Human Reproduction Update Advance Access originally published online on July 8, 2005
Human Reproduction Update 2005 11(5):513-525; doi:10.1093/humupd/dmi019
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Noncontraceptive health benefits of combined oral contraception
The ESHRE Capri Workshop Group** A meeting was organized by European Society of Human Reproduction and Embryology (ESHRE) in Capri, September 2, 2004, with financial support from Schering S.p.A. to discuss the above subjects. The speakers included D.T.Baird (Edinburgh), J.Collins (Hamilton), P.G.Crosignani (Milan), A.Glasier (Edinburgh), H.S.Jacobs (London), C.La Vecchia (Milan), I.Milsom (Gothemburg), B.Reid (Kingstone, Canada), A.Templeton (Aberdeen) and P.Vercellini (Milan). The discussants included G.Benagiano (Rome), G.C.Frigerio (Milan), G.Hoppe (Pointe-Claire, Canada), S.Liberatore (Milan), K.Schmidt-Gollwitzer (Berlin), S.O.Skouby (Copenhagen) and A.Volpe (Modena). The report was prepared by J.Collins (Hamilton) and P.G.Crosignani (Milan).
To whom correspondence should be addressed at: P.G.Crosignani, Department of Obstetrics and Gynecology, University of Milano, Via Commenda 12, 20122 Milano, Italy. E-mail: piergiorgio.crosignani{at}unimi.it
 |
Abstract |
|---|
Contraception is one of the keystones of reproductive health. The availability of effective contraception has helped to change dramatically the structure of the world’s population during the last 50 years, through a demographic transition involving lower fertility rates and longer survival. As the transition evolves more slowly in developing countries, different effects on population structures contribute to civil strains. Oral contraception (OC) is an extremely effective method of contraception that also confers health benefits beyond pregnancy prevention. Notable effects on the reproductive system include relief from troublesome symptoms associated with menstruation such as heavy periods, painful periods and irregular bleeding. Many women also have improvement in acne and hirsutism. Moreover, OCs may be used to treat menorrhagia or symptomatic endometriosis. Use of OCs is associated with a long lasting reduction in the risk of developing cancer of the ovary and the endometrium. The effects on benign breast disease (BBD), bone health and colon cancer are less clear and merit further investigation.
Key words: oral contraception/dysmenorrhea/menorrhagia/acne/hirsutism/ovarian / endometrial and colon cancers
|
Introduction |
|---|
Every year the Capri Workshop has opened with some remarks by Egon Diczfalusy on the state of the world’s population. These scholarly statements highlight some of the important changes and trends that have occurred in different societies as they progress through the demographic transition. Above all, Egon has always pleaded for a rational ‘evidence based’ appraisal. ‘We are living in a world of increasing complexity, characterized by serious problems such as population, food and energy shortage, pollution etc., and natural solutions on a global scale are not always available or acceptable. In such a world, the passionate search for the emotion-free truth becomes exceptionally important and I would like to close my remarks by re-emphasizing the importance of the scientific method when dealing with the major problems confronting mankind’ (Diczfalusy, 1981
). In 2004, for the first time in 20 years, Egon was unable to attend the Capri Workshop. The discussants thought it important, however, to re-emphasize the crucial role that contraception plays in shaping the structure of societies. We are living in a time of immense demographic changes. On the one hand, the encouraging reduction in the rate of growth of the world’s population is easing some of the pressing problems identified by Egon 25 years ago. On the other hand, the rapid changes in the age structure of populations of countries at different stages of the demographic transition can lead to economic problems and social tensions (Figure 1).
|
Contraception is one of the keystones of reproductive health. Without contraception the individual is unable to choose when and how many children he/she has (Baird, 1965, 2000). Increasing population puts strain on the world’s finite resources such as land, water and clean air. In the last decade, the prevalence of contraceptive usage has increased worldwide. Together with the increase in life expectancy this has led, in many countries, to profound changes in the population structure. As societies progress through the demographic transition from high fertility and mortality to one of low fertility and low mortality, there are considerable social stresses which may lead to conflict. Thus patterns of reproduction have a major impact on the social and economic structure of a country (Lee, 2003).
At a global level, contraception is important in helping reduce overcrowding, pressures on resources, pollution and global warming, and a loss of animal species due to loss of habitat. It is just as important at an individual level—providing the means for individual couples to have the number of children they wish when they wish them. The health benefits of avoiding unplanned pregnancy and the ability to extend interbirth intervals have been well documented (Burkman et al., 2001; Petitti, 2003). There is an increasing awareness of the opportunity that many contraceptive interventions may provide for additional health benefits. Thus, the combined oral contraceptive pill in the short term reduces many of the troublesome side effects associated with menstruation. In the long term, the risk of cancer of the ovary, uterus and probably also of the colon and rectum is reduced (Burkman et al., 2004). Long-term follow-up studies have found no evidence of increased or decreased mortality in women who have taken a combined oral contraceptive (OC; Beral et al., 1999; Vessey et al., 2003). Societal benefits also accrue from use of OCs, such as reduced disability due to dysmenorrhea and menorraghia, which seriously affect approximately 2.5 x 106 women annually in the USA and cost US industry 8% of the total wage bill (Thomas and Ellertson, 2000). In the future, it may be possible to devise strategies to reduce the risk of other life threatening disease such a breast cancer.
