5.1.1 Estrogens

In the UK, the majority of COC as well as the CTP and CVR, contain between 20 μg and 35 μg of the synthetic estrogen, EE. Current ‘low-dose’ COC was developed to reduce the health risks associated with the high estrogen content of COC in the 1960s and 1970s.3 A COC product containing the synthetic estrogen, mestranol (metabolised to EE) is also available (50 μg mestranol roughly equates to 35 μg EE).4

COC has been introduced that contains 17β-estradiol, structurally identical to that which occurs naturally in humans. Theoretically, estradiol COC could have improved safety profiles compared to formulations containing EE or mestranol due to reduced thrombotic and metabolic effects.5 This has not yet been established: although limited evidence suggests that estradiol COC is safe to use and highly effective in preventing unintended pregnancy,6 further research will be required to assess the safety profile of estradiol COC relative to that of EE COC.

5.1.2 Progestogens

Progestogens are synthetic steroids designed to have some of the properties of progesterone. The synthetic progestogen component of CHC allows convenient dosing intervals, potent suppression of ovulation, and prevents over-proliferation of the endometrium in response to estrogen. Newer progestogens were developed to have fewer androgenic and glucocorticoid effects; some are anti-androgenic and have potentially favourable anti-mineralocorticoid effects.7

However, different progestogens may modify the effect of EE on hepatic clotting factors differently; CHC containing some newer progestogens in combination with EE appear to be associated with greater risk of VTE than COC containing other progestogens. See Section 10.1.1.

The progestogens that are components of CHC are sometimes grouped by ‘generation’, according to the time they were first marketed as constituents of COCs.7,8 (note that classification varies).

• First: norethisterone (NET)

• Second: levonorgestrel (LNG)

• Third: desogestrel, gestodene, norgestimate*

• Newer/other: drospirenone (DRSP), dienogest, nomegestrol acetate

(*sometimes classified as second generation: LNG is one of its active metabolites).

Norelgestromin is a metabolite of norgestimate; etonogestrel is the active metabolite of desogestrel.

Co-cyprindiol (containing 35 μg EE with cyproterone acetate, an anti-androgen) is indicated for management of moderate to severe acne and hirsutism. Women using co-cyprindiol for these indications do not require additional contraception.9

6.1.1 Combined oral contraception

The majority of COC in the UK is designed to be taken as 28-day cycles, with 21 consecutive daily active pills followed by a 7-day HFI prior to starting the next packet of pills. The first seven pills inhibit ovulation29 and the remaining 14 pills maintain anovulation. Traditionally women have then either had seven pill-free days or taken seven placebo tablets; during this HFI, most women will have a withdrawal bleed due to endometrial shedding. It should be made clear to women that this bleed does not represent physiological menstruation and that it is has no health benefit.

6.1.2 Combined transdermal patch

One patch is applied to the skin and worn for 7 days to suppress ovulation. Thereafter the patch is replaced on a weekly basis for two further weeks. The fourth week is patch-free to allow a withdrawal bleed. A new patch is then applied after seven patch-free days.

6.1.3 Combined vaginal ring

One ring is inserted into the vagina and left in place continuously for 21 days. After a ring-free interval of 7 days to induce a withdrawal bleed, a new ring is inserted.

6.2.1 Suggested tailored CHC regimens

A variety of regimens have been studied; however, data are currently too limited to recommend one approach over another. In some countries (but not the UK) COC are available that are intended to be taken as an 84/7 regimen, with a 3-monthly HFI. In the UK, some monophasic 24/4 COC regimens are marketed30,31 and one quadriphasic COC with a 2-day HFI is available.32 The FSRH supports off-label use of tailored CHC regimens such as those detailed in Table 1 using monophasic CHC that are licensed to be taken as a 21/7 regimen. If the preparation includes placebo pills, these should be omitted. Multiphasic COC should not be used in tailored regimens.

6.2.2 Contraceptive effectiveness of tailored CHC regimens

How does contraceptive effectiveness of tailored regimens compare with standard regimens?

It has been suggested that reducing the frequency of the HFI, abolishing the HFI and/or shortening the HFI could reduce the risk of escape ovulation and resulting contraceptive failure.

The findings of studies that consider the contraceptive effectiveness of tailored CHC regimens compared to one another and to standard 21/7 regimens are difficult to bring together because they consider different CHC formulations (different doses of estrogen, different progestogens) and different regimens. In addition to very limited direct data from studies with pregnancy as an outcome, indirect evidence is available from studies that consider ovarian activity during HFI of various lengths as a marker of potential risk of pregnancy.

On the basis of the available evidence (summarised below), the GDG considers that extended or continuous CHC regimens and shortened HFI could theoretically reduce risk of escape ovulation compared with standard CHC regimens, but that there is not conclusive evidence of greater contraceptive effectiveness.

Studies considering ovarian activity in the HFI as an indication of risk of pregnancy

A standard 7-day HFI is associated (in most studies) with more hypothalamic-pituitary-ovarian axis activity than a shortened, 3- or 4-day HFI.24,25,33–36 Ovarian activity has been observed to be significantly lower with continuous COC than with standard 21/7 regimens.37,38 It has therefore been suggested that a regimen including a HFI shorter than 7 days could reduce the risk of pregnancy resulting from escape ovulation.

Imperfect use of CHC in the days around the HFI could in effect extend the HFI. A systematic review39 identified 10 biomedical studies that reported on risk of ovulation with a deliberately extended HFI. Findings varied: five studies observed no ovulations with extended HFI up to 14 days; in the other five studies, presumed ovulations were documented with HFI between 8 and 11 days long. All studies involved small numbers of women and used different definitions and indicators of ovulation. A shorter HFI could theoretically reduce the risk of ovulation if CHC use around the HFI is imperfect.

One small RCT found that endogenous estradiol levels increased more quickly during the HFI after extended (84/7) COC use than in a 21/7 regimen.40 The authors postulated, but did not demonstrate, that escape ovulation could occur earlier with extended than with standard COC regimens.

There is limited evidence that ovarian activity in the HFI could be more pronounced in obese women. In one study41 hormone profiles at the end of a 7-day HFI in 10 women with normal BMI and 10 women with obesity noted estradiol levels consistent with dominant follicles and progesterone levels consistent with ovulation in more women with obesity than women with normal BMI. A further study42 of women with obesity who were initially taking 20 μg EE COC in a 21/7 regimen and were then randomised to either a continuous regimen of 20 μg EE COC or to 30 μg EE COC as a 21/7 regimen noted that both options decreased evidence of a follicle-like structure compared with 20 μg EE taken in a 21/7 regimen.

Studies with pregnancy as an outcome

A systematic review39 identified no direct evidence for risk of pregnancy associated with a HFI extended beyond 7 days.

The RCTs identified by a 2014 Cochrane Review27 were not powered to detect significant differences in pregnancy risk between continuous/extended and cyclical CHC regimens; only one reported a significantly lower risk of pregnancy with continuous compared to cyclical (vaginal) administration of COC (odds ratio (OR) 0.14, 95% confidence interval (95% CI) 0.02–0.97).43 One RCT published after the Cochrane Review (and designed to investigate continuation rather than effectiveness) reported similar pregnancy rates among 358 women randomised to either cyclical or continuous use of the same COC.44

Some observational studies suggest that risk of pregnancy associated with extended CHC regimens45 or CHC regimens with shortened HFI46 could be lower than that with traditional 21/7 regimens. A study47 examining pregnancy rates during an extended 91-day CHC regimen with 20 μg EE noted no reduction in contraceptive efficacy in women with obesity compared to women without obesity.

