Effectiveness of dienogest in improving quality of life in Asian women with endometriosis (ENVISIOeN): interim results from a prospective cohort study under real-life clinical practice

The study was carried out at 36 sites in Thailand (five centres), Indonesia (ten centres), Republic of Korea (twelve centres), Malaysia (four centres), Philippines (three centres) and Singapore (two centres) in accordance with the amended version of the Declaration of Helsinki (Oct 2013) and complied with Good Clinical Practice. For more information, see Clinicaltrials.​gov (NCT02425462).

The inclusion criteria were: Asian female ≥18 years old; clinical or surgical diagnosis of endometriosis; presence of EAPP; uninfluenced decision taken by the physician to newly prescribe dienogest; written informed consent. Women were excluded in case of participation in an investigational program with interventions outside of routine clinical practice; any contraindication listed in the local summary of product characteristics. All eligible patients were treated according to standard medical guidelines or usual care of the participating institute.

The observation period of 24 months for each patient enrolled in this study consisted of the treatment phase and, in case of discontinuation of the treatment, the follow up period. Subjects were followed whether or not they remained on the treatment with dienogest. Information on the date and reasons for discontinuation were collected.

The primary endpoint was to assess the effectiveness of dienogest to change HRQoL with special attention to the dimension of pain. Patients were asked to complete the Endometriosis Health Profile-30 (EHP-30) at baseline and at the visits 6 months after start of treatment. The EHP-30 is a disease-specific, reliable and valid instrument to measure the effects of endometriosis on HRQoL, especially on physical, psychologic and social aspects, from the patients’ perspective, as described previously [17‐19]. Secondary effectiveness endpoint variables comprised the assessment of changes in the other core as well as modular HRQoL domains and changes in the severity of EAPP from baseline to 6-months visits measured by patient-reported numeric rating scale (NRS) with a 4-week recall period at every visit.

All TEAEs, including the ones leading to discontinuation were recorded at each visit throughout the study and summarized using the MedDRA coding system (not to be confused with the bleeding profile terminology). The bleeding profile was evaluated based on the following categories (definition within the figure): normal bleeding, irregular bleeding cycle, amenorrhea, intermenstrual bleeding/spotting. Further secondary endpoints analyzed within the scope of the interim analysis were continuation rate as well as patients’ and physicians’ satisfaction rate.

The treating physician collected historical data (demographic and clinical characteristics) from medical records if available and treatment-related data during visits that took place in routine practice.

To circumvent the general bias susceptibility inherent to observational studies, e.g. selection bias, it ensured that the study population represented routine patient profiles treated with dienogest in Asia. The patients were selected only based on inclusion and exclusion criteria in a consecutive manner.

The calculation of the sample size focused on the pain dimension in the EHP-30 score. Assumptions for the sample size have been based on earlier studies that collected information on quality of life in patients with endometriosis who have been treated daily with a progestin [20]. For this planned study similar results were expected. With a sample size of 696 subjects the 2-sided 99% confidence interval was expected to extend at most nominal +/− 2.25 in the score (which corresponds to +/− 4.5% of the expected mean change) with a probability of 95%. With a drop-out rate of about 20% it was advisable to include at least 870 subjects into this study.

In terms of safety this study was sufficiently large to capture uncommon adverse events. With the planned sample size at least one adverse event with a relative frequency of 0.3% was expected to be observed with a probability of 90%.

Effectiveness analyses were conducted on the efficacy analysis set (EFF) that included all patients with evaluable EHP-30 questionnaire at baseline and at least one evaluable EHP-30 questionnaire post-baseline between week 12 and 36 after start of treatment. Measurements after discontinuation of treatment were excluded. Safety analyses were performed on the full analysis set (FAS) and comprised all patients that took at least one dose of dienogest.

The HRQoL parameters were assessed as described previously. The score ranged between 0 (best health status) and 100 (worst health status) [18]. For EAPP measured by patient-reported NRS, the patients were asked to circle one number in a range between 0, referring to “absence of pain” and 10, corresponding to “unbearable pain”. Changes in the severity of EAPP from baseline visit to 6-months visit, measured by patient-reported NRS at every visit were evaluated overall and separately for subgroups stratified by baseline severity of EAPP, use of rescue medication, prior treatment and method of diagnosis. The patient numbers for the conducted analyses vary due to variable numbers of complete and evaluable questionnaires, especially for analyses of mean changes.

