Is the multinational, surveillance PRO-E2 study informative for all countries? The Italian data on VTE and contraceptive effectiveness

Abstract

Purpose

To evaluate whether the thromboembolic risk and contraceptive effectiveness of NOMAC-E2 observed in the PRO-E2 study can be extended to each participating country, as lifestyle, cardiovascular risk factors and prescribing habits may differ geographically. This analysis was performed on the PRO-E2 Italian subpopulation, where smoking habit and women over 35 years were more prevalent compared with the overall study population.

Materials and methods

Data from NOMAC-E2 or levonorgestrel-containing COCs (COC_LNG) new users were descriptively analysed. Incidence rates of thrombosis (events/10,000 women-years [WY]) and the Pearl Index (pregnancies/100 WY) were calculated.

Results

Overall, 11,179 NOMAC-E2 and 8,504 COC_LNG users were followed up to 2 years (34,869 WY). The NOMAC-E2 cohort included more women over 35 vs. COC_LNG (37.7% vs. 31.8%; p = 0.001). A comparable low risk of combined deep venous thrombosis of lower extremities (DVT) and pulmonary embolism (PE) was observed in NOMAC-E2 (1.7/10,000 WY; 95% CI: 0.21–6.2) and COC_LNG users (6.6/10,000 WY; 95% CI: 2.4–14.4). Similar results were obtained by considering all thromboembolic events (VTE). Unintended pregnancies did not differ between NOMAC-E2 (0.12/100 WY; 95% CI: 0.06–0.21) and COC_LNG (0.15/100 WY; 95% CI: 0.08–0.26) cohorts.

Conclusion

Despite the higher age and tobacco use, findings from the Italian subpopulation were broadly consistent with overall PRO-E2 results, confirming a similar low thromboembolic risk and high contraceptive effectiveness of NOMAC-E2 and COC_LNG.

Short condensation

This subgroup analysis of the PRO-E2 study provides comprehensive epidemiological data on the use of combined oral contraceptives in a large Italian cohort, with a higher prevalence of women over 35 years and smokers. The study confirms the low thromboembolic risk and high contraceptive effectiveness of NOMAC-E2 pill.

摘要

目的：为了评估在PRO-E2研究中观察到的NOMAC-E2的血栓栓子风险和避孕效果是否可以推广到每个参与国家, 因为生活方式、心血管风险因素和处方习惯可能存在地域差异。本研究是在PRO-E2的意大利亚组人群上进行, 与总体研究人群相比, 意大利人群吸烟和35岁以上的女性更普遍。

材料和方法：描述性分析NOMAC-E2或新应用含左旋诺孕酮的COCS(COC_LNG)者的数据。计算血栓发生率(事件/10,000妇女年[WY])和珀尔指数(怀孕/100WY)。

结果：总体而言, 11,179名NOMAC-E2和8504名COC_LNG应用者获得了长达2年的随访(34,869 WY)。NOMAC-E2队列纳入了更多35岁以上的女性(37.7% vs 31.8%；p=0.001)。NOMAC-E2 (1.7/10,000 WY；95%CI：0.21-6.2)和COC_LNG应用者(6.6/10,000 WY；95%CI：2.4-14.4)合并下肢深静脉血栓形成(DVT)和肺栓塞(PE)的风险相对较低。考虑所有血栓栓塞性事件(VTE)也得到了类似的结果。意外妊娠在NOMAC-E2 (0.12/100WY；95%CI：0.06-0.21)和COC_LNG (0.15/100WY；95%CI：0.08-0.26)队列中没有差异。

结论：尽管年龄更大、烟草者更多, 但来自意大利人群的研究结果与总体PRO-E2结果大体一致, 证实了NOMAC-E2和COC_LNG具有相似的低血栓栓塞率和高避孕效果。

概括

PRO-E2研究的这一亚组分析提供了有关在35岁以上妇女和吸烟者较多的意大利队列中复方口服避孕药的全面流行病学数据。这项研究证实了NOMAC-E2的低血栓栓塞率和高避孕效果。

Keywords:

• NOMAC-E2
• nomegestrol acetate
• oestradiol
• levonorgestrel
• venous thromboembolism
• unintended pregnancy
• Italy

Introduction

Combined oral contraceptives (COCs) remain one of the most common, safe and effective methods of reversible birth control. Nonetheless, COC use is not devoid of side effects and health-related risks [1].

