NovoSeven® in Severe Postpartum Haemorrhage - Experiences From UK, DK, FR, NL

This is a retrospective non-interventional cohort study of women with sPPH who were treated with NovoSeven® or other standard of care. The analyses will be done utilising data that were collected for other purposes in four countries. The data include four previously established cohorts from the UK, France, Denmark and The Netherlands. In short, these population-based studies collected data, using enhanced systems, from the medical records of the women who met the case definition for each of the respective studies. Data are available from approximately 1260 (Netherlands, 39 exposed to NovoSeven®), 1655 (France, 18 exposed to NovoSeven®), 245 (DK, 28 exposed to NovoSeven®) and 162 (UK, 13 exposed to NovoSeven®) patients with persistent PPH.

Information regarding patient characteristics, concomitant treatment and outcomes in PPH women treated with NovoSeven® is limited in many countries. The current study will therefore include a description of women with PPH treated with NovoSeven®, or standard of care in each of the four cohorts. This will also include an assessment of the comparability of the cohorts and the overall rate of clinical outcomes in the study population. To aid in this assessment propensity scores will be calculated and used to match the women with an event of sPPH treated with NovoSeven® and those that are not.

The UK and Danish datasets were national and the datasets from Netherlands, and France were from multiple regions. The TeMpOH-1 study in the Netherlands collected data from 75% of national births. The EPIMOMS study in France collected data from six regions: Alsace, Lorraine, Auvergne, Rhone-Alpes, Ile-de-France and Basse-Normandie and these regions represented 18% of national births.

Data come from the below described cohorts.

Denmark: The Danish data were identified and collected using the Danish Medical Birth Registry (DMBR) and the Danish Trans-fusion Database (DTD). During the period 2001-2009, women who had a birth identified in the DMBR and had transfused 10 or more units of red blood cells (RBC) within 24 hours identified from DTD were included in study. The DMBR collects information on all births in Denmark and this includes clinical and demographic data from each birth. The DTD is a national database and contains information of the quantity and type of blood products transfused and the time of each transfusion. Additional data were extracted from the patient charts (Study II in the PhD thesis(PubMed ID 29510809)).

France: This study included data from the EPIMOMS study (PubMed ID 33423750). This is a population-based study that prospectively identified cases of severe maternal morbidity from 6 regions in France: Alsace, Lorraine, Auvergne, Rhone-Alpes, Ile-de-France and Normandy. Cases were identified for a year between 2012-2013 from 113 intensive care units and 118 maternity units. Selected maternal morbidities were identified by a local reporter in each hospital. Data concerning demographics, previous medical history, current pregnancy, causes, management and maternal and perinatal outcomes were collected from the patient charts of each woman. Major obstetric haemorrhage (MOH) was one of the severe maternal morbidities in the EPIMOMS study. To be included as a MOH case, women had to be at least 22 weeks of gestation or more and had to meet one of the following criteria: 1. ≥1500mls blood loss; 2. blood transfusion ≥ 4 RBC; 3. uterine artery embolisation, ligation or compressive uterine sutures; 4. have an emergency peripartum hysterectomy; 5. Haemorrhage with organ dysfunction as per EPIMOMS definition.

The Netherlands: TeMpOH-1 was a national retrospective cohort study conducted between January 2011 and January 2013 in the Netherlands, which included 75% of all hospitals. Women were identified using records from blood transfusion services and birth registries of participating hospitals. Data were collected from medical records and included information about previous medical history, current pregnancy, causes of haemorrhage, haematological parameters, blood components, fluid resuscitation and surgical and medical management(PubMed ID31730187 and 30784827). Women who experienced an obstetric haemorrhage and had received at least one of the following were included: 1. four units or more of RBC ; 2. multicomponent blood transfusion (RBC and fresh frozen plasma and/or platelet concentrates); 3. plasma in addition to RBCs as a result of obstetric haemorrhage.

