Variations across Europe in hospitalization and management of pregnant women with SARS‐CoV‐2 during the initial phase of the pandemic: Multi‐national population‐based cohort study using the International Network of Obstetric Survey Systems (INOSS)

Abstract Introduction The majority of data on COVID‐19 in pregnancy are not from sound population‐based active surveillance systems. Material and methods We conducted a multi‐national study of population‐based national or regional prospective cohorts using standardized definitions within the International Network of Obstetric Survey systems (INOSS). From a source population of women giving birth between March 1 and August 31, 2020, we included pregnant women admitted to hospital with a positive SARS‐CoV‐2 PCR test ≤7 days prior to or during admission and up to 2 days after birth. The admissions were further categorized as COVID‐19‐related or non‐COVID‐19‐related. The primary outcome of interest was incidence of COVID‐19‐related hospital admission. Secondary outcomes included severe maternal disease (ICU admission and mechanical ventilation) and COVID‐19‐directed medical treatment. Results In a source population of 816 628 maternities, a total of 2338 pregnant women were admitted with SARS‐CoV‐2; among them 940 (40%) were COVID‐19‐related admissions. The pooled incidence estimate for COVID‐19‐related admission was 0.59 (95% confidence interval 0.27–1.02) per 1000 maternities, with notable heterogeneity across countries (I 2 = 97.3%, P = 0.00). In the COVID‐19 admission group, between 8% and 17% of the women were admitted to intensive care, and 5%–13% needed mechanical ventilation. Thromboprophylaxis was the most frequent treatment given during COVID‐19‐related admission (range 14%–55%). Among 908 infants born to women in the COVID‐19‐related admission group, 5 (0.6%) stillbirths were reported. Conclusions During the initial months of the pandemic, we found substantial variations in incidence of COVID‐19‐related admissions in nine European countries. Few pregnant women received COVID‐19‐directed medical treatment. Several barriers to rapid surveillance were identified. Investment in robust surveillance should be prioritized to prepare for future pandemics.


| INTRODUC TI ON
In 2021, a World Health Organization (WHO) systematic review concluded that SARS-CoV-2 infection in pregnancy was associated with an increased risk of admission to an intensive care unit (ICU) for the mother, as well as preterm birth, and admission to a neonatal unit for the infant.Maternal age over 35 years, high body mass index (BMI), non-white ethnicity, preexisting diabetes, chronic hypertension and preeclampsia were associated with severe COVID-19 in pregnancy, although the definition of severe disease was not always clear. 1 Another systematic review on medical treatment reported that only few pregnant women received medication for COVID-19. 2However, the majority of included studies were conducted outside of Europe, most were hospitalbased studies or case series, and only a small number of women with severe outcomes were included.Most studies were unclear regarding the timing and reason for the SARS-CoV-2 test and hospital admission.Additional issues encountered in these systematic reviews were the variation in the data items collected and the definitions used. 1,2ereas several ad-hoc case-based data collections were set up, there was a lack of population-based studies using standardized definitions of exposures and outcomes.Countries participating in the International Network of Obstetric Survey Systems (INOSS) collaboration conduct national or regional population-based surveillance studies of severe complications or rare diseases in pregnancy. 3The INOSS collaboration was established in 2010, and later developed common protocols with uniform definitions and core data items. 4][10][11][12][13][14][15] The main aim of this study was to assess incidence of hospitalization with SARS-CoV-2 infection among pregnant women, across multiple European countries, during the first months of the pandemic.A secondary aim was to describe the COVID-19-directed medical treatment of these women hospitalized with SARS-CoV-2.