6.2.3 Safety of tailored CHC regimens

Available information regarding safety of extended compared to standard CHC regimens is reassuring. RCTs comparing extended and standard CHC regimens record very few serious adverse events associated with either type of regimen.27 Observational studies report similar (small) numbers of serious adverse events for extended regimens and for regimens with shortened HFI compared to 21/7 regimens.48,49

Direct data comparing risk of cardiovascular events and cancer between extended and standard CHC regimens are lacking. However, indirect evidence regarding cardiovascular risk is reassuring. In a Phase 3 trial of 20 μg EE/DRSP COC, 1067 women were randomised to standard cyclical, flexible extended or fixed extended pill-taking regimens for 1 year. A total of 755 women then entered an extension phase, taking the flexible extended regimen. Metabolic and haemostatic parameters, serum hormone levels and blood pressure were similar in all groups. Numbers of serious adverse events were very low in all groups.50 A smaller trial which randomised 78 women to use extended or standard cyclical regimens of 30 μg EE/DRSP COC found no statistically significant differences in carbohydrate or lipid profiles between the two groups over 6 months of use.51 In a third trial, 174 women were randomised to cyclical or continuous use of 20 μg EE/LNG COC (the dose of LNG was different in the two groups). The authors concluded that after 13 months of use, carbohydrate metabolism, lipid profile and haemostatic variables were broadly similar between the groups, but that further studies would be required to assess long-term continuous CHC.52 Haemostatic parameters were reported to be similar for 187 women randomised to extended or cyclical use of 30 μg EE/LNG COC for 6 months.53

Several studies have assessed the endometrium during continuous or extended CHC; endometrial thickening and histological abnormalities were not observed.27,50,54,55

6.2.4 Bleeding patterns with extended CHC regimens

A Cochrane Review of RCTs27 reported that in most studies bleeding patterns with extended CHC regimens were equivalent or improved compared to standard regimens.37,43,51,56–64 A systematic review28 that included both RCTs and observational studies concluded that overall, the total number of days of bleeding was lower with continuous or extended regimens than with cyclical use of CHC. Although there was an increase in breakthrough bleeding during the first months of use of continuous or extended regimens, its frequency and intensity subsequently decreased over time. Limited evidence suggests that bleeding patterns with continuous or extended use of the CTP and CVR show a similar reduction in bleeding/spotting days over time to that seen with extended use of COC.59,62,65–67

One study68 compared bleeding patterns in 139 existing cyclic COC users who were randomised to continuous use for 180 days of COC containing 30 μg EE/100 μg LNG, 20 μg EE/100 μg LNG, 30 μg EE/1000 μg NET or 20 μg EE/1000 μg NET; it should be noted that only the first of these is a COC preparation available in the UK. The study suggested more favourable bleeding patterns (more amenorrhoea and fewer spotting days) with continuous use of NET COC than with continuous use of LNG COC. The authors noted that the study findings did not support use of higher EE doses to prevent breakthrough bleeding during continuous COC use. It is not known how bleeding patterns with continuous use of other COC would compare.

6.2.5 Tailored CHC regimens and symptoms associated with the HFI

Symptoms such as headache, pelvic pain, bloating and breast tenderness are more frequent during the HFI than during the time that CHC hormones are being taken.69,70 Women may choose extended CHC regimens to avoid or reduce the frequency of symptoms associated with the HFI.

A Cochrane Review27 of RCTs identified studies that reported improvement in menstrual-related headache, bloating, tiredness and menstrual pain with extended COC regimens.

Observational studies similarly suggest benefit; a cohort study71 of 111 women who reported cyclical symptoms with two cycles of use of a 21/7 COC regimen found that mood, headache and pelvic pain scores improved significantly after the women switched to an extended COC regimen and were followed up for a year. Some 80% of the women continued the extended regimen for the full year, and 6 months after that most women reported that they had continued the extended regimen on their own. A prospective cohort study72 of 109 women given 30 μg EE/DRSP COC for two 21/7 cycles, followed by two 84/7 cycles (two-thirds completed all cycles) reported a significantly reduced incidence of heavy menstrual bleeding (HMB), intermenstrual bleeding, dysmenorrhoea, abdominal bloating, depressed mood and irritability at the end of the second 84/7 cycle compared to at enrolment.

6.2.6 Return to fertility after use of extended CHC regimens

One study73 found that of 187 women aged 18–49 years who had used continuous 20 μg EE/LNG COC for at least 6 months, 98.9% returned to spontaneous menstruation or became pregnant within 90 days. In another study, amongst 47 women who had used 20 μg EE/LNG COC continuously for 84 days, ovulation was observed within 37 days of stopping treatment in all but one case (98%).74

6.2.7 Acceptability of tailored CHC regimens

Evidence indicates that continuous or extended CHC regimens offer an acceptable alternative to standard CHC regimens for many women.

A UK trial56 randomised 503 women to use 30 μg EE/LNG COC either as a 21/7 regimen or as a regimen in which pills were taken consecutively until a 3-day bleed triggered a 3-day break in pill-taking. Of the 252 women using the tailored regimen, 179 women were followed up at 12 months; 54% were continuing the tailored regimen. This compared to 67% of women followed up in the 21/7 regimen arm of the study. Of note, there was no significant difference between standard and tailored regimens in terms of the proportion of women who continued taking any COC regimen at 12 months. The authors concluded that “tailored use can provide a suitable alternative to standard COC use for selected women” and that “offering the tailored regimen is unlikely to increase the risk of discontinuation of the COC”. Qualitative research carried out as part of this trial showed that some women using tailored COC regimens much preferred the reduced bleeding while others disliked the unpredictability of bleeding.56

A recent trial in Australia65 randomised 172 women to continuous use for a year of either a CVR or 20 μg EE/LNG COC. Women were instructed to take a 4-day break if they bled for four consecutive days. The study reported high satisfaction rates with continuous use of both methods; women liked the ease of use and infrequent bleeding, and for some women the absence of dysmenorrhoea, headaches and premenstrual symptoms were particular advantages. However, many women who were otherwise satisfied with the regimen disliked the unpredictability of bleeding.

A Cochrane Review27 identified six other RCTs that considered participant satisfaction with extended regimens; satisfaction was high and similar for both traditional and extended regimens.

A retrospective review75 of 318 women with unwanted hormone withdrawal symptoms on COC, who were counselled about using extended COC regimens, reported that of the 292 women followed up, 267 women chose to start an extended regimen and 172 continued to use the extended regimen at the time of their last follow-up. Various extended durations of pill-taking and shortened HFI were used.

As far back as 1977, a UK cohort study76 of 202 women started on a tricycling COC regimen reported that 82% of participants welcomed the reduction in bleed frequency. A total of 107 women completed 12 months of the tricycling regimen; of these 91% chose to continue the regimen after the trial had finished. Surveys of women in Holland,77 Germany,78 the USA79,80 and Italy81 found that more than half of the women of reproductive age would prefer to bleed less than once a month. In an Australian survey 54% of women preferred a monthly bleed.82

A Cochrane Review27 identified three RCTs in which study discontinuation rates were higher for extended or continuous regimens than for traditional 21/7 cycles; 10 RCTs reported no significant difference in discontinuation between regimens.

6.2.8 Offering tailored CHC regimens

In a UK study 112 women requesting COC (most already COC users) were given information about tailored COC regimens by a HCP using a structured script and patient advice sheet. The information was welcomed and understood by women and HCPs did not find the information-giving too time consuming.83

6.3.1 Starting CHC at the beginning of a natural menstrual cycle

Five studies are identified by a recent systematic review29 that consider the risk of ovulation (but not of pregnancy) if CHC is started at a time other than Day 1 of the menstrual cycle. Two small studies87,88 (n=14 and n=22) compared a group of women starting COC on Day 1 with a group starting on Day 5. No ovulations were observed, but ovarian suppression was greater if COC was started on Day 1. In two larger studies89,90 comparing commencement of COC on Day 1 and Day 7, ovarian activity was greater in the Day 7 group; risk of ovulation was significantly greater in the Day 7 group with a 20 μg EE, but not a 30 μg EE COC. However, markers used for occurrence of ovulation are not consistent across studies and frequency of ovulation may have been overestimated. A study91 of 40 women randomised to start COC on Day 1 or CVR on Day 5 observed no ovulations, but there was less ovarian suppression in the group who started CVR on Day 5.

In a study considering the effects of oral contraceptives administered at defined stages of ovarian follicular development, no ovulations were observed when COC was commenced at a follicle diameter of 10 mm (mean cycle Day 7.6).92 However 5/14 women ovulated when COC was initiated at 14 ± 1 mm (mean Day 11.7 ± 0.7; range Day 5–20) and 14/15 ovulated when COC was initiated at 18 ± 1 mm (mean Day 13.6 ± 0.8; range Day 7–20).

No studies are identified that consider ovarian activity when starting the CTP on different cycle days therefore advice is extrapolated from evidence for COC and CVR. Estradiol-containing COC should be started on Day 1 of a natural menstrual cycle. If started at any other time, additional contraceptive precautions are advised as per the manufacturer’s recommendations.