This review summarizes the evidence on noncontraceptive benefits associated with use of the OC. It will draw on data from randomized controlled trials (RCTs) and epidemiological studies among women using OCs for contraception. Trials of the OC as a treatment for symptomatic women (with dysmenorrhea, for example) are also relevant to whether a corresponding noncontraceptive benefit can be expected among women using the OC for contraception.
|
Dysmenorrhea |
|---|
Dysmenorrhea is the commonest form of menstrual disorder with a reported prevalence of 50–90% among young women (Andersch and Milsom, 1982
; Sundell et al., 1990). In Sweden, 15% of young women have been reported to suffer from dysmenorrhea, which causes absenteeism from school or work every month. Recent reports from the USA have stated that dysmenorrhea accounted for 600 x 106 lost hours and 2 billion dollars in lost productivity annually (Thomas and Ellertson, 2000). In the UK, OC users had 26 hospital admissions for dysmenorrhea per 100,000 women per annum, compared with 50 among nonusers (Vessey et al., 1996).
The pain of dysmenorrhea arises from the release of prostaglandins, which cause an increase in myometrial activity (Lundström and Gréen, 1978). Several studies have documented that combined oral contraceptives (OCs) reduce menstrual prostaglandin release and thereby decrease uterine contractility and alleviate dysmenorrhea (Lundström and Gréen, 1978; Hauksson et al., 1989). OCs have been reported to provide effective pain relief for 70–80% of women with primary dysmenorrhea (Lundström and Gréen, 1978; Hauksson et al., 1989; Milsom et al., 1990).
The impact of low-dose OC use on dysmenorrhea was assessed in a longitudinal survey of a random sample of 19-year-old women (n = 596) (Milsom et al., 1990). Among nonusers, prevalence of dysmenorrheal was 72% at age 19 and 67% at age 24; both prevalence and severity of dysmenorrhea were 10 to 20% lower at age 19 and at age 24 among women who were current users of low-dose OCs compared to women of the same age and from the same population who were not using an OC. Women who started taking an OC after age 19 during the course of the 5-year survey experienced a reduction in the severity of dysmenorrhea with OC use.
In a noncomparative trial, Larsson et al. (1992) reported that treatment with a low-dose OC significantly reduced dysmenorrhea. Fourteen of the 20 young women included in the trial complained of dysmenorrhea on inclusion before starting with a low-dose OC. After 6 months, treatment with a low-dose OC only 4 women reported dysmenorrhea.
A RCT involved 52 women less than 32 years of age who completed 4 months of treatment with placebo or an OC containing 20 µg of ethinyl estradiol (EE) and desogestrel 150 µg with an additional 20 µg of EE in the last 5 days of the usual tablet-free period (Hendrix and Alexander, 2002). Menstrual cramping was significantly reduced (P < 0.001) in OC users compared to placebo users. In another RCT, among women age 17–45 years (average 28), dysmenorrhea prevalence decreased approximately from 56 to 39% during 6-month use of oral contraceptives containing 20 µg of EE and either desogestrel 150 µg or levonorgestrel 100 µg (Winkler et al., 2004).
Thus, there are data from RCTs, cross-sectional surveys and a non-comparative trial that support the value of OCs, including those with low dosage, in reducing the pain associated with menstrual periods.
|
Menorrhagia |
|---|
Approximately, 10% of fertile women suffer from menorraghia, defined as a menstrual blood loss of >80 ml. The prevalence of menorrhagia increases with increasing age. Excessive blood loss may lead to iron deficiency anaemia and ultimately necessitate hysterectomy (Hallberg et al., 1966
). Menstrual blood loss has been demonstrated to be reduced by approximately 50% in users of both high-dose OCs (Nilsson and Solvell, 1967; Nilsson and Rybo, 1971) and low-dose OCs (Fraser and McCarron, 1991; Larsson et al., 1992). In the UK, OC users had 267 hospital admissions for dysmenorrhea per 100,000 women per annum, compared with 611 among nonusers (Vessey et al., 1996).
The influence of a low-dose OC on menstrual blood loss was assessed by Larsson et al. (1992) in a noncomparative trial where menstrual blood loss was assessed objectively using the alkaline haematin technique. Twenty healthy young women with above average blood loss before commencing OC medication (60.2 ± 5.6 ml) were treated with an OC containing 30 µg EE/150 µg desogestrel. At 3 and 6 months on low-dose OC, the average blood loss had decreased to 36.5 ± 5.2 ml and 33.7 ± 4.1 ml, respectively (P < 0.001). The duration of menstruation was also reduced (P < 0.01) during OC use. An Australian group reported similar results from a randomized trial involving 45 women with menorrhagia who were treated during an 8-cycle crossover treatment trial with mefenamic acid or naproxen or a low-dose OC or danazol (Fraser and McCarron, 1991). Table I summarizes three studies evaluating the influence of combined oral contraceptives on menstrual blood loss. A significant reduction in menstrual blood loss (43%) was recorded during the use of the low-dose OC in a US-RCT involving 201 women with menorrhagia and dysfunctional bleeding who were allocated to placebo or an oral contraceptive containing EE 35 µg and triphasic norgestimate, 87% of oral contraceptive users reported reduced menstrual bleeding compared with 45% of those allocated to placebo (Davis et al., 2000). These studies demonstrate that low-dose OCs reduce menstrual blood loss and are effective in the treatment of menorrhagia.
