1. General Information
National Perinatal Epidemiology Unit,
University of Oxford,
Old Road Campus,
Oxford, OX3 7LF
Dr Catherine Deneux-Tharaux
INSERM researcher epidemiologist
53 Avenue de l'Observatoire
Dr Marie-Pierre Bonnet
Doctor anaesthetist /Epidemiologist
INSERM 53 Avenue de l'Observatoire
Dr Griet Vandenberghe
Dept. of Obstetrics & Gynaecology
Ghent University Hospital
De Pintelaan 185, BE-9000 Gent, Belgium
Virginie Van Leeuw
Perinatal Epidemiology Center (CEpiP), Non-profit organisation
Route de Lennik, 808 BP 5971070 Bruxelles, Belgium
Dr. Wei-Hong Zhang
Research laboratory in human reproduction, Université libre de Bruxelles (ULB) /
Perinatal Epidemiology Center (CEpiP), Non-profit organisation
& Route de Lennik, 808 BP 597
1070 Bruxelles, Belgium
Professor of Perinatal Epidemiology,
Director, National Perinatal Epidemiology Unit,
National Perinatal Epidemiology Unit,
Nuffield Department of Population Health,
University of Oxford,
Old Road Campus, Headington,
Oxford OX3 7LF
Professor Marian Knight
NIHR Professor of Maternal and Child Population Health
National Perinatal Epidemiology Unit,
University of Oxford,
Old Road Campus,
Oxford, OX3 7LF
An international collaborative study examining anaphylaxis in pregnancy using the International Network of Obstetric Survey Systems (INOSS)
Prospective cohort study
Anonymised data collection only – no active participants
Planned Sample Size
Varied, depending on the number of included countries. Aim to collect 150 cases.
Planned Study Period
July 2016 – March 2018
Using the INOSS collaboration this study will assess the incidence, causative agents, management and outcomes of anaphylaxis in pregnancy. It further aims to assess the factors associated with poor outcomes.
Anaphylaxis in pregnancy is a potentially fatal systemic hypersensitivity reaction, which is rapid in onset. It is characterised by a life-threatening airway, breathing or circulatory problems often with skin or mucosal change. Exposures resulting in anaphylaxis in the general population include allergens such as medication, insect stings or food. Anaphylaxis commonly occurs when allergens trigger an IgE mechanism causing mast cell activation, resulting in an anaphylactic reaction (Simons and Schatz, 2010).
There is very little literature regarding the incidence of anaphylaxis in pregnancy; a study in Texas estimated an incidence of 2.7 cases per 100,000 deliveries (Mulla et al.). It has, however, been proposed that anaphylaxis is increasing in the general population (Simons and Schatz, 2010, Ewan P.W., 2006). This is particularly concerning as the outcomes of anaphylaxis in pregnancy are very serious, as shown by case reports of maternal and neonatal mortality (Cooper et al., 2005, Berenguer et al., 2013). Outcomes have also been reported to include maternal cardio-respiratory compromise resulting in cardiac arrest in some instances, and hypoxic brain injury in the neonate. However, due to possible publication bias these case reports may not be representative of all the cases of anaphylaxis in pregnancy. As a result, they may over emphasise the poor outcomes associated with this condition. This accentuates the need for a population-based study to be undertaken to provide a reliable estimate of the maternal and perinatal outcomes associated with anaphylaxis in pregnancy.
Anaphylaxis can be caused by a number of agents such as food, latex, insect stings and medication (Sengupta and Kohli, 2008, Khan et al., 2008, Shingai et al., 2002, Sheikh, 2002, Cuciti et al., 2005). Anaphylaxis is a severe life-threatening condition in the general population, however in pregnancy it can result in catastrophic outcomes for both mother and child. Exposure to antibiotics is increasing in the pregnant population through the use of prophylactic antibiotics one hour before an elective caesarean section and for treatment of group B streptococcal carriage to prevent neonatal transmission (Sengupta and Kohli, 2008, Khan et al., 2008). As a result, it is important to assess the causative factors of anaphylaxis in pregnancy and whether the increased antibiotic usage is associated with an increased incidence of anaphylaxis in pregnancy
A study within the UK has identified only a small number of cases. Therefore, this condition would benefit from the collection of cases internationally in order to increase the sample size. This would be beneficial for two reasons; firstly, it would increase the number of cases included in the study thus allowing us to characterise this population more accurately. Secondly, it would potentially allow inter-country comparisons of the presumed causal factors and management of anaphylaxis of pregnancy. Different policies for prophylactic antibiotic prophylaxis across countries may explain different incidence rates, therefore making an international study extremely insightful.
