This is retrospective observational study on premenopausal women, 18 to 50 years of age, attending university hospital outpatient contraception services, each of whom received dienogest-based hormonal contraceptives for at least 6 months. All patients in our clinic routinely sign informed consent to the anonymous use of their clinical data in scientific publications. Data refer to patient management (diagnostic procedures and treatments), performed in accordance with good clinical practice. Diagnostic procedures and treatments are paid partially or completely by the Italian national health system. No incentive is given to the patient. Data are recorded in an electronic database, from where they can be retrieved for retrospective analyses as necessary. Publication of results is approved, as in this case, by the hospital’s internal review board.

For the purposes of this study, retrospective data pertaining to 173 women consulting the outpatient service for contraception in the period December 2016–October 2018 was retrieved. Among the women in question, 95 had been suffering from menstrual pain or abnormal uterine bleeding and had been referred for an outpatient ultrasonographic evaluation to be performed in the follicular phase (days 4 to 9) of the menstrual cycle (Fig. 1). For these women, the following information were retrieved from the electronic database: age, number of births, previous uterine surgery, presence of abnormal uterine bleeding, duration of menstrual cycle in days, menstrual pain scored on a 0–10 cm visual analogue scale (VAS) [28], ultrasound uterine features, presence of pelvic endometriosis, and type of contraceptive prescribed. On the basis of their request and medical eligibility, each woman received either a combined hormonal contraceptive, a levonorgestrel intrauterine system (IUS), an etonorgestrel implant, or a progestin only pill (desogestrel). Data from a follow-up visit and ultrasound investigation performed after 6 months of treatment were also retrieved. On the basis of baseline ultrasound investigation, the women were divided into those with ultrasound features of adenomyosis (n = 55) and those without ultrasound features of adenomyosis (n = 40). Women already on hormones at baseline were excluded (n = 3 with and n = 2 without adenomyosis). Similarly, women who did not have a follow-up ultrasound at six months were not considered in this analysis (n = 4 with and n = 8 without adenomyosis). Among those with follow-up ultrasound, the vast majority of women received combined hormonal contraception containing dienogest in association with ethinylestradiol or in a quadriphasic association with 17$\beta$-estradiol. In order to ensure greater homogeneity of data, we excluded women who received levonorgestrel-IUS (n = 4 with and n = 4 without adenomyosis), etonorgestrel implant (n = 3 with and n = 3 without adenomyosis), vaginal ring (n = 1 with and n = 3 without adenomyosis) or progestin-only pill (n = 2 with adenomyosis). Accordingly, analysis was performed on 38 women with ultrasound features of adenomyosis and 20 women without ultrasound features of adenomyosis, all receiving dienogest-based combined hormonal contraception (Fig. 1).

Fig. 1.

Flow-chart of study design. DNG, dienogest.

Ultrasound investigations were performed in an outpatient service for gynaecological ultrasound investigation by one experienced sonographer (A.X.), who was blind to the type of treatment the woman was receiving. In our clinical practice, blindness is routine so that ultrasound investigations are as objective as possible [30]. Specifically, ultrasound was performed by a trans-vaginal ultrasound (TVS) probe during the proliferation phase of the menstrual cycle (days 5–10), using a GE E6 (GE Medical Systems, Zipf, Austria) ultrasound machine and a wideband 5–9 MHz transducer. Power Doppler was performed using fixed, preinstalled settings (frequency, 6–9 MHz; pulse repetition frequency, 0.6–0.3 kHz) to distinguish myometrial cysts from blood vessels. Uterine volume was measured by means of the ellipsoid formula (uterine longitudinal diameter $\times$ transverse diameter $\times$ anterior-posterior diameter $\times$ 0.5223), without considering the cervix. Endometrial thickness was also measured.