United Kingdom : Data from UK comes from a national surveillance system used to identify cases of massive transfusion from consultant led obstetric units nationally in the UK. Reporters were asked to complete a data collection form using the medical records of the women. Data were collected anonymously. During the period between June 2013 to July 2014, information on women who fulfilled the following criteria were included: 1. received 8 or more units of RBC transfused within 24 hours ; 2. 20 or more weeks of gestation. The United Kingdom Obstetric Surveillance System was used to identify the patients (https://www.npeu.ox.ac.uk/downloads/files/ukoss/forms/UKOSS-Major-Obstetric-Haemorrhage.pdf).

The primary objective will be answered using propensity score matching to ensure exchangeability between sPPH patients treated with NovoSeven® and those that were not across data from the Netherlands, Denmark and France (if these are comparable).

For exposed women: Time0 is defined as time of first administration of NovoSeven®. It occurs x minutes after onset of sPPH. For every exposed woman, all unexposed and eligible women are evaluated at this time=time0. The ones with similar propensity score are selected for matching (up to four will be selected). Thus, for unexposed women (matched controls): Time0 is the timepoint used when they were selected for matching. It is equal to the period from onset of sPPH to time of first administration of NovoSeven® for the patient for which they are a matched control. In this way exposed and unexposed women are comparable because they have similar propensity score at the same (specific) Time0.

Generally, the following approach will be taken:

• Propensity score models and matching will be performed separately by country
• Matching will be done within patients with the same delivery mode
• Pooling will be done after matching including all countries which allows formal analysis - this will be the primary population
• The primary analysis will be based on the pooled data - meaning there we will have 1 (primary) p-value for the primary endpoint
• Subgroup-analyses of the primary endpoint will be done by country - there we will have 1 (supportive) p-value per country

Propensity score matching will be performed to ensure comparability between women with an event of sPPH being exposed to NovoSeven® with women with an event of sPPH that are not exposed to NovoSeven at the matching timepoint. In this approach, the propensity score will reflect the estimated probability of being administered with NovoSeven® during the course of severe PPH. A propensity score for every woman with sPPH will be estimated by an appropriate analytic model with NovoSeven® as the dependent variable. Covariates associated with initiation of NovoSeven® administration will be included in a model to calculate propensity scores. Characteristics considered to be potential confounders for the association between use of NovoSeven® and the outcome or characteristics considered to be risk factors for the occurrence of the primary outcome measure alone will be included as covariates in the propensity score model.

Characteristics that could be included as covariates are (not an exhausted list): gestational age, multiple pregnancy, cause of PPH, volume of blood loss at time of intervention, and haemostatic drugs used (other than NovoSeven®) that had already been applied at the time of intervention. A matching algorithm for the propensity score will be used to match women with an event of sPPH exposed to NovoSeven® with women with an event of sPPH that are not at that timepoint. Women will be censored when they have been hysterectomised, dies, or stopped having PPH.

Matching will be done by up to 1:4 with a calliper of 0.1. If this is not possible a calliper of 0.2 will be used instead. The criteria for switching to 0.2 will be described in the SAP.

The estimand will address the relative effect of NovoSeven® compared to propensity score matched controls on the occurrence of invasive procedures in women with an event of sPPH. The five attributes of the estimand can be seen below. The estimand is defined from the following five elements outlined in the ICH E9(R1):

1. The treatment conditions are NovoSeven® vs. propensity scored matched controls not receiving NovoSeven®.
2. The treatment effect will be estimated for women with an event of sPPH defined by the inclusion criteria in the four cohorts exposed to NovoSeven® and their matched controls.
3. The treatment effect is assessed by the occurrence of invasive procedures within 20 minutes to 24 hours following time0.
4. Intercurrent events will be handled by a hypothetical strategy or a treatment policy strategy (further details in protocol Table 1).
5. The treatment effect will be quantified by the odds ratio.