Women with short menstrual cycles

Fewer than 5% of women aged 15–44 years and fewer than 2% of women aged 20–39 years have menstrual cycles shorter than 20 days.93 Even smaller numbers (<1%) of women aged 14–42 years have cycle lengths shorter than 15 days.94 However, if there is concern about earlier ovulation associated with very short or variable cycles, women can be given the option to use additional contraceptive precautions when starting CHC after Day 1.

6.3.2 Quick starting CHC

Quick starting CHC, rather than waiting for the start of the next menstrual period, has the potential advantage of reducing the time during which a woman is at risk of pregnancy; it could also reduce barriers to commencing effective contraception that could result if a woman has to wait to start her contraceptive method.86 CHC can be quick started for medically eligible women in either of the following circumstances:

1. If pregnancy can be reasonably excluded (see Box 1) CHC can be quick started immediately.

2. If pregnancy cannot be reasonably excluded, a high-sensitivity urine pregnancy test (HSUPT) (able to detect human chorionic gonadotrophin (hCG) levels around 20 mIU/ml) should be taken; if negative, CHC can be quick started despite a risk of pregnancy from very recent UPSI. Almost all the available evidence suggests no adverse impact of fetal exposure to contraceptive hormones on pregnancy outcomes or risk of fetal abnormality. Refer to the FSRH Guideline Quick Starting Contraception for details of studies assessed. A further HSUPT should be taken 21 days after the last UPSI. Note that quick starting of CHC should be delayed for 5 days after ulipristal acetate (UPA) oral emergency contraception (EC).95

Bleeding patterns associated with quick starting CHC

A systematic review found that bleeding patterns were similar regardless of when in the cycle CHC was started.29

6.3.3 Summary of advice for starting CHC

Table 2 summarises when combined hormonal methods can be started and the requirements for additional contraceptive precaution. For further details on use of CHC after pregnancy (childbirth, abortion, miscarriage, ectopic pregnancy or gestational trophoblastic disease) see the FSRH Guideline Contraception After Pregnancy.97

6.3.4 Switching to CHC from other contraceptive methods

FSRH guidance98 on switching from hormonal contraception and non-hormonal contraception to CHC is summarised in Table 3a and Table 3b, respectively. See FSRH Guidance Switching or Starting Methods of Contraception.98

7.2.1 Is contraceptive effectiveness of CHC affected by obesity/weight?

Key information

• C.  Most evidence suggests no association between weight/BMI and effectiveness of COC.

• D.  Limited evidence suggests a possible reduction in patch effectiveness in women ≥90 kg.

Note that use of CHC is UKMEC 3 for use by women with BMI ≥35 kg/m². UKMEC recommendations relate to safety of use rather than to effectiveness.

Evidence relating to the effect of body weight/BMI on effectiveness of CHC is limited to observational studies in which height, weight and pregnancy are often self-reported and potential confounding factors such as contraceptive compliance and frequency of intercourse are unknown. Most studies include relatively few women in the highest weight/BMI categories. There is limited evidence that ovarian activity in the HFI could be more pronounced in obese women.41,42 See Section 6.2.2 regarding studies considering ovarian activity in the HFI as an indication of risk of pregnancy.

A 2017 systematic review109 reported that 10 out of 14 studies of COC identified did not report a difference in effectiveness by body weight or BMI; in the remaining four studies, the magnitude of the reported difference in COC failure was very small. Differences between different COC formulations could not be distinguished. A 2016 Cochrane Review110 concluded that in general the evidence identified did not indicate an association between increasing body weight or BMI and effectiveness of COC.

The limited evidence identified109 suggests that increasing body weight and BMI may contribute to decreasing effectiveness of the CTP. No direct evidence regarding body weight/BMI and effectiveness of the contraceptive ring was identified.

The GDG (in line with the Summary of Product Characteristics for the CTP)10 notes that contraceptive effectiveness of the CTP could be decreased in women weighing ≥90 kg, therefore it is recommended that additional precautions or an alternative method should be advised for women ≥90 kg.

7.2.2 Is contraceptive effectiveness of CHC affected by bariatric surgery?

Clinical recommendation

• D.  Women who have had bariatric surgery should be advised that the effectiveness of COC could be reduced.

There are theoretical concerns that both malabsorptive and restrictive bariatric procedures could decrease the absorption of oral contraceptives.111,112 A systematic review113 identified one small prospective cohort study114 in which two of nine women using COC after biliopancreatic bypass surgery (both with significant diarrhoea) became pregnant in the 2 years following surgery; a second descriptive study115 reported no pregnancies among COC users (number unknown) in the 2 years after a gastric banding procedure.

One pharmacokinetic study116 reported lower serum progestogen levels after administration of oral progestogen for morbidly obese women (BMI ≥40 kg/m²) after bypass surgery compared with healthy non-obese women; the difference could however be attributable to body weight. In contrast, a second pharmacokinetic study117 reported higher serum progestogen levels after administration of an oral combined estradiol/progestogen preparation in 12 morbidly obese women after bypass surgery than in six morbidly obese women who had not had surgery.

The evidence is too limited to make a definite recommendation regarding the effectiveness of COC after bariatric surgery. The GDG recommends that women who have had bariatric surgery should be advised of potential reduced effectiveness of COC and should consider a non-oral method of contraception.

7.2.3 What drug interactions are important to consider in relation to CHC?

Serum levels of contraceptive hormones may be altered by concomitant use of other drugs; contraceptive effectiveness could be affected. Hormonal contraceptives may themselves alter serum levels of other drugs that a woman is taking, with potential adverse effects. Drug interactions should therefore always be considered when prescribing CHC and when prescribing other drugs for women who are using CHC.

Refer to the FSRH Guideline Drug Interactions with Hormonal Contraception.118

Enzyme-inducing drugs

Clinical recommendations

• D.  Women using enzyme-inducing drugs should be informed that the contraceptive effectiveness of CHC could be reduced during use of the enzyme-inducer and for 28 days after stopping.

• D.  Women using enzyme-inducing drugs should be offered a reliable contraceptive method that is unaffected by enzyme-inducers.

FSRH Guideline Drug Interactions with Hormonal Contraception118 notes that hepatic enzyme-inducing drugs increase the metabolism of estrogens and progestogens, which could reduce the contraceptive effectiveness of all CHC methods. Women using enzyme-inducing drugs should be advised to switch to a contraceptive method (e.g. intrauterine methods or the progestogen-only injectable) that is unaffected by enzyme-inducers.

It is established practice that if, after advice to switch contraceptive method, a woman wishes to use COC concomitantly with an enzyme-inducing drug (with the exception of rifampicin or rifabutin which are potent enzyme-inducers) use of a minimum 50 µg (30 µg + 20 µg) EE monophasic combined pill may be considered during use of the enzyme-inducer and for a further 28 days after stopping. A continuous or tricycling regimen plus a shortened pill-free interval of 4 days should be used. Breakthrough bleeding could indicate low serum EE concentrations. Dose of EE can exceptionally be increased up to a maximum of 70 µg after specialist advice. It is not known how such usage affects risk of VTE. The use of two patches or two rings is not recommended.

Short-term use of enzyme-inducing drugs (<2 months) can be managed more flexibly than longer-term use. Continuing use of CHC with consistent and careful use of condoms may be appropriate in this situation.

Lamotrigine

Clinical recommendation

• D.  Women taking lamotrigine should be advised that CHC may interact with lamotrigine; this could result in reduced seizure control or lamotrigine toxicity. The risks of using CHC could outweigh the benefits.

Serum levels of lamotrigine can be reduced by CHC.119–124 A case series122 reported increased seizure frequency in four women with reduced lamotrigine levels following the initiation of COC. Data120 demonstrate increased lamotrigine levels during the HFI and increased lamotrigine side effects have been reported on cessation of CHC.119

Lamotrigine is not thought to be an enzyme-inducing drug, but the manufacturer advises that contraceptive effectiveness of CHC could be reduced by concurrent use of lamotrigine. This advice is based on a study125 of 16 women using 30 µg EE/150 µg LNG and lamotrigine for 6 weeks which found a modest increase in levonorgestrel clearance and changes in serum FSH and LH. The clinical significance in terms of contraceptive effectiveness is unknown. An earlier study of women using 30 µg EE/150 µg LNG and lamotrigine for 10–14 days reported a statistically non-significant decrease in the mean plasma concentration of EE but no change in LNG levels; no ovulation occurred and no change in menstrual patterns was observed.126

The risk of using CHC may outweigh the benefit for women using lamotrigine, given the potential risk of reduced seizure control whilst taking CHC, and potential for lamotrigine toxicity in the HFI. It is advised that alternative contraception should be considered.118

Antibiotics (non enzyme-inducing)

Key information

• D.  Additional contraceptive precautions are not required when antibiotics that do not induce enzymes are used in conjunction with CHCs.