Ultrasound diagnosis of pelvic endometriosis was performed, in line with the IDEA consensus opinion, by 2D-TVS [31]. Ultrasound diagnosis of diffuse adenomyosis was defined, according to MUSA criteria [11], as the contemporaneous presence of at least two or more of the following features: (1) heterogeneous myometrium, seen as irregular myometrial echo-texture with decreased or increased echogenicity; (2) hypo-echoic striation in the myometrium (fan-shaped shadowing), seen as a pattern of thin acoustic shadows not arising from echogenic foci and/or leiomyoma (Fig. 1); (3) myometrial anechoic lacunae or cysts, seen as a round anechoic areas within the myometrium, hyper-echogenic islands, and sub-endometrial lines and buds (Fig. 2); (4) asymmetrical myometrial thickening of the uterine wall (Fig. 2); (5) globular uterine configuration; and/or; (6) increased uterine volume; (7) presence of a poorly defined, thickened, irregular, interrupted endo-myometrial junctional zone (JZ) [8] (Fig. 2). Information regarding other additional signs such as uterine tenderness and the question-mark or the ear sign were not available in the records [30]. Asymmetrical thickening of uterine walls was defined as thicker posterior or anterior uterine wall, unrelated to leiomyoma. To this end, the ratio between the anterior and posterior wall thickness was calculated. A ratio around 1 indicates that the myometrial walls are symmetrical, while a ratio well above or below 1 indicates asymmetry [11] (Fig. 2). In the selected sample there was no case of focal adenomyosis. The concomitant presence of fibroids was also assessed [11], and the volume of each fibroid was measured by means of the ellipsoid formula.

The study was focused on detecting modification of ultrasound parameters of adenomyosis during treatment. A secondary outcome was the clinical response in terms of VAS score for menstrual pain and length of menstrual cycles.

Fig. 2.

Ultrasund features of adenomyosis. (A) Hypoechoic striation in the myometrium (parallel shadowing), seen as a pattern of thin acoustic shadows. (B) Myometrial anechoic lacunae or cysts, seen as round anechoic areas within the myometrium. (C) Asymmetrical myometrial thickening of the uterine wall. (D) Poorly defined endo-myometrial junctional zone (JZ).

Statistical analysis was performed using StatView statistical software (version 5.01.98, SAS Institute Inc., Cary, NC, USA). Descriptive analysis was applied. Comparisons within groups were performed by paired t test. The chi-squared test was used to compare changes in prevalence over time. As applicable, means or prevalence in the two groups were compared via the Student’s t test or chi-squared test, respectively, as a secondary statistical analysis. A two-tailed p value 0.05 was considered as significant. Numerical data are expressed as mean $\pm$ standard deviation (SD).

The study was powered to document a within-group difference in volume between prior to and after 6 months of treatment that was greater than or at least equal to one standard deviation (SD) of the difference. In such a setting, if type I error is 0.05 and type II error is 0.2, 11 cases are sufficient to document an intragroup difference.

A secondary analysis evaluated other ultrasound features and compared data between women with and without concomitant pelvic endometriosis or with and without adenomyosis.

Among the 95 women whose records were retrieved, 55 had ultrasound features consistent with adenomyosis, while 40 women displayed inhomogeneous echo-textured myometrium with no ultrasound features of adenomyosis (Fig. 1). Analysis was performed only on those women with a baseline and 6-month scan who had received dienogest-based hormonal contraceptives (n = 58) (Fig. 1). Among these 58, 38 had ultrasound features of adenomyosis and 20 did not. Women received either estradiol valerate and dienogest (n = 22 women with adenomyosis and n = 8 women without adenomyosis) or ethinyl estradiol and dienogest (n = 16 with adenomyosis, and n = 12 without adenomyosis). The mean age was 36.2 $\pm$ 7.3 years and 40.3 $\pm$ 7.8 years (p = 0.06) in women with and without adenomyosis, respectively, and the number of to-term pregnancies in these groups was 0.8 $\pm$ 0.9 and 1.0 $\pm$ 1.1, respectively. Eleven women with adenomyosis (29%) and none of those without had undergone previous uterine surgery, such as laparotomic myomectomy or caesarean section (Table 1).

Uterine volume was 84.509 $\pm$ 53.695 mm${}^3$ and 80.513 $\pm$ 38.424 mm${}^3$ (Table 1), and endometrial thickness was 7.7 $\pm$ 0.5 mm and 7.5 $\pm$ 0.7 mm in women with and without adenomyosis, respectively. Twenty-six percent (n = 10) and 20% (n = 4) of women with and without adenomyosis had sub-serous or intramural uterine fibroids (Fig. 2). Mean fibroid volume was 12666 $\pm$ 35023 mm${}^3$, (diameter range 10–68 mm) and 6110 $\pm$ 3102 mm${}^3$ (diameter range 25–35 mm), respectively. No ultrasound feature of adenomyosis was detected in the uteri of women without a diagnosis of adenomyosis. Inhomogeneous myometrial echo-texture was present in 38/38 (100%) and globular configuration in 33/38 (87%) of the uteri of women with adenomyosis. Hypo-echoic striation of the myometrium (parallel shadowing) was present in 36/38 (95%) (Fig. 3), intra-myometrial cysts in 3/38 (8%), asymmetric uterine walls in 8/38 (21%), and poorly defined endo-myometrial junctional zone (JZ) in 34/38 (89%) of adenomyotic uteri. Pelvic endometriosis was detected at ultrasound in 22/38 (58%) women with adenomyosis and in no woman without adenomyosis.

