- Top of page
- Supporting Information
Please cite this paper as: Knight M, Pierce M, Seppelt I, Kurinczuk J, Spark P, Brocklehurst P, McLintock C, Sullivan E, on behalf of the UK’s Obstetric Surveillance System, the ANZIC Influenza Investigators, and the Australasian Maternity Outcomes Surveillance System. Critical illness with AH1N1v influenza in pregnancy: a comparison of two population-based cohorts. BJOG 2011;118:232–239.
Objective To compare admissions to intensive care units (ICUs) with confirmed AH1N1v influenza in pregnancy in Australia, New Zealand and the UK.
Design National cohort studies.
Setting ICUs in Australia, New Zealand and the UK.
Population Fifty-nine women admitted to ICUs in Australia and New Zealand in June–August 2009, and 57 women admitted to ICUs in the UK in September 2009–January 2010.
Methods Comparison of cohort data.
Main outcome measures Incidence of ICU admission, comparison of characteristics and outcomes.
Results There was a significantly higher ICU admission risk in Australia and New Zealand than in the UK (risk ratio 2.59, 95% CI 1.75–3.85). Indigenous women from Australia and women with Maori/Pacific Island backgrounds from New Zealand had the highest admission risk (29.7 admissions per 10 000 maternities, 95% CI 17.9–46.3). Women admitted in Australia and New Zealand were significantly more likely to have a pre-existing medical condition (51% versus 21%, P = 0.001), but were less likely to receive antiviral treatment (80% versus 93%, P = 0.038) than women admitted in the UK. Women admitted in the UK had a longer length of hospital stay (median 21 days, range 3–128 days) than women admitted in Australia and New Zealand (median 12 days, range 3–66 days), but there were no other differences in maternal or pregnancy outcomes.
Conclusions The difference in admission risk may reflect a second phase effect from successful clinical and public health interventions, as well as differences in population characteristics between the countries. The overall severity of the AH1N1v influenza infection in pregnancy is evident, and emphasises the importance of an ongoing immunisation programme in pregnant women in both northern and southern hemispheres.
- Top of page
- Supporting Information
Women in pregnancy and the puerperium are known to be at higher risk of complications of seasonal influenza than the non-pregnant population.1,2 With the emergence of the new pandemic AH1N1v influenza strain in April 2009, concern was raised that pregnant women would be particularly severely affected, with rapid reports of small case series of maternal hospitalisations and deaths associated with the disease.3–5 In response to these reports, specific surveillance of AH1N1v infection in pregnant women was undertaken in several countries,6–8 with most focussing on women with severe illness, i.e. those admitted to hospital or to intensive care units (ICUs). This approach of sentinel surveillance has been advocated in the pandemic situation, where it becomes impractical to identify all infected pregnant women on a population basis.9 In addition, systems were put in place to make antiviral drugs rapidly available to those with suspected infection, as well as accelerating the development and delivery of immunisation programmes.
The peak of the first wave of infection was reported in July 2009, with high numbers of infected individuals reported in southern hemisphere countries where the emergence of the infection coincided with the normal winter influenza season.10 During this time, national surveillance of all pregnant and postpartum women admitted to an ICU with confirmed AH1N1v infection was undertaken in Australia and New Zealand.6 At the outset it was widely believed that a second wave of infection would be more severe than the first wave, as this was the pattern observed in the 1918 pandemic.11 Northern hemisphere countries experienced this second wave of infection, with a peak occurring between November and December 2009, during the Northern hemisphere winter,12 and national surveillance of women hospitalised with confirmed AH1N1v influenza in pregnancy in the UK identified that admission to ICU in the UK was associated with both obesity and delayed treatment with antiviral medication.8 However, no comparisons have been undertaken between affected women from the first and second waves of infection in order to formally investigate differences in the patterns of disease. This information is important to inform planning for future pandemics, as well as to inform ongoing immunisation programmes and service planning for the forthcoming Southern hemisphere influenza season, and subsequent Northern hemisphere influenza season later in the year, during which AH1N1v is expected to remain the predominant circulating influenza strain.
