Introduction

Increased mortality and morbidity in pregnant women have been observed during previous pandemics and influenza epidemics [1]–[4] and it is widely accepted that pregnancy is a risk factor for hospitalisation following influenza infection. [5], [6] The effects of maternal influenza infection are, however, not fully understood, though postulated to result from a combination of immunological and physiological changes. [7], [8].

In April 2009, an outbreak of novel influenza A(H1N1)pdm09 began in Mexico and on June 11^th the World Health Organization (WHO) declared a pandemic. The Department of Health, England, established an Influenza Clinical Information Network (FLU-CIN) at the outset in order to gather detailed clinical and epidemiological information from hospitalised A(H1N1)pdm09 cases in the UK. [9], [10] Previous reports from other, similar case series have demonstrated an increased risk of hospitalisation among pregnant women [11] but there are inconsistencies in the apparent severity of outcomes. Whilst some suggest that hospitalised pregnant women have a decreased risk of admission to intensive care [11]–[14], others report an increased risk of severe illness. [15], [16].

This study describes the characteristics of pregnant women in the FLU-CIN cohort hospitalised with confirmed A(H1N1)pdm09 infection. It compares pandemic influenza outcomes with non-pregnant women of the same age, adjusting for co-morbidities and illness severity at admission, which have not always been taken fully into account by other studies.

Methods

The FLU-CIN project is described in further detail in an earlier paper. [9] FLU-CIN was an ‘emergency’ study with a purposive sampling frame based around 13 sentinel hospitals in five clinical ‘hubs’ (Nottingham, Leicester, Imperial College London, Sheffield and Liverpool), with contributions from a further 45 non-sentinel hospitals in England and 17 in Scotland, Wales and Northern Ireland. This included five children’s hospitals and five respiratory tertiary referral centres (three with facilities for Extra Corporeal Membrane Oxygenation (ECMO)). Children’s hospitals and tertiary referral centres were not mutually exclusive; one of three ECMO centres was a children’s hospital. Trained staff reviewed the hospital case notes and electronic records of patients with A(H1N1) pdm09 confirmed by real-time reverse-transcriptase polymerase chain reaction (PCR), and completed a common, standardised data extraction form. [9], [10] From this source cohort, all women aged 15–44 years were selected (childbearing age as defined by the National Association for Public Health Statistics and Information Systems). [17] Data collected included demographic characteristics, past medical history (including co-morbidities as denoted by the Charlson’s co-morbidity index [18], [19] and obesity) vaccination history, pre-admission care, clinical presentation, requirement for high dependency (Level 2) or intensive care (Level 3)admission, gestational age at admission, and maternal and foetal outcomes. High dependency (Level 2) care was defined as patients requiring more detailed observation or intervention (including support for a single failing organ system and those ‘stepping down’ from higher levels of care),and intensive (Level 3) care was defined as patients requiring advanced respiratory support alone or basic respiratory support together with support of at least two organ systems (including all complex patients requiring support for multi-organ failure). Any missing data were coded as dummy variables and reporting of information otherwise assumed to be complete.

Anonymised data were analysed using STATA (Version 11.0, StataCorp Inc.) and univariate analyses performed using logistic regression to compare factors associated with length of hospital stay, need for higher level care (admission to Level 2 or Level 3 care), death and combined severe outcomes (defined as Level 2/Level 3 admission or death) in pregnant and non-pregnant women of similar age. Multivariable regression analysis was conducted to adjust for factors that could potentially confound the relationship between pregnancy and pandemic influenza outcomes, using a conceptual model based on a combination of expert opinion and existing evidence. [10] Potential confounding variables selected for inclusion were co-morbidities (using Charlson’s Co-morbidity Index) [18], [19], recorded obesity (not classified under Charlson), in-hospital antiviral use and severity of illness at admission (indicated by severe respiratory distress and defined as severe breathlessness (e.g. unable to complete sentences in one breath, use of accessory muscles, supra-clavicular recession, tracheal tug or feeling of suffocation), and specified in the FLU-CIN questionnaire as CAT triage criteria A [20]). Missing values were coded separately as a dummy variable for inclusion in the multivariable logistic regression model. Sensitivity analyses were performed adding pre-admission antiviral use as a covariate in the model and restricting analysis to women without underlying co-morbidities.

Results

Characteristics of Pregnant Cohort

The FLU-CIN cohort consisted of 1,520 cases with PCR confirmed A(H1N1)pdm09 infection. Three hundred and ninety-five (26%) were women aged 15–44 years, 82 of whom (21%) were pregnant, compared with an expected prevalence of 6% based on the source population. [9], [10] One additional pregnant female identified in the full FLU-CIN cohort was excluded from the current analysis as she was aged 14 years.

