Severity of maternal SARS-CoV-2 infection and perinatal outcomes of women admitted to hospital during the omicron variant dominant period using UK Obstetric Surveillance System data: prospective, national cohort study

Design, data sources, and study period

A national, prospective cohort study was conducted using data from the UK Obstetric Surveillance System.14 This system entails active surveillance with reporting from all 194 hospitals in the UK with a consultant led maternity unit and includes well established routines to secure complete reporting.15 Information about women who died or who had stillbirths or neonatal deaths was crosschecked with data from the organisation responsible for maternal and perinatal death surveillance in the UK (MBRRACE-UK).16 As individual level data for SARS-CoV-2 variants were not recorded in medical records, we restricted inclusion to the period in which the omicron SARS-CoV-2 variant was the dominant circulating strain in the UK (15 December 2021 to 14 March 2022). Pregnancy outcomes were included in the primary report and in supplementary data received by UK Obstetric Surveillance System before 19 April 2022. The start cut-off date of 15 December was chosen because the variant then represented 50% or more of sequenced new infections.13

Study population and study groups

Women were included if they were admitted to hospital during pregnancy and had a positive SARS-CoV-2 polymerase chain reaction (PCR) test in seven days or fewer before admission, during admission or up to two days after giving birth. Hospital admission was defined as a stay in hospital overnight or longer for any cause, or an admission of any duration to give birth. Women who did not meet this case definition were excluded (figure 2). We categorised women who were included into two mutually exclusive groups based on presence or absence of covid-19 symptoms.

Request permission

The symptomatic group were women who were reported to be admitted to hospital due to covid-19 disease or symptoms, or who were reported to be symptomatic, or received respiratory support of any kind. The asymptomatic group were women admitted for labour, obstetric care, or other reasons, and who were not reported to have symptoms related to SARS-CoV-2 and who did not receive respiratory support, or who were reported to be asymptomatic if the reason for admission was not known. The vaccination status of these women was categorised by the number of doses before admission: unvaccinated, one dose, two doses, or three doses, or vaccination status unknown.

Measures

A composite measure indicating moderate to severe SARS-CoV-2 infection was based on the WHO criteria of covid-19 disease severity.17 Women were classified as having moderate to severe respiratory disease if one or more of the following was reported: oxygen saturation of <95% on admission, need for respiratory support, evidence of pneumonia on imaging, admission to an intensive care unit, or death. Respiratory support was recorded as the maximum level of support in one of the following categories: oxygen therapy (supplementation by nasal prongs or non-rebreathe mask of <15 L/min), high flow nasal cannula of ≥15 L/min or continuous positive airway pressure, mechanical ventilation, or extracorporeal membrane oxygenation. Mode of birth was categorised as follows: caesarean section before or in labour, assisted vaginal birth, or unassisted vaginal birth. Gestational age was categorised by weeks+days as <22 weeks, 22+0-27+6 weeks, 28+0-33+6 weeks, 34+0-36+6 weeks, and ≥37 weeks for age at admission and age at childbirth. We used the following pregnancy outcomes: pregnancy loss (either miscarriage before 24 weeks or termination of pregnancy), total births, live births, stillbirths, admission to neonatal unit, and early neonatal death in the first week.

The following sociodemographic and medical risk factors were included: maternal age, body mass index, occupation (woman or partner in paid work v neither in paid work), ethnic background (Asian, black, Chinese, other, or mixed ethnic minorities v white), smoking (current smoker v non-smoker), medical conditions before or during pregnancy (asthma, hypertension, cardiac disease, and diabetes before or in pregnancy), parity (nulliparous v multiparous), and plurality (single v multiple). Descriptions of drug treatments were based on national guidance issued on 1 July 2020 using the latest edition at the time of admission.18

The study was registered with ISRCTN (number 40092247) and the protocol is available at https://www.npeu.ox.ac.uk/ukoss/current-surveillance/covid-19-in-pregnancy.