|
Several of the modern hormonal forms of contraception [e.g. OCs, progestogen only pills (POPs), implants, medicated IUDs and injectables] induce amenorrhea in a varying proportion of women and OCs can be taken in long cycle use reducing the number of annual menstruations. Eliminating menstruation or reducing its frequency has a profound impact on the occurrence of menorrhagia and the consequences of this common condition. A 1-year RCT compared four cycles of an extended cycle regimen taken daily for 84 days with 7 days of placebo (n = 397) and 13 cycles of a standard 21-day regimen with 7 days of placebo (n = 195; Anderson and Hait, 2003). The extended cycle regimen involved more unscheduled breakthrough bleeding than the control 28-day cycle regimen, but during the 364 days, bleeding and/or spotting occurred on a median of 35 days with the extended cycle regimen compared with 53 days for the conventional regimen.
In a study of 220 women who were offered the extended regimen of oral contraception (OC), 174 tried to use it and 121 (67%) continued to do so for more than 1.6 years of follow-up. They did so for a variety of reasons: premenstrual syndrome (PMS; 45%) dysmenorrhoea/pelvic pain (40%), heavy withdrawal bleeding (36%), and menstrual-associated headaches (35%) etc. The majority of patients who initiated an extended regime reported a reduction in their menstrual-associated complaints, and 99% of women still using the extended regimen noted improvement in their quality of life (Sulak et al., 2004). Although monthly menstruation is natural and normal, it is not necessarily beneficial for all women. When indicated, menstruation can be minimized with safe and widely available forms of hormonal contraception.
|
Endometriosis |
|---|
Drugs used in the treatment for endometriosis are not cytoreductive (Prentice, 2001
). Quiescent implants have been demonstrated in nearly all women treated with danazol, gonadotrophin-releasing hormone agonists (GnRH Ag) and progestogens. At restoration of ovulation and of physiologic levels of estrogens, both the eutopic and ectopic endometrium resumes its metabolic activity. As a consequence, medical therapy is symptomatic and pain relapse at treatment suspension is the rule (Vercellini et al., 1993).
Drugs to be administered only for some months due to poor tolerability, severe metabolic side effects or high cost, do not greatly benefit women with symptomatic endometriosis (Vercellini et al., 2003c). Contraceptive progestogens alone or combined with estrogens are instead generally well tolerated, have a more limited metabolic impact as opposed to danazol or GnRH Ag, are inexpensive and may be used on a long-term basis (Moore et al., 2003; Prentice et al., 2003). Cyclic hormonal contraception is the only treatment for endometriosis that permits a monthly uterine bleeding. Dysmenorrhoea is known as the most frequent and most severe complaint in women with this disease. The symptom may therefore not subside completely during administration of an OC. Women affected by dysmenorrhoea during cyclic use of an OC may benefit from the shift to a continuous administration. A monophasic OC (desogestrel 0.15 mg and EE 0.02 mg) was prescribed continuously in 50 patients with dysmenorrhoea relapsing after conservative surgery for endometriosis and not responding to the cyclic use of the same OC (Vercellini et al., 2003b). During the 2-year study period, 38% of women reported amenorrhoea, 36% spotting and 26% breakthrough bleeding. The mean score of menstrual pain, evaluated according to a 100-mm visual analogue scale, showed a reduction from 75 ± 13 to 31 ± 17. Moderate or severe side effects were reported by 14% of the women.
Progestogens may prevent implantation and growth of regurgitated endometrium inhibiting expression of matrix metalloproteinases and angiogenesis, and show several anti-inflammatory in vitro and in vivo effects that may reduce the phlogistic status generated by the metabolic activity of the ectopic endometrium and the consequent immune response. The effect of OC on regulation of in vitro endometrial cell growth has recently been studied by Meresman et al. (2002) in biopsy specimens performed in 13 women with endometriosis and in 13 controls. Apoptosis of the endometrium is regulated by steroid hormones and is controlled by the expression of several regulatory genes. The proto-oncogene B cell lymphoma/leukemia-2 (Bcl-2) blocks apoptosis, whereas the Bax protein antagonizes its survival activity. The relation Bcl-2/Bax predetermines the cells’ susceptibility to apoptotic inducers. Eutopic epithelial and stromal endometrial cells from patients with endometriosis show an augmented survival capability probably caused by an abnormally high effect of Bcl-2. Exposure to a monophasic OC for 30 days significantly increases endometrial apoptosis in comparison with pretreatment levels, both in epithelial and stromal cells, determining values similar to those observed in the endometrium from control women without endometriosis. In fact, OC administration reverses the abnormal increment in Bcl-2 expression in patients with endometriosis and induces a major increase in the expression of Bax. Furthermore, endometrial cell proliferation in subject with the disease is significantly reduced by exposure to an OC.
Moreover, anovulation, decidualisation, amenorrhoea and the establishment of a steady estrogen–progestogen milieu contribute to disease quiescence.
