If ovulation is not induced with the initial dose of 50 mg, the dose of CC can be increased in 50-mg increments in subsequent cycles up to a maximum dose of 250 mg/day. Approximately 70% of anovulatory women treated with CC respond to a dose of 100–150 mg/day (Gysler et al 1982). In practice, doses of 250 mg are rarely used. Women who do not ovulate while taking a dose of 150 mg are considered to be ‘clomiphene-resistant’ (Vandermolen et al 2001). The required CC dose is correlated to body weight, as there is a significant association between CC treatment failure and BMI greater than 27 kg/m². Women who are overweight should be advised that a 5% reduction in weight may improve endocrine and ovarian function, increase spontaneous conception rates and improve the response to CC treatment. Lower doses of 25–50 mg/day should be considered for those women who are exceptionally sensitive to treatment with CC.

Monitoring of CC treatment is recommended. Ovulation is confirmed either with serum progesterone measurements or with the use of transvaginal ultrasonography. Serum progesterone concentrations are measured during the luteal phase. Ultrasound monitoring should be offered at least during the first treatment cycle to ensure that the woman is on the lowest necessary dose to achieve ovulation (National Institure for Health and Clinical Excellence 2004). Monitoring will minimize the risk of multiple pregnancy and ovarian hyperstimulation syndrome (OHSS), and is essential in patients in whom ovulation does not seem to have been induced.

Results

Studies with CC have shown an ovulation rate of 60–85% and a pregnancy rate of 30–40%. The discrepancy between ovulation and fecundity rates has not been explained, but may be due to other coexisting fertility factors, such as endometriosis or tubal damage. Some argue that it may be associated with possible antioestrogenic effects that CC exhibits on the endometrium and cervical mucus (Milson et al 2002).

A systematic review of 12 randomized controlled trials supported the effectiveness of CC therapy for anovulatory infertile women (Beck et al 2005). Clomiphene was found to be effective in increasing pregnancy rate compared with placebo [fixed odds ratio (OR) 5.8, 95% confidence interval (CI) 1.6–21.5; number needed to treat (NNT) 5.9, 95% CI 3.6–16.7].

Cumulative conception rates continue to rise after six cycles of treatment, reaching a plateau after 12 cycles. However, prolonged treatment with CC may be associated with an increased risk of ovarian malignancy. Therefore, after completion of 12 cycles of treatment, alternative therapies ought to be considered (National Institute for Health and Clinical Excellence 2004). Factors associated with increased risk of CC treatment failure include increased maternal age, raised BMI and history of severe oligomenorrhoea (Milson et al 2002).

Side-effects

Side-effects with CC treatment are common, but usually mild and transient. Hot flushes and other vasomotor symptoms may occur in up to 10% of patients. Other symptoms include abdominal pain, nausea, vomiting, breast tenderness, headache, insomnia and hair loss. Rarely, visual disturbances (scotomata, blurred or doubled vision) may occur. In the latter case, CC should be discontinued and ophthalmology assessment arranged if symptoms do not cease.

Risks of treatment

With CC therapy, the risk of multiple gestation is approximately 2–13% (Venn and Lumley 1994). OHSS may occur in approximately 13% of cases, whereas severe OHSS is rare. Counselling of women undergoing ovulation induction with CC regarding the above risks is, therefore, essential.

The incidence of spontaneous miscarriage is 13–25% (Milson et al 2002). This figure is no different from the incidence of miscarriage in spontaneous conceptions. In addition to this, there is no evidence to indicate that CC treatment is associated with a higher incidence of congenital abnormalities.

Adjuvant treatment

A systematic review of 12 randomized controlled trials (Beck et al 2005) examined the concomitant administration of CC with other agents to infertile anovulatory women. The use of clomiphene in combination with tamoxifen did not provide any evidence of increased effect on pregnancy rate compared with clomiphene alone. The results were similar when CC plus ketoconazole was compared with CC alone, and CC plus bromocriptine was compared with CC alone.

However, clomiphene plus dexamethasone treatment resulted in a significant improvement in pregnancy rate (fixed OR 11.3, 95% CI 5.3–24.0; NNT 2.7, 95% CI 2.1–3.6) compared with clomiphene alone, as did clomiphene plus pretreatment with combined oral contraceptives (fixed OR 27.2, 95% CI 3.1–235.0; NNT 2.0, 95% CI 1.4–3.4).

Tamoxifen

Tamoxifen is a triphenylethylene derivative structurally similar to CC. It has been widely used in the treatment of breast cancer. In ovulation induction, tamoxifen is administered in doses of 20–80 mg/day, from day 2 to day 6 of the cycle. Tamoxifen, unlike CC, does not exhibit any antioestrogenic effects on the endometrium or vaginal mucosa. Therefore, it was thought that its use might be superior to CC.

Preliminary research has shown that tamoxifen has similar ovulation (44% with tamoxifen vs 45% with CC) and pregnancy rates (22% with tamoxifen vs 15% with CC) (Boostanfar 2001, Gerhard and Runnebaum 1979) compared with CC.

