Influenza hospitalizations during childhood in children born preterm

Abstract Objective The objective is to determine if children born preterm were at increased risk of influenza hospitalization up to age five. Methods National registry data on all children born in Norway between 2008 and 2011 were used in Cox regression models to estimate adjusted hazard ratios (aHRs) for influenza hospitalizations up to age five in children born preterm (<37 pregnancy weeks). HRs were also estimated separately for very preterm (<32 weeks), early term (37–38 weeks), and post‐term (≥42 weeks) children. Results Among 238,628 children born in Norway from January 2008 to December 2011, 15,086 (6.3%) were born preterm. There were 754 (0.3%) children hospitalized with influenza before age five. The rate of hospitalizations in children born preterm was 13.8 per 10,000 person‐years (95% confidence interval [CI] [11.3, 16.7]), and 5.9 per 10,000 person‐years (95% CI [5.5, 6.4]) in children born at term (≥37 weeks). Children born preterm had a higher risk of influenza hospitalization before age 5: aHR 2.33 (95% CI [1.85, 2.93]). The risk increased with decreasing gestational age and was highest among those born extremely/very preterm; aHR 4.07 (95% CI [2.63, 6.31]). Compared with children born at 40–41 weeks, children born early term also had an elevated risk of influenza hospitalization; aHR (37 weeks) 1.89 (95% CI [1.43, 2.50]), aHR (38 weeks) 1.43 (95% CI [1.15, 1.78]). Conclusion Children born preterm had a higher risk of influenza hospitalizations before age five. An elevated risk was also present among children born at an early term. Children born preterm could benefit from influenza vaccinations.

such as lung and heart disease, diabetes, neurological conditions, renal or liver failure or immuno-compromised disorders are recommended the annual influenza vaccination. Being born preterm is not identified as a risk condition in Norway. 6 The WHO defines preterm birth as children born before 37 completed weeks of gestation, and globally more than 10% of live births, around 15 million children every year, are estimated to be born preterm. 7,8 These births are associated with substantial morbidity, and one million deaths a year have been attributed preterm birth. 9 There is a disproportionately higher burden of preterm births in low-and middle-income countries. High-income countries tend to have lower rates, though in some high-income countries, such as Austria and the United States, the proportion of preterm births is higher than 10%. 10,11 Norway has one of the lowest preterm birth rates in the world (5.6%). 12 Established maternal risk factors for preterm birth include African-American/African-Caribbean ethnicity, low or high maternal age, multiple pregnancy, maternal infections and low socioeconomic status. 13 Improved treatment has increased the chance of survival in preterm infants, particularly for those born extremely preterm (before 28 weeks). 14,15 Increased survival, combined with a continuing high-and for some increasing-proportion of preterm births in many countries, has led to an growing number of children in the population who were born preterm. 10,16,17 For all children, and especially those born preterm, respiratory tract infections cause a substantial number of hospitalizations and fatalities during the first years of life. 9,18-21 Children born preterm have higher overall morbidity and mortality 22 and an increased risk of hospitalizations due to infections. 23,24 Previous studies focusing on preterm birth as a potential risk factor for infections have mostly focused on hospitalizations with respiratory infections overall or respiratory syncytial virus (RSV) infections, and not on influenza. 18,[25][26][27][28] There is a scarcity of studies investigating influenza hospitalizations among preterm children beyond infancy, and currently insufficient evidence to implement vaccine recommendations for this group. In addition, it is not known to what extent risk differs with different gestational lengths, and whether children born early term also may be at higher risk.
Our aim was to assess whether children born preterm or early term were at increased risk of being hospitalized with influenza in the first 5 years of life. By including being born preterm as a risk factor for severe influenza beyond infancy could inform decisions on including these children in annual influenza vaccination programmes around the world.

| METHODS
We included all children born in Norway between 1 January 2008 and 31 December 2011, and followed them to age five. We used data from two national health registries in Norway: the Medical Birth Registry of Norway (MBRN) 29

| Gestational age
Completed gestational weeks at birth are recorded by the MBRN and based on routine ultrasound measurements when these were available (for 98% of the children), or last menstrual period when ultrasound estimations were missing. We categorized gestational age at birth into term (≥37 completed weeks) or preterm (<37 weeks), with the latter category additionally subdivided into extremely/very preterm (<32 weeks) and moderately/late preterm (32 to <37 weeks).

| Influenza hospitalizations
An influenza hospitalization was defined as any hospitalization recorded on the NPR with the codes J09 ('Influenza due to identified zoonotic or pandemic influenza virus'), J10 ('Influenza due to identified seasonal influenza virus') or J11 ('Influenza, virus not identified') as listed in the 10th revision of the International Statistical Classification of Diseases (ICD-10). In ICD-10, laboratory confirmation is required for the J09 and the J10 diagnoses, but not for the J11 diagnosis. To reduce potential misclassification, we excluded hospital admissions outside the yearly influenza surveillance period in Norway (running from October to May). Details of data sources and categorizations are presented in Table S1.
We used Cox regression modelling to assess associations between gestational age at birth and hospital admission for influenza up to age 5, using age in days from birth as the underlying time metric.
Children were followed from birth until the first influenza hospitalization, death, emigration, or their fifth birthday. In additional analyses, we estimated the risk of influenza hospitalization in three different age groups (<1 year, 1 year and 2-4 years). A sensitivity analysis excluded multiple births. As potential confounders, we included child sex, maternal parity, maternal age, multiple birth, season of birth, parental education and maternal smoking. We conducted complete case analyses and adjusted for these covariates in our models. Dependency between siblings was taken into account by the use of robust standard errors. The proportional hazards assumption was evaluated by visual inspection of cumulative hazard curves and by testing