The analysis was primarily of explorative and descriptive nature. All variables were analyzed descriptively with appropriate statistical methods. SAS release 9.4 (SAS Institute, Cary, NC, USA) was used for data analysis. Continuous variables were described by visit (baseline and 6-months) and as mean change from baseline to 6 months for each item of the questionnaire, if applicable. For change from baseline in EHP-30 scores descriptive statistics as well as confidence intervals have been provided. Frequency distributions of patients with improvement, deterioration or no change in EHP-30 and EAPP scores were summarized. Missing data was not replaced and was given as “missing” in the tables. All adverse events were presented within incidence tables as preferred terms by system organ class. The descriptive analysis of adverse events was the basis to assess tolerability and safety of dienogest in the study population. The analyses described were performed on TEAEs. Non TEAEs were tabulated without further stratification. Partially missing TEAE onset dates were imputed following a worst case approach. The earliest possible date was used (as described above), but in case the AE onset could be before or after start of dienogest, TEAE onset was imputed by the first day of dienogest treatment.

Overall, a total of 895 patients were enrolled at 36 sites from 04/2015 to 08/2016. The FAS comprised 865 patients who were eligible, gave their informed consent and took at least one dose of dienogest. Furthermore, 510 patients with evaluable primary outcome were included into the EFF. The mean ± SD duration of treatment with end of observation was 14.1 ± 7.5 months (n = 97 evaluable patients; EFF). Leading reasons for end of observation were patient lost to follow-up (n = 37/97; 38.1%) and regular end of study (n = 35/97; 36.1%).

Baseline characteristics are summarized in Table 1. The most common symptoms of endometriosis were dysmenorrhea (n = 684/865; 79.1%) and chronic pelvic pain (n = 279/865; 32.3%). While endometriosis was diagnosed within 1 year before initial visit in the majority of the patients (n = 644/864; 74.5%), only less than half of the women reported onset of first symptoms within this time span (n = 402/865; 46.2%). Moreover, 29.9% of women (n = 259/865) documented appearance of first symptoms between one and 5 years before and 23.0% (n = 199/865) more than 5 years before. In fact, 24.6% of patients diagnosed within 1 year before initial visit experienced their first symptoms 1–5 years earlier and 12.4% of patients even > 5 years before, amounting to approximately 37% of patients with delayed diagnosis of endometriosis. The time period from appearance of first symptoms to first diagnosis tended to be shorter in surgically diagnosed patients than in patients with clinical only diagnosis (mean, 21.3 ± 43.9 months vs. 29 ± 44.8 months, respectively). The majority of women (n = 449/514; 87.4%) underwent prior surgery to treat endometriosis-associated pain. The proportion of women with pain recurrence despite surgical treatment was 27.4% (n = 123/449) and the mean ± SD duration until pain recurrence after surgery was 19.3 ± 21.0 months (duration was evaluable for 114 women). A total of 179 women received previous hormonal treatment, with 56.4% of those women (n = 101/179) suffering from pain recurrence after a mean ± SD time of 9.9 ± 13.9 months (evaluable for 97 women). Pain medications were given to 89 patients (of 514 with prior endometriosis treatment; 17.3%), predominantly 2–3 days per week (n = 50/89; 56.2%), with a mean ± SD pain recurrence after 11.9 ± 37.2 days (evaluable for 79 patients) after or during pain therapy.

In total, there were nine serious adverse events (SAEs) reported in eight patients (Table 5). Thereof, anaemia (2/9 events) represented the most common. In eight cases the reason for seriousness was “Hospitalization necessary or prolonged”, in the remaining case “Important medical event”. All SAEs were recovered or resolved. Drug-related TEAEs were documented for 272 of 865 patients (31.5%), with amenorrhea (5.9%), metrorrhagia (5.1%) and vaginal haemorrhage 4.5%) representing the most common. A minority of patients (n = 45/865; 5.2%) discontinued dienogest therapy due to drug-related TEAEs. Abnormal uterine bleeding (preferred terms: vaginal haemorrhage, menorrhagia and metrorrhagia) was the most prominent reason for discontinuation (0.9%; n = 8/865). TEAEs irrespective of causality were documented in 35.5% of patients (n = 307/865) and were predominantly of mild-to-moderate intensity. Most common TEAEs are presented in Table 6.