Methods and materials

The study protocol was published after ethical approval of the study documents [12]. In brief, a large cohort of women starting OC treatment (first-ever users or switchers to a new product) was actively monitored for the occurrence of rare or unexpected adverse outcomes possibly related to exposure to OCs.

The primary focus of the active study monitoring was to assess the occurrence of new cardiovascular events. These events were monitored continuously for the duration of the study.
In addition, all other serious adverse events (i.e., adverse events that result in death, a life-threatening experience, inpatient hospitalization, persistent or significant disability/incapacity, or require medical/surgical intervention to prevent one of said outcomes) were fully documented and reviewed by both the investigator and the Advisory Council at least twice a year.

2.1 Study objectives
see objectives

2.2 Study population
Recruitment of the cohort members was conducted via a network of physicians who prescribe OCs.
Plans called for the combined cohort to include at least 50,000 women recruited in seven European countries: Austria, Belgium, Denmark, France, Germany, the Netherlands, and the United Kingdom.
At the participating centers, all women who received a prescription for a new OC were asked by their physicians whether they were willing to participate.
These women could be either OC starters (first-ever users) or OC switchers. More specific inclusion or exclusion criteria were not applied because of the non-interference approach of the study design.
The objective was to avoid influencing preference for specific OCs, while at the same time to undertake significant efforts to ensure standardized, comprehensive, and reliable documentation of all baseline characteristics and adverse events during the follow-up period.

2.3 Baseline survey and follow-up
Baseline data were recorded on a questionnaire that addressed the participants' state of health and potential risk factors. Participants provided their medical history, including medication history and history of OC use.
They also provided their addresses and phone numbers, those of relatives or friends who could serve as back-up contacts, and those of their primary care physicians/gynecologists.
Follow-up assessments for each woman were scheduled every 6 months. The self-administered follow-up questionnaires addressed the occurrence of adverse events. Reasons for discontinuing OC use or switching to another OC were requested if applicable.
The questionnaires were reviewed for completeness and plausibility/consistency of the responses. Missing or inconsistent information was clarified directly with the women by phone.
A low “loss to follow-up rate” was essential for the validity of the study. In order to minimize loss to follow-up, a four-level follow-up process was established. Level1 activities included mailing the follow-up questionnaire and --- in case of no response --- two reminder letters.

If level1 activities did not lead to a response, multiple attempts were made to contact the woman, her friends, relatives, and gynecologist/primary care physician by phone. In parallel to these level2 activities, searches in national and international telephone and address directories were started (level3 activities).
If this was not successful, an official address search via the respective governmental administration was conducted (in some countries centralized, in others decentralized at community level). This level4 activity usually yielded information about a new address (or information that the respondent had moved abroad or died).

2.4 Validation
A self-administered questionnaire used by patients at six-month intervals is a very sensitive tool that captures almost all serious clinical outcomes. From a methodological point of view, it captures a much higher proportion of these outcomes than methods that rely solely on the prescribing gynecologist who often is not directly involved in the diagnosis and treatment of these outcomes (e.g., cardiovascular events).
However, it must be considered that there is a significant difference between the rates of reported and validated events, because laypersons often misclassify adverse events (e.g., pain in the legs after standing for a long period of time as “thrombosis”, or migraine attacks as “stroke”, even if modern imaging procedures do not provide any indication of the perceived event). Therefore, validation of the self-reported events was of great importance.
Follow-up questionnaires that contained information about such events were immediately passed on to the medical reviewer group. In case of unclear or missing information, the women were contacted by phone, e-mail or other means. For many events it was necessary to contact the diagnosing and/or treating physician for clarification and validation of the information received from the patient. This procedure was mandatory for all serious adverse events (including all main outcomes of interest).

