Multi‐decade national cohort identifies adverse pregnancy and birth outcomes associated with acute respiratory illness hospitalisations during the influenza season

Abstract Background Despite the World Health Organization (WHO) recommendation that pregnant women be prioritised for seasonal influenza vaccination, coverage in the Western Pacific Region remains low. Our goal was to provide additional data for the Western Pacific Region about the value of maternal influenza vaccination to pregnant women and their families. Methods We conducted a 16‐year retrospective cohort to evaluate risks associated with influenza‐associated maternal acute respiratory infection (ARI) in New Zealand. ARI hospitalisations during the May to September influenza season were identified using select ICD‐10‐AM primary and secondary discharge codes from chapter J00–J99 (diseases of the respiratory system). Cox proportional hazards models were used to calculate crude and adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs). Results We identified 822,391 pregnancies among New Zealand residents between 2003 and 2018; 5095 (0.6%) had ≥1 associated ARI hospitalisation during the influenza season; these pregnancies were at greater risk of preterm birth (aHR 1.50, 95% CI 1.39–1.61) and low birthweight (aHR 1.64, 95% CI 1.51–1.79) than pregnancies without such hospitalisations. We did not find an association between maternal ARI hospitalisation and fetal death (aHR 0.96, 95% CI 0.69–1.34) during the influenza season. Maternal influenza vaccination was associated with reduced risk of preterm birth (aHR 0.79, 95% CI 0.77–0.82), low birthweight (aHR 0.87, 95% CI 0.83–0.90) and fetal death (aHR 0.50%, 95% CI 0.44–0.57). Conclusion In this population‐based cohort, being hospitalised for an ARI during the influenza season while pregnant was a risk factor for delivering a preterm or a low birthweight infant and vaccination reduced this risk.


| BACKGROUND
Pregnant women with influenza are at increased risk of hospitalisation and intensive care unit (ICU) admission than their non-pregnant counterparts. [1][2][3][4] In addition, pregnant women with influenza may be at increased risk of pregnancy-related complications and poor birth outcomes compared with pregnant women without influenza. 5,6 Two studies from the 2009 pandemic suggested that hospitalised maternal cases were at increased risk of preterm birth. 7,8 Whereas these studies observed a significant risk of influenza-associated hospitalisation among pregnant women, several others did not. 9,10 In 2017, Fell et al 11 systematically reviewed studies that reported on maternal influenza and birth outcomes for the World Health Organization (WHO). They concluded that additional high-quality studies about the risk of preterm birth, small for gestational age and fetal death following maternal influenza virus infection were needed. 12 Insufficient information about the value proposition for influenza vaccination of pregnant women hampers many vaccination programmes. For example, despite the WHO recommendation that pregnant women be prioritised for seasonal influenza vaccination, coverage in the Western Pacific Region remains suboptimal. 13 Some of the most significant barriers to immunisation during pregnancy are a lack of information about the benefits and safety of vaccination, insufficient buy-in from pregnant women's healthcare providers and barriers to preventive care. 13,14 To substantiate the value of influenza vaccination in preventing more than respiratory illnesses among pregnant women and their infants, we sought to (1) quantify the association between acute respiratory infection (ARI) hospitalisation during the influenza season and adverse birth outcomes and (2) measure the effects of influenza vaccination on pregnancy outcomes. This study was designed as a retrospective cohort study to minimise the potential for bias and to recruit a large enough sample size to obtain statistical power for analysis.
Findings from this study can be used by maternity services and antenatal healthcare providers in risk communication about maternal influenza vaccination.  (Table 1). Immunisation data from Proclaims, a dataset that holds fee-for-service payments made to general practitioners for patient visits in New Zealand, were also used. 13  For the mothers in the study, we first identified all hospital admissions at any gestation with primary and secondary ICD-10-AM discharge codes from chapter J00-J99 (diseases of the respiratory system). Second, we excluded hospitalisations due to chronic respiratory conditions (ICD-10-AM codes J30-39, J60-70, J80-84, J90-94, J95 and J97-99) 16 that might not have been precipitated by ARI.
Finally, the exposed cohort consisted of pregnancies with at least one hospitalisation with primary and secondary ARI ICD-10-AM discharge codes that took place during the May to September influenza season when ARIs were most likely to be associated with influenza illness. 17 For simplicity, we will henceforth refer to a maternal ARI hospitalisation that occurred during the influenza season as a 'maternal ARI hospitalisation'. The unexposed referent cohort was selected from the same population as the exposed cohort but consisted of pregnancies with no history of ARI hospital admission (per select ICD-10-AM