Statistical analysis

Numbers and proportions are presented with 95% confidence intervals calculated by using the Agresti-Coull method, and where data were missing, proportions are presented out of known cases. Statistical analyses were done using STATA version 17 (Statacorp, TX). In this national cohort study, the study sample size was governed by the disease incidence, thus, no formal power calculation was carried out.

Patient and public involvement

Patients and public were part of the UK Obstetric Surveillance System steering committee and were involved in study oversight but not in the design, reporting, conduct, or dissemination of this study. Preliminary results of the study have been published open access on MedRxiv and the final results will be publicly disseminated from the National Perinatal Epidemiology Unit and collaborating organisations.

Respiratory support and medical treatment for women with symptomatic covid-19

Nine (0.3%) of 2713 asymptomatic women were admitted to an intensive care unit for indications unrelated to their SARS-CoV-2 infection. Overall, 144 (14.6%) of women with symptoms had at least one indicator of moderate to severe infection (table 2) and 99 (10.4%) required respiratory support. However, the proportion of women with symptoms who received any SARS-CoV-2 specific drug treatment was low (n=50, 5.1%, 95% confidence interval 3.9% to 6.6%); 0.3% (n=3) received antivirals, 0.9% (n=9) received tocilizumab, 4.4% (n=43) received corticosteroids for maternal indication, and 0.6% (n=6) received monoclonal antibodies. Five women were recruited to the RECOVERY trial. Among the 30 women with symptoms who were admitted to intensive care, 13 (43.3%) received specific drug treatment; one woman (3.3%) received antivirals, six (20.0%) received tocilizumab, 11 (36.7%) received corticosteroids for a maternal indication, two (6.7%) received monoclonal antibodies, and one (3.3%) was recruited to the RECOVERY trial.

Vaccination status

The proportion of women with symptoms who had received no vaccine, one, two or three vaccine doses was 58.9% (n=489), 12.7% (n=107), 23.0% (n=194), and 6.5% (n=55), respectively (table 3). A total of 129 (15.3%) women with symptoms whose vaccination status was known had a composite measure of moderate to severe infection. Approximately one in five (93/489, 19.0%) women with symptoms who were admitted to hospital and who had not been vaccinated had moderate to severe infection, reducing to one in twenty (3/55, 5.5%) among women who received three doses. Of 447 women in the two dose group for whom date of vaccination was available, 353 (79.0%) were known to have received their second vaccine dose more than 3 months before admission; this included 11 (57.9%) of the 19 women who had a composite indicator of moderate to severe infection and two who were admitted to intensive care. Overall, only three (2.3%) of the 129 women who had a composite indicator of moderate to severe infection and known vaccination status had completed the three vaccine doses advised to protect the pregnant population against severe omicron infection at the time of hospital admission.

Pregnancy outcomes

Pregnancy outcome was known for 2841 pregnancies, 556 (56.4%) in the symptomatic group and 2285 (84.2%) in the asymptomatic group (table 4). Almost a third of women with symptoms (n=314, 31.9%) were known to have been discharged while still pregnant. The proportion of births at gestational weeks 22-36+6 was 16.0% (n=87) among women with symptoms versus 10.8% (n=238) in women without symptoms. Birth was expedited due to covid-19 for 23 (4.2%) women in the symptomatic group; none of these women was known to have received three vaccine doses (table 5).

Among 2821 infants, 19 stillbirths were reported; 10 (1.8%) in the symptomatic group and 9 (0.4%) in the asymptomatic group (table 6). Six (60%) of the ten stillbirths among women with symptoms occurred to women who were unvaccinated or had received one dose. From the data available, we were unable to assess the the role of SARS-CoV-2 in the stillbirths. Admission to a neonatal unit was reported for 71 (13.0%) of 549 infants born to women with symptoms and 209 (9.3%) of 2253 infants born to women without symptoms.

Principal findings

This national, prospective cohort study has identified that, among pregnant women admitted to hospital with SARS-CoV-2 infection during the period when the omicron variant of concern was dominant, around one in four were symptomatic. One in ten of these pregnant women with symptoms needed respiratory support. Few women with moderate to severe infection received covid-19 specific drug treatments, notably only around half of the women admitted to an intensive care unit.