Under routine medical conditions, diagnosis of a serious adverse event is not always confirmed by a diagnostic method with high specificity. Therefore, investigators classified all serious adverse events as “confirmed” or “not confirmed”. Events that were confirmed by diagnostic measures with high specificity (e.g., phlebography for DVT or cerebral MRT for cerebrovascular accidents) or a clinical diagnosis supported by a diagnostic test with low specificity (such as D-dimer for VTE) were considered confirmed.
Events were considered not confirmed if the diagnosis reported by the patient was excluded by diagnostic measures, if a different medical condition was diagnosed by the attending physician, or if the reporting woman did not contact a health professional to clarify her symptoms and no diagnostic measures were performed that could have clarified the diagnosis.

2.5 Blinded adjudication
For the purpose of continuously monitoring safety data during the study, classification of reported serious adverse events was performed by the investigators. For the final analysis, classification of the primary outcome of interest --- VTE --- was verified by a process of independent blinded adjudication.
In order to minimize classification bias, the decisions made by the investigators were reassessed by three independent medical experts specializing in radiology/nuclear medicine, cardiology, and internal medicine/phlebology. These specialists reviewed all available information on the reported VTE.
Brand names, dose, regimen, and composition of the OC(s) used by the reporting woman were rendered anonymous for this process. These adjudicators performed their reviews independently of each other and without knowing the judgment of the other adjudicators or the investigators.
Using a conservative approach for the final analysis, a reported VTE was classified as confirmed if at least one adjudicator had considered the event as confirmed.

2.6 Evaluation
The EURAS study was powered to demonstrate the non-inferiority of DRSP-containing OCs regarding VTE risk in comparison to LNG-containing OCs and ‘Other OCs’. Cox regression analyses of the main clinical outcomes were carried out in accordance with the final analysis plan, which was approved by the Advisory Council before the first interim analysis.
The following pre-defined confounder variables were included in the Cox regression model: age, BMI, duration of use, and VTE history for VTE; and age, BMI, smoking, and hypertension for arterial thromboembolism (mainly, acute myocardial infarction and ischemic stroke).
Based on the rather small number of outcomes, the Advisory Council limited the number of confounder variables to these well-established risk factors for the outcomes.
The analyses focused on comparisons among the three OC cohorts (DRSP-containing OCs, LNG-containing OCs, ‘Other OCs’).
Over the years of follow-up, however, many women:

  1. changed their OC (20.9%; e.g., switched from LNG-containing OCs to DRSP-containing OCs);
  2. switched to a non-oral hormonal contraceptive (3.1%; e.g., switched from LNG-containing OC to a fertility control patch); or
  3. stopped all forms of hormonal contraception either temporarily or permanently (44.2%; e.g., because of planned pregnancy or
         separation from partner).

All study participants --- including pregnant women --- were included in the follow-up until the end of the study, unless they withdrew their informed consent. This means that data on two additional cohorts --- users of non-oral hormonal contraceptives (NOHC) and non-users --- were also available.
Exploratory analyses of these data are also presented in this publication wherever they contribute to better understanding of the results. However, it must be noted that those women who stopped OC use might conceivably have done so due to factors that may influence the incidence of AEs (e.g., planned pregnancy, diagnosis of risk factors). It is also conceivable that women who are susceptible to specific AEs (e.g., VTE) experienced this outcome during OC use, stopped OC use, and started prophylactic treatment to avoid recurrence of the event. Therefore, these results must be considered as tentative.
The interim results of the EURAS study [13] were based on intention-to-treat (ITT) analyses. That is, all data from an individual patient were assigned to the treatment she used at study entry.
ITT analyses can be used to calculate the incidence for the outcome of interest fairly easily at any point in time. They are therefore well suited for monitoring patient safety during the early phase of the study, before many study participants switched their OC.
In the subsequent course of the study it became increasingly important to know precisely which preparation was actually being taken and the actual extent of total exposure to this OC (or OC group) at the time of an AE. The final analyses included both an “as treated” (AT) and an ITT analysis.
The safety conclusions of the study are based on the AT analyses because the ITT approach potentially dilutes differences between treatments. In this study, however, conclusions based on ITT results would not differ from the conclusions based on AT results.

All analyses were performed with the statistical package STATA 9.0.