A meta-analysis, involving four trials, compared tamoxifen with CC in anovulatory women and showed that the two drugs were equally effective. Both drugs resulted in similar ovulation rates (OR 0.755, 95% CI 0.513–1.111). No benefit of tamoxifen over CC in achievement of pregnancy per cycle (OR 1.056, 95% CI 0.583–1.912) or per ovulatory cycle (OR 1.162, 95% CI 0.632–2.134) was demonstrated (Steiner et al 2005).

One randomized controlled trial showed that combination therapy of tamoxifen with CC did not improve pregnancy rates [8.6% with tamoxifen/CC vs 4.8% with CC alone; relative risk (RR) 1.80, 95% CI 0.20–16.21] (Suginami 1993).

In most units, tamoxifen is utilized as a second-line therapy in patients who experience side-effects on CC or who appear non-responsive.

Letrozole

Letrozole, an aromatase inhibitor, may be an alternative option for clomiphene-resistant anovulatory women. It has even been suggested that letrozole could be used as a first-line ovulation induction agent instead of CC. However, experience with this medication remains limited, and further research is needed to establish its role, utility and safety.

Pharmacology and mechanism of action

Aromatase inhibitors were initially introduced for the treatment of breast cancer. Aromatase is a cytochrome P-450 haemoprotein that catalyses the rate-limiting step in the production of oestrogens using androgens as a substrate. Letrozole is a highly potent, selective and reversible aromatase inhibitor (Bayar et al 2006). Letrozole inhibits peripheral oestrogen production and, therefore, stimulates endogenous FSH secretion. Moreover, the accumulated androgens may increase follicular sensitivity to FSH (Mitwally and Casper 2004). Its high oral bioavailability and short half-life make it a suitable agent for ovulation induction. One potential advantage of letrozole over CC is its lack of direct antioestrogenic effects on the cervical mucus, the endometrium, uterine blood flow and embryo development (Barroso et al 2006).

Results

In a Cochrane review, Cantineau et al (2007) identified five trials comparing aromatase inhibitors with antioestrogens used as ovulation induction agents in ovarian stimulation protocols for intrauterine insemination (IUI) in subfertile women. No benefits of letrozole use were identified. Other studies involving women with anovulation or unexplained infertility have shown comparable pregnancy rates between letrozole and CC (Barroso et al 2006, Bayar et al 2006, Davar et al 2006, Jee et al 2006). One study reported higher pregnancy rates in PCOS patients treated with letrozole (Atay et al 2006). Results from larger randomized controlled trials are needed to determine the role of letrozole in ovulation induction.

Dopamine Agonists

Pharmacology and mechanism of action

Prolactin secretion from the anterior pituitary is mainly regulated by the inhibitory control of dopamine. Therefore, drugs that exhibit a dopaminomimetic activity reduce prolactin serum concentrations and restore gonadal function in hyperprolactinaemic women, with or without a pituitary adenoma. If a prolactinoma is present, dopamine agonists may also reduce tumour size (see Chapter 16).

The two most commonly used dopamine agonists are bromocriptine and cabergoline. Both medications are ergot alkaloids that mimic dopamine’s actions as they bind to dopamine receptors. Bromocriptine’s serum concentrations peak 1–3 h after oral administration, but very little remains in the circulation after 14 h. Cabergoline is a longer-acting dopamine agonist with a higher affinity for the dopamine receptors than bromocriptine. A single dose of cabergoline inhibits prolactin secretion for 7 days.

A newer dopamine agonist licensed for the treatment of hyperprolactinaemia is quinagolide. Quinagolide is a non-ergot dopamine D₂ agonist. It has a longer biological half-life than bromocriptine.

Indications for treatment

Dopamine agonists are the treatment of choice for infertile women with hyperprolactinaemia. Nevertheless, if ovulation is not achieved despite establishment of euprolactinaemia, adjuvant therapy with CC or gonadotrophins must be considered.

It has been observed that approximately 30% of patients with PCOS show hyperprolactinaemia. In these patients, adjuvant therapy with dopamine agonists may be considered. The pathophysiology behind this remains unclear.

Regimen, monitoring and results

Bromocriptine is the most widely used preparation. Treatment starts at a low dose and is increased gradually. The initial dose should be low in order to minimize gastrointestinal and cardiovascular side-effects. Usually, therapy is initiated with 1.25 mg bromocriptine given at bedtime to suppress the nocturnal prolactin secretion. The dose is gradually increased until the dose required to maintain normal prolactin levels has been established. The usual daily dose is 7.5 mg in divided doses, increased if necessary to a maximum of 30 mg/day. Serum prolactin concentrations must be monitored regularly, and ovulation is monitored with serum progesterone concentrations. Ovulation is usually monofollicular, and ultrasound monitoring of follicle size and numbers is unnecessary if ovulation and regular menses are induced.

Women with a microprolactinoma who are planning to conceive should be prescribed bromocriptine. If pregnancy occurs, bromocriptine may be discontinued as the risk of tumour growth is very small (2%). Nevertheless, in the case of macroprolactinomas, the risk of growth is significant (25%) and treatment with bromocriptine should be continued throughout pregnancy (Balen