We identified cases of invasive H. influenzae disease during a 10-year period (October 1, 2008‒September 30th, 2018) from the hospital laboratory records of Auckland City Hospital, North Shore Hospital, Waitakere Hospital, and Middlemore Hospital, which provide free healthcare to the population of the Auckland region (resident population ≈1.7 million persons in 2018). We searched the computerized records of the 3 microbiology laboratories serving these hospitals to identify all patients who fulfilled the US Centers for Disease Control and Prevention criteria for H. influenzae invasive disease: isolation of H. influenzae from >1 samples collected from a normally sterile site (e.g., blood, cerebrospinal fluid, placental tissue) (14). In New Zealand, maternity care is delivered through a network of primary, secondary, and tertiary birthing facilities that, in the Auckland region, are served by these 3 microbiology laboratories. Home births are uncommon, accounting for 3.4% of births (15). All neonatal hospital-level care, such as care that would be required for neonates who have H. influenzae disease, is delivered in the study hospitals.

We reviewed electronic health records for all cases of invasive disease to identify maternal invasive H. influenzae infections, defined as case-patients from whom H. influenzae was isolated from samples from pregnant women, and neonatal invasive H. influenzae infections, defined as case-patients from whom H. influenzae was isolated from samples from infants in the first 28 days of life. Taken together, these case-patients constituted the pregnancy-associated invasive H. influenzae study population. Neonatal cases were considered early onset if H. influenzae was identified from samples taken within 48 hours of birth.

We extracted prioritized ethnicity, area-level New Zealand deprivation index (NZDep2013; index of socioeconomic deprivation based on maternal location of residence at the time of delivery) (16), maternal age, gestation, microbiologic and prespecified clinical outcome data (pregnancy outcome, death at 30 days, and death at 12 months) from the electronic health records. We grouped NZDep2013 index data into quintiles (1 = least socioeconomic deprivation area, 5 = most socioeconomically deprived area). We defined term birth as delivery at >37 weeks’ gestation. Whether H. influenzae isolates were Hib was determined by testing performed at the Invasive Pathogens Laboratory at the Institute of Environmental Science and Research (Porirua, New Zealand).

Antimicrobial susceptibility testing of all isolates was performed in the hospital laboratories by using accredited methods from the European Committee on Antimicrobial Susceptibility Testing (https://www.eucast.org) or Clinical Laboratory Standards Institute (https://www.clsi.org). We extracted data on susceptibility test results from the laboratory records for each isolate.

We categorized cases as intraamniotic infection when H. influenzae was cultured from placental tissue, products of conception, or high vaginal swab specimens for case-patients who had concurrent H. influenzae bacteremia. We categorized cases as pneumonia if the clinical diagnosis was pneumonia or if chest radiography during the same hospital admission was reported as demonstrating pneumonia. We categorized cases were as meningitis if the clinical diagnosis was meningitis or if H. influenzae was isolated from cerebrospinal fluid. We categorized case-patients who had >1 positive blood culture as having primary bacteremia when the documented clinical impression did not specify an alternative clinical syndrome (such as meningitis or pneumonia) and cases could not be otherwise categorized by other microbiologic culture results.

Birth rate and demographic data for the Auckland region during the study period were provided by Statistics New Zealand as a customized data extract to enable calculation of incidence rates. We used relative risk from univariate and multivariate Poisson regression with ethnicity, age, and NZDep2013 in a regression model to look for an association between ethnicity, age, or deprivation and pregnancy-associated invasive H. influenzae disease. We performed statistical analyses by using SAS version 9.4 (SAS Institute Inc., https://www.sas.com). The study was approved by the Auckland Health Research Ethics Committee (AHREC 000103).

We identified 54 cases of pregnancy-associated invasive H. influenzae disease: 38 (70.4%) maternal cases and 16 (29.6%) neonatal cases. In 2 pregnancies, the mother and the neonate both had invasive H. influenzae disease; therefore, the 54 index cases resulted from 52 unique pregnancies.

H. influenzae has recently been recognized as a rare but major cause of pregnancy-associated invasive disease. In this retrospective study in the Auckland region, accounting for more than one third of the New Zealand population, the overall incidence of pregnancy-associated invasive H. influenzae disease was 19.9 cases/100,000 births. Our findings build on those from England and Wales, where the incidence of invasive NTHi infection in pregnant women was 17-fold higher than in nonpregnant women and was strongly associated with preterm birth and a high case-fatality rate (7,9). However, the mechanism of adverse pregnancy outcomes was unclear; chorioamnionitis was noted in only 7.3% of cases of early-onset neonatal NTHi infection (9). Our data indicate that IAI is the probable cause of preterm birth and fetal loss; we found clinical or microbiologic evidence of IAI in 66.7% of our cases. IAI caused by H. influenzae has been noted in case reports/series previously (10,11), but not in a study of this size.