The majority of women with anaphylaxis in pregnancy are treated in the UK using the same algorithm as those in the general population. Management is characterised by a rapid assessment of life threatening cardio-respiratory compromise, which is usually treated with adrenaline, oxygen and fluids (Soar et al., 2008). It is unclear whether this best practice algorithm is optimal for treatment of anaphylaxis in pregnancy and whether it yields the best outcomes for both mother and child. Cross-country difference in managements and associated outcomes may enable development of guidance on optimal management for anaphylaxis in pregnancy.
3.1 Rationale for the use of the International Network of Obstetric Survey Systems (INOSS)
An INOSS study will advance the evidence on causal factors, managements and outcomes of women with anaphylaxis in pregnancy and by increasing the number of cases it will improve study power thus allowing the robust examination of this rare complication. INOSS also offers the unique opportunity to assess whether there are any country level differences in the causal factors, managements and outcomes of rare complications. The use of uniform definitions, common data sets and data collection methods across nations will help reduce information bias. Furthermore, the use of national data collection methods with a high response rate will reduce the likelihood of selection bias.
The following definition will be applied across participating countries:
The cases will be all pregnant women in the INOSS region identified as having anaphylaxis according to the following definition (Resuscitation Council, 2008).
Anaphylaxis is defined as a severe, life-threatening generalised or systemic hypersensitivity reaction. The following two criteria must be met for a diagnosis of anaphylaxis to be made:
1. A life-threatening airway problem and/or breathing problem and/or circulatory problem
2. Sudden onset and rapid progression of symptoms
1. A life-threatening airway problem is taken to include:
- Laryngeal or pharyngeal oedema
- Hoarse voice
2. A life-threatening breathing problem is taken to include:
- Shortness of breath and raised respiratory rate (or dyspnea)
- Wheeze (laryngospasm or bronchospasm)
- Decreased oxygen saturations
- Confusion secondary to hypoxia
- Respiratory exhaustion or respiratory arrest
3. A life-threatening circulatory problem is taken to include:
- Signs of shock such as faintness, pallor or clammy
- Tachycardia >100bpm
- Systolic BP <90mmHg or MBP<60mmHg or measured hypotension
- Decreasing level of consciousness
- Signs of ischaemia on ECG
- Cardiac arrest
Previous definitions have included skin and/or mucosal features as a core criteria. However, these may not be evident if treatment is rapidly implemented, so this study shall include all women in whom the final clinical diagnosis is anaphylaxis, irrespective of the presence or absence of skin/mucosal changes.
5. Research questions
What is the incidence of anaphylaxis in pregnancy amongst the included nations?
What are the presumed causative agents for anaphylaxis in pregnancy?
How is anaphylaxis in pregnancy managed?
What are the maternal and perinatal outcomes?
Are there any factors associated with poor maternal or infant outcomes?
6. Study design
This will be the first prospective international cohort study using obstetric surveillance systems at an international level to examine anaphylaxis in pregnancy. Data collection has been completed in the UK using the UKOSS data collection system, so the UKOSS methodology will be used as the model for data collection in other INOSS countries. Cases from participating countries and the already collected UKOSS cases will be combined in an international collaborative study.
6.1 Case notification
Cases from France countries will be prospectively identified from obstetric centres on a three-monthly basis via email. In France, cases will be identified through two sources. The first is the national French network of anaesthesiologists working in maternity units (“Club des Anesthésistes Réanimateurs en Obstétrique CARO”). The second is a network of laboratories in which serum tryptase levels are tested (“Allergobionet”); here cases will be identified by identifying women with an elevated tryptase level and a clinical suspicion of anaphylaxis whose blood samples were sent from an obstetric unit. The subsequent obstetric unit will be contacted for data collection. Six months after the completion of the study, the national French hospital discharge data will be examined to identify cases and further check case ascertainment. If there are further cases of anaphylaxis in pregnancy the obstetric unit in which they occurred will be contacted and asked to complete the online data collection form.
In Belgium, the cases will be notified through the existing Belgium Obstetric Surveillance System (B.OSS). A monthly email will be sent to the nominated reporting clinicians. The clinician will report a case of anaphylaxis in pregnancy on the online B.OSS case notification system.