Progestogens and estrogen–progestogens are effective in the control of pain symptoms in approximately three in four women affected by endometriosis (Table II). Their effect does not seem to be inferior to that of other drugs used for the disease (Vercellini et al., 1993; Moore et al., 2003; Prentice et al., 2003). Different compounds can be administered by the oral (Cosson et al., 2002; Vercellini et al., 2002, 2003b), intramuscular (Vercellini et al., 1996), subcutaneous, intravaginal and intrauterine route (Vercellini et al., 1999, 2003a; Fedele et al., 2001), with specific advantages and disadvantages.
|
Medical treatment plays a role in the therapeutic strategy when administered over a prolonged period of time. Given their good tolerability, minor metabolic effects and low cost, contraceptive steroidal preparations must therefore be considered drugs of choice and are currently the only safe and economic alternative to surgery (The Royal College of Obstetricians and Gynaecologists, 2000).
|
Premenstrual syndrome |
|---|
During the reproductive years, up to 80–90% of menstruating women will experience symptoms (breast pain, bloating, acne and constipation) that forewarn them of impending menstruation, so-called premenstrual molimina. Over 60% of women report swelling or bloating (Lee and Rittenhouse, 1991
), although objective documentation of weight gain is lacking in most of these women (Faratian et al., 1984). Cyclic breast symptoms affect 70% of women with 22% reporting moderate to extreme discomfort (Ader et al., 2001). Available data suggest that as many as 30–40% of these women are sufficiently bothered by symptoms to seek relief. The term PMS should be reserved for a more severe constellation of symptoms—mostly psychiatric, that lead to major interference with day-to-day activities and interpersonal relationships (Reid and Yen, 1981). Women with this degree of symptoms probably comprise 3 to 5% of women in their reproductive years (Wood et al., 1982; Johnston et al., 1988; Rivera-Tovar and Frank, 1990).
PMS has been defined as ‘the cyclic recurrence in the luteal phase of the menstrual cycle of a combination of distressing physical, psychological and/or behavioural changes of sufficient severity to result in deterioration of interpersonal relationships and/or interference with normal activities’ (Reid and Yen, 1981).
|
The American Psychiatric Association in an appendix to their Diagnostic and Statistical manual have described Premenstrual Dysphoric Disorder (PMDD) (Table III).
It is likely that PMS has emerged as a twentieth-century phenomenon in part due to the fact that women’s increasing control over reproduction has eliminated the cycle of repeated pregnancy and lactation that formerly characterized the lives of women from puberty to menopause (Reid, 1991). PMS-like behaviour has been reported both in humans and in nonhuman primates as long as they demonstrate menstrual cyclicity. In the nonhuman primate, zoologists have noted premenstrual changes in behaviour and appetite similar to those reported by women with PMS (Gilbert and Gillman, 1956; Sassenrath et al., 1973).
PMS may affect women at any stage of reproductive life. PMS sufferers often relate that symptoms become progressively worse over time and since women have increasing contact with health care providers for nonpregnancy-related concerns in their later reproductive years, this may account for the preponderance of older women seeking help for PMS.
PMS disappears during suppression of the ovarian cycle (for example during hypothalamic amenorrhea due to excessive physical or nutritional stress, during lactational amenorrhea, during pregnancy and after menopause—either natural or induced) (Reid, 2004). Support for a beneficial effect of menstrual cycle suppression comes from historical evidence that PMS disappears during pregnancy and lactation, in women with amenorrhea due to hypothalamic causes and following natural or surgical menopause (Casson et al., 1990) as well as from studies involving such as Danazol (Hahn et al., 1995) and GnRH Ag (Mezrow et al., 1994) which can temporarily and reversibly block menstrual cyclicity.
Despite being one of the most commonly used medications among women of reproductive age, the OC has been given to women with premenstrual complaints for 40 years without a clear understanding of what to expect in terms of symptom duration or severity. We know that OCs will reduce menstrual cramps and flow, and by so doing may improve the entire premenstrual/menstrual experience for some women. However, many women will discontinue OC use due to distressing premenstrual symptomatology, and of those women using the OC, some will report the onset of premenstrual symptoms at an earlier stage than in nonOC cycles.
Observational studies in the 60s and 70s suggested that premenstrual complaints were reduced in many oral contraceptive users (Moos, 1968), although symptoms could be exacerbated by OC use in some women.
The first systematic studies to examine the effects of the OC on PMS found little difference in PMS symptoms between OC users and nonusers (Bancroft and Rennie, 1993), nor were there significant differences between agents with differing progestational potencies (Andersch, 1982). Monophasic and triphasic preparations showed similar rates of symptomatology (Graham and Sherwin, 1992).