One in five pregnant women who had not been vaccinated and were admitted to hospital with symptoms had moderate to severe infection, reducing to one in ten with two vaccine doses and one in twenty with three doses. No women with three doses were admitted to intensive care, and most pregnant women with symptoms and moderate to severe respiratory disease, intensive care unit admission, or who died, had not been vaccinated according to the recommended schedule for the pregnant population for the omicron variant (two doses and a third dose if the interval from the second dose exceeded three months).

Strengths and weaknesses of this study

To our knowledge, this study is the first national, prospective cohort study to describe pregnancy and perinatal outcomes during the period when the omicron SARS-CoV-2 variant was dominant. A key strength of these data is the existing UK Obstetric Surveillance System mechanism for national case identification of all women admitted to hospital across the UK, resulting in low risk of selection bias. Of note, the Obstetric Surveillance System cannot provide information about pregnant women in the general population with mild or asymptomatic disease who are not admitted to hospital. In the UK, universal SARS-CoV-2 testing for all obstetric admissions was implemented from May 2020. Pregnant women without symptoms in whom SARS-CoV-2 infection was detected by screening on admission to hospital, were most commonly admitted to give birth.19 Therefore, we categorised the included women by symptoms to avoid misclassification bias and increased adverse outcomes being incorrectly attributed to SARS-CoV-2.20

Lag periods for vaccines were not included in the current analyses and this could lead to underestimation of protection. Some of the pregnant women who had received two vaccine doses or fewer might also have delayed the second dose due to covid-19 infection; information about previous infection was not available in this study. These women could potentially be misclassified into a category with lower expected protection while having a reduced risk due to post-infection immunity, and this misclassification could result in overestimation of the protective effect of different vaccine doses. As with previous analyses,5 SARS-CoV-2 variant sequencing data were not available for individual women and a proxy time period for the data collection was used instead; this proxy is a limitation. Additionally, more women in the symptomatic group were still pregnant or had not known pregnancy outcome at the latest data retrieval compared with the asymptomatic group, which is likely to affect the observed rates of key neonatal outcomes.

Interpretation and comparison with related studies

In this study during the omicron dominant period, the proportion of women with symptoms and moderate to severe infection was 14.6%, which is lower than in the wild type (24.5%), alpha (36.2%), and delta (42.8%) variant periods in the UK.5 15 However, a greater proportion of pregnant women with symptoms had received one or more vaccine doses than in previous variant periods. This vaccination prevalence needs to be taken into account when comparing outcomes across variant periods, recognising that previous vaccination would likely confer some degree of protection from both severe illness and symptomatic infection. When solely unvaccinated pregnant women admitted to hospital with symptomatic infection are considered, maternal outcomes are similar to those observed during the initial wild type infection period.5 Among those in need of respiratory support, irrespective of vaccination status, the use of mechanic ventilation or extracorporeal membrane oxygenation was 13.3% and thus lower than in previous periods (30.7% in wild type, 23.5% in alpha, and 21.4% in delta periods).5

Covid-19 specific drug treatments, which are now standard care for patients who are not pregnant,21 22 were used infrequently, even for women who needed respiratory support. The proportion of patients that received any drug treatment for covid-19 (one or more of an antiviral, tocilizumab, maternal corticosteroids, and monoclonal antibodies) was lower (5.1%) in our study than in the alpha (14.9%) and delta periods (13.6%). Although this finding might partly reflect a lower severity of illness, only about half of pregnant women admitted to an intensive care unit due to covid-19 received any covid-19 specific drug treatment. The Royal College of Obstetricians and Gynaecologists guidelines issued on 19 June 2020 recommended that corticosteroid treatment should be considered for all women who were clinically deteriorating due to covid-19.18 In our study, maternal corticosteroid treatment was reported for 4.4% of women with symptoms during the omicron period, compared with 12.7% during the alpha and 12.0% during delta periods. Approximately a third (37%) of women admitted to intensive care received corticosteroids.