Most broad-spectrum antibiotics are non-enzyme-inducing and it is established FSRH guidance118 that no additional contraceptive precaution is required unless the antibiotics (and/or illness) cause vomiting or severe diarrhoea (see Section 7.2.4).127,128

Progestogen receptor modulators

Clinical recommendation

• D.  Women should be advised to wait 5 days after taking UPA-EC before starting CHC. Women should be made aware that they must use condoms reliably or abstain from sex during the 5 days waiting and then until their contraceptive method is effective.

UPA is a selective progesterone receptor modulator. Limited evidence from biomedical studies129,130 suggests that effectiveness of oral hormonal contraception is not reduced by concomitant use of UPA-EC. Biomedical studies have demonstrated that starting an EE/LNG COC131 or a desogestrel POP129 soon after UPA 30 mg given for emergency contraception (UPA-EC) reduces the ability of UPA-EC to delay ovulation and could therefore reduce the effectiveness of UPA-EC. The FSRH Guideline Emergency Contraception recommends that after UPA-EC, commencement of CHC is delayed for 5 days (at least 120 hours) after UPA-EC has been given.95 This ensures that the UPA-EC is as effective as possible in preventing pregnancy resulting from the episode(s) of UPSI for which it was taken. After the 5 days waiting, CHC can be started with advice to use additional contraceptive precautions for the following 7 days.

Theoretically, there could be an interaction between CHC and UPA taken for management of fibroids.

7.2.4 Severe diarrhoea or vomiting

Clinical recommendation

• D.  Women using COC should be advised that contraceptive effectiveness could be reduced by vomiting or severe diarrhoea.

The general advice for women using oral contraceptives is to follow the instructions for missed pills if vomiting occurs within 3 hours of taking COC or severe diarrhoea occurs for >24 hours. Women should be advised to consider non-OC if diarrhoea or vomiting persist.118 Some drugs have the potential to reduce the effectiveness of contraceptives indirectly by causing diarrhoea or severe vomiting.

9.4.1 EE/DRSP COC for management of premenstrual symptoms

Addressing EE/DRSP COC specifically, a 2012 Cochrane Review of RCTs156 concluded “Drospirenone with EE 20 µg COC may help treat premenstrual symptoms in women with premenstrual dysphoric disorder (PMDD). However, a powerful placebo effect was evident: in one trial, symptoms were reduced for 48% of the drospirenone COC group versus 36% of the placebo group. The difference may not be clinically significant. Furthermore, effectiveness after three cycles is unknown. Little evidence exists for treating less severe symptoms or whether a drospirenone COC works any better than other COCs.”

Subsequent to the Cochrane Review a small RCT demonstrated no significant difference between placebo and EE/DRSP COC taken either cyclically or continuously for management of PMDD.157 The authors commented on the marked placebo response that was observed; a second small RCT suggested that COC containing either DRSP or desogestrel could reduce PMS symptoms, but there was no placebo arm.158

9.4.2 Other COC for management of premenstrual symptoms

Evidence from RCTs for CHC other than EE/DRSP is sparse and inconclusive, but suggests that other COC formulations could also be associated with symptom improvement in women with PMDD.159,160 The findings of the observational studies identified161–166 are limited by small study size, short duration of observation, lack of a verified symptom assessment tool, and lack of a comparator group that was not using hormonal contraception.

9.4.3 Extended COC regimens for management of premenstrual symptoms

A cohort study (and its extension) commenced 114 women on an EE/DRSP COC for two 21/7 cycles and then changed them to a continuous regimen of the same COC.71,162 102 women completed 168 days and 80 completed a year of extended use. The study reported that PMS symptoms improved significantly with the extended regimen compared to cyclical use.

9.9.1 Endometrial cancer

Systematic review and meta-analyses of observational studies196,197 indicate that OC use is associated with a reduced risk of endometrial cancer that correlates with duration of use and persists for many years after cessation. A recent meta-analysis197 of 36 international epidemiological studies found that every 5 years of OC use is associated with a relative risk of 0.76, resulting in a 50% reduction in risk of endometrial cancer with 10–15 years of use. A meta-analysis reported a persistent protective effect for as long as 30 years after cessation of OC; no significant differences were observed when comparing OC doses and formulations.197

In line with this, the RCGP Oral Contraception Study195 reported that compared to never-users of OC, the risk of endometrial cancer for ever-OC-users is reduced by 34% (IRR 0.66; 99% CI 0.48–0.89). Similarly, recently published data from a large US prospective cohort study198 indicate a significant 34% reduction in risk of endometrial cancer associated with ≥10 years OC use compared to never-use or use for less than a year; the greatest risk reduction was observed amongst smokers and women with obesity.

9.9.2 Ovarian cancer

A systematic review of observational studies199 reported a reduction in risk of ovarian cancer in ever-users of OC compared to never-users. The meta-analysis199 found that the protective effect is duration-dependent, with women who have used OC for at least 10 years having a 50% reduction in incidence of ovarian cancer. A review of published meta-analyses200 reported that the protective effect has been observed in women both with and without a genetic predisposition to ovarian cancer, increases with increasing duration of OC use and persists for at least 30 years after cessation of use. A subsequently published case-control study201 which included 1632 cases and 2340 controls suggested that there could be greater protection conferred to those who start OC before the age of 35 years than those who start after age 35 years. The UK RCGP Oral Contraception Study195 reported that, compared to never-users of OC, the risk of ovarian cancer of ever-OC-users is reduced by 33% (IRR 0.67; 99% CI 0.5–0.89). Recently published data from a large US prospective cohort study198 indicate a significant 40% reduction in risk of ovarian cancer associated with ≥10 years of OC use compared to never-use or use for less than 1 year.

BRCA gene mutation carriers

On the basis of evidence from case-control studies that include small numbers of cases, three systematic reviews with meta-analysis202–204 have concluded that amongst BRCA carriers, use of OC (COC or POP not specified, but majority likely to be COC) is associated with reduced risk of ovarian cancer with use, proportional to the duration of use. The evidence is stronger for BRCA1 carriers but exists for both BRCA1 and BRCA2. See Section 10.4.

9.9.3 Colorectal cancer

Evidence from three meta-analyses196,205,206 of data from observational studies suggests that ever-OC-users have a reduced risk of colorectal cancer compared to never-OC-users. In line with the findings of the other two studies, the most recent meta-analysis205 which included 12 cohort studies and 17 case-control studies with a total of 15 790 cases of colorectal cancer reported a summary relative risk (RR) for colorectal cancer of 0.82 (95% CI 0.76–0.88) for ever-users versus never-users of OC after adjustment for confounding lifestyle factors. More recently published data198 from a large prospective cohort study suggested no significant association between OC use and risk of colorectal cancer. The UK RCGP Oral Contraception Study195 reported that compared to never-users of OC, the risk of colorectal cancer of ever-users is reduced by 19% (IRR 0.81; 99% CI 0.66–0.99).

9.9.4 Mortality

A meta-analysis207 of eight cohort studies involving 217 868 women and 40 570 deaths did not find a significant association between ever-use of OC and all-cause mortality (hazard ratio (HR) 0.94; 95% CI 0.87–1.02) regardless of duration of OC use and time since last use. The UK RCGP Oral Contraception Study, a large prospective UK cohort study with 1.3 million woman-years of observation, found that women who had ever used OC had a 12% lower risk of all-cause mortality when compared with those who had never used OC.208 In addition to reduced mortality, ever-users of OC were observed to have a non-significant reduction in overall cancer risk, and for some cancers this protective effect persisted for decades following cessation of OC.195 Reassuringly, over a 44-year period of follow-up, no increased cancer risk was observed later in life for women who had ever used OC.