6.2 Data collection methods
Following the report of a case, the notifying clinician will be asked to complete an online data collection form in the OpenClinca web-based data collection system. Investigators in the UK will send a web-link to the OpenClinca form in an email to the reporting clinician. The online data collection form will ask for confirmation of the case definition, information on the causative agent, management and immediate outcome. Clinicians will complete the online data collection forms using information from the patient’s medical records. Personal identifiers will not be collected thus only anonymous data obtained after the event has occurred will be used. Reporting clinicians will be asked to keep their own record of the unique study number and patient identifiers in order to facilitate elimination of duplicate reports. If no online form is received within 6 weeks of the case notification, the reporting clinician will be contacted by email. If there is still no report after a further 4 weeks, a further email will be sent to remind them to fill in the online form.
The study will last for a period of 1.5 years with a proposed start date of September 2016.
The analysis will describe the incidence, characteristics, causative agents, managements and outcomes of anaphylaxis in pregnancy in the combined cohort.
We will additionally investigate whether there are any factors association with poor maternal and perinatal outcomes.
Factors to be explored:
Causative agents, previous exposure to the agent, previous reaction to other agents, prophylactic antibiotic treatment and location where anaphylaxis occurred.
Managements used: time taken for delivery to occur and drugs used in resuscitation period.
Outcomes to explore:
Perinatal outcomes considered separately and as a composite outcome: perinatal mortality, perinatal morbidity (neonatal encephalopathy), preterm birth (<37 weeks gestation and <32 weeks gestation), stillbirth, early neonatal death, neonatal intensive care admission.
Maternal outcomes considered as a composite outcome: major maternal morbidities (cardiac arrest, organ dysfunctions, hypoxic brain injury and critical care admission) and maternal mortality.
7. Data requested:
Please see attached draft data collection form.
8. Statistical procedures
8.1 Data exploration
All analyses will be conducted using Stata v13(SE). The data will be examined and cleaned for outlying and implausible values. The distribution of continuous variables will be assessed. Tests for departure from linearity will be undertaken in continuous variables. Where appropriate, continuous variables will be categorised into clinically meaningful ranges for ease of interpretation.
8.2 Descriptive analysis
The incidence of anaphylaxis during pregnancy in each country will be calculated with 95% confidence intervals.
The characteristics, causative agents, managements, and perinatal and maternal outcomes will be described for cases of anaphylaxis during pregnancy in each country. Histograms will be used to assess normality in continuous variables. Categorical variables will be summarised as frequencies and percentages. Normally distributed continuous variables will be summarised as means with standard deviations; skewed continuous variables will be summarised as medians with interquartile ranges.
8.3 Multivariable analysis
If the study has a sufficient sample size, a multivariable analysis will examine the risk factors associated with poor maternal/perinatal outcomes. Collinearity will be checked between all plausible linear associations prior to multivariable analysis, using Pearson’s correlation coefficient. Variables that are highly correlated will be explored and the most appropriate variable will be included in the final model. Credible interactions will be identified using previously published literature and these will be tested using likelihood ratio tests. Each maternal/perinatal outcome will be assessed in a separate multivariable model while controlling for specific confounders. These models will be presented as odds ratios and 95% confidence intervals. A sensitivity analysis will be undertaken including only cases confirmed by the presence of allergen-specific IgE.
8.4 Missing data
Previous UKOSS studies have shown that the distribution of missing data across variables is not random; for this reason multiple imputation is unlikely to be used although this will be explored further in this specific dataset (Lindquist et al., 2013). In order to account for missing data a separate ’proxy’ category for missing data will be created for variables with >2% missing values.
9. Ethics committee approval
Ethics approval has been sought and approved for the UK study. Ethics approval will be sought from each individual country for the collection of anonymised data.
France: Ethics approval will be sought from the SFAR’s ethics committee (Société Française d’Anesthésie Réanimation). In addition, each hospital where cases will be identified will make a declaration to the National Data Protection Agency (CNIL, Commission Nationale de l’informatique et des Libertés) that clinical data will be used for research.
Belgium: The B.OSS study methodology received approval from the Ghent University Ethics Committee (Reference Number 2015/1470, Belgian Registration number B670201526875) as central ethics committee. Ethical committees from the participating hospitals have been informed of the study and the central ethics approval. The Ghent University Ethics Committee will be informed of the start of the study of Anaphylaxis in Pregnancy.