Fig. 3.

Modification of Ultrasound Features of Adenomyosis. Ultrasound of the uterus before (A) and after 6 months (B) of therapy.

During the six-month period, in women with adenomyosis uterine volume decreased significantly by –10.601 $\pm$ 18.513 mm${}^3$ (p = 0.001) (–13.1 $\pm$ 22.1%) (Table 2). In women without adenomyosis, uterine volume increased, but not significantly so, by +24.367 $\pm$ 53.073 mm${}^3$ (p = 0.07, +28.0 $\pm$ 35.0%). Upon secondary statistical analysis, volume changes in uteri with and without adenomyosis were found to be significantly different (p = 0.001). Changes in uterine volume were not conditioned by fibroid modifications. Indeed, fibroid volume did not change significantly during the six months of treatment (from 12.666 $\pm$ 35.023 mm${}^3$ to 14.273 $\pm$ 33.368 mm${}^3$—a net change of +1.606 $\pm$ 4.203 mm${}^3$; p = 0.16). Mean endometrial thickness decreased to 3.4 $\pm$ 0.3 mm and 3.2 $\pm$ 0.3, in women with and without adenomyosis features with no difference between them.

Some sonographic features of adenomyosis were modified by treatment. Hypo-echoic striation of the myometrium (parallel shadowing) was present in 95% of cases and myometrial cysts in 8% at baseline, but both had disappeared completely after 6 months of hormonal contraception (p = 0.0001) (Table 2). The prevalence of uterine-wall asymmetry (8/38; 21%) was unchanged after treatment, but mean asymmetry decreased from a ratio of 2.8 $\pm$ 0.8 to 1.9 $\pm$ 0.7 (p = 0.0001). The prevalence of poorly defined endo-myometrial junctional zone (JZ) (34/38; 89%), inhomogeneous myometrial echo-texture (38/38; 100%), and globular uterus configuration (33/38; 87%) were not altered by treatment.

The mean VAS score for menstrual pain was 7.3 $\pm$ 3.2 and 1.3 $\pm$ 0.8 in women with and without adenomyosis (p = 0.0001). In women with adenomyosis, there was no difference in VAS score between those with and without pelvic endometriosis (7.6 $\pm$ 2.9 vs. 6.7 $\pm$ 3.8, p = 0.433). The duration of menstrual bleeding was 5.3 $\pm$ 1.2 days and 4.2 $\pm$ 0.3 days in women with and without adenomyosis, respectively (p = 0.0002).

During treatment, VAS score for menstrual pain decreased to 1.9 $\pm$ 2.8 (–4.0 $\pm$ 3.6; p = 0.0001) in women with adenomyosis, and to 0.5 $\pm$ 0.4 in women without ultrasound features of adenomyosis. In women with adenomyosis the reduction of menstrual pain was similar in women with and without pelvic endometriosis (–5.0 $\pm$ 3.6 vs. –5.6 $\pm$ 3.7; p = 0.655), and reached similar post-treatment values in both, specifically 2.57 $\pm$ 3.2 vs. 1.13 $\pm$ 1.95, respectively (p = 0.133).

Menstrual flow duration declined significantly in women both with (from 5.3 $\pm$ 1.2 to 4.1 $\pm$ 0.5 days; p = 0.0001) and without (from 4.2 $\pm$ 0.3 to 3.8 $\pm$ 0.2 days; p = 0.0001) ultrasound features of adenomyosis.

Side effects were few and minor among women who completed the 6 months of therapy. Spotting and breakthrough bleeding occurred in 9.6% of women, with no need for discontinuation. Twelve women on dienogest-based contraceptive did not return at follow-up. Considering the worst-case scenario that all these women had withdrawn due to side effects related to the treatment, continuation rate at 6 months of dienogest-based hormonal contraceptives was 83%.