assessment Monitoring System survey demonstrated that pregnant women who received Medicaid or whose household income was at or below the federal poverty level were significantly less likely to report receiving the seasonal influenza vaccine.²¹ Pregnant women living in areas classified as rural or frontier or who received their care in the Urban South or Southwest Intermountain regions were also significantly less likely to report immunization. The Intermountain Urban South region includes Utah County, 1 of 30 US counties with the largest estimated numbers of unvaccinated children from 1995-2001 CDC National Immunization Surveys (NIS) data.²² It is possible that factors leading to parental vaccine hesitancy in children may similarly affect pregnant women considering maternal immunization during pregnancy. We also documented that women with underlying medical conditions were more likely to receive influenza immunization. These women may have been recognized as having high risk for influenza complications by their providers and therefore been encouraged to receive influenza vaccine. Interventions that target both healthy pregnant women and those with chronic conditions are needed to increase vaccine uptake.

Infants born to women who reported receiving the influenza vaccine during pregnancy in our cohort were significantly less likely to have a diagnosis of or be hospitalized with ILI or laboratory-confirmed influenza. Several previous retrospective studies have not successfully demonstrated protection against influenza disease in infants. A retrospective study over 5 influenza seasons (1997–2002) concluded that infants born to women who received influenza immunization during pregnancy had an equal risk for influenza or pneumonia admissions and ILI compared with infants born to unimmunized women.¹⁶ Another retrospective cohort study conducted from 1995 to 2001 by the CDC Vaccine Safety Datalink found no reduction in medical visits for respiratory illness among infants of women receiving influenza immunization during pregnancy compared with infants born to unimmunized women.^2,17 These previous studies were performed in smaller cohorts over fewer influenza seasons and did not examine laboratory-confirmed influenza diagnoses in infants. Earlier studies have also demonstrated that ILI can be an insensitive predictor of laboratory-confirmed influenza infection, particularly in young children, leading to misclassification bias.²³ ILI at Intermountain facilities is reported with an influenza-specific ICD-9-CM diagnostic code algorithm that correlates well with laboratory-confirmed influenza, which may reduce misclassification in our study.²⁴ More importantly, the widespread use of respiratory viral diagnostic testing in young infants for many years at Intermountain outpatient facilities allowed us to examine the impact of maternal influenza immunization status on infant laboratory-confirmed influenza and hospitalization for laboratory-confirmed influenza. Strikingly, the receipt of influenza vaccine had no effect on RSV outcomes, increasing confidence that the administration of influenza vaccine during pregnancy resulted in specific benefits for infant health outcomes related to influenza.

The results of this large retrospective study support the conclusions of prospective studies regarding the protective benefit of maternal influenza immunization during pregnancy. The morbidity and mortality associated with influenza in infants have been well described.^25,26 Our study strengthens the evidence that maternal immunization provides passive protection against influenza to infants during the vulnerable period before they are old enough to receive active immunization.

One of the strengths of our study is our ability to systematically identify self-reported influenza vaccine in a large retrospective cohort of pregnant women presenting for delivery. It is possible that self-reported influenza immunization at time of delivery is limited by recall bias. We believe that maternal recall of receipt of influenza vaccine is likely to be accurate. In an earlier study conducted in Intermountain facilities, we demonstrated that maternal recollection of receiving intrapartum antibiotics was 85% accurate and that the 15% of women whose recall differed from the medical record were as likely to overreport as underreport receiving treatment.²⁷ In our study of influenza vaccine, if women who did or did not receive the vaccine were misclassified, the results would have been biased toward the null hypothesis.

Another strength of our study was the success of the maternal–infant record linkage. Our ability to link most of the maternal–infant pairs in our cohort allowed us to capture infant influenza-related outcomes through 6 months by using ICD-9-CM codes. The widespread use of viral diagnostic testing in the Intermountain system is another strength of our study and allowed accurate classification of infants with influenza.

Our study has several limitations. Maternal influenza immunization was self-reported and overall low during the study period compared with current immunization rates in pregnant women. Because maternal immunization status could be confirmed only by self-report at time of delivery, we do not know the exact timing of immunization, which limited our ability to accurately predict the exact timing and duration of protection afforded to the infant by influenza season or year of maternal immunization. Furthermore, we did not know the specific influenza strains women were vaccinated against in our cohort and were therefore unable to match the maternal vaccine strain with the influenza strain observed in the infants, as is typically done in prospective vaccine trials designed to measure vaccine efficacy. However, despite low maternal immunization rates and the inherent limitations of retrospective self-reported immunization data, ours was 1 of the largest retrospective maternal–infant cohorts to date.

Finally, many potential unmeasured factors may have influenced the magnitude of the risk of influenza infection in the infants. It is possible that some of the benefits to the infant attributed to maternal immunization may have resulted from the benefit of cocooning from household contacts immunized against influenza or from disease prevention and care-seeking behaviors in women who did and did not report immunization. Health care provider decision-making regarding testing infants for influenza may also have been influenced by the provider’s awareness of maternal immunization status. Variations in vaccine effectiveness over time may also have influenced the risk of influenza infection in infants. Despite these limitations, the significant decrease in the burden of influenza illness in infants born to women who reported receiving influenza immunization during pregnancy is reassuring. Our results also support the need to incentivize system-level efforts to improve influenza immunization coverage rates in pregnant women, particularly in those with government insurance or no insurance.