UK: The UKOSS study followed the Health Research Authority Confidentiality Advisory Group’s (HRA CAG) guidelines to use only anonymised data. The UKOSS study methodology received approval from the London Multi-Centre Research Ethics Committee (study reference 04/MR02/45), REC Reference 10/H0717/20. No additional approval is required for secondary analysis of previously collected anonymous data.
10. Data procedures and study management
This study seeks to describe anaphylaxis in pregnancy internationally. This non-interventional study will analyse information collected from the patient’s medical notes, therefore, this study will not influence the management of any individual women whose data are included. This is a descriptive study and will collect information only. Information must be collected on all cases occurring in the population in order to obtain accurate incidence information. This means that it will not be practical to obtain individual patient consent. The data collected from the surveillance systems or registries will not include any names, addresses, dates of birth, hospital or health service or social security number or any information that would directly identify the patient. Therefore, the investigators will not have access to any identifiable data. This study will be conducted at the NPEU, University of Oxford, and the NPEU information governance and data security policies will apply.
In the UK, the HRA CAG have judged that the collection of information for the purpose of studying incidence and identifying means to improve patient care, which is not individually identifiable and does not lead to any change in management for the individual patient is acceptable without requiring individual patient consent.
11. Confidentiality and privacy
All data to be transferred to the researchers is already de-identified. All study information will be maintained in strictest confidence in accordance with the Data Protection Act 1998, United Kingdom of Great Britain and Northern Ireland. Data will be stored in a password-protected area of a secure University network, accessible only to named members of the research team. All results will be published in a form that will not allow individuals to be identified, that is, in tabular, aggregate form only.
The NPEU is covered by the University of Oxford’s Data Protection Act registration (registration number: Z575783X) and all aspects of this project will comply with the Data Protection Act. The Information Guardian for the NPEU is the Director, Professor Jenny Kurinczuk.
12. Data transfer, storage, access and record retention
If data is collected in each participating country and not using the OpenClinica platform it will be transferred to researchers in the University of Oxford using the University of Oxford Oxfile secure transfer service at the institution. The data will be encrypted before being deposited into this transfer service at the University of Oxford. The password for the file will be sent separately to the researchers at the University of Oxford. From here a single copy of the file will be stored on the secure server at the National Perinatal Epidemiology Unit, University of Oxford (NPEU).
The data will be stored in a secured project folder located on a NPEU, University of Oxford storage system managed by the Information Technology (IT) team. Access to the secured project folder will be limited to the named investigators. Access to the project folder is controlled by the NPEU IT team. The storage network is protected by firewalls and intrusion prevention systems managed centrally by the IT team. The NPEU network is part of the University of Oxford internal network. The storage system will snapshot the archive regularly to prevent against accidental deletion, and backups that are undertaken are located in a secured location. No copy of the data will be made and the data will not be stored or transferred using portable data devices, for example, USB drives, CDs and DVDs. Only named researchers will have access to the data. The NPEU maintains a comprehensive information security protocol that aims to ensure that confidentiality, integrity and availability of information is maintained at all times.
All the data collected using the OpenClinca platform will be transferred to the country of origin at the end of the study. Any data retained at the NPEU will be deleted two years after the publication of the findings. All electronic files will be permanently deleted from the computer server and the research data management systems at the NPEU, University of Oxford.
13. Definition of the end of the study
This project is part of a DPhil/PhD, so the end of the study will be defined as the date of acceptance of the DPhil thesis into the Examination School, High Street, University of Oxford, or the publication of the final manuscript from the study, whichever is the later.
14. Study size
The UKOSS study estimated the incidence of anaphylaxis as 1.6 (95% CI: 1.1-2.1) cases per 100,000 maternities. In a cohort of 37 cases collected from the UK, there were 2 maternal deaths giving a case fatality rate of 5.4% (95 CI: 0.7-18.2). Using the UKOSS study as an estimate of the case fatality rate in cases of anaphylaxis of pregnancy, as an example, table 1 presents the confidence intervals around a point estimate of 5% for different sample sizes. Table 2 shows the estimated number of cases for the included countries.
Table 1. The variation in confidence intervals according to sample size using case fatality as an example.
Number of cases
Number of maternal deaths
95% Confidence interval
1.4 – 12.3
1.6 – 11.3
2.3 – 10.2
Table 2. The estimated number of cases for France and Belgium in the study period.