The Norwegian Regional Committee for Medicine and Health
Research Ethics approved this study and provided a waiver of consent for participants.

| RESULTS
There were 245,281 children registered in the birth registry of Norway between 1 January 2008 and 31 December 2011. Less than 3% were excluded due to missing covariate information (n = 4171) or invalid linkage (n = 2482), leaving 238,628 children for analyses ( Figure S1). In total, 15,086 (6.3%) of the children were born preterm (<37 weeks). Of these, 12,941 (85.8%) were moderate/late preterm (born between Weeks 32 and 36), and 2145 (14.2%) were extremely/very preterm (born before 32 weeks gestation). Preterm children were more likely to be multiple births, male, first-born, and have a younger or older mother, and a mother who smoked during pregnancy; and less likely to have at least one parent with college or university education (Table 1).
Overall, 754 (0.3%) children were hospitalized with influenza below the age of five (Table S2). Of these, 101 (13.4%) were born preterm. Compared with children born at term, the cumulative incidence of hospital admission with influenza was higher in children born preterm ( Figure 1A), and even higher among very preterm infants ( Figure 1B). The rate in preterm children was 13 Our results support previous studies finding that lower gestational age is associated with an increased risk of hospital admissions due to respiratory infections, and that risk of infection increase with lower gestational age. 19,34 A systematic review investigated a range of risk factors for children hospitalized with influenza and concluded that prematurity was one of the most important risk factors. 35 However, five of the seven studies included in this review did not define prematurity in terms of gestational age, and they were unable to assess the risk beyond the first 2 years of life. With the lack of supporting evidence, preterm children are still not defined as a risk group and prioritized for influenza vaccination.
Increased susceptibility to severe influenza could partly be explained by sequelae and comorbidities associated with preterm birth. 22 Neonates and preterm infants even more so, have an immature immune system in the first months of life. Foetal lungs develop gradually, and preterm birth interrupts the normal maturing process, but also additional factors contribute to the increased susceptibility for infections. 36,37 An estimated 40% of children born extremely preterm develop sequelae including bronco-pulmonary dysplasia (BPD), 38 which is observed to cause reduced respiratory function that persist into late adolescence and adulthood. 39,40 Increasingly lower gestational age is associated with reduced lung function. 41 Previous research primarily reports an increased risk for severe influenza for children born moderately to extremely preterm. These studies lacked sufficient detail to assess the association between the extent of prematurity and risk of hospital admission, and the need for further studies has been emphasized. 35 Some researchers have addressed overall respiratory morbidity among those born moderately/late preterm and early term. Findings support an increased risk for respiratory hospitalization for those both born late preterm (35-36 weeks) and born even early term (37-38 weeks), but these studies have not addressed risk of influenza hospitalizations in particular. [42][43][44][45][46] Sequelae (such as reduced lung function) is known to increase the risk of severe influenza, but is only present in a minority of preterm children. The majority, also those born extremely preterm, have no known bronchopulmonary dysplasia later in life. 38 The increased risk of influenza hospitalization we found in this study among the early term and late preterm is therefore not likely to be explained by long-term respiratory-related sequela.
All registry-based studies are prone to some misclassification.
According to the ICD-10 diagnostic guidelines, laboratory testing for influenza was not required for all the ICD-10 diagnoses we used to define hospitalizations with influenza. In the absence of laboratory confirmation, recording of an influenza diagnosis depends on the clinician's judgement of the clinical symptoms and patient history and therefore may be incorrect. However, 83% of the influenza hospitalizations included in our study were associated with ICD-10 J09 or J10 diagnosis, which requires laboratory confirmation. Previous studies have found that ICD-10 data underestimate rather than overestimate the numbers of influenza cases, with a high specificity for influenza diagnoses. 47 Data on laboratory testing for seasonal influenza is not recorded in national registries, and we did not have information on laboratory testing in our study. We aimed to reduce potential misclassification of influenza diagnoses by excluding hospitalizations outside the influenza surveillance period. It may be that children who were born preterm and are admitted to hospitals with influenza symptoms are disproportionately likely to be tested and diagnosed with influenza, causing a potential inflation of association. However, we believe that preterm birth is less likely to be considered a key patient characteristic when treating children for influenza beyond the infant period and also that most children with symptoms will have been tested (irrespective of preterm status).
We were able to stratify preterm births into extremely/very preterm (<32 weeks) and moderately/late preterm. However, we were unable to further stratify the former category into extremely or very preterm for anything other than descriptive analysis due to small cell counts.
Although preterm birth is highly correlated with low birth weight, our aim was to address the total effect of being born preterm (which includes lower birth weight) rather than assessing the independent effects of gestational age and birth weight, which are difficult to disentangle. 48 We did not include data on the vaccination status of the child or the mother in our study.

CONFLICT OF INTEREST
The authors have no conflict of interest to disclose.

FUNDING INFORMATION
Funded/supported by the Norwegian Institute of Public Health and by the Norwegian Research Council's Centre for Excellence (scheme #262700).

ETHICS STATEMENT
The study was approved the Regional Ethics Committee of South-Eastern Norway, #2010/2583 which waived the need for consent.
No material has been reproduced from other sources.

PEER REVIEW
The peer review history for this article is available at https://publons. Registry is intended nor should be inferred.