The aims of this study were to formally compare the incidence of admission to ICU with confirmed AH1N1v influenza in pregnancy between Australia, New Zealand and the UK, to describe the characteristics of the cohorts of infected women, and compare the maternal and pregnancy outcomes between first-wave Southern hemisphere cases and second-wave Northern hemisphere cases.
- Top of page
- Supporting Information
Between June and August 2009 in Australia and New Zealand, 59 women were identified with confirmed AH1N1v infection, with onset in pregnancy, who were subsequently admitted to an intensive care unit in an estimated 86 449 maternities, equating to an estimated 6.8 women per 10 000 maternities, based on 2008 birth figures (95% CI 5.2–8.8).16,17 Over the 5 months between September 2009 and January 2010 in the UK, 57 women were identified with confirmed AH1N1v infection with onset during pregnancy who were subsequently admitted to an ICU, in an estimated 332 569 maternities. This equates to 1.7 women per 10 000 maternities (95% CI 1.3–2.2).14 The peak of admissions in Australia and New Zealand occurred in the week beginning 13 July 2009; the peak in the UK occurred in the week beginning 12 October 2009. In order to account for the different length of study periods, incidence was also estimated in the UK data for the 13 weeks centred on the peak week of admissions. During this time period, 51 women were admitted to ICU in the UK with confirmed AH1N1v in pregnancy in an estimated 193 796 maternities, representing an incidence of 2.6 admissions per 10 000 maternities (95% CI 2.0–3.5), a significantly lower estimated admission incidence than that in Australia/New Zealand (risk ratio 2.59, 95% CI 1.75–3.85).
Characteristics of the women admitted to ICU in Australia, New Zealand and the UK are shown in Table 1. Women admitted in Australia and New Zealand were significantly more likely to be indigenous (Aboriginal, Torres Strait Islander Maori, or Pacific Islander), and this is reflected in the population group-specific admission risks (Table 2). Women from both cohorts were of higher parity than the population of women giving birth,8 but there were no differences in parity between the cohorts. Australian and New Zealand women were more likely to be older or younger (Figure 1) than women admitted in the UK, and were more likely to have a reported pre-existing medical problem, although this difference was entirely explained by a difference in the reported diagnoses of pre-existing asthma. Women admitted in Australia and New Zealand had a greater number of reported co-existing illnesses than those admitted in the UK (Table 1). This was not solely the result of differences in the reported diagnoses of asthma: 7% of women in Australia and New Zealand had two or more reported co-morbidities other than asthma, compared with 2% of women in the UK (P = 0.065). Women admitted in Australia and New Zealand were less likely to be in their third trimester of pregnancy at admission, and were less likely to be treated with antiviral medication. However, there were no significant differences in the proportions of women managed with antiviral medication within 2 days of symptom onset, nor in the median symptom onset–antiviral treatment interval, amongst those who received antiviral medication, between women from Australia or New Zealand and the UK.