| MATERIAL AND ME THODS
This was a multi-national population-based study within the INOSS (EUPAS40489).The INOSS countries collect ongoing prospective data on maternal mortality, severe maternal morbidity and rare diseases in pregnancy.As the obstetric surveillance systems are ongoing, topics can be swiftly introduced.Therefore, at the start of the pandemic, most countries had national permission in place to start retrieving clinical data promptly.10][11][12][13][14][15] In each participating country, standardized case report forms were used to retrieve prospectively recorded information from clinical records.Cases were identified by reporting clinicians or by combining data from perinatal/birth registries, and hospital discharge records.National quality control by linkage to other data sources were performed if possible; this process is summarized by country in Appendix S2.
The source population comprised all pregnant women giving birth in participating countries and regions between March 1, 2020 and August 31, 2020.The estimated total number of maternities was based on notified births to national perinatal/birth registries systems if available, or alternatively on hospital data. 3Women were included if they were admitted to hospital with a positive SARS-CoV-2 polymerase chain reaction (PCR) test within 7 days prior to or during admission in pregnancy, and up to 2 days after giving birth.This definition did not distinguish between different indications for SARS-CoV-2 testing.To discriminate between the symptomatic women who needed COVID-19-related healthcare and pregnant women who were screened at admission for labor and obstetric care, admission was categorized as COVID-19-related or non-COVID-19-related.If the cause of admission was unknown, women were classified as COVID-19-related if they were reported to have any symptoms and as non-COVID-19 if reported to be asymptomatic.

| Statistical analyses
7][18] The incidence of hospitalization per 1000 maternities was reported by admission group with 95% confidence intervals (CI) for each country separately.Heterogeneity across studies was assessed using the I 2 statistic.Secondary outcomes were rare and consequently pooling for further analyses was not feasible.The secondary outcomes are described by country to show the variation in characteristics across countries.Proportions are presented as percentages with the range across countries.If information about risk factor or outcome was missing, the analysis was performed based on the total number with complete information.Numbers of missing data are presented in Tables S2-S4.
National analyses were performed using IBM SPSS Statistics (IBM Corp.) or STATA (STATA Corp. LLC).Aggregate national results were analyzed using STATA.

| Ethics statement
This study was evaluated by the medical review committee at University Medical Center, Utrecht, and was deemed exempt from ethical review due to the use of anonymous national level data (protocol no.21/682, October 4, 2021).All relevant guidelines have been followed and necessary ethics committee approvals have been obtained according to national regulations in each INOSS country (Appendix S3).Data-sharing agreements for anonymous, aggregated data were signed if required with participating countries.Data were managed and stored in accordance with national regulations and the General Data Protection Regulation.Numbers from national datasets from Denmark, Finland, Iceland and Norway were merged to avoid reporting of small numbers.

| RE SULTS
The estimated number of maternities captured by the pregnancy surveillance systems from March 1, 2020 and August 31, 2020 were 816 628.In this source population, 2338 pregnant women had a positive SARS-CoV-2 PCR test and were admitted to hospital (Figure S1).
The testing strategies varied in the different areas, and Figure 1 indicates the approximate timing of testing strategy changes in each of the participating countries.Across all countries and the whole study period, symptomatic women admitted to hospital were tested.
However, screening procedures varied between countries, from no screening during the whole period, to early implementation of routine testing of asymptomatic women admitted for labor or obstetric care.
Of the 2338 pregnant women, 940 (40.2%) had COVID-19related admissions, whereas 1398 (59.8%) tested positive for SARS-CoV-2 but were admitted for non-COVID-19 healthcare, most frequently labor or obstetric care (Figure S1).Table 1 shows the characteristics of the study population by admission group in each country.In most countries, the number of women with risk factors was low and pooled assessment of risk factors was not feasible.In all countries and regions, most women in the COVID-19 admission group were in the third trimester of pregnancy.
The pooled incidence estimate for COVID-19-related admission was 0.59 (95% CI 0.27-1.02)per 1000 maternities, ranging from no admissions in Iceland to 1.8 per 1000 maternities in the UK (Figure 2).There was notable heterogeneity across countries (I 2 = 97.3%,P = 0.00).The incidence of non-COVID-related admission varied even more widely, with similarly high heterogeneity (I 2 = 98.6%,P = 0.00) (Figure 2), and was likely to be influenced by screening policies.Severe maternal outcomes and COVID-19-directed medical treatment in the COVID-19-related admission group is shown in Table 2.
Hydroxychloroquine was prescribed more frequently in Italy (58%) and Belgium (45%) than in the Netherlands (11%); there were no prescriptions in the Nordic countries and the UK.Administration of corticosteroids for maternal indication varied from no prescriptions in the Netherlands to a few prescriptions in the other countries.Thromboprophylaxis was administered less frequently in the UK (14%) than in Belgium (36%), Italy (36%) and the Nordic countries (55%).Treatment of thromboembolism was rare, with only three prescriptions in total.