The major serious adverse effect associated with COC use is venous thromboembolism (VTE). VTE risk is driven by both genetic and acquired risk factors [2]. COC-related increase of VTE risk seems to be associated with the total oestrogenic potency of the formulation, which depends on both oestrogen and progestin contribution. COC oestrogenic component was shown to promote a prothrombotic state by stimulating oestrogen-sensitive haemostatic factors and several hepatic proteins. This is particularly relevant for the synthetic oestrogen ethinylestradiol (EE), which entails a higher oestrogenic potency than natural oestrogens given its resistance to hepatic inactivation [3–5]. Progestins in COCs modulate the oestrogen-mediated procoagulant effects in a variable manner according to their different activity on hormonal receptors, especially androgen receptors. Androgenic progestins are able to partially counteract oestrogens’ stimulatory effects, while non-androgenic or antiandrogenic progestins were demonstrated to elicit low or no modulatory effects [1,4,6].

EE continues to be the most frequently used oestrogen in COCs. As a clear relationship between the magnitude of VTE risk and EE dose was proven, moving from 50 to 30 μg EE, EE dose has been further reduced up to 15 μg [1,2].

The current use of low-dose COCs (≤35 μg EE) is associated with an overall two-fold to three-fold increased residual VTE risk [3,7]. The risk may vary depending on the progestin in the formulation [2]. Levonorgestrel (LNG) is assumed to entail the lowest VTE risk [2,8]. Indeed, COCs with less or no androgenic (desogestrel and gestodene), or anti-androgenic progestins (cyproterone acetate, drospirenone, dienogest and chlormadinone acetate) seem to be associated with a higher VTE risk compared with LNG-containing COCs (COC_LNG) [2,7–9].

More recently, attempts to replace EE with natural oestradiol to further improve COC safety have been accomplished [1,6,10].

A monophasic COC containing 2.5 mg nomegestrol acetate (NOMAC) and 1.5 mg micronized 17β-oestradiol (E2) in a 24/4 regimen was authorised for oral contraception by the European Medicines Agency (EMA) in 2011. Micronized E2 yields good bioavailability, and has weak oestrogenic effects and a mild metabolic impact on oestrogen-sensitive hepatic proteins [10]. NOMAC has strong anti-gonadotropic and anti-oestrogenic activities, and a moderate anti-androgenic activity [10–12]. NOMAC-E2 contraceptive efficacy was demonstrated to be maintained even when one active pill is missed any time and two active tablets are missed mid-cycle, possibly due to NOMAC long elimination half-life (about 46 h); all other cases entail a loss of effectiveness [11,13].

NOMAC-E2 demonstrated a lower metabolic impact [10,14] and fewer changes on coagulation and fibrinolysis markers compared with COCs containing LNG/EE [14,15].

NOMAC-E2 safety and effectiveness were further supported by the large international active surveillance study PRO-E2, being conducted to fulfil a post-marketing requirement to EMA [11,12]. NOMAC-E2 was not associated with a higher VTE risk compared with COC_LNG, demonstrating also a slightly lower incidence of deep venous thrombosis (DVT) and pulmonary embolism (PE) [12]. Moreover, the unintended pregnancy risk was significantly lower in NOMAC-E2 vs. COC_LNG users, especially in women under 25 years [16]. NOMAC -E2 entailed a comparable VTE risk and contraceptive effectiveness vs. COC_LNG in perimenopausal women [17].

Environmental and genetic risk factors for VTE in women using COCs presumably vary among countries, and this may account for a different VTE incidence in different geographic areas [18]. Women’s lifestyles and physician prescribing habits may also have a role in this geographic variance. Whether PRO-E2 results can be extended to each participating country or not remains to be verified.

A significant number of Italian women were recruited in the PRO-E2 study (i.e., 20,444), comprising the second largest subpopulation of PRO-E2 and including more women over 35 years and current smokers compared with the overall study population. This is particularly relevant as advancing age and smoking habit are important contributors to VTE risk.

Notably, no real-world data on contraception and COC-associated VTE risk have been previously reported in a large Italian cohort. As Italy has still one of the lowest rates of oral contraceptive use in Europe, country-level real-world data putting COC risks and benefits into local perspective may support contraceptive counselling, especially regarding new COC combination safety, and help improve awareness among Italian women, fostering in turn COC use [19].