Table 1. Characteristics of women admitted to critical care with AH1N1v in pregnancy in Australia, New Zealand and the UK
|Characteristic||Australia and New Zealand n, %* (n = 59)||UK n, %* (n = 57)||P (χ2)|
|<20||7 (12)||3 (5)||0.009|
|20–34||36 (61)||49 (86)|| |
|≥35||16 (27)||5 (9)|| |
|White||34 (58)||49 (86)||0.001**|
|Other (excluding indigenous)||6 (10)||8 (14)|| |
|Indigenous (Aboriginal, Torres Strait Islander, Maori or Pacific Islander)||19 (32)||N/A|| |
|<25 kg/m2||17 (31)||17 (33)||0.85|
|25–29.9 kg/m2||14 (26)||15 (29)|| |
|≥30 kg/m2||23 (43)||19 (37)|| |
|Pre-existing medical conditions|
|Asthma||18 (31)||5 (9)||0.003|
|Asthma requiring inhaled or oral steroids||5 (9)||4 (7)||0.77|
|Chronic lung disease||1 (2)||1 (2)||1.00***|
|Haematological disorder||4 (7)||1 (2)||0.36***|
|Cardiac disease||5 (9)||0 (0)||0.057***|
|Metabolic disorder||1 (2)||4 (7)||0.20***|
|Neurological condition||1 (2)||2 (4)||0.62***|
|Gastrointestinal disorder||3 (5)||0||0.24***|
|Endocrine disorder||1 (2)||0||1.00***|
|Essential hypertension||2 (3)||1 (2)||1.00***|
|Other||5 (8)||2 (4)||0.44***|
|Any pre-existing medical problem|
|No||29 (49)||45 (79)||0.010|
|Yes||30 (51)||12 (21)|| |
|Number of reported co-morbidities|
|0||29 (49)||45 (79)||0.003****|
|1||22 (37)||8 (14)|| |
|2||6 (10)||4 (7)|| |
|3||1 (2)||0 (0)|| |
|4||1 (2)||0 (0)|| |
|0||18 (31)||19 (33)||0.95|
|1||17 (29)||17 (30)|| |
|2+||23 (40)||21 (37)|| |
|Trimester of admission|
|1st/2nd||17 (30)||7 (13)||0.021|
|3rd||39 (70)||49 (88)|| |
|Yes||1 (2)||0 (0)||1.00***|
|No||58 (98)||57 (100)|| |
|Treated with antiviral medication|
|Yes||47 (80)||53 (93)||0.038|
|No||12 (20)||4 (7)|| |
|Treated within 2 days of symptom onset|
|Yes||6 (10)||9 (17)||0.29|
|No||53 (90)||44 (83)|| |
|Symptom onset–antiviral treatment interval (median, range)||7 (0, 37)||6 (0, 67)||0.095*****|
Table 2. Population group-specific admission risks in Australia, New Zealand and the UK
|Population group||Estimated denominator number of maternities in study period||Number of women admitted||Admission risk per 10 000 maternities (95% CI)||Risk ratio [95% CI]|
|White (UK)||157 925||43||2.7 (2.0–3.7)||Ref|
|Other (UK)|| 35 871||8||2.2 (1.0–4.4)||0.82 (0.33–1.76)|
|White (Aus/NZ)|| 67 324||34||5.1 (3.5–7.1)||1.85 (1.15–2.98)|
|Other excluding indigenous* (Aus/NZ)|| 12 719||6||4.7 (1.7–10.3)||1.73 (0.60–4.09)|
|Indigenous* (Aus/NZ)|| 6406||19||29.7 (17.9–46.3)||10.9 (6.00–19.1)|
Figure 1. Age at admission for women admitted to intensive care with AH1N1v during pregnancy in Australia, New Zealand and the UK.
Download figure to PowerPoint
The outcomes for women and their infants were very similar between the two cohorts (Table 3). The only statistically significant difference was in the length of overall hospital stay, which was longer in the UK cohort. Six women died in the Australian and New Zealand cohort following ICU admission with AH1N1v in pregnancy, a rate of 6.9 deaths per 100 000 maternities (95% CI 2.6–15.1). In the peak 13 weeks of the study period, four women died in the UK cohort following ICU admission with AH1N1v in pregnancy, a rate of 2.1 deaths per 100 000 maternities (95% CI 0.6–5.3). This was not statistically significantly different from the rate in Australia and New Zealand (risk ratio 3.36, 95% CI 0.80–16.2, P = 0.11), but note that small numbers give this comparison limited power to detect statistically significant differences.