| Supplemental analysis
We conducted a supplemental analysis to examine the potential Time contributed for the exposed cohort (influenza vaccinated) started at whichever of the following events occurred last: start of influenza season, date of vaccination or start of pregnancy. Time contributed for the exposed cohort ended at whichever of the

| Risk of maternal ARI hospitalisation during influenza season
Only 0.6% (n = 5095) of the pregnancies had at least one maternal ARI hospitalisation during the influenza season and 99.4% (n = 817,296) did not (Figure 1). The risk of maternal ARI varied significantly across the study population by baseline characteristics (

| Maternal ARI hospitalisations and risk of adverse pregnancy and birth outcomes
Maternal ARI hospital admission during influenza season was associated with a greater risk of adverse pregnancy and birth outcomes in multivariate adjusted models compared with pregnancies without maternal ARI hospitalisation during the influenza season (

| Possible protective effect of maternal influenza vaccination in the prevention of adverse outcomes
We found an association between influenza vaccination status and adverse pregnancy and birth outcomes ( Our finding that influenza-vaccinated mothers had almost half the incidence of fetal death compared with unvaccinated mothers is consistent with that observed in Australia. 31 The proportion of stillbirths that can be attributed to influenza infection remains unknown so it is unclear how many fetal deaths could be prevented with maternal vaccination. Influenza vaccination during pregnancy decreased the risk of preterm birth and low birthweight in our cohort. In a systematic review and meta-analysis of five studies, Nunes et al observed a similar protective effect between seasonal influenza vaccination during pregnancy and adverse birth outcomes. 32 A pooled analysis of three randomised control trials found no association between maternal vaccination and stillbirth, preterm birth, low birthweight and small for gestation age although it did find that maternal influenza immunisation was protective against low birthweight (RR 0.85, 95% CI 0.74-0.96) in one of the sites. 33 Our findings can be used by maternity services and antenatal healthcare providers to develop risk communication messages about the value of maternal influenza vaccination in preventing influenza illnesses and their complications among pregnant women and their infants. 34 Despite the WHO recommendation that pregnant women be prioritised for seasonal influenza vaccination, coverage in the Western Pacific Region remains suboptimal. 13,35,36 A retrospective cohort study using National Collection datasets found that although influenza vaccination seems to be increasing among pregnant women in New Zealand, only 31% had been vaccinated in 2018. 13 Studies have documented that influenza vaccines are safe and effective at preventing influenza in pregnant women and infants. 33 This highlights the importance of conveying the value of maternal influenza vaccination to families in New Zealand and ensuring easy access to vaccines.
This study used data on a national cohort of pregnant women over the span of 16 years available from New Zealand's Ministry of Health National Collection datasets. Our findings from this retrospective cohort analysis indicate that maternal ARI hospitalisations during the influenza season are associated with preterm birth and low birthweight and that influenza vaccination may help prevent these outcomes. Moreover, we found that maternal influenza vaccination was linked to lower rates of fetal death. Findings that support the morbidity, and fetal and infant mortality associated with ARI and influenza infection, along with the benefits of influenza vaccine in this group, are useful to better establish the value proposition of influenza vaccines in preventing respiratory illnesses and related complications.
The value of influenza vaccination, in this case, goes beyond preventing mostly self-limited ARIs and seems to extend to the prevention of expensive and potentially deadly birth outcomes during influenza epidemic periods. Improving understanding, access and provider recommendations relating to maternal influenza immunisation in New Zealand could prove useful in improving pregnancy and birth outcomes.