Our data confirm that outcomes of pregnancy-associated invasive H. influenzae disease for the fetus or neonate are poor. Adverse pregnancy outcomes (fetal loss, preterm birth, or birth of an infant requiring care in NICU/SCBU) occurred in 94% of pregnancies for which outcome data were available. Only 6% of pregnancies resulted in live birth of an infant not requiring NICU/SCBU care. In contrast, outcomes of pregnancy-associated invasive H. influenzae disease for the pregnant woman are generally good. There were no deaths among the 50 women for which data was available, suggesting that this condition can be readily treated by delivery of the fetus and placental tissues, plus administration of antimicrobial drugs.

We found that pregnancy-associated invasive H. influenzae infection disproportionately affected Māori persons, who experience a higher burden of many infectious diseases in New Zealand (17). Pacific women also had a higher incidence of disease than women of European or other ethnicities, but this difference did not reach statistical significance by multivariable analysis. Potentially relevant to our study are the high rates of sexually transmitted infections in young Māori and Pacific women (18). Large ethnic disparities in the incidence of common sexually transmitted infections in New Zealand have persisted, relatively unchanged, in recent years (19). In the light of increasing evidence that H. influenzae may cause nongonococcal urethritis in men (20), one possible hypothesis is that sexually acquired vagino‒cervical H. influenzae infection was the immediate precursor of IAI in the women we studied. A high incidence of sexually transmitted infections (21) and a high incidence of pregnancy-associated invasive H. influenzae disease (22) has also been observed in indigenous women in Australia, supporting this proposed mode of infection.

Our study supports IAI as the mechanism by which H. influenzae mediates poor pregnancy outcomes. Our findings suggest that IAI is responsible for the major manifestations of pregnancy-associated invasive H. influenzae disease. H. influenzae is rarely isolated from the genital tracts of pregnant women, having been found in <0.5% of samples from healthy pregnant women (23–25). We presume that H. influenzae infection of the lower genital tract of pregnant women, perhaps acquired as a sexually transmitted infection, or by some other mode of acquisition, places pregnant women at risk for ascending infection; placental infection would then be the route of infection for the fetus. Alterations in hormonal, metabolic, and immune regulation that occur during pregnancy to enable healthy fetal development might result in spread of H. influenzae infection from the vagina to the uterine cavity and increase the risk for placental infection (26).

Strengths of this study include the analysis of 10 years of data from 4 hospitals and 3 microbiology laboratories caring for demographically diverse populations, with linked outcomes of neonatal and maternal cases. Weaknesses include the retrospective study design, heterogeneity in the quality of the clinical documentation, and missing data. The true burden of pregnancy-associated H. influenzae disease, both in our study and in clinical practice, might be underestimated, given that this diagnosis relies on appropriate collection and testing of microbiologic samples. We applied a strict definition of microbiologically confirmed invasive H. influenzae disease. It is likely that some women who have signs and symptoms consistent with pregnancy-associated invasive H. influenzae disease during the study period did not have adequate microbiologic sampling to enable this diagnosis to be made. Maternal bacteremia was commonly accompanied by concurrent isolation of H. influenzae from genital tract or placental specimens, indicating that clinicians were suspicious of IAI as the cause for bacteremia in these women. In contrast, isolation of H. influenzae from maternal specimens was uncommon in cases of neonatal bacteremia, perhaps suggesting that the potential for IAI had not been recognized at the time of delivery, resulting in failure to collect appropriate specimens.

Future work should further examine the epidemiology of pregnancy-associated invasive H. influenzae disease, assessing whether incidence varies in specific populations, including other indigenous or socioeconomically deprived populations. Rates of genital tract colonization with H. influenzae should be quantified in high-risk populations, particularly Māori and Pacific women in New Zealand and indigenous women in Australia. Larger prospective studies should seek to identify factors that predispose to pregnancy-associated invasive H. influenzae disease. This approach might identify associations with other sociodemographic variables that our study lacked power to detect. Invasive H. influenzae disease should be also considered for pregnant women with signs of chorioamnionitis. Empiric antimicrobial drug treatment in this setting should be with an agent active against H. influenzae and other major maternal/perinatal pathogens. H. influenzae bacteremia in pregnancy should prompt clinicians to consider intraamniotic infection.

In our study, the overall incidence of pregnancy-associated H. influenzae invasive disease was 19.9 cases/100,000 births, similar to the national rate of early-onset neonatal group B Streptococcus sepsis in New Zealand (23 cases/100,000 live births) during 2011–2013 (27) and higher than the national rate of pregnancy-associated listeriosis in New Zealand (12.3 cases/100,000 live births) during 1997–2016 (28). The rates of early-onset group B Streptococcus and of pregnancy-associated listeriosis were not higher in those of Māori descent than in persons of European descent (27,28).

In conclusion, the rates of adverse outcomes in pregnancy-associated invasive H. influenzae disease we found were comparable with those for pregnancy-associated listeriosis; fetal loss occurred in 35.4% of cases in our study and in 34% of pregnancy-associated listeriosis cases in New Zealand (28). Comparisons across studies using different methods require caution. Nonetheless, our data indicate that, in New Zealand, the burden of H. influenzae in pregnancy might be comparable to, or higher than, that seen for pregnancy-associated listeriosis. In addition, the risk for this condition is particularly high for persons of Māori ethnicity.