Main outcome measures
We defined the overall term “Maternal health” and divided all relevant International Classification of Diseases (ICD) codes into four main groups, namely “Pregnancy”, “Delivery”, “Pregnancy loss” and “Death”, and further into subgroups (Figure 1/ Table S1). Supplementary table 1 shows how we defined the main groups and subgroups using the ICD 8th and 10th edition. Unless clearly stated elsewise, the results are only given in those Marfan syndrome women with at least one pregnancy registration (n=91) and in their matched controls who all had at least one pregnancy registration (n=6,801). For each registration within the main groups as well as the subgroups, we divided the women with Marfan syndrome into two categories according to whether the Marfan syndrome diagnosis was known at the time of registration or not. Thus, a woman with Marfan syndrome could have multiple registrations related to maternal health before being diagnosed and multiple registrations after being diagnosed with Marfan syndrome. For all main groups and subgroup registrations, we only utilized the first registration in each group, regardless of time of diagnosis.
To report the total number of new-borns, we excluded all registrations less than 270 days after a delivery registration, as each delivery could have multiple delivery related registrations within few weeks’ intervals.
In order to study the relationship between maternal health and aortic disease, we defined aortic dissection and aortic operation as previously described.7 In brief, aortic disease is either a registration of an aortic dissection, defined as a registration within the ICD system with a diagnosis of aortic dissection, or a registration of an aortic operation, defined as a registration in the Nomesco Classification of Surgical Procedure (NCSP) related to aorta (for details see Table S2).
The women with Marfan syndrome were diagnosed from 1977 to 2014 and the registrations of maternal health were from 1977 to 2017. Data were accessed using a secure remote access to Statistics Denmark. In order to avoid any possibility of personal identification of cases, Statistics Denmark prohibits specification of the exact number of cases with a given condition if less than three, and we therefore report these as “<3”.