Thus, this analysis aims at evaluating the safety and efficacy of NOMAC-E2 compared with COC_LNG among Italian COC users in the PRO-E2 study to provide further insights on COC risks and effectiveness in Italy.

Materials and methods

The PRO-E2 study methodology has been previously reported [11,12]. Briefly, new users (starters and restarters) of NOMAC-E2 and COC_LNG were followed in a real-life setting. Starters were first-ever users of any COC, while restarters were restarting hormonal contraceptive (HC) use with a COC after at least a 2-month break. PRO-E2 primary objective was to assess the risk of DVT of lower extremities and PE in NOMAC-E2 vs COC_LNG users. Secondary objectives included all VTE risk and pregnancy outcome assessments. Ethical approval was obtained as required by local law and an independent Safety Monitoring and Advisory Council monitored the study.

Study population

The study was conducted in Europe, Australia, and Latin America between 2014 and 2020 by health care professionals (HCPs).

All women newly prescribed an eligible COC (i.e., NOMAC-E2, COC_LNG monophasic preparation containing 20–30 μg EE, or COC_LNG multiphasic preparation containing up to 40 μg EE) could participate if they had not used a COC in the past 2 months, signed an informed consent form, and completed a baseline questionnaire [11,12].

Overall, 101,498 women were enrolled in the PRO-E2 study and 91,313 were followed. Of these 101,498 women, 20,444 were enrolled and 19,683 were followed at Italian study sites.

Baseline survey and follow-up

Participants completed a baseline questionnaire to record demographic data, reason(s) for COC prescription, medical history, concomitant medications, previous contraceptive use and lifestyle factors. At 6, 12 and 24 months participants received via mail/e-mail follow-up questionnaires on contraceptive use, pregnancy and occurrence of other outcomes of interest. Self-reported outcomes of interest were classified by investigators as ‘confirmed’ or ‘not confirmed’, and serious cardiovascular outcomes were further subjected to blinded adjudication, as described in detail previously. During the follow-up period women could stop or switch to other COC or HC for any reason; thus, additional sub-cohorts other than NOMAC-E2 and COC_LNG formed [12].

Evaluation

Statistical analyses on the PRO-E2 Italian subpopulation, focusing on NOMAC-E2 and COC_LNG sub-cohorts, are descriptive and based on the ‘as treated’ (AT) population (outcomes of interest were assigned to the COC used at the time of the event). Incidence rates (IR) per 10,000 women-years (WY) and corresponding 95% Confidence Intervals (CI) are provided for VTE and other safety outcomes. Unintended pregnancies are described by the Pearl Index (PI, i.e., contraceptive failure number per 100 WY of exposure), and corresponding 95% CI.

As the number of confirmed events in the Italian subpopulation was too low for allowing a meaningful time-to-event analysis, only IR and PI are presented. The chi squared test and Student’s ‘t’ test were used to compare frequencies and means, respectively, related to baseline characteristics, when considered relevant.

Results

Overall, 20,444 women were enrolled in the PRO-E2 study in Italy. Of these, 19,683 started using an eligible COC and were followed (34,869 WY of observation): 11,179 NOMAC-E2 and 8,504 COC_LNG users. Most COC_LNG users (93.5%) were prescribed at study entry COC_LNG monophasic preparation containing 20 μg EE. Starters comprised 48.6% and 51.4% of NOMAC-E2 and COC_LNG cohorts, respectively.

During the follow-up period, 1,257 women switched to other COCs (COC_Other sub-cohort), 332 switched to other HC (OHC sub-cohort) and 10,499 stopped using HCs (ex-user sub-cohort). This is line with data on previous HC use duration at study entry, showing a general trend whereby HCs were most frequently stopped between 1 and 5 years (40%) and during the first year of use (29.3%), with similar percentages between NOMAC-E2 and COC_LNG cohorts.

