Estimation of the Endometriosis Fertility Index prior to operative laparoscopy

STUDY QUESTION Can the Endometriosis Fertility Index (EFI) be estimated accurately before surgery? SUMMARY ANSWER The EFI can be estimated accurately based on mere clinical/ultrasound information, with some improvement after adding data from diagnostic laparoscopy. WHAT IS KNOWN ALREADY The EFI is a validated clinical instrument predicting the probability of pregnancy after endometriosis surgery without the use of ART. Being an end-of-surgery-score, it implies the decision for operative laparoscopy to be made in advance—hence, its role in the pre-surgical decision-making process remains to be established. STUDY DESIGN, SIZE, DURATION Single-cohort prospective observational study in 82 patients undergoing complete endometriosis excision (between June and December 2016). Two methods were used to estimate the final EFI: type A based on non-surgical clinical/ultrasound findings only, and type B based on the combination of non-surgical clinical/ultrasound findings and diagnostic laparoscopy data. To calculate EFI type A, an algorithm was created to translate non-surgical clinical/imaging information into rASRM (revised American Society of Reproductive Medicine)—and EFI points. EFI type A and type B estimates were assessed for their clinical and numerical agreement with the final EFI score. Agreement was defined as clinical if EFI scores were within the same range (0–4, 5–6, 7–10), and numerical if their difference was ≤1. PARTICIPANTS/MATERIALS, SETTING, METHODS All 82 patients underwent complete laparoscopic CO2-laser excision of any rASRM stage of endometriosis in the Leuven University Fertility Centre (LUFC) of University Hospitals Leuven, a tertiary referral centre for both endometriosis and infertility. An anonymized clinical research file was created. For each patient, three different data sets were created, in order to allow the estimation of the (surgical part) EFI and of the rASRM scores, defined as follows: ‘Estimated type A’ contained only non-surgical clinical/imaging data, ‘Estimated type B’ included type A information plus the information of the diagnostic laparoscopy and ‘Final EFI’ included information of type A, type B and all intra-operative information required to calculate the final EFI. To calculate EFI type A without surgical information, a set of rules was used to translate pre-surgical clinical/imaging information into (rASRM and EFI points). Scoring was done by one person (C.T.), with a time interval of 4 weeks between sessions for each EFI type. Next to the EFI, also rASRM score and stage were calculated. MAIN RESULTS AND THE ROLE OF CHANCE Agreement rate between estimated EFI type A and final EFI was high for both the clinical (0.915; 95% CI 0.832–0.965) and numerical definition (0.878; 95% CI 0.787–0.940). Agreement rates between estimated EFI type B and final EFI were even higher (clinical (0.988; 95% CI 0.934–1.000), numerical (0.963; 95% CI 0.897–0.992)). LIMITATIONS, REASONS FOR CAUTION Type A estimation is dependent on high-level gynaecological ultrasound expertise, which may not be available in all clinics. A small number of patients had no prior clinical, ultrasound (hard markers) or surgical confirmation of the diagnosis of endometriosis. When applying the estimated EFI type A in clinical practice, a priori assumptions of the presence or absence of endometriosis will need to be made in adjunct to the estimation of the estimated type A EFI when counselling patients on the potential benefit of an (at least diagnostic) laparoscopy. The level of agreement for type A or B should also be taken into account when counselling patients on the type of efforts undertaken to attempt to diagnose or rule out endometriosis. WIDER IMPLICATIONS OF THE FINDINGS As this study reports, the EFI can be estimated accurately based on clinical/ultrasound data only without the need for any surgical data. This means that the EFI could be used as an instrument to guide joint physician–patient decision-making between surgery, ART or other fertility management options for the individualized treatment of women with endometriosis-related infertility. STUDY FUNDING/COMPETING INTEREST(S) During this study period, C.T. was supported by FWO (Research Fund Flanders, Grant number 1700816N) and UZ Leuven KOF (University Hospitals Leuven, Klinisch Onderzoeksfonds). The LUFC received unrestricted research grants from Ferring Pharmaceuticals and Merck SA. Gedeon Richter and MSD sponsored travel to and attendance at scientific meetings. C.M. received consultancy fees from Lumenis (paid to KU Leuven, no private revenue). T.D. has been vice-president and head of global medical affairs infertility for the multinational pharmaceutical company Merck (Darmstadt, Germany) since 1 October 2015. He continues his academic appointment on a part-time basis as Professor of Reproductive Medicine at the University of Leuven (KU Leuven). T.D. has been vice-president and head of global medical affairs infertility for the multinational pharmaceutical company Merck (Darmstadt, Germany) since October 2015. He is also a Guest Professor in Reproductive Medicine and Biology at the Department of Development and Regeneration, Group Biomedical Sciences, KU Leuven (University of Leuven), Belgium, and an Adjunct Professor at the Department of Obstetrics and Gynecology in the University of Yale, New Haven, USA. This work was initiated before he joined Merck KGaA in October 2015, and completed during the subsequent years. TRIAL REGISTRATION NUMBER study registration number at UZ Leuven Clinical Trial Centre: S59221.