| DISCUSS ION
In this multi-national study, using population-based cohorts from nine European countries or regions during the first months of the pandemic, we found substantial variations in the incidence of COVID-19-related admissions.Additionally, the low number of women who received COVID-19-directed medical treatment was striking.
To the best of our knowledge, this is the largest multi-national  11,15 Additionally, the use of uniform case report forms with similar definitions aided comparability across countries; however, it was not possible to compile completely uniform datasets.
Although the INOSS collaboration enabled collection of real-time data during the pandemic readily, analysis and reporting of the data was delayed by the lack of funding and staff.The collaboration between clinicians and epidemiologists was key to the relevance of the study information, but as the pandemic progressed, clinical workloads increased.Alongside this, even though the countries planned national quality control by linkage to other national registries, not all countries have yet been able to assess completeness at the time of analysis.We observed a substantial lag time in availability of registry data in some countries and further delays due to the pandemic have also been described for publication of European routine perinatal data. 20e difference in admission rate due to COVID-19 between countries may be related to different public health strategies to contain and limit viral transmission as well as other factors including population density, social security and public trust.The multi-national approach allowed for comparison of the impact of varying timing and nature of public health measures implemented in the countries.According to the Oxford Coronavirus Government Response tracker, strict national public health measures were implemented from the beginning of the pandemic in Italy, which coincided with the decline in the hospitalization rate from March 2020. 21The public health measures in the UK, the Netherlands and Sweden were less strict at the beginning and were enforced later, which may explain the peak in hospitalization rates in April 2020. 21This may provide evidence to confirm that the timing and nature of public health measures does have an impact on the risk of infection and thus also on the risk of COVID-19-related admission among pregnant women.This knowledge could aid consideration of measures for future SARS-CoV-2 waves or emerging infections, especially in a population that is reluctant to vaccinate. 22rthermore, because of the large variation in incidence of COVID-19-related admission and severe disease, the current dataset had too few severe events in most countries to assess risk factors reliably.Despite this, we confirm overall high frequencies of previously described risk factors associated with COVID-19.[27][28][29] Despite this, only a small number received thromboprophylaxis.
The low levels observed may have been caused by concerns about the association between COVID-19 and thrombocytopenia. 27The low utilization of corticosteroids for maternal indication could reflect the concern for disease progression when using steroids. 30 June 2020, the RECOVERY trial demonstrated that the use of corticosteroids was effective in reducing mortality and the duration of invasive ventilation. 31Following that, it was implemented in national guidelines. 26,27,29Likewise, the use of hydroxychloroquine varied between countries, with frequent prescription during the initial months of the pandemic in Italy and Belgium.These prescriptions were based on available safety data without evidence of efficacy, mostly within a research setting. 32,33In June 2020, the RECOVERY trial demonstrated that it was not effective and guidance against use of these treatment was issued. 34We also observed large variation in the use of antibiotics between countries.However, the indication for antibiotic use was unknown, and therefore they may have been given for other reasons.The overall utilization of COVID-19 medical treatments during the first months was low, consistent with the review of Giesbers et al. 2 Pregnant women are a vulnerable population that, with the exception of the RECOVERY trial, is often excluded from large international trials and early phase research on medicines and vaccines, contrary to pre-pandemic policy guidance.
Women who were discharged with ongoing pregnancies were followed up until they gave birth.The primary outcome was the incidence of admission with a SARS-CoV-2 infection per 1000 maternities.Secondary outcomes included severe maternal and perinatal outcomes.The severe maternal outcomes were maternal death (COVID-19-related during admission), maternal ICU admission due to COVID-19, and maximum level of respiratory support (mechanic ventilation or extracorporeal membrane oxygenation [ECMO], continuous positive airway pressure [CPAP] or high-flow nasal cannula, and oxygen supplementation).COVID-19-directed medication included: antibiotics, antivirals, hydroxychloroquine, anti-interleukin 6, intravenous (i.v.) immunoglobulins, corticosteroids for fetal or maternal indication, and low molecular weight heparin (LMWH) for thromboprophylaxis or treatment of thromboembolic disease.Severe neonatal outcomes included stillbirth (intrauterine death prior to or during labor after 22 weeks' gestation, or after 24 weeks' gestation in the UK), admission to neonatal unit, and neonatal death (before discharge following birth).These neonatal outcomes were assessed with the infant as the unit of analysis to account for multiple pregnancies.Pregnancy loss and termination of pregnancy were not addressed in this study.The covariates included maternal age at positive SARS-CoV-2 test, body mass index (BMI, pre-pregnancy weight or earliest recorded weight in pregnancy), obesity (BMI ≥30), migrant or minority background, parity (previous births after 22 weeks' gestation or 24 weeks in the UK), gestational age at infection and at birth (based on last menstrual period or ultrasound according to national guidelines) and mode of birth (vaginal birth or cesarean section).Migrant background was defined as maternal country of birth outside of Europe.Minority background was defined as black, Asian, Chinese, Mixed, or other ethnic minorities (BAME).