Baseline characteristics

There were no substantial differences between cohorts with respect to most baseline characteristics (Table 1). NOMAC-E2 users had a higher mean age (31.8 ± 9.53 years) than COC_LNG users (30.7 ± 9.30 years) (p < 0.001). More NOMAC-E2 users were aged 35 years or older compared with COC_LNG users (37.7% vs. 31.8%, respectively; p = 0.001). Participants aged 40 years or older were 24.3% and 19.5% (p = 0.001) in the NOMAC-E2 and COC_LNG cohorts, respectively, while COC users younger than 20 years were 10.1% and 12.5% (p = 0.001), respectively. Weight and body mass index were comparable between both cohorts. More NOMAC-E2 users reported having more than a university entrance level education vs. COC_LNG users (25.1% vs. 22.5%, p = 0.001).

Table 1. Selected baseline characteristics by user cohort.

	NOMAC-E2	COCLNG	Total
Number (%) of women	11,179 (56.8)	8,504 (43.2)	19,683 (100)
WY (%)	12,360 (35.4)	9,653 (27.7)	34,869 (100)^a
Age (years)
Mean (SD)	31.8 (9.53)	30.7 (9.30)	31.3 (9.45)
Age category, n (%)
<20 years	1,127 (10.1)	1,059 (12.5)	2,186 (11.1)
20–<25 years	2,267 (20.3)	1,818 (21.4)	4,085 (20.8)
25–<30 years	1,967 (17.6)	1,607 (18.9)	3,574 (18.2)
30–<35 years	1,608 (14.4)	1,312 (15.4)	2,920 (14.8)
35–<40 years	1,493 (13.4)	1,052 (12.4)	2,545 (12.9)
≥40 years	2,717 (24.3)	1,656 (19.5)	4,373 (22.2)
Weight (kg)
Mean (SD)	60.6 (11.18)	60.9 (11.70)	60.7 (11.41)
BMI (kg/m^2)
Mean (SD)	22.51 (3.98)	22.64 (4.13)	22.57 (4.05)
BMI category, n (%)
<20	2,988 (26.7)	2,244 (26.4)	5,232 (26.6)
20–<25	5,688 (50.9)	4,214 (49.6)	9,902 (50.3)
25–<30	1,756 (15.7)	1,441 (16.9)	3,197 (16.2)
30–<35	431 (3.9)	382 (4.5)	813 (4.1)
≥35	137 (1.2)	115 (1.4)	252 (1.3)
Education level, n (%)
Less than university entrance level	2,482 (22.2)	2,136 (25.1)	4,618 (23.5)
University entrance level	5,874 (52.5)	4,448 (52.3)	10,322 (52.4)
More than university entrance level	2,805 (25.1)	1,915 (22.5)	4,720 (24.0)

^a Describes WY from the whole study population, including women who stopped using hormonal contraceptives (ex-user) or switched to other contraceptives (i.e., progestin-only method, injections, implants, LNG-releasing intrauterine devices, contraceptive patches) during the follow-up period.

NOMAC-E2: nomegestrol acetate and 17β-oestradiol; COC_LNG: levonorgestrel-containing combined oral contraceptives; WY: women-years; SD: standard deviation; BMI: body mass index.

Cohorts were also comparable in relation to other cardiovascular risk factors (Table 2).

Table 2. Selected cardiovascular risk factors at study entry by user cohort.

	NOMAC-E2	COCLNG	Total
Number (%) of women	11,179 (56.8)	8,504 (43.2)	19,683 (100)
DVT, n (%)	9 (0.08)	8 (0.09)	17 (0.09)
PE, n (%)	0 (0.00)	2 (0.02)	2 (0.01)
Myocardial infarction, n (%)	0 (0.00)	4 (0.05)	4 (0.02)
Stroke, n (%)	0 (0.00)	1 (0.01)	1 (0.005)
Treated high blood pressure, n (%)	123 (1.1)	67 (0.8)	190 (1.0)
Thrombophilia, n (%)	27 (0.2)	33 (0.4)	60 (0.3)
Other severe diseases, n (%)	290 (2.6)	219 (2.6)	509 (2.6)
Any surgery, n (%)	3,548 (31.7)	2,765 (32.5)	6,313 (32.1)
Current smokers, n (%)	3,208 (28.7)	2,625 (30.9)	5,833 (29.6)
Heavy smokers^a, n (%)	381 (3.4)	248 (2.9)	629 (3.2)
Family history of VTE, n (%)	273 (2.4)	201 (2.4)	474 (2.4)
Family history of ATE, n (%)	281 (2.5)	196 (2.3)	477 (2.4)

^a >15 cigarettes per day.