A new OC preparation containing a novel progestin with diuretic effects (drospirenone) has undergone the most rigorous testing in normal women and women with strictly defined PMDD. Though the evidence supporting a role for fluid retention as an etiologic component of PMS is lacking (Faratian et al., 1984), many women are distressed by feelings of bloating and edema. In an RCT involving 82 women with PMDD, patients treated with the drospirenone and EE oral contraceptive had fewer luteal phase symptoms than those treated with placebo, although between-group differences were statistically significant only for appetite, acne and food cravings, P = 0.027 (Freeman et al., 2001). The drosperinone OC offers relief for some physical and some psychological manifestations of PMS and may improve health-related quality of life (Borenstein et al., 2002; Freeman, 2002). For prolonged treatment among patients with moderate symptoms, OC is preferable, more acceptable and cheaper than selective serotonin reuptake inhibitors, but with severe symptoms premensrual dysphoric disorder (PPDD) more specific treatment is required.
|
Hyperandrogenism |
|---|
By far, the commonest ovarian cause of hyperandrogenisation causing acne and hirsutism is polycystic ovary syndrome (PCOS). The main source of androgens in PCOS is the ovary, although there may also be a contribution from the adrenal (Yen, 1999
).
Acne is the most common skin condition treated by physicians, affecting up to 40% of adolescents and 10% of adult women (Redmond et al., 1997; Thiboutot et al., 2001). Acne and seborrhoea are mild clinical manifestations of hyperandrogenization which may improve during OC use as a noncontraceptive benefit. Androgens are related to the development of acne, whereas many combined oral contraceptive pills and in particular those which contain anti-androgens are clearly beneficial for the treatment of acne. A recently published review considered OC among effective treatments for acne such as topical retinoids and topical or oral antibiotics (Haider and Shaw, 2004). A recent systematic review reported on five placebo-controlled trials that evaluated oral contraceptive treatment for women with moderate to severe acne (Arowojolu et al., 2004). OCs reduced acne lesion counts and severity as assessed by clinicians and self-assessment. The trial data indicate that 10 to 20% more OC users than placebo users were improved by clinician assessment when the OC was taken expressly for the treatment of acne. It seems likely that women using OCs for contraception who also suffer from acne would have a similar likelihood of benefit. The RCT evidence may not be relevant, however, to OC users with mild acne, which might involve different etiologic factors and a correspondingly different response to OCs.
The effect of OC on acne depends on their progestogen activity. Theoretically progestogen dominance would exacerbate acne, whereas estrogen dominance would exert an anti-androgenic effect on the sebaceous glands. Several OCs have been investigated in comparative studies in the treatment of acne in women, including products containing cyproterone acetate (CPA; Erkkola et al., 1990), norgestimate (Redmond et al., 1997), desogestrel (Volpe et al., 1994), gestodene (Piérard-Franchimont et al., 2000), chlormadinone acetate (Worret et al., 2001) and drospirenone (van Vloten et al., 2002). The systematic review found 14 trials that compared the effectiveness of various OCs in the management of acne (Arowojolu et al., 2004). OCs that contained chlormadinone acetate or CPA were better than those containing levonorgestrel or desogestrel and similar to a OC containing drospirenone; however, data were limited and some studies of similar OC comparisons produced conflicting results (Arowojolu et al., 2004).
With respect to treatment of hirsutism, a systematic review of treatment for hirsutism found one small study that compared CPA in combination with EE to placebo. There was a significant subjective reduction in hair growth with CPA therapy, but the confidence limits were wide (Van der Spuy and le Roux, 2003).
Side effects of CPA include weight gain, oedema, headache, fatigue, mood changes, reduced libido and breast tenderness. In some countries, it is approved not as a contraceptive, but only for the treatment of androgenization, and in some countries, regulatory authorities recommend liver function tests be carried out before and during treatment with cyproterone-containing hormone preparations.
Combined hormonal contraception and bone mineral density
There is now a body of literature on the effects of hormonal contraception on bone. The research falls into two broad categories; studies investigating bone mineral density (BMD) and epidemiological data on fracture risk. High-dose progestogen-only contraception, particularly Depo Provera, inhibits ovarian activity causing hypo-oestrogenism and loss of BMD (Cundy et al., 1998; Scholes et al., 1999). There is no good evidence, however, that lower bone density is associated with any clinically significant increase in the risk of fracture (Reproductive Health and Research World Health Organization, 2004). Low-dose progestogen only contraceptives (pills, implants and intrauterine system (IUS) do not inhibit ovarian oestrogen production in the majority of women and appear to have no significant measurable effect on BMD (Petitti et al., 2000).
Noncontraceptive benefits of the combined pill (Kaunitz, 1999; Sulak, 1999; Burkman et al., 2001) all mention a possible beneficial effect of OC on BMD. The combined pill—and the patch, ring and injection—suppresses ovulation and the mid-cycle rise of oestrogen, substituting a constant dose of estrogen and progestogen. Nevertheless, average monthly concentrations of estrogen are lower during use of the combined pill than in a typical ovulatory cycle. Studies on the effect of the combined oral contraceptive pill on BMD are difficult because BMD is influenced by many factors which are difficult to control for (including inheritance, diet, exercise and smoking). The studies are not easy to interpret because it is hard to find a control group that has never been exposed to contraceptive hormones. Moreover, women who start a particular method of contraception (even during a clinical trial) often discontinue after a relatively short time resulting in a high dropout rate. It is particularly difficult to compare studies since they involved different types of OC and measured BMD at different sites using different techniques and different time intervals. A systematic review (Kuohung et al., 2000) concluded that there is fair evidence that OC use has a favourable effect on BMD. The review, however, included trials which involved perimenopausal and postmenopausal women, and more recent studies (Petitti et al., 2000; Reed et al., 2003) have found no effect of OC on BMD among women under 40. The most recent studies have also measured markers of bone turnover among women using the combined pill (Nappi et al., 2003; Endrikat et al., 2004). Although there is evidence for an increase in markers of bone formation and a reduction in markers of resorption, even among pills containing 20 µg or 15 µg EE (Nappi et al., 2003; Endrikat et al., 2004), no changes in BMD have been observed; however, it is possible that the study period is insufficient to show any effect of the pill. Data on the effect of the combined pill on the BMD of adolescents are conflicting; while some studies suggest a beneficial effect on BMD (Berenson et al., 2001; Cromer, 2003), others show no measureable effect (Lloyd et al., 2000). One study (Berenson et al., 2001) suggested a differential effect of the type of progestogen.
