Endocrine Abstracts

Introduction: Endometriosis is a complex reproductive disease without clear etiology, thought to be on the increase, and possibly related to endocrine disruptor (EDs) exposure such as PCBs, dioxins, phthalates, or pesticides. It is part of a national project on reproductive health indicators related to EDs in France. We aimed to build a nationwide indicator of endometriosis incidence based on routinely collected data, usable to study temporal and spatial trends for environmental purposes.

Methods: A multidisciplinary expertise gathering clinicians, epidemiologists, statisticians and experts in health data mining was composed to elaborate a strategy for tracking new cases of endometriosis from the French national health care data system. We built corresponding algorithms, using ICD-10 codes for diseases and/or medical acts and/or drugs. We performed a review of literature on the epidemiology of endometriosis and links with EDCs exposure. We analyzed the usefulness of the indicators regarding our environmental purposes.

Results: Monitoring endometriosis nationwide was only possible using hospital discharges. We built 3 indicators:

– The first one involved only diseases codes of endometriosis (N80);

– The second one involved also codes for the main surgical acts, thought to enable histological diagnosis, observed nationwide in hospital discharges (a posteriori approach);

– The third indicator was issued from an a priori approach: the experts deciding which diseases codes and acts were the most relevant to identify specific forms of endometriosis, i.e. endometriomas. For each indicator, incident cases were defined as the first stay recorded without occurrence at least in the 5 prior years. In the period 2006 to 2017, we identified nationwide 30 600, 23 600 and 7500 new annual operated cases, with indicators 1 to 3, respectively. For comparison, we estimated the annual crude incidence rate with the first indicator to 12.9/10 000, which is of the same order of magnitude as in European countries using similar methods.

Conclusion: All three indicators are usable to monitor hospital cases of endometriosis nationwide, and analyze temporal trends and spatial trends at the postcode scale for environmental purposes, to explore environmental issues that are geographically determined. The indicator one reflects the better the hospital incidence, the other ones could be used for sensibility studies. The main limits of this monitoring method are the undervaluation of the public health problem, because only hospital cases are included, and possible variations of treatment or coding. A close collaboration with clinicians is critical to overcome these limits.