NOMAC-E2: nomegestrol acetate and 17β-oestradiol; COC_LNG: levonorgestrel-containing combined oral contraceptives; DVT: deep venous thrombosis; PE: pulmonary embolism; VTE: venous thromboembolism; ATE: arterial thromboembolism.

VTE

DVT of lower extremities and PE

In women without pre-defined risk factors at study entry (i.e., pregnant within 3 months of treatment initiation, a history of cancer/chemotherapy, or an increased genetic risk of VTE), nine confirmed DVT of lower extremities and PE were observed over the course of 32,791 WY of follow-up. No additional event was observed when including women with pre-defined risk factors at study entry.

The incidence rate per 10,000 WY of confirmed DVT of lower extremities and PE was lower in the NOMAC-E2 cohort (2 DVT and PE over 11,680 WY of follow-up; 1.7/10,000 WY, 95% CI, 0.21–6.2) compared with the COC_LNG cohort (6 DVT and PE over 9,065 WY of follow-up; 6.6/10,000 WY 95% CI, 2.4–14.4) (Table 3). Of note, all events observed in the COC_LNG cohort occurred among users of 20 μg EE/LNG monophasic COCs (8,448 WY).

Table 3. Confirmed DVT of the lower extremities and PE in women without pre-defined risk factors^a at study entry by user (sub-)cohort.

	NOMAC-E2	COCLNG^b	COCother^c	OHC^c	Ex-user^c	Total
Number of WY	11,680	9,065	1,854	550	9,642	32,791
Confirmed DVT of the lower extremities and PE
Number	2	6	0	1	0	9
Incidence per 10,000 WY (95% CI)	1.7 (0.21–6.2)	6.6 (2.4–14.4)	0.00 (0.00–19.9)	18.2 (0.46–100.9)	0.00 (0.00–3.8)	2.7 (1.3–5.2)
Confirmed DVT of the lower extremities
Number	1	5	0	0	0	6
Incidence per 10,000 WY (95% CI)	0.86 (0.02–4.8)	5.5 (1.8–12.9)	0.00 (0.00–19.9)	0.00 (0.00–66.9)	0.00 (0.00–3.8)	1.8 (0.67–4.0)
Confirmed PE
Number	1	2	0	1	0	4
Incidence per 10,000 WY (95% CI)	0.86 (0.02–4.8)	2.2 (0.27–8.0)	0.00 (0.00–19.9)	18.2 (0.46–100.9)	0.00 (0.00–3.8)	1.2 (0.33–3.1)

^a Pregnant within 3 months of study entry, history of cancer/chemotherapy, increased genetic risk of VTE (e.g., Factor V Leiden, Protein S or C deficiency).

^b COC_LNG cohort includes 4 WY of >40μg EE use.

^c Data on these sub-cohorts are reported for completeness of information, but were not subjected to the present subgroup analysis.

NOMAC-E2: nomegestrol acetate and 17β-oestradiol; COC_LNG: levonorgestrel-containing combined oral contraceptives; COC_Other: combined oral contraceptives other than NOMAC-E2 or COC_LNG; OHC: other hormonal contraceptives; WY: women-years; DVT: deep venous thrombosis; PE: pulmonary embolism; CI: confidence interval.

Note. One woman in the COC_LNG cohort experienced both a DVT and a PE.

All VTE

Overall, 14 out of 35 VTEs reported over 34,869 WY of follow-up (including, but not restricted to, DVT of lower extremities and PE) were confirmed. In particular, 4 confirmed VTEs occurred in NOMAC-E2 users (3.2/10,000 WY; 95% CI, 0.88–8.3) and 7 in COC_LNG users (7.3/10,000 WY; 95% CI, 2.9–14.9) (Table 4). Of note, all VTE reported in the COC_LNG cohort were observed among users of 20 μg EE/LNG monophasic preparations (9,026 WY).

Table 4. All confirmed VTE by user (sub-)cohort.