10.1.1 Effect of progestogen type on VTE risk

A systematic review identified no relevant RCTs.230 Meta-analysis230 of the data from observational studies (which may be subject to confounding and prescribing bias) indicated that use of low-dose COC (<50 μg EE) containing cyproterone acetate, desogestrel, gestodene or DRSP was associated with a significant 1.5- to 2-fold risk of VTE compared to use of COC containing LNG. Norgestimate COC was found to be associated with similar VTE risk to LNG COC. These findings are consistent with those of other recent meta-analyses which suggest that use of some third-generation COC is associated with greater VTE risk than use of second-generation COC (differences are not always statistically significant).209,210,231,232

In 2014, the EMA published estimated figures for absolute risk of VTE in users of CHC (see Table 5).

Note that estimates of background VTE risk in women of reproductive age vary. The EMA estimates that the incidence of VTE among women who are not using CHC and are not pregnant is about 2 per 10 000 woman-years233; a large Danish cohort study234 reported a VTE incidence of 3.7 per 10 000 woman-years for non-users of CHC; a 2007 literature review235 concluded that the overall VTE incidence in all women of reproductive age is likely to be in the range of 5–10 per 10 000 woman-years.

10.1.2 Effect of dose of EE on VTE risk

Available studies are limited by their observational nature and the small numbers of VTE events in some study groups. COC preparations that have been studied vary with respect to progestogen as well as the dose of EE; it is therefore difficult to compare the effect of EE dose on VTE risk.

A systematic review and meta-analysis suggested that VTE risk among users of COC is dependent on the dose of EE, with higher EE dose being associated with greater VTE risk.209,236 The meta-analysis209 concluded that a 50 μg EE/LNG COC was associated with RR for VTE of 2.1 (95% CI 1.4–3.2) and 2.3 (95% CI 1.3–4.2) compared with a 30 μg and 20 μg EE/LNG COC, respectively. VTE risk was significantly greater for 30 μg EE than 20 μg EE COC containing gestodene, but not those containing LNG or desogestrel.

Subsequent to the systematic review, a large French database study indicated a lower risk of pulmonary embolism associated with COC containing 20 μg EE than COC with 30–40 μg EE (adjusted RR 0.75; 95% CI 0.67–0.85).237

10.1.3 Effect of estrogen type on VTE risk

Most COC contain EE. However, COC are now available that have estradiol as their estrogen component, in combination with either dienogest or nomegestrol acetate. Direct evidence regarding risk of VTE associated with use of these COC is limited. A large cohort study in the USA and Europe which observed 47 VTE events amongst 50 203 new COC users followed for a mean of 2.1 years reported adjusted HRs of 0.4 (95% CI 0.2–1.0) and 0.5 (95% CI 0.2–1.5) for users of estradiol valerate/dienogest compared to users of “other COC” and EE/LNG COC, respectively.238 Studies that have compared the effect of estradiol COC on haemostatic variables to that of EE/LNG COC suggest that use of estradiol-containing COC (particularly estradiol/nomegestrol acetate) is associated with a haemostatic profile similar to or more favourable than that with EE/LNG COC. The potential impact of this on VTE risk remains to be clarified.5,239–243

10.1.4 Effect of route of administration on VTE risk

Data for VTE risk associated with the EE/norelgestromin CTP and EE/etonogestrel CVR are limited and conflicting.

Combined transdermal patch

A systematic review244 identified conflicting evidence from seven observational studies that compared the VTE risk associated with use of the CTP to that with use of COC containing LNG or norgestimate. One retrospective cohort study245 and one case-control study246 reported significant, 2-fold greater VTE risk among CTP users compared to COC users. One cohort study247 considered only risk of cerebral venous thrombosis and recorded very few events (two in COC users and none in CTP users). Of the remaining four case-control studies, one248 reported a non-significantly increased VTE risk for CTP users compared to COC users (OR 2.0; 95% CI 0.9–4.1) and three248–250 did not find significant differences.

Combined vaginal ring

A systematic review244 identified three studies that examined VTE among vaginal ring users compared to EE/LNG COC users. Two studies (one cohort105 and one case-control250 study) did not find a statistically significant difference between CVR and EE/LNG COC users. The third (cohort) study245 reported an increased risk of VTE (RR 1.9; 95% CI 1.3–2.7) in women using the CVR compared to EE/LNG COC users.

10.1.5 Women with inherited thrombophilias

Known thrombophilia is an absolute contraindication to CHC use (UKMEC 4).96

Women with inherited thrombophilias who use CHC are at significantly greater risk of VTE than CHC users without thrombophilia. A systematic review and meta-analysis251 which included data from 14 observational studies concluded that COC users with mild thrombophilias (e.g. factor V Leiden heterozygosity and prothrombin G20210A mutation) had a risk of VTE almost six times that of COC users without thrombophilia (RR 5.89; 95% CI 4.21–8.23); severe thrombophilias (e.g. antithrombin, protein C and protein S deficiencies and the homozygous forms of factor V Leiden and prothrombin G20210A) were associated with a more than 7-fold increase in VTE risk (RR 7.15; 95% CI 2.93–17.45).

Cohort studies included in the systematic review reported absolute risks of VTE as high as 4.3 (95% CI 1.4–9.7) and 4.62 (95% CI 2.5–7.9) per 100 pill years among COC users with antithrombin, protein C or protein S deficiency who also had a family member with both a thrombophilia and a VTE event. The authors of the review conclude that compared to the general CHC-using population (VTE risk of 6 per 10 000 woman-years), CHC users with a mild thrombophilia and a family history of VTE have an 8- to 33-fold increase in VTE risk; for CHC users with a severe thrombophilia and a positive family history, VTE risk is increased 70-fold. It is suggested that there could be co-inheritance of other thrombophilic defects.251

Thrombophilia screening

Screening for inherited thrombophilia in women with no known family history of thrombophilia is considered to be neither cost effective nor necessary.252,253 Screening women with a first-degree relative who has a hereditary thrombophilia is of uncertain benefit; studies evaluating the cost-effectiveness are of generally low quality.254 One limitation of these studies is that they consider only 1 year of COC use rather than the longer durations that reflect many women’s use. Furthermore, they do not take into account other situations that increase VTE risk – pregnancy, surgery, air travel – for which knowing thrombophilia status could be beneficial.

Due to the lack of high-quality evidence, the GDG advises that a woman with a first-degree relative who has an inherited thrombophilia can be counselled that a negative thrombophilia screen does not necessarily exclude thrombophilia (particularly if the relative has had a VTE event). A contraceptive method other than CHC that is not associated with increased VTE risk should be considered.

10.2.1 CHC use and risk of myocardial infarction

A Cochrane Review256 of 24 observational studies found a significantly increased risk of MI for current users of COC compared with non-users (RR 1.6; 95% CI 1.2–2.1). The review included studies of users of older COC with higher EE content. Three earlier meta-analyses reported similar significant results (ORs 1.70 (95% CI 1.2–2.3), 1.84 (95% CI 1.38–2.44) and 2.48 (95% CI 1.91–3.22) for current use compared to non-use).257–259 One further meta-analysis210 found a non-significant increased risk of MI associated with current COC use (OR 1.34; 95% CI 0.87–2.08).

10.2.2 CHC use and risk of ischaemic stroke

A Cochrane Review256 found a RR for ischaemic stroke of 1.7 (95% CI 1.5–1.9) for current COC use compared to non-use. This is consistent with the findings of three previous meta-analyses which reported OR for ischaemic stroke in current COC users of 1.8 (95% CI 1.2–2.8),259 2.12 (95% CI 1.56–2.86)257 and 1.90 (95% CI 1.24–2.91)210 compared to non-users. A 2015 meta-analysis260 found a higher OR (OR 2.47; 95% CI 2.04–2.99) for first-time ischaemic stroke with current COC use compared with non-current use, however the risk declined significantly with decreasing estrogen dose.

10.2.3 Estrogen dose and risk of ATE

A Cochrane Review256 and a meta-analysis260 concluded that risk of ATE seemed to increase with higher doses of estrogen in COC.

Myocardial infarction: A large cohort study255 (data extracted from Danish databases) – which was included in the Cochrane Review – reported RR of MI of 1.4 (95% CI 1.07–1.81) for users of COC containing 20 µg EE and 1.88 (95% CI 1.66–2.13) for users of COC containing 30–40 µg EE compared to non-users. One meta-analysis258 of 19 case-control studies and four cohort studies found a significantly increased risk of MI with preparations containing 30–49 μg EE (OR 1.97; 95%CI 1.43–2.71) but no increased risk for 20 μg EE preparations (OR 0.92; 95% CI 0.21–4.08). A later (2016) large French cohort study237 reported a significantly lower risk of MI with use of COC containing 20 μg EE than with use of COC containing 30–40 μg EE (RR 0.74; 95% CI 0.67–0.82).