Estimated number of live births in study period
Estimated number of cases
Note: used the incidence rate from the UK, which was constructed using number of maternities rather than the number of live births.
Amendments are changes made after the research has received approval from all the collaborating countries. Any amendments and deviations from the protocol will be notified to the participating countries and to the relevant ethics
This study is being completed as a part of DPhil degree at the University of Oxford. The student Stephen McCall has received a small allocation of funds from the Medical Research Council to cover research related costs (£5000 per annum).
17. Public disclosure
It is important that the clinicians who collected the data receive feedback of the outcomes of the study. The anonymised study results will be disseminated in an internationally peer reviewed journal. The initial results will be feedback to the INOSS members. Furthermore, they will be presented at specialist internationally relevant conferences. The NPEU has a close working relationship with the Department of Health and is known to have influenced health policy in the UK and worldwide.
This study will form a part of a DPhil (PhD) project and will be analysed and written up by the DPhil student. Authors (and the order of authors) on the published paper(s) will be discussed and agreed prior to the initiation of the study, according to the INOSS policy. All authors must meet ICMJE criteria.
19. Impact of research
The topic of anaphylaxis in pregnancy is an under researched area and lacks strong evidence based guidelines. Evidence has suggested that current guidelines have had limited impact in reducing poor maternal and perinatal outcomes. Observational evidence will be able to identify the risk factors that are associated with poor outcomes. Thus, helping to target better care at this high risk demographic.
BERENGUER, A., COUTO, A., BRITES, V. & FERNANDES, R. 2013. Anaphylaxis in pregnancy: a rare cause of neonatal mortality. BMJ Case Reports, 2013.
COOPER, G. M., MCCLURE, J. H. & BOARD, O. B. O. T. E. 2005. Maternal deaths from anaesthesia. An extract from Why Mothers Die 2000–2002, the Confidential Enquiries into Maternal Deaths in the United Kingdom: Chapter 9: Anaesthesia. British Journal of Anaesthesia, 94, 417-423.
CUCITI, C., MAYER, D. C., ARNETTE, R. & SPIELMAN, F. J. 2005. Anaphylactoid reaction to intravenous sodium ferric gluconate complex during pregnancy. International Journal of Obstetric Anesthesia, 14, 362-364.
EWAN P.W., F. T. F. T. B. S. F. A. A. C. I., . 2006. The nature and extent of allergyin the United Kingdom. A report to the Department of Health Review of Allergy Services. London: Department of Health.
KHAN, R., ANASTASAKIS, E. & KADIR, R. A. 2008. Anaphylactic reaction to ceftriaxone in labour. An emerging complication. Journal of Obstetrics and Gynaecology, 28, 751-753.
MULLA, Z. D., EBRAHIM, M. S. & GONZALEZ, J. L. Anaphylaxis in the obstetric patient: analysis of a statewide hospital discharge database. Annals of Allergy, Asthma & Immunology, 104, 55-59.
RESUSCITATION COUNCIL 2008. Emergency treatment of anaphylactic reactions; guidelines for healthcare professionals. London: Resuscitation Council (UK).
SENGUPTA, A. & KOHLI, J. K. 2008. Antibiotic prophylaxis in cesarean section causing anaphylaxis and intrauterine fetal death. Journal of Obstetrics and Gynaecology Research, 34, 252-254.
SHEIKH, J. 2002. Intrapartum anaphylaxis to penicillin in a woman with rheumatoid arthritis who had no prior penicillin allergy. Annals of Allergy, Asthma & Immunology, 99, 287-289.
SHINGAI, Y., NAKAGAWA, K., KATO, T., FUJIOKA, T., MATSUMOTO, T., KIHANA, T., NODA, K. & MORI, T. 2002. Severe Allergy in a Pregnant Woman after Vaginal Examination with a Latex Glove. Gynecologic and Obstetric Investigation, 54, 183-184.
SIMONS, F. E. R. & SCHATZ, M. 2010. Anaphylaxis during pregnancy. Journal of Allergy and Clinical Immunology, 130, 597-606.
SOAR, J., PUMPHREY, R., CANT, A., CLARKE, S., CORBETT, A., DAWSON, P., EWAN, P., FOËX, B., GABBOTT, D. & GRIFFITHS, M. 2008. Emergency treatment of anaphylactic reactions—guidelines for healthcare providers. Resuscitation, 77, 157-169.