	NOMAC-E2	COCLNG^a	COCother^c	OHC^c	Ex-user^c	Total
Number of WY	12,360	9,653	1,951	590	10,314	34,869
All confirmed VTEs
Number	4	7	1	1	1	14
Incidence per 10,000 WY (95% CI)	3.2 (0.88–8.3)	7.3 (2.9–14.9)	5.1 (0.13–28.5)	16.9 (0.43–94.0)	0.97 (0.02–5.4)	4.0 (2.2–6.7)
Idiopathic VTE^b
Number	3	7	1	1	1	13
Incidence per 10,000 WY (95% CI)	2.4 (0.50–7.1)	7.3 (2.9–14.9)	5.1 (0.13–28.5)	16.9 (0.43–94.0)	0.97 (0.02–5.4)	3.7 (2.0–6.4)

^a COC_LNG cohort includes 4 WY of >40μg EE use.

^b VTEs excluding those cases with acute risk factors such as pregnancy, delivery, trauma, immobilisation, long-haul travel, surgery, chemotherapy.

^c Data on these sub-cohorts are reported for completeness of information, but were not subjected to the present subgroup analysis.

NOMAC-E2: nomegestrol acetate and 17β-oestradiol; COC_LNG: levonorgestrel-containing combined oral contraceptives; COC_Other: combined oral contraceptives other than NOMAC-E2 or COC_LNG; OHC: other hormonal contraceptives; WY: women-years; VTE: venous thromboembolism; CI: confidence interval.

All confirmed VTEs except one event in the NOMAC-E2 cohort were considered idiopathic VTEs. No confirmed VTE was fatal.

Unintended pregnancies

Overall, 32 confirmed unintended pregnancies were observed in all study sub-cohorts, and specifically 13 in the NOMAC-E2 cohort (0.12/100 WY; 95% CI, 0.06–0.21) and 13 in the COC_LNG cohort (0.15/100 WY; 95% CI, 0.08–0.26) (Table 5). All unintended pregnancies reported in the COC_LNG cohort occurred among users of 20 μg EE/LNG monophasic preparations (7,895 WY).

Table 5. Confirmed unintended pregnancies (contraceptive failure) by user (sub-)cohort.

	NOMAC-E2	COCLNG^a	COCother^c	OHC^c	Total
Number of WY	10,677	8,430	1,775	541	34,829
Confirmed unintended pregnancies
Number	13	13	5	1	32
Incidence per 100 WY (95% CI)	0.12 (0.06–0.21)	0.15 (0.08–0.26)	0.28 (0.09–0.66)	0.18 (0.005–1.0)	0.09 (0.06–0.13)
'Perfect HC use’^b
Number	3	5	1	1	10
Incidence per 100 WY (95% CI)	0.03 (0.006–0.08)	0.06 (0.02–0.14)	0.06 (0.001–0.31)	0.18 (0.005–1.0)	0.03 (0.01–0.05)
'Non-perfect HC use’ ^b
Number	10	8	4	0	22
Incidence per 100 WY (95% CI)	0.09 (0.04–0.17)	0.09 (0.04–0.19)	0.23 (0.06–0.58)	0.00 (0.00–0.68)	0.06 (0.04–0.10)

^a COC_LNG cohort includes 3 WY of >40 μg EE use.

^b ‘Perfect HC use’ – the reported use was compliant with the instructions on the package insert and no extenuating circumstances were present (e.g., diarrhoea, vomiting or the use of antibiotics). ‘Non-perfect HC use’ – the reported use was not compliant with the instructions on the package insert.

^c Data on these sub-cohorts are reported for completeness of information, but were not subjected to the present subgroup analysis.

NOMAC-E2: nomegestrol acetate and 17β-oestradiol; COC_LNG: levonorgestrel-containing combined oral contraceptives; COC_Other: combined oral contraceptives other than NOMAC-E2 or COC_LNG; OHC: other hormonal contraceptives; WY: women-years; CI: confidence interval.

The confirmed unintended pregnancies that occurred despite self-reported ‘perfect use’ (i.e., HC use was compliant with instructions on the package insert and no extenuating circumstances, such as diarrhoea, vomiting or antibiotic use, were present) were 3 out of 13 in NOMAC-E2 users (0.03/100 WY; 95% CI, 0.006 − 0.08) and 5 out of 13 in COC_LNG users (0.06/100 WY; 95% CI, 0.02 − 0.14).