Table 3. Maternal and pregnancy outcomes of women admitted to intensive care with AH1N1v in pregnancy in Australia/New Zealand and the UK
|Outcome||Australia and New Zealand n, %* (n = 59)||UK n, %* (n = 57)||P (χ2)|
|No||53 (90)||53 (93)||0.74**|
|Yes|| 6 (10)|| 4 (7)|| |
|Additional maternal morbidity***|
|No||33 (63)||31 (58)||0.69|
|Yes||20 (38)||22 (42)|| |
|Extracorporeal membrane oxygenation|
|No||50 (85)||47 (82)||0.81|
|Yes|| 9 (15)||10 (18)|| |
|Length of hospital stay (days)***|
|1–3|| 8 (16)|| 1 (2)||0.039|
|4–7|| 9 (18)||11 (22)|| |
|7–14||14 (27)|| 9 (18)|| |
|15–21|| 7 (14)|| 6 (12)|| |
|22+||13 (25)||24 (47)|| |
|Median [range]||12 [3, 66]||21 [3, 128]||0.034*****|
|Length of ICU stay (days)****|
|1–3||13 (25)||15 (37)||0.48|
|4–7|| 7 (14)|| 8 (20)|| |
|7–14||13 (25)|| 6 (15)|| |
|15–21|| 8 (16)|| 7 (17)|| |
|22+||10 (20)|| 5 (12)|| |
|Median [range]||10 [1, 65]|| 6 [1,62]||0.078|
|Livebirth||51 (89)******||45 (88)||0.50*|
|Miscarriage|| 2 (4)|| 0 (0)|| |
|Stillbirth|| 4 (7)|| 5 (10)|| |
|Termination|| 0 (0)|| 1 (2)|| |
|Data not received (ongoing pregnancy at discharge)|| 3|| 6|| |
|No||47 (96)||45 (100)||0.50*|
|Yes|| 2 (4)|| 0 (0)|| |
|<2500g||17 (40)||21 (42)||0.88|
|≥2500g||25 (60)||29 (58)|| |
|Gestation at delivery|
|<32 weeks||11 (24)||14 (28)||0.65|
|<37 weeks||19 (49)||31 (62)||0.21|
|≥37weeks||20 (51)||19 (38)|| |
|Other infant morbidity***|
|No||39 (80)||37 (82)||0.75|
|Yes||10 (20)|| 8 (18)|| |
- Top of page
- Supporting Information
There was a clear difference between the risk of admission to ICU with confirmed AH1N1v in pregnancy in Australia and New Zealand during the first wave of the pandemic and the risk of admission to ICU with confirmed AH1N1v in pregnancy in the UK during the second wave of infection in the Northern hemisphere winter, as estimated from these two national cohort studies. The immunisation programme for pregnant women in the UK was rolled out after the peak of infections occurred among the women in this study, and therefore differences in immunisation status between the women cannot account for the differences. There are several other potential explanations for this observed difference.
The difference between the two groups is an artefact, resulting from differential case ascertainment methods in the two cohorts.
There was a higher overall incidence of AH1N1v infection in the Australian and New Zealand population, hence a greater proportion of pregnant women were infected, and thus a greater number developed severe disease, although this was the same proportion of those infected.
Population differences between Australia, New Zealand and the UK have led to a higher susceptibility to infection and/or greater disease severity among Australasian women.
The threshold for admission to ICU may differ between Australia, New Zealand and the UK.
The difference results from a second-phase effect: i.e. direct and indirect public health campaigns raised awareness of the potential for infection and the severity in pregnancy among the public and practitioners, leading to earlier diagnosis and more aggressive intervention in the second-phase, Northern-hemisphere epidemic, resulting in a lower incidence of severe cases in the northern hemisphere.
A change in virulence of the infecting organism has occurred between the first and second waves of the pandemic, such that the disease is less severe in the second wave.
Cases were identified using different mechanisms in the two cohorts, thus differential case ascertainment is a possibility. The Australian and New Zealand cases were identified through a network of reporting clinicians in ICUs, whereas the UK cases were reported through a network of reporting midwives, obstetricians and anaesthetists in consultant-led maternity units. There is thus the possibility that women admitted to ICUs were under-ascertained in the UK cohort, although preliminary cross-checking with the Intensive Care National Audit and Research Centre (ICNARC) in the UK suggests that under-reporting, although present to some extent, is very unlikely to fully account for the greater than two-fold difference in admission risk. This therefore suggests that there is a true difference in admission risk between the two cohorts.