Ischaemic stroke: A meta-analysis260 found that the OR for ischaemic stroke was 1.56 (95% CI 1.36–1.79) in women using COC containing 20 µg EE, 1.75 (95% CI 1.61–1.89) for 30–40 µg EE and 3.28 (95% CI 2.49–4.32) for ≥50 µg EE compared with non-current COC use. The Danish cohort study reported a RR of ischaemic stroke of 1.60 (95% CI 1.37–1.86) for users of 20 µg EE and 1.75 (95% CI 1.61–1.92) for users of EE 30–40 μg compared to non-users. The 2016 French cohort study237 reported a significantly lower risk of ischaemic stroke with use of a COC containing 20 µg EE than with use of a COC containing 30–40 µg EE (RR 0.82; 95% CI 0.70–0.96).

10.2.4 Progestogen type and risk of ATE

A Cochrane Review256 concluded that risk of MI and ischaemic stroke did not vary clearly according to progestogen type.

Myocardial infarction: One meta-analysis259 found no significant association between progestogen type and risk of MI. A Danish cohort study reported RR for MI of 1.33–2.28 for use of 30–40 μg EE COC containing different progestogens (compared with non-use of CHC), but the differences in risk between progestogens were not significant.255 Two meta-analyses reported OR for MI compared with CHC non-use of 1.13 (95% CI 0.66–1.92)261 and 1.34 (95% CI 0.91–1.98)210 for use of third-generation COC, while for second-generation COC the OR were 2.18 (95% CI 1.62–2.94)261 and 1.79 (95% CI 1.16–2.75)210 and for first-generation COC, 3.37 (95% CI 2.04–5.54).210 One meta-analysis258 reported a significantly increased risk of MI for users of first- and second-generation COC compared to non-users (OR 2.21; 95% CI 1.30–3.76; p=0.004) and 2.17 (95% CI 1.76–2.69; p<0.0005), respectively), but not for users of third-generation COC (OR 1.27; 95% CI 0.96–1.67; p=0.094).

Ischaemic stroke: Two systematic reviews257,259 reported similar risk of ischaemic stroke for current users of second- and third-generation COC.

DRSP-containing COC: One large database cohort study found that women using COC containing EE 30 μg and DRSP had an increased risk of ATE compared to women using preparations of COC containing 20–35 μg EE and LNG, norethisterone or norgestimate but this risk was only statistically significant in women aged 35–55 years.262 Conversely, data from the EURAS study224 suggested a non-significant reduction in ATE risk with EE/DRSP compared to other COC, and the continuation of this study reported a significantly lower risk of ATE with EE/DRSP compared to other COC (adjusted HR 0.4; 95% CI 0.2–0.8).263 An Israeli cohort study reported similar ATE risks for users of EE/DRSP and second- and third-generation COC.264

10.2.5 Estrogen type and risk of ATE

The limited data relating to risk of ATE associated with use of COC containing estradiol suggest that risk is similar to that associated with EE COC.238

10.2.6 Route of administration and risk of ATE

A recent systematic review244 identified one cohort study265 and one case-control study246 that compared ATE risk for use of the CTP with that of COC containing norgestimate; neither study reported a significant difference in risk of MI or ischaemic stroke. Two prospective cohort studies were not included in the systematic review because the comparator group comprised users of multiple different COC; one reported similar risk of ATE for use of the CVR to that for use of (various) COC.263 The other study262 found no significant differences in ATE risk between users of the CTP, CVR and (various) COC.

10.2.7 Additional risk factors for ATE

Risk of ATE is greater for COC users with additional risk factors such as hypertension266 and smoking96 and may be increased by hyperlipidaemia.267

Migraine and risk of ischaemic stroke

The evidence (summarised below) suggests that CHC users with migraine with aura are at greater risk of ischaemic stroke than CHC users without migraine. UKMEC 201696 recommends that use of CHC by women who have migraine with aura is UKMEC 4. For women who have migraine without aura, initiation of CHC is UKMEC 2, but if there is new-onset migraine after starting CHC, the risks of continuation generally outweigh the benefits (UKMEC 3).

It is important to note that although relative risk of ischaemic stroke is increased, the number of ischaemic strokes in women of reproductive age is small and the absolute risk remains low. Using data from the UK General Practice Research Database the incidence of ischaemic stroke in all women aged 15–49 years was estimated at 3.56 (95% CI 3.05–4.07) per 100 000 per year.268

Meta-analysis of evidence from observational studies indicates that the risk of ischaemic stroke in individuals who have migraine is approximately twice that of those who do not have migraine.269–273 Four meta-analyses269,270,272,273 find a significant association between migraine with aura and ischaemic stroke, but only one272 reports a significant association for migraine without aura. Use of CHC independently increases the risk of ischaemic stroke (see Section 10.2.2).

Evidence is limited regarding risk of ischaemic stroke amongst women who have migraine and are also users of CHC.270,274,275 A 2016 systematic review274 concluded that the fair to poor quality evidence identified suggests an increased risk of stroke amongst CHC users with migraine compared with non-users of CHC who have migraine, and that the effect on ischaemic stroke risk of migraine and CHC are additive. A recent nested case-control study276 which considered data from US databases identified 25 887 ischaemic strokes amongst women aged 15–49 years between 2006 and 2012. Compared to women who did not have migraine and did not use CHC, the OR for ischaemic stroke was 1.8 (95% CI 1.1–2.9) for women with migraine without aura who also used CHC. This is similar to the OR observed for women with migraine without aura who did not use CHC (OR 2.2; 95% CI 1.9–2.7); CHC was not observed to further increase risk. For women who had migraine with aura, the OR for ischaemic stroke was 2.7 (95% CI 1.9–3.7) with no CHC use and increased to 6.1 (95% CI 3.1–12.1) if they also used CHC compared to women without migraine or CHC use.

10.4.1 CHC use and risk of breast cancer

Current CHC use

A recent large Danish cohort study reported a relative risk of breast cancer of 1.19 (95% CI 1.13–1.26) for current or recent users of COC compared to never-users of hormonal contraception.277 Risk appeared to increase with duration of use. No major differences in risk were observed between COC containing different progestogens. The study adjusted for potential confounding factors, but data for alcohol intake, age at menarche and breastfeeding were not available and those for BMI were incomplete.

Meta-analyses of data from observational studies,278,279 sometimes including use of older COC with higher estrogen content, have reported a slight but significant increased risk of breast cancer among women currently using COC compared with women who have never used CHC, the RR ranging from 1.19–1.24. One meta-analysis of five cohort studies found a very small but significant increase in breast cancer risk for every 5 years (RR 1.07; 95% CI 1.03–1.11) and 10 years (RR 1.14; 95% CI 1.05–1.23) of use; however, many of the individual studies did not have statistically significant findings.280 Three more recent studies, including the Oxford-Family Planning Association contraceptive study and a meta-analysis of 44 observational studies, found no link between duration of COC use and breast cancer risk.196,281,282

Past CHC use

Increased breast cancer risk associated with current COC use has been found to decline gradually after cessation of COC, with no significant increased risk of breast cancer after 10 years of non-use.196,278,283,284 Compared with never-use of CHC, reported RRs for breast cancer for ever-users of COC are close to unity (range 1.0–1.08) and non-significant.196,278,280,281,283,284 Very long-term follow-up of the RCGP Oral Contraception Study cohort indicates that the increased breast cancer risk that is seen in current and recent OC users appears to be lost within approximately 5 years of stopping, with no evidence of increased risk in later life.195 The evidence is that there is no significant association between use of COC and mortality from breast cancer.208,285–287

10.4.2 Family history of breast cancer

Women with a family history of breast cancer have an increased background risk of breast cancer compared to women with no such family history.288 Several studies do, however, suggest that women with a family history of breast cancer who have ever used CHC are at no higher risk of breast cancer than women with a family history who have never used CHC.204,278,289,290 UKMEC does not restrict use of CHC for women with a family history of breast cancer.96

10.4.3 Genetic mutations relevant to breast cancer risk

UKMEC states that for carriers of known gene mutations associated with breast cancer (e.g. BRCA mutations) the risks of using CHC generally outweigh the advantages.96 Women with such inherited mutations have a higher background risk for breast cancer.291–294 The evidence regarding whether CHC use further increases the risk of cancer for these women is inconsistent, with some systematic reviews/meta-analyses reporting conflicting findings among the studies considered295–297 and others finding no increased risk with CHC use.202–204

Three meta-analyses202–204 based on observational studies conclude that carriers of BRCA mutations who use COC have a reduced risk of ovarian cancer compared with never-users (RR 0.50, OR 0.57 and OR 0.58). This advantage would need to be weighed against the potential increased risk of breast cancer.