Changes in BMD may not accurately reflect actual fracture risk. Only a handful of studies have reported on the relationship between OC use and fracture. Two cohort studies from the UK documented a 20% increase in the risk of fracture among ever users compared with never users of the pill (Cooper et al., 1993; Vessey et al., 1998). The relative risk (RR) of fracture was 1.2 (95% CI 1.08–1.34) in the Cooper et al. study and 1.2 (95%CI 1.1–1.4) in the Vessey et al. study. Both studies, however, included few women over age 50 and almost none over age 65 when osteoporotic fracture becomes more common. Both groups of authors concluded that a direct effect of the OC on fracture risk seemed unlikely and hypothesized an effect of confounding variables including smoking, alcohol use and risk-taking behaviour. In contrast, in a case–control study of 1327 Swedish women aged over 50 with hip fracture, ever use of the OC was associated with a 25% reduction in the risk of hip fracture (Michaelsson et al., 1999). The beneficial effect was more marked with use of pills containing 50 µg EE and confined to women with a BMI <25 (Table IV). It is difficult to extrapolate these data to the twenty-first century when women use low-dose pills, take them for many years and tend to have a significantly higher body mass index than women in the 1960s and 1970s.
|
For women with hypo-estrogenism, combined hormonal contraception may be of clinical benefit. The evidence that OC improves BMD in perimenopausal women (compared with controls not using hormonal contraception) is more persuasive. OC seems to be as good as menopausal hormone therapy (HT) in preserving or increasing BMD in this group (Taechakraichana et al., 2000). It is almost certainly appropriate to extrapolate these findings to women with premature menopause. The combined pill may indeed be preferable to HT for these women who find that the need to use HT at a relatively young age reminds them daily of their condition.
For women with stress-related conditions (hypogonadotrophic or normogonadotrophic amenorrhoea), there is quite good evidence that OC use maintains or improves BMD (Castelo-Branco et al., 2001).
A number of trials have demonstrated, however, that there is no beneficial effect on BMD among women with anorexia nervosa (Gordon et al., 2002; Munoz et al., 2002). The bones of underweight women with poor nutrition simply do not respond to estrogen. Since estrogen increases bone resorption as well as deposition, very underweight women can theoretically lose more bone if they are treated with estrogen. Preliminary data do suggest that if OC is given together with recombinant human insulin-like growth factor (IFG-I) BMD is significantly increased (Grinspoon et al., 2002) (Figure 2).
|
In conclusion, combined hormonal contraception probably has little clinically significant effect on BMD and fracture risk in women of reproductive age and normal weight who use modern low-dose pills. For women with oligo-amenorrhoea and normal weight, the OC is probably beneficial in respect of BMD and simultaneously provides contraception if required. OC will not benefit BMD among women with anorexia nervosa and may even be detrimental.
|
Benign breast disease |
|---|
Use of the combined oral contraceptive pill is associated with an increased risk of breast cancer that returns to normal within 10 years after cessation of use (The Collaborative Group on Hormonal Factors in Breast Cancer, 1996
). Paradoxically, the use of OC reduces the risk of benign breast disease (BBD; Kaunitz, 1999; Burkman et al., 2001). BBD is a general term which incorporates a wide spectrum of disorders. Despite numerous attempts, there is no internationally agreed classification of these breast disorders. The only genuine point of agreement seems to be that the term excludes malignant breast pathology. There seems to be no clear consensus either as to whether BBD is associated with an increased risk of breast cancer with the single exception of benign tumors with atypical change.
Two major conditions dominate the literature on BBD; fibrocystic disease and fibroadenoma (Goehring and Morabia, 1997). Fibrocystic breast disease is an ill-defined condition characterized by the presence of palpable lumps in the breast, usually with pain and tenderness. The lumps and the symptoms fluctuate with the menstrual cycle. The maximum incidence is among women in their early forties (Figure 3) and the symptoms tend to worsen until the menopause. The histological changes are in the breast epithelium which often contains micro- or macro-cysts and adenosis (development of new lobular or ductal structures). The epithelium may display metaplasia into apocrine epithelium such as that found in sweat glands, or hyperplasia, with or without atypia. Fibroadenoma is a benign tumour arising from a single breast lobule, usually a well-defined mass containing an epithelial and a stromal component. Fibroadenomas appear to result from hyperplasia and distortion originating in a single lobule involving normal epithelial components and abnormal stromal components. The maximum incidence is among women aged 20–25 years (Figure 3). The histological changes are confined to the stroma, the epithelial cells are normal.