Figure 3
Figure 3 shows the incidence of COVID-19-related admission per 1000 maternities by month of first positive test.In Belgium and Italy, the hospitalization rate was highest in March 2020.There was initially an increase in incidence in the Netherlands, Sweden and the UK, showing a peak in April 2020.The incidence of COVID-19related hospitalization in the Nordic countries Denmark, Finland and Norway was low between March and August 2020.The incidence of admission per 1000 maternities by month of first positive test, stratified by reason of admission, is presented in Figure S2.

19 F I G U R E 2
containing population-level data on pregnant women with SARS-CoV-2.The national INOSS studies took into account more than 800 000 women giving birth.The population-based nature provided estimates of incidence, and in comparison with hospital or case-based studies, these estimates are less vulnerable to bias such as selection and reporting.The clinical information was retrieved from medical records and enabled discrimination between asymptomatic and symptomatic women.This reduced the risk of misclassification bias, attributing severe outcomes to COVID-19 rather than to obstetric complications, and enabled comparison across countries with different testing policies.Still, the testing policies varied over time, and stratification of the analyses by COVID-19-related admission may not fully account for this heterogeneity in testing.The varying testing strategies also affected the incidence of infection in the general population across Europe.Pooled risk of hospitalization per 1000 maternities by cause of admission.The use of a comparator group of pregnant women admitted to hospital for non-COVID-19-related reasons is a limitation of this study.Data on pregnant women who were not admitted to hospital as a comparator group would allow further exploration of the consequences of mild/asymptomatic infection.Such data are available in some of the INOSS countries but are not included in the present study.

F I G U R E 3
Admission due to COVID-19 per 1000 maternities per month of first positive PCR test, March to August 2020, by country. of preterm birth.Consistent with our study, non-COVID-19-related admissions were excluded from their analyses. 23Variation in the clinical management between countries may arise from different thresholds for interventions for COVID-19, as well as for obstetric complications.ICU admission may vary depending on local capacity, experience and national medical practice and guidelines, and it is important to discriminate between COVID-19and non-COVID causes related to other obstetric emergencies.24Furthermore, early in the pandemic, knowledge was lacking about the effect of COVID-19 in pregnancy, and safety data of COVID-19 medicines in pregnant women were not yet available.As a result, national treatment guidelines varied and treatment might have depended on the discretion of the clinician or institution.

BOSS ItOSS NethOSS NOSS UKOSS National, Belgium National, Italy National, the Netherlands National, Denmark, Finland, Iceland & Norway Regional, Sweden National, United Kingdom
Abbreviations: BAME, black, Asian and minority ethnic; BMI, body mass index; COVID-19, coronavirus disease 2019; NA, not available.a These cases were reported with a check-box obesity (yes/no), resulting in fewer cases missing than for numerical BMI value.

National, Belgium National, Italy National, the Netherlands National, Denmark, Finland & Norway National, United Kingdom
Medical treatment and maternal outcomes among pregnant women admitted due to COVID-19 from March to August 2020 by country.