11.2.1 Differences in bleeding pattern between routes of CHC administration

The nearly constant rate of release of contraceptive steroids administered transdermally or transvaginally, together with the theoretical advantages in respect of compliance, suggest that cycle control could be better than that with COC.313

OC versus CTP: A 2013 Cochrane Review99 identified four RCTs which compared cycle control among women using the EE/norgestimate patch to that among women using a variety of COC. None showed any differences in cycle control despite compliance being significantly better among patch users than among women using the COC in all three studies reporting compliance. More recently published studies have explored different formulations of patches, none as yet licensed.

OC versus CVR: In the 2013 Cochrane Review99 seven trials with the CVR reported bleeding data. Significant differences were found in four trials, all tending towards a lower incidence of breakthrough bleeding (at least in some cycles) among CVR users than among women using a variety of COC. One Chinese study published since the Cochrane Review311 compared the CVR with COC containing 30 μg EE/3mg DRSP. The authors reported significantly less bleeding/spotting in the CVR group than in the COC group for all 13 cycles (p<0.05) These differences are biologically plausible since there is greater bioavailability of steroids when administered through the vagina than through other routes.313

11.2.2 Differences in bleeding patterns between formulations of oral contraceptives

Many studies have attempted to show differences in bleeding patterns between different formulations of COC, however seldom do these studies compare like with like. For example, COC containing different progestogens are compared but they often also contain different doses of EE;314 monophasic formulations are compared with triphasic formulations;314 and regimens with 7 days of placebo pills are compared with a regimen with only two 2 days of placebo.308 The majority of such studies are undertaken by pharmaceutical companies as Phase 3 trials leading to licensing of a new product. The comparator pill could be chosen to maximise the chance of the new product showing non-inferiority. Moreover while the differences in the duration of unscheduled bleeding may be statistically significant, the effect is often small and unlikely to influence acceptability of the method; a difference between 12 and 15 days of unscheduled bleeding over 90 days may not be clinically significant.308

The evidence for an effect of the dose of estrogen on the incidence of unscheduled bleeding is more persuasive than that for an effect of different types of progestogen.

Effect of estrogen dose: A 2013 Cochrane Review315 of 22 RCTs concluded that women using 20 μg of EE experienced higher rates of unscheduled bleeding than women using >20 μg. One study merits specific mention. Although criticised in the review for not specifying who was blinded, Akerlund et al (1993)316 undertook an RCT comparing women using a COC containing 150 μg desogestrel with either 20 μg or 30 μg EE. Thus the progestogen component was the same in both groups. Women taking 20 μg EE were more likely to report irregular bleeding (OR 1.56; 95% CI 1.10–2.20). The difference between the incidence of unscheduled bleeding was statistically significant for 8 of the 12 months of follow-up.

Effect of progestogen type: A 2004 systematic review317 of 22 studies included two RCTs that reported less unscheduled bleeding with third- compared with second-generation progestogens (RR 0.71; 95% CI 0.55–0.91). These RCTs were limited by short study periods. Six RCTs found that cycle control was better with COC containing second-generation versus first-generation progestogens, both for monophasic (RR 0.69; 95% CI 0.52–0.91) and triphasic formulations (RR 0.61; 95% CI 0.43–0.85).

Effect of estrogen type: A number of studies have attempted to compare bleeding patterns among women using COC containing natural estrogens as opposed to those containing EE. Once again the type of progestogen is also different between the formulations, making the results difficult to interpret. Two RCTs comparing use of 1.5 mg 17β-estradiol/2.5 mg nomegestrol acetate30 and 30 μg EE/3 mg DRSP found that the incidence of unscheduled bleeding was comparable between the two formulations.318,319 One RCT comparing estradiol valerate/dienogest32 with 20 μg EE/100 μg LNG over seven cycles also found no difference in the proportion of women experiencing unscheduled bleeding, but reported that women in the estradiol group experienced fewer bleeding/spotting days (median 16 vs 21 during Days 1–90; 12 vs 15 during Days 91–180; p<0.0001).308

11.2.3 Differences in bleeding patterns between CHC regimens

See Section 6.2.4.

Unscheduled bleeding with CHC: summary

In summary, HCPs should inform women that unscheduled bleeding commonly occurs with CHC use but is likely to improve over the first 3–4 months of use. Formulations with higher doses of EE are probably associated with a lower risk of unscheduled bleeding when used in a traditional regimen, and pills containing second-generation progestogens could offer better cycle control than those containing norethisterone. Blood levels of exogenous hormones may vary 10-fold between individuals320 and intra-individual differences also occur, so what works for one woman may not work for another. Trial and error may therefore be the best approach for an individual woman to find a formulation that suits her needs. Extended or continuous regimens could offer less unscheduled bleeding. HCPs should consider other causes of unscheduled bleeding. For guidance on the management of unscheduled bleeding in women using hormonal contraception, refer to the FSRH Guideline Problematic Bleeding with Hormonal Contraception.312

12.1.1 Assessment of medical eligibility for CHC

Medical history and lifestyle factors

Medical eligibility must be assessed prior to prescription of any contraceptive method including CHC. The UK Medical Eligibility Criteria for Contraceptive Use 2016 (UKMEC 2016)96 provides recommendations for the safe use of CHC by women with different personal characteristics and medical conditions. Each of the personal characteristics or medical conditions considered by the UKMEC is assigned to one of four categories as defined in Table 6. UKMEC 3 and UKMEC 4 characteristics and conditions for CHC use are listed in Appendix 2. If a woman has multiple conditions that are UKMEC 2 for CHC use and relate to the same risk, the suitability of CHC should be carefully considered.

Specific attention should be given to enquiring about:

• Thrombophilia or previous VTE

• Ischaemic heart disease, stroke or transient ischaemic attack, peripheral vascular disease

• Additional risk factors for venous or arterial thromboembolism (e.g. smoking, obesity, recent childbirth, immobility, hypertension, migraine, diabetes, hyperlipidaemia, antiphospholipid antibodies, arrhythmia, complicated congenital/valvular heart disease or cardiomyopathy)

• Personal history of breast cancer/known breast cancer-related gene mutation

• Hepatobiliary disease

• Recent childbirth, current breastfeeding.

Use of a self-completed checklist to assess medical eligibility for CHC

The use of a suitable self-completed checklist may be appropriate for identifying personal characteristics, medical conditions or use of medication that could affect medical eligibility for CHC. Four studies (two from the USA356,357 and one each from the UK358 and Tanzania359) which examined assessment of medical eligibility for CHC using questionnaires self-completed by women compared to clinical assessment by trained providers reported high levels of agreement between women and HCPs. All four studies found that women were more likely than providers to identify contraindications. The UK study concluded that no clinically important information relevant to a particular woman’s use of CHC was missed and none of the women would have been wrongly prescribed CHC based only on their self-completed questionnaires.

If self-assessment checklists are used they should be developed and validated to ensure that they are effective.

Measurements and tests

Blood pressure and body mass index (BMI) should be documented for all women before prescription of CHC. The prescriber must be confident that measurements are recent and accurate.

Blood pressure: Women with severe hypertension (systolic pressure ≥160 mmHg or diastolic pressure ≥100 mmHg) should not use CHC (UKMEC 4). Women with less severe hypertension (systolic pressure 140–159 mmHg or diastolic pressure 90–99 mmHg), or with adequately controlled hypertension should not use CHC (UKMEC 3). Blood pressure should therefore be evaluated before initiating CHC.

Evidence from a systematic review360 suggests that risk of MI and ischaemic stroke is higher among women who do not have their blood pressure measured before initiating COC.

Women with raised blood pressure measured in non-primary care settings (e.g. at a pharmacy or specialist services) should be provided with suitable alternative effective contraception and advised to see their general practitioner for blood pressure assessment.