Analysing NOMAC-E2 and COC_LNG contraceptive failure by age (Figure 1), there were 8 confirmed unintended pregnancies among women under 25 years in NOMAC-E2 users (0.23/100 WY; 95% CI, 0.10–0.46) and 4 in COC_LNG users (0.13/100 WY; 95% CI, 0.04–0.34). Among women aged between 25 and 40 years, 5 confirmed unintended pregnancies occurred in NOMAC-E2 (0.11/100 WY; 95% CI, 0.04–0.26) and 9 in COC_LNG users (0.24/100 WY; 95% CI, 0.11–0.45). No confirmed unintended pregnancy was observed in women over 40 years.

Figure 1. Contraceptive failure by user (sub-)cohort and age category.

Discussion

Findings and interpretation

In the PRO-E2 Italian subpopulation, the risk of DVT of lower extremities and PE was comparable between NOMAC-E2 and COC_LNG cohorts, with a lower incidence, albeit not statistically significant, in NOMAC-E2 vs. COC_LNG users. Similar results were observed for all VTE risk.

The contraceptive failure rate among Italian study participants was comparable between the two cohorts. The contraceptive effectiveness of NOMAC-E2 and COC_LNG did not differ even after stratification by age group (i.e., women younger than 25 years and women aged between 25 and 40 years).

Results in the context of what is known

PRO-E2 study was primarily aimed at assessing the non-inferiority of NOMAC-E2 in terms of the risk of DVT of lower extremities and PE vs COC_LNG, which are assumed to entail the lowest VTE risk and thus were considered as the best comparator.

Patient characteristics and demographics for the Italian subpopulation were generally similar to those of the overall PRO-E2 population, including the high proportion of women who stopped or switched the prescribed COC once during follow-up [11,12]. However, a greater proportion of Italian women in the study were restarters (i.e., 50.2% vs. 36.5% in the overall population [12]) and current smokers (29.6% vs. 21.2% [11], respectively). Moreover, the Italian subpopulation included more women aged 35 years or older (35.1%) compared with the overall population (29.6%) [11], among whom 22.2% and 15.1% [11], respectively, were over 40 years.

Safety results on VTE risk in the Italian subpopulation were generally aligned with the overall PRO-E2 results, in spite of a higher prevalence of women over 35 years of age and with smoking habit. NOMAC-E2 thromboembolic risk did not differ from that of COC_LNG in both the Italian and overall population. Of note, results of the overall PRO-E2 population excluding Russia were considered for comparison due to the particularly low VTE incidence observed in Russian participants, which is presumably ascribable to Russia’s peculiar features in terms of behaviour and health care system, as previously discussed by Reed and colleagues [12]. In particular, the incidence of thromboses was lower in NOMAC-E2 vs. COC_LNG users, in both the Italian subpopulation (1.7 vs. 6.6 per 10,000 WY for DVT of lower extremities and PE, and 3.2 vs. 7.3 per 10,000 WY for all VTE, respectively) and the overall study population excluding Russia (3.3 vs. 4.7 per 10,000 WY for DVT of lower extremities and PE, and 4.1 vs. 5.8 per 10,000 WY for all VTE, respectively) [12].

The observed VTE risk for COC_LNG in the Italian subpopulation is slightly higher than that of the overall study population [12], but consistent with the previous EURAS-OC study (8/10,000 WY) [20] and what is currently reported as the VTE incidence for COC_LNG (5–7/10,000 WY) [21,22].

Moreover, a similar trend of non-inferiority of NOMAC-E2 in comparison with COC_LNG was previously observed for another natural oestrogen-based COC (dienogest/oestradiol valerate) in the INAS-SCORE study [23]. By pooling data from PRO-E2 and INAS-SCORE studies the cumulative VTE incidence rate for E2-containing COC resulted 3.2/10,000 WY vs. 4.5/10,000 WY for COC_LNG [24].

As regards contraceptive failure, a comparable risk was observed between the NOMAC-E2 and COC_LNG cohorts in the Italian subpopulation. Conversely, the overall PRO-E2 results showed a significantly lower risk of unintended pregnancy in NOMAC-E2 users (0.15/100 WY; 95% CI, 0.11–0.19) vs. COC_LNG users (0.41/100 WY; 95% CI, 0.35–0.47) (adjusted HR 0.45, p < 0.0001), particularly in women aged under 25 years [16], while no substantial difference was observed for women over 40 [17]. These results may reflect the fact that a higher contraceptive failure rate was observed in COC_LNG users in the overall population vs. the Italian subpopulation. This in turn may be ascribed to a higher rate of unintended pregnancies related to non-perfect COC use that occurred in the COC_LNG cohort in the overall population (0.24/100 WY; 95% CI, 0.20–0.29) vs. the Italian subpopulation (0.09/100 WY; 95% CI, 0.04–0.19). Of note, this incidence rate in NOMAC-E2 users was comparable between the overall (0.10/100 WY; 95% CI, 0.07–0.14) and the Italian populations (0.09/100 WY; 95% CI, 0.04–0.17).