On admission, six of the 82 pregnant women (7%) were in the first trimester, 33 (40%) in the second and 40 (49%) in the third (missing data  = 3). Fifty-nine (72%) pregnancies were reported to be single, and one multiple (missing data  = 22). Twenty-seven (33%) were admitted to hospital during the first wave of the pandemic and 55 (67%) during the second wave. The interval between onset of symptoms and admission ranged from 0–17 days (median 2.0; interquartile range 1–4; missing data  = 29). The main findings of the univariate analysis comparing the characteristics of pregnant and non-pregnant women aged 15–44 are presented in Table 1. Pregnant women were significantly less likely to exhibit severe respiratory distress at initial hospital assessment or require supplemental oxygen on admission than non-pregnant women of a similar age.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Comparison of patient characteristics for pregnant and non-pregnant women of child-bearing age from the FLU-CIN cohort.

https://doi.org/10.1371/journal.pone.0041638.t001

Discussion

This is the only study of which we are aware to present findings amongst pregnant women hospitalised during both waves of the 2009 pandemic in the UK. In addition, it is unique in its comparison of severe outcomes for pregnant and non-pregnant hospitalised cases of childbearing age with confirmed A(H1N1)pdm09, following careful adjustment for severity of illness at admission and other confounding variables.

The data collected were not independently verified and there was no attempt to obtain missing fields or follow-up cases, which may have resulted in some reporting bias. However, there was full extraction from case records, and comparison with non-pregnant women of the same age that also had A(H1N1)pdm09 infection, which illuminates the course of A(H1N1)pdm09 in pregnancy.

The high prevalence of pregnancy in this cohort of women aged 15–44 years, aligns with previous studies which suggest a marked increase in the likelihood of admission with A(H1N1)pdm09 when compared with non-pregnant women of similar age. [11], [15] In accordance with other published case series, [12], [13], [16], [21], [22] the majority of women were in the second and third trimesters of pregnancy, suggesting an increased risk of hospitalisation in the later stages of pregnancy, although this may be confounded by a failure to recognise some early pregnancies. Pregnant women were significantly less likely to have underlying co-morbidities than non-pregnant women of childbearing age, emphasising the importance of pregnancy itself as a risk factor for hospitalisation. But despite the increased likelihood of admission for pregnant women, we found little evidence of greater illness severity at initial hospital assessment, as suggested by our findings that pregnant women were less likely to present with respiratory distress or require supplemental oxygen on admission. This does not appear to be confounded by the use of antiviral medication, since pregnant women were, if anything less likely to receive this pre-admission. Other factors such as a lower threshold, or alternative primary indication for hospital admission (such as obstetric complication) may be of more importance. The reason for admission was, however, not always reported making it difficult to explore this further by stratified analysis. Three quarters (74%) of the pregnant cohort required a hospital stay of two days or more, suggesting that ‘front door’ anxiety for the pregnant patient may not fully explain the increased admission rate.

There was no evidence of a greater likelihood of severe outcomes in pregnant women as indicated by length of stay, requirement for Level 2/3 care or mortality. This held true even when adjusted for severity of illness on admission and underlying co-morbidities, or when comparing outcomes in women with no co-morbidities, through sensitivity analysis. The burden of mortality in other cohorts of pregnant women hospitalised with A(H1N1)pdm09 ranges from 1% [14] to 6% [13] and is not dissimilar from that observed in this case series. The risk of admission to intensive care does, however, appear to vary, some case series reporting this to be lower [11]–[14] and others higher [15] when compared with infected non-pregnant women of a similar age. This may be explained by different criteria for admission to Level 2/3 care which exist in different countries, although even when considering only the most seriously ill women who required mechanical ventilation, there was still no apparent increase in the risk of death for pregnant women in this cohort. Other studies have found pregnant women with confirmed A(H1N1)pdm09 in critical care experience higher rates of viral pneumonia despite a lack of difference in mortality [23], but no statistically significant difference was observed in this cohort. Although this suggests that once hospitalised, outcomes for women of childbearing age in the FLU-CIN cohort were similar regardless of pregnancy, caution must be exercised due to the smaller number of observations and potential lack of power with which to detect a true underlying effect.

Inconsistency in the estimated size of effect of pregnancy on severity of outcomes following A(H1N1) pdm09 infection may also reflect a marked variation in the prevalence of underlying co-morbidities and residual confounding; not all reports adjusted for presence of co-morbidities in the same way.[11]–[15] It is possible that differences in the ethnic composition of the populations considered may also have a role to play.

Despite the apparent lack of effect of pregnancy on severe outcomes once hospitalised, it seems that influenza infection in pregnancy nevertheless has serious implications for maternal health since ten of the 11 pregnant women admitted to Level 2/3 care in this cohort required mechanical ventilation, and only three had co-morbid illnesses. Further studies with greater power are however required in order to definitively conclude this. Where co-morbidities were reported for pregnant women, asthma was most prevalent, which again aligns with observations in other case series. [22], [24] Other data also suggest that A(H1N1)pdm09 cases with asthma experienced a reduced likelihood of severe outcome after admission. [9], [10].