Few pregnant women who had received two or more doses of vaccine were admitted with symptomatic SARS-CoV-2, and few of the women with a composite indicator for moderate to severe infection had received three vaccine doses according to current recommendations to protect pregnant women against severe omicron infection. Vaccination for all pregnant women regardless of risk group was recommended in the UK from 16 April 2021, and all adults were eligible to receive vaccination from mid-June 2021.23 Pregnant women were identified as a risk group and prioritised for vaccination from mid-December 2021, this included recommendation for a third booster dose if the interval after the second dose exceeded 3 months.10 Vaccine uptake for the second dose by females ranged from 68% to 87% in the age categories from 18 to 45 years in England by 22 May 2022.24 Vaccine coverage surveillance among women who gave birth in England up to 31 January 2022 reported that the proportion of women who had received two doses of vaccine increased from 38.4% in November 2021 to 50.6% in January 2022, while 40.5% were unvaccinated in January 2022.25 Similarly, vaccine coverage has been low in Scotland where 32.2% of women who gave birth in October 2021 had received two doses of vaccine during pregnancy compared with 77.4% of women of reproductive age (18-44 years), and 98.1% of women admitted to the intensive care unit were unvaccinated.6 In the current study, 62% of the women with information about vaccination status were unvaccinated.

In the general adult population, effectiveness against symptomatic disease with omicron variant of concern after the second dose declines from 60-75% three weeks after vaccination to 20% at 15 weeks and 10% after 25 weeks,26 and three doses have been shown to give better protection against severe disease.27 Among women who had received two doses, 79% were known to have received their second dose three or more months prior to admission. The number of pregnant women who had received a third booster dose was low in our study, but few severe infections in this group indicates the importance of the third dose to protect pregnant women from both hospital admission with symptomatic covid-19 and need for respiratory support.

Disproportionate admissions due to covid-19 among pregnant women with ethnic minority backgrounds were less prominent in the current study than previously described during the wild type period.15 National guidance has emphasised the importance of addressing this inequality and advised active healthcare seeking in these groups.18 The observation time in the current study is short and the findings cannot yet reliably indicate if the smaller differences can be attributed to better communication, prevention, healthcare seeking strategies or previous infection. Preliminary surveillance results indicated that the omicron variant of concern has a secondary attack rate of 10-13% and therefore factors that increase transmission, such as multi-occupancy housing and public facing occupations, are important also for this variant.28–30 Since socioeconomic deprivation is also a known independent risk factor for adverse pregnancy outcome, this could be a source of residual confounding in this study.

Neonatal outcomes were purposely not compared between omicron and other periods as a high proportion of pregnancies were continuing at the time of analysis. However, the available data suggest that the risk of stillbirth during this period could be lower than observed during the delta period.5 Further follow-up is required to clarify the effect of infection during the omicron dominant period on perinatal outcomes such as stillbirth.

Implications for clinicians and policy makers

The findings of this study indicate that the risk of severe respiratory failure in unvaccinated pregnant women with omicron variant of concern is similar to that observed in the UK during the initial wild type variant wave of the pandemic.15 Few women with moderate to severe disease received covid-19 specific drug treatments and understanding this persistently low use of evidence based treatments among severely ill pregnant and postpartum women is an increasingly urgent priority.

Although severe outcomes were less frequent in the current period than in the previous alpha and delta variant dominant periods, the risk of hospital admission due to covid-19 was higher in the UK than in other European countries during the initial months of the pandemic.15 ,31 32 This higher risk could be associated with factors such as early implementation of public health measures to limit viral transmission in the other countries. If public health interventions could, to some extent, protect pregnant women during the first wave of covid-19, individual protection through vaccination is now available. Our results indicate that most current instances of respiratory failure among pregnant women are preventable, yet vaccine uptake among pregnant women remains low compared with the general female population of reproductive age. Continued, strong efforts to improve uptake of the vaccine during pregnancy are still needed. This effort is of even greater importance because infection continues to rapidly rise in both high and low resourced settings.33