Indeed, COC contraceptive failure in a real-life setting is known to largely depend on non-compliance (e.g., missed pills) and exhibit geographical variations [11,25]. PRO-E2 overall contraceptive failure rate (0.20/100 WY; 95% CI, 0.18–0.22) [11] was slightly lower than that of the European cohort of INAS-SCORE (0.4/100 WY; 95% CI, 0.4–0.5) [26], as well as the unintended pregnancy rate reported for COCs containing drospirenone, LNG, chlormadinone acetate, desogestrel, and dienogest in the EURAS-OC (0.48/100 WY; 95% CI, 0.44–0.53) [27].

Clinical implications

Large, prospective, controlled, cohort studies have been previously demonstrated to be suitable for safety monitoring of contraceptives, reliable identification of relevant clinical outcomes and providing robust estimates of their incidence [20,28].

This analysis on the large PRO-E2 Italian subpopulation (i.e., 19,683/91,313, 21.5% of the overall population) provides detailed information about the safety and efficacy profile of a COC containing natural oestrogen and EE/LNG COCs in an Italian real-life setting, where notably women over 35 years and smoking habit were more prevalent. Of note, 35 years corresponds to the age restriction for COC use when other cardiovascular risk factors are present, as per current practice guideline [29].

These data further suggest that NOMAC-E2 entails a very low VTE risk and high contraceptive effectiveness, which are both comparable to those of COC_LNG.

Research implications

This analysis provides information on the safety and efficacy of COCs, including newer formulations, in a population of Southern Europe, like Italy, for which the COC-associated risk of VTE and efficacy were unknown. Thus, it may support contraceptive counselling in the Italian clinical practice context, possibly fostering in turn COC use. Additional site-specific dedicated studies may help to apply knowledges developed in a restricted number of populations and countries.

Strengths and limitations

Strengths and limitations of the PRO-E2 study have been previously addressed in detail [11,12]. Further limitations of this Italian analysis are that relatively few women were included in some study sub-cohorts, and a small number of events occurred in some instances (e.g., VTEs). Moreover, incidence rates of primary/secondary endpoints may be susceptible to selection bias, as the study was powered based on the global sample, accounting for potential imbalances between countries regarding cofactor-outcome relation in the primary statistical analysis model on the overall study population.

Conclusions

Overall, the safety and contraceptive effectiveness of NOMAC-E2 and COC_LNG in the PRO-E2 Italian subpopulation were broadly consistent with those observed in the overall study population.

In conclusion, this analysis provides for the first time comprehensive epidemiological data on COC thromboembolic risk and effectiveness in a large Italian cohort and, in line with PRO-E2 study objectives, further confirms the safety and contraceptive effectiveness of NOMAC-E2 in clinical practice.

Acknowledgements

The authors would like to thank the Berlin Center for Epidemiology and Health Research (ZEG) whose contribution to the study and the present cohort analyses was invaluable. The authors would also like to acknowledge Eleonora Comi, PhD of PRINEOS S.r.l. for medical writing support.

Disclosure statement

Franca Fruzzetti has served as speaker and advisory board member for Theramex, Gedeon-Richeter, Bayer, Organon, Exeltis, Italfarmaco. Angelo Cagnacci has served as speaker and advisory board member for Theramex, Gedeon-Richetr, Bayer, Organon, Exeltis, Italfarmaco. Mitra Boolell is a full-time employee at Theramex HQ Ltd. Costantino Di Carlo had past financial relationships (lecturer, member of advisory boards and/or consultant) with: Theramex, Bayer, Organon, Gedeon Richter, Shionogi and Exeltis. Vincenzina Bruni has served as speaker and advisory board member for Theramex, Gedeon Richter, Exeltis and Bayer.

Data availability statement

The data supporting the findings of this study cannot be made available as containing proprietary information.