|
A number of studies published in the 1970s and 1980s (McGonigle and Huggins, 1991; Goehring and Morabia, 1997) suggested that OCs used in the 1970s exerted a protective effect against benign breast disease. In many of these studies, however, the diagnosis of BBD was not confirmed by biopsy but was based on self-reporting. The three more recent studies have demonstrated somewhat inconclusive results. One small prospective study of OC users in Greece suggested a reduced incidence of BBD diagnosed by mammography and ultrasound (Tzingounis et al., 1996). One hospital-based case–control study from France (Charreau et al., 1993) of women with BBD confirmed by biopsy demonstrated a reduced risk among OC users, but only of nonproliferative disease and only for pill use before first full term pregnancy. In contrast a case-cohort analysis within a large cohort of Canadian women undergoing breast screening and biopsy (Rohan and Miller, 1999) showed an inverse relationship between OC use and proliferative BBD without atypia. In this study, there was no relationship between BBD with atypia and OC use. More modern pills contain lower doses of both estrogen and progestogen than those used in the 1970s, and it is possible that the benefit for BBD is diminished, or lost, with lower doses of steroids.
All the research suggesting a beneficial effect of the combined pill on benign breast disease comes from observational studies with potential bias. It is likely that a great deal of benign breast disease goes unrecognized or certainly does not reach the attention of the medical profession. The finding that BBD is more common among better-educated women of higher socioeconomic status suggests probable selection bias (Goehring and Morabia, 1997). The finding that BBD is less common in obese women suggests a likely detection bias (Goehring and Morabia, 1997). At least one study demonstrated a prescribing bias (Janerich et al., 1997), suggesting that women with a history of BBD are less likely to be prescribed oral contraceptives. The definitions of BBD vary between the published studies, some of which are of rather poor quality and in many the numbers involved are small. There is good evidence that OC use stimulates epithelial cell proliferation in the breast (Anderson et al., 1989; Isaksson et al., 2001) casting some doubt on the biological plausibility of a protective effect on benign disease.
In conclusion, although there is some evidence that high-dose oral contraceptives may reduce the risk of BBD, there are significant problems with bias, study design and interpretation, and there is no evidence of biological plausibility (Burkman et al., 2004). While an increased risk of BBD has never been demonstrated, the beneficial effects of the pill on the breast may have been overstated.
|
Fibroids |
|---|
It has been suggested that OC use is associated with a reduced risk of fibroids (Ross et al., 1986; Parazzini et al., 1988
). However, in these early studies, the diagnosis was limited to women with surgical confirmation of fibroids. Two large studies from the USA in which fibroids were diagnosed either at hysterectomy or by ultrasound (Marshall et al., 1998; Wise et al., 2004) suggest no clear pattern of association, although both report a weak association between fibroids and use of the combined pill before the age of 17. A smaller study of 335 randomly selected Swedish women aged 25–40 reported a rather low prevalence of (ultrasound detected) fibroids (5.4%; 95% CI 3.0–7.8%) in contrast to earlier studies, which suggested a prevalence of 20 to 25% in women over age 30 (Borgfeldt and Andolf, 2000). Although the sample size was too small to detect an effect of the OC on fibroids, uterine size was significantly smaller among women on the combined oral contraceptive pill than in women with natural cycles. The authors suggest that the high prevalence of OC use might explain the low prevalence of fibroids in their study. |
Benign ovarian tumours |
|---|
Just as use of the combined pill is associated with a reduced risk of ovarian cancer, it has been suggested that the risk of benign ovarian tumours is also reduced. A number of studies have shown a reduced risk of functional ovarian cysts which is unsurprising since ovulation is inhibited (Holt et al., 1992
). However, a large case–control study from the USA demonstrated that ever use of the combined pill was associated with a decreased risk of nonfollicular benign tumours, including serous and mucinous adenoma, teratoma and endometrioma (OR 0.79, 95% CI 0.6–1.05) (Westhoff et al., 2000). The reduction in risk was associated with duration of use. Although OC use is associated with a lower prevalence of benign ovarian tumours, trials have failed to show that the use of OC can prevent the development of ovarian cysts or treat existing ovarian cysts (The ESHRE Capri Workshop Group, 2001). |
Sexually transmitted infections |
|---|
 TOPAbstractIntroductionDysmenorrheaMenorrhagiaEndometriosisPremenstrual syndromeHyperandrogenismBenign breast diseaseFibroidsBenign ovarian tumours Sexually transmitted infections CancerConclusionsReferences |
|---|
With an apparently ever increasing rate of sexually transmitted infections (STIs) (British Medical Association Board of Science and Education, 2002
), several European governments have expressed concerns about the current poor state of sexual health (Great Britain Department of Health, 2003).
Currently, the most prevalent infection in Western Europe is Chlamydia trachomatis (CT). The prevalence of CT infection among asymptomatic European women ranges from 1.7 to 17%, depending on the setting (Wilson et al., 2002). The mode was 6% for women seeking contraception and 4% for women having cervical smears. The difference in prevalence is determined mainly by age, rather than by clinical setting, with the highest prevalence in teenagers and the lowest one in women over the age of 30. Prevalence rates may be as high in men, although not necessarily following the same age distribution, calling into question any screening initiative that focuses solely on women (Fenton et al., 2001).