Pregnant women were quickly identified as a priority group for both early anti-viral treatment and vaccination in the UK, as elsewhere. There was widespread availability of antiviral drugs in the community following establishment of a National Pandemic Flu Service in July 2009 [25], and whilst 74% of pregnant women in this cohort were prescribed anti-viral medication only 10% (8/82) received these in the community prior to admission. It is difficult to determine whether earlier treatment would have altered the observed outcomes, but delayed treatment has been associated with an increased risk of severe illness in pregnant women in previous hospitalised case series. [14], [16], [22] Although differences in antiviral drug uptake between pregnant and non-pregnant women are small, this suggests an ongoing need to emphasise the use of early anti-viral therapy on suspicion of influenza and promote its acceptability for use during pregnancy in a pandemic situation. [26] Evidence regarding the safety of Oseltamivir in pregnancy is, however, limited [27], [28] and further studies which focus on the risk of potential adverse events would clearly be of value.

Vaccination status was inconsistently recorded in medical notes and may in part be explained by the fact that the pandemic vaccine was not introduced in the UK until late October 2009, so 54 of the pregnant women in the cohort would either not have had the opportunity to be vaccinated, or would not have seroconverted (if vaccinated), prior to illness onset. Although there were large amounts of missing data, the fact that only 7% of those for whom the pandemic vaccine would have been available in time to have offered protection were noted to have had it, compared with 25% noted not to have had it, suggests that coverage was low, as observed in other hospitalised cohorts. [29], [30] This further highlights the need to emphasise existing public health policy, and for healthcare workers, especially GPs and midwives, to advocate for vaccination against influenza during pregnancy.

Most women did not deliver during admission and were therefore not followed up as part of the FLU-CIN protocol. Of the sixteen women who did deliver during admission, three of the four who were admitted primarily with respiratory illness delivered pre-term babies and two mothers died post-partum. In comparison there was only one pre-term delivery and one maternal death amongst the ten women admitted for predominantly obstetric reasons. Adverse foetal outcomes following maternal influenza infection have previously been described, [30], [31] and although these data must be interpreted with caution, they suggest a potential short-term impact of influenza infection on foetal as well as maternal health.

Supporting Information

Acknowledgments

We gratefully acknowledge the teams who helped identify cases and collated clinical data: Alison Booth, Margaret Charlesworth, Sarah Rodenhurst, Angela Ballard and Alison Holmes at Imperial College Healthcare NHS Trust, London, UK; Sally Batham, Phayre Parkinson, Tracy Kumar and Aiden Dunphy at the University Hospitals of Leicester NHS Trust, Leicester, UK; Anne Tunbridge, Patty Hempsall, Joyce Linskill, Aimee Turner, Sharon Grindle, Dawn Shevlin and Eric Moulds at Sheffield University Hospitals NHS Trust, Sheffield, UK; Elvina White, Elaine Scott, Jennifer Cater, Erica Sergi and Helen Hill at Alder Hey Children’s Hospital NHS Foundation Trust, Liverpool, UK; Deborah Fleetwood, Lorna Roche, Sarah Dyas, and Maria Boswell at the Royal Liverpool and Broadgreen University Hospital’s Trust, Liverpool, UK; Gillian Vernon, Gillian Houghton, Heather Longworth and Angela Kerrigan at Liverpool Women’s Hospital, Liverpool, UK; Sonia Greenwood, Gemma Thompson, Emily Jarvis and Charlotte Minter at the Nottingham University Hospitals NHS Trust, Nottingham, UK; Kristina Lum Kin, Jacqueline Daglish, Sam Hayton, and Gemma Slinn at Birmingham Children’s Hospital, Birmingham, UK; Michelle Lacey, Kevin Rooney, Karen Duffy, Anne Gordon, Eleanor Anderson, Hilary Davison, William Carman, Mark Cotton, Arlene Reynolds, Heather Murdoch, Karen Voy, Rosie Hague and Ali McAllister for their contribution to FLU-CIN in Scotland. Brian Smyth and Cathriona Kearns from the National Public Health Agency, Northern Ireland for identifying cases and facilitating data collection; Teresa Cunningham at the Southern Trust and Leslie Boydell at the Belfast Trust for facilitating data collection. Alemayehu Amberbir, Safaa Al-Badri, Baraa Mahgoob and Nachi Arunachalam at the University of Nottingham for data entry and obtaining background population data; also Graham Watson for database development and support; and Tom Bewick at the Nottingham University Hospitals NHS Trust, Nottingham, UK for manually reviewing the records to identify additional pneumonia cases. We also thank Professor Sir Gordon Duff, Co-Chair of the Scientific Advisory Group for Emergencies, Professor Janet Darbyshire, who Co-Chaired the Influenza Clinical Information Network Strategy Group and Elaine Gadd and Colin Armstrong at the Department of Health, London, UK for their support and constructive remarks; Dr Shona Kelly of the University of Nottingham (now at University of South Australia) for help in developing the initial data collection tool; and Dr Patrick O’Brien of University College London Hospitals NHS Foundation Trust for assistance in further developing the data set. We thank those Chief Executive Officers, clinicians, virologists and managers, too numerous to mention, who were active in notifying cases to FLU-CIN. SJB and PJMO wish to acknowledge the support of the UK NIHR Biomedical Research Centre scheme.