Weight (BMI): Women with BMI <35 kg/m² generally can use CHC (UKMEC 2). Women with BMI ≥35 kg/m² generally should not use CHC (UKMEC 3), although CHC may be prescribed by a specialist provider. BMI should be documented before starting CHC. Baseline weight could additionally be helpful for monitoring any changes and counselling women who might be concerned about later weight change perceived to be associated with their contraceptive method.

Pelvic examination: A consultation regarding contraception may be used as an opportunity for health screening but screening should not be a condition for prescribing. Pelvic examination is not necessary before initiation of CHC because it does not affect the decision to prescribe or withhold hormonal contraception.361

Clinical breast examination: Although women with current breast cancer should not use CHC (UKMEC 4), screening asymptomatic women with a clinical breast examination before initiating CHC is not necessary because of the low prevalence of breast cancer among women of reproductive age. A systematic review361 did not identify any evidence regarding outcomes among women who were screened versus not screened with a breast examination before initiation of hormonal contraceptives. Breast self-examination can be encouraged.

Laboratory testing: A systematic review362 conducted to evaluate the evidence regarding health outcomes among women with and without laboratory testing to identify certain medical conditions, including diabetes, hyperlipidaemia, liver disease, cervical cancer, STIs or human immunodeficiency virus (HIV), prior to initiating contraceptives did not identify any relevant evidence.

The GDG (in line with the WHO Selected Practice Recommendations for Contraceptive) advise that routine screening and examination other than blood pressure and BMI are unnecessary prior to prescription of CHC and do not contribute substantially to the safe and effective use of CHC.

12.1.2 Assessment of factors that could affect contraceptive effectiveness

Drug interactions

Some medications could reduce the contraceptive effectiveness of CHC by induction of hepatic enzymes. Contraceptive hormones can affect the action of certain medications. Refer to the FSRH Guideline Drug Interactions with Hormonal Contraception118 for further information. (See also Section 7.2.3).

Malabsorption

The effectiveness of COC (but not the CTP or CVR) could be reduced by malabsorption resulting from, for example, vomiting and severe diarrhoea (see Section 7.2.4), bariatric surgery, small bowel resection (see Section 7.2.2) or inflammatory bowel disease (see FSRH Guideline Sexual and Reproductive Health for Individuals with Inflammatory Bowel Disease).363

12.1.3 Women for whom CHC is unsuitable

Women assessed as medically ineligible for CHC should be told why, and the risk to their health should be explained. Alternative contraceptive methods which can be safely used should be discussed and offered.

Women with medical conditions or using medications that could reduce contraceptive effectiveness should be given advice regarding (and where possible provided with) alternative methods that would be effective for them.

12.6.1 Combined oral contraception

A systematic review364 suggested that provision of a greater number of packs of COC was associated with increased rates of continuation. Studies that compared provision of 1 vs 12 packs, 1 vs 12 or 13 packs, or 3 vs 7 seven packs found increased continuation of pill use among women provided with more pill packs.365–367 Provision of more pills packs was associated with fewer pregnancy tests, fewer pregnancies, and lower costs for users. However, a greater number of pill packs (i.e. 13 packs vs 3 packs) was also associated with increased pill wastage in one study.365

12.6.2 Combined transdermal patch

No evidence was identified on which to base recommendations for the number of packs of patches provided but once a woman is established on CTP, provision of a 12-month supply is reasonable.

12.6.3 Combined vaginal ring

Following manufacture, supplies of Nuvaring have to be kept refrigerated. After dispensing, Nuvaring can be stored at room temperature and used within a maximum of 4 months. Therefore no more than one pack of three Nuvaring can be provided at any one time. Arrangements should be made for repeat prescription to prevent users routinely having to return to the provider more than once per year. SyreniRing does not required refrigeration and a one-year supply can be dispensed.

13.1.1 What should be done at CHC follow-up?

Clinical recommendation

• ✓  Medical eligibility, drug history, method adherence and method satisfaction should be reassessed at follow up. BMI and blood pressure should be recorded.

Routine follow-up has long been considered an essential aspect of ensuring the safety of contraceptive use and a way to ensure adherence. At follow-up, medical eligibility should be rechecked, drug history updated, method adherence and method satisfaction assessed, and alternative contraceptive options considered. Women should be reminded about signs, symptoms, health events and changes in medication that should prompt them to seek medical review.

Weight and blood pressure have been the two parameters considered most important with respect to routine follow-up.

Weight/BMI: A systematic review339 identified no large effect of CHC use on weight gain. In one matched cohort study of adolescents using COC published since this review,369 BMI increased by a mean of 0.04 kg/m² per month among COC users versus 0.025 kg/m² per month among controls. If BMI is ≥35 kg/m² at follow-up, women should be advised to switch to an alternative method of contraception.

Blood pressure: A systematic review370 of five studies, with methodological limitations, examining blood pressure after CHC initiation, generally showed that only a small percentage of women (2% in the largest study) develop hypertension between 1 and 24 months after starting hormonal contraception. One matched cohort study published since this review found no significant changes in blood pressure among adolescents using COC during 18 months of follow-up when compared with controls.369 No studies were identified that examined changes in blood pressure with time among patch users. A small study of 18 women suggested that in normotensive women, the CVR slightly increased 24-hour ambulatory blood pressure after six cycles of use.371 The increase was statistically significant but small (mean 24-hour blood pressure increased by 2.69 ± 5.35 mmHg) and clinical relevance is uncertain.

Since hypertension increases the risk of MI and stroke, the GDG recommend that a woman wishing to continue using CHC should have her blood pressure checked annually to ensure that it remains within the acceptable range.

14.1.1 VTE risk during travel

Long duration travel is a weak risk factor for VTE. The risk is not confined to air travel (although most of the evidence relates to flying), increases with the duration of travel and with the presence of pre-existing risk factors for VTE. The British Society for Haematology bases its recommendations on a systematic review of the evidence.374 Data from prospective studies which evaluated air travellers for deep vein thrombosis (DVT) before and after their journeys suggest an incidence of DVT of 1 in 560 people travelling by air for 8 hours. Retrospective studies suggest that pulmonary embolism is extremely rare in flights of <8 hours but that the risk is associated with the duration of travel (5/million in flights >12 hours) and with pre-existing risk factors for VTE. There is indirect evidence that maintaining mobility during travel is a reasonable precaution, while for people without pre-existing risk factors, compression stockings and anti-platelet drugs (aspirin) are not indicated.

Case reports have suggested that CHC use could be associated with further increased risk of VTE associated with travel.375 The GDG recommends that women using CHC should minimise immobility during travel, but existing evidence is inadequate to allow any recommendation to be made about use of compression stockings or anti-platelet drugs.

14.1.2 Adherence to CHC when crossing time zones

Women travelling through different time zones should be reminded of the importance of taking their COC pill approximately 24 hours after their most recent pill (i.e. using the time of day in the time zone in which the last pill was taken as a reference point rather than local time).

A COC is missed when it has been more than 24 hours since the pill should have been taken. Two pills have been missed when it has been more than 72 hours since the last pill was taken. Refer to Section 8.1 for advice on missed CHC.

15.1.1 Use of CHC by women aged over 40 years

Women may use CHC for contraception until age 50 years provided there are no contraindications as outlined in UKMEC.96 After age 50 years, the risks of CHC use generally outweigh the contraceptive benefits and women should be advised to switch from CHC to a POP, subdermal implant, LNG-releasing IUS or a non-hormonal contraceptive method.378 Women taking CHC for non-contraceptive benefits who wish to continue use after the age of 50 years should be considered on an individual basis. See FSRH Guideline Contraception for Women Aged Over 40 Years.378

15.1.2 Duration of CHC use

For medically eligible women under the age of 50 years, there is no arbitrary maximum length of time for which CHC can be used. Repeated starting and stopping of CHC should be discouraged because VTE risk is highest in the months after initiation or when restarting after a break of at least 1 month (see Section 10.1).

15.3.1 Return of fertility

Women who wish to conceive after stopping contraception should be advised that after stopping CHC there is no significant delay in return to fertility and conception rates are comparable to those among women stopping other contraceptive methods or using no contraception (see Section 11.6). Folic acid supplementation can be started prior to stopping CHC.

15.3.2 Preconception care

Refer to the FSRH Clinical Statement Preconception Care.379

15.3.3 Unplanned pregnancy during CHC use

The available evidence suggests that fetal exposure to CHC is not associated with fetal abnormality (see Section 6.3.2). See FSRH Guideline Quick Starting Contraception.86