Although earlier studies suggested a decrease in hospitalizations for pelvic infection, in other studies oral contraceptive use was associated with an increased risk of chlamydial infection and gonorrhoea (Cottingham and Hunter, 1992). Most of the studies have been cross-sectional in design, have used insensitive tests for chlamydia infection and most have failed to control for potential confounding factors such as sexual behaviour and age. Of course, hormonal contraceptive use is common and so are STIs, but the increased prevalence of cervical ectropion associated with use of the pill might enhance the transmission of infection. A recent observational study involving 814 women in the USA who were using the pill, Depo Provera or a nonhormonal method of contraception carefully documented sexual behaviour (Morrison et al., 2004). The authors concluded that the use of OC was not significantly associated with an increased risk of chlamydia or gonorrhoea after adjusting for other risk factors (HR 1.5; 0.6–3.5). No association was found between the presence of cervical ectopy and infection either with oral contraceptive or injectable contraceptive use. The risk of acquiring CT infection during 1 year of follow-up was 1.9-fold higher in oral contraceptive users (95% CI 0.7, 4.8), compared with users of nonhormonal contraceptives after adjusting for sexual behaviour. Although this adjusted risk is not significant, the point estimate corresponds to the risk estimates from the earlier studies and underlines the need for promoting barrier methods of contraception (Morrison et al., 2004). Clearly, prevention of cervical infection is not a noncontraceptive benefit from OC use.
An indirect benefit may be the opportunity to conduct screening for STDs in family planning clinics, now that highly sensitive nucleic acid-based tests can allow for noninvasive testing for CT in a variety of settings (Watson et al., 2002). The prevalence of infection among women attending three Family Planning Clinics in Scotland is shown in Table V. The overall rate was 5.2% and infection rates were not apparently associated with contraceptive use, including condom use, number of partners or reasons for attending the clinic (Noone et al., 2004). Teenagers and women undergoing abortion represent the most cost-effective group for testing, although the evidence supports the offer of testing to younger women attending Family Planning Clinics (Norman et al., 2004).
|
However, there are a number of unanswered questions surrounding the offer of CT testing at Family Planning Clinics, not least the frequency of testing and the possibility of reinfection. However, the majority of women agree with a policy of regular testing for CT and expressed a preference for the Family Planning Clinic over other healthcare settings. Although the need to obtain contraceptive advice provides an opportunity for screening women for STIs, only the male condom has been shown to have a protective effect (Cates, 2001).
|
Cancer |
|---|
Oral contraceptives convey a protection against ovarian, endometrial and perhaps colorectal cancer. The inverse relation between OC use and ovarian and endometrial cancers is one of the most consistent epidemiological findings and one of the most important examples—on a public health level—of large-scale chemopreventive intervention (IARC, 1999
; La Vecchia et al., 2001).
An indication of the long-term favourable impact of OCs on ovarian carcinogenesis comes from descriptive epidemiology. In several developed countries, in fact, young women showed substantial declines in ovarian cancer mortality over the last few decades. Cohort analysis of trends in mortality from ovarian cancer showed that women born after 1920—i.e. from the generations who had used OCs—had reduced ovarian cancer rates, and the downwards trends were greatest in countries where oral contraceptives have been more widely utilized (La Vecchia et al., 1998; IARC, 1999).
The protection was similar for modern, low-dose estrogen–progestin pills (Ness et al., 2000), and for various histotypes of ovarian cancer (Tung et al., 2003), whereas it is unclear whether the protection is similar for women with familial ovarian cancer(s).
Moreover, quantification of the long-term effect of the OC on ovarian carcinogenesis remains open to discussion. Although a duration effect has been reported from most studies, time since first use (latency) and last use (recency) has been less frequently considered. The overall estimates of protection for ever use is approximately 40%, and the favourable effect of OC on epithelial ovarian cancer seems to persist for at least 10 years after stopping use according to the CASH study and most likely up to 15–20 years (La Vecchia et al., 1998; IARC, 1999). Thus, in a pooled analysis of European studies, the RR was 0.8 for use stopped for either 
10 years or 20 years (Bosetti et al., 2002), in a large multicentric US case–control study, the RR was 0.5 for use stopped for 15–19 years, and 0.8 for 20 years (Rosenberg et al., 1994) (Table VI). In the Oxford Family Planning Association (FPA) cohort study, the RR of death from ovarian cancer was 0.4 at the 30 year follow-up (Vessey et al., 2003).
|
OC use reduces the risk of endometrial cancer by approximately 50% (La Vecchia et al., 1998; IARC, 1999), but the relatively limited number of elderly women who had used OCs does not allow an accurate estimation of the protection afforded after long periods. Still, the reduced risk of endometrial cancer seems to persist at least 15–20 years after stopped use. In the CASH study, the RR was 0.5 for 10–14 years since stopping, in the WHO study the OR was 0.2 for high progestogen content pills 10 years since stopping, and in a multicentric US study the OR was 0.3 for 15–19 years and 0.8 for 20 years since stopping OC use (IARC, 1999; La V