Reporting systems for influenza and influenza-like illness vary between the three countries. In New Zealand and the UK, sentinel surveillance systems report consultation rates with influenza-like illness. In both countries at the peak of the AH1N1v epidemics, the reported consultation rate for influenza-like illness was just under 300 per 100 000 population per week.20,21 At the peak of the epidemic, an estimated 50% of influenza-like illnesses were caused by AH1N1v,20 but it has also been reported from serological studies that the actual number of infections occurring in the UK was up to ten times that estimated through surveillance of influenza-like illness.22 In contrast to the UK and New Zealand, influenza is a notifiable disease in Australia. There were 1274 confirmed influenza cases reported per 100 000 population at the height of the epidemic in July 2009.23 This equates to an approximate weekly rate of 290 cases per 100 000 population, very similar to the reported rate of influenza-like illness in New Zealand and the UK. It does not appear, therefore, that the difference in ICU admission risk in pregnancy can be solely the result of differential population infection rates.
Are there then any population differences that might account for the higher admission risk in Australia and New Zealand? The clearest difference in demographic characteristics lies in which vulnerable populations were most affected. Over 40% of the women admitted to intensive care units in Australia and New Zealand were from Aboriginal, Torres Strait Islander, Pacific Islander or Maori groups. Other studies have highlighted a between three- and six-fold increase in the risk of severe disease and death among indigenous populations,24,25 and the relative increase in risk of hospitalisation, admission to ICU and death among indigenous Australians with AH1N1v has been estimated to be very similar to the relative increases in risk associated with pregnancy.26 This difference alone is therefore likely to explain a significant proportion of the observed difference in incidence of admission.
The presence of co-existing illnesses has been shown to be associated with severe disease and death from AH1N1v in pregnancy.7,27 The women admitted to ICU in Australia and New Zealand appeared to have a higher number of co-existing illnesses than those admitted to ICU in the UK. Indigenous women are known to have more co-existing illnesses and pregnancy complications than other Australian and New Zealand women, which may partly account for this.28 In the UK population obesity was associated with an increased risk of admission to ICU with confirmed AH1N1v in pregnancy,8 but there are no clear differences in body mass index (BMI) between the Australasian and UK cohorts in this analysis, and no national data to suggest a difference in obesity in pregnancy at a population level. A greater proportion of women were older (35 years or above) or younger (<20 years old) in the Australian and New Zealand cohort. Indigenous women who are pregnant are more likely to be younger,29 which may explain this bimodal age distribution.
The enormous media and medical coverage of the pandemic resulting from the direct and indirect public health campaigns led by the World Health Organisation and governments resulted in a culture of heightened awareness of the potential for infection, the course of the illness, and the severity in pregnancy among the public and health professionals.30 These factors are likely to have influenced both pregnant women’s health-seeking behaviour and the earlier diagnosis and more aggressive management of influenza by clinicians in the second-phase, northern-hemisphere epidemic. The suggestion of more interventionist management by UK clinicians is supported by our observation that a higher proportion of women in ICU in the UK were managed with antiviral medication compared with the southern-hemisphere cohort. It is possible that a number of infections in pregnant women in the UK were reduced in severity as a result of the policy of early antiviral administration, effectively preventing ICU admissions.31 Alternatively, the stressors on the health sector associated with the early phases of the epidemic in the southern hemisphere may have led to a lower threshold effect for admission to ICU of pregnant women before the natural history of the H1N1 strain of influenza was fully described. Australia and New Zealand have more than twice the number of ICU beds per million population compared with the UK, which would allow for a lower threshold for admission. This would have resulted in a lower incidence of severe cases, as defined by admission to ICU in the northern hemisphere, and may explain the shorter length of stay in Australian and New Zealand ICUs compared with the UK, with the southern-hemisphere cohort including a wider spectrum of morbidity, with a greater proportion of women staying 1–3 days in hospital in Australia and New Zealand compared with the UK. If a threshold effect was an important explanation for the difference in admission risk, however, we would expect to see worse outcomes among the UK cohort, and we have not demonstrated this. We thus believe a threshold effect is unlikely, whereas a second-phase effect, with the northern hemisphere benefitting from clinical and public health guidance introduced as a result of the first-phase experience, is very likely to be part of the explanation for the observed difference in admission risk.
There are clear population and management differences between Australia, New Zealand and the UK that may lead to a greater severity of AH1N1v infection to account for the observed higher admission risk. However, despite the presence of factors associated with greater disease severity in the Australian and New Zealand cohort, there were no significant differences in either maternal or pregnancy outcomes between the cohorts, and this would argue against an increase in virulence of the infecting organism between the first and second waves of infection. There remains the possibility, however, that the virus changed in some way between the first wave and second wave of infection. There are no current data to support this assertion,32 but nevertheless we are unable to exclude it as a possibility.
Overall, nearly 10% of women admitted to ICU with AH1N1v in pregnancy died, over 10% suffered pregnancy losses and more than 50% of the remaining women delivered preterm, emphasising the severity of the disease in pregnancy. Women in both cohorts experienced lengthy stays in both ICU and hospital, reflecting a considerable burden to hospital services. Women in the UK cohort tended to remain in hospital for a longer period, although the reasons for this remain unclear. It is unlikely to reflect differences in the health systems, as most ICUs are located in public hospitals in all three countries, and is similarly unlikely to reflect geography, as women in Australia and New Zealand are more likely to live a longer distance away from a medical facility, and may thus be less likely to be discharged early. It is possible that the observed difference in length of stay reflects the difference in trimester of admission in the UK and Australian and New Zealand women. Women in the UK cohort were more likely to be admitted in the third trimester, and therefore were more likely to have been delivered during their admission, which may have effectively lengthened their stay, particularly if they were delivered preterm. Nevertheless, the observed difference in length of stay still merits further exploration to investigate the reasons behind it, and any associations with disease outcomes, in order to help inform future service planning.
- Top of page
- Supporting Information
The first wave of the AH1N1v influenza pandemic among pregnant women in Australia and New Zealand was associated with a higher risk of admission to intensive care compared with the second wave of the pandemic in the UK in late 2009. This difference may reflect successful public health interventions in the UK, including earlier and more comprehensive use of antivirals, introduced after the severity and clinical course of the disease in pregnancy became clear from the southern-hemisphere experience. It is unlikely to result from differences in case ascertainment or overall infection rate. It is most likely to be explained by differences in the characteristics of vulnerable populations between the nations, notably in women with pre-existing morbidity and indigenous women. These findings have clear implications for the planning of services in Australia and New Zealand in both the forthcoming influenza season and for future pandemics.
This analysis did not suggest that there was any change in disease severity between the first and second waves of the pandemic. However, the study reiterates the overall severity of AH1N1v influenza infection in pregnancy, and emphasises the importance of an ongoing immunisation programme in pregnant women in both northern and southern hemispheres.
Disclosure of interests
The authors declare that they have no conflicts of interest.
Contribution to authorship
MK and MP undertook the analysis and wrote the first draft of the paper. All members of the writing committee contributed to the study design, writing and editing of the paper, and the decision to submit for publication.
Details of ethics approval
The ANZICII/AMOSS study was approved by the institutional ethics committee of each centre, and the requirement for individual subject informed consent was waived at all sites. The UKOSS study was approved by the County Durham & Tees Valley 1 Research Ethics Committee (study reference 09/H0905/66).
The ANZIC Influenza Investigators registry is supported by: the Department of Health and Ageing, Commonwealth Government of Australia; New South Wales Health, Government of New South Wales; Department of Health, Government of Victoria; the Australian and New Zealand Intensive Care Research Centre; the Australian and New Zealand Intensive Care Society; and an unrestricted grant from CSL, Melbourne, Victoria. The Australasian Maternity Outcomes Surveillance System is supported by National Health and Medical Research Council (Australia) project grant, no. 510298. The UK Obstetric Surveillance System study was funded by a grant from the National Institute for Health Research Health Technology Assessment Programme. MK is funded by the National Institute for Health Research National Coordinating Centre for Research Capacity Development. None of the funders had any role in the study design and the collection, analysis, and interpretation of data, or in the writing of the article and the decision to submit it for publication. The researchers confirm their independence from funders and sponsors.
This study would not have been possible without the contribution and enthusiasm of the ANZIC, AMOSS and UKOSS reporting clinicians, who notified cases and completed the data-collection forms. We would also like to acknowledge the UKOSS Steering Committee, ANZIC Influenza Investigators and AMOSS investigators (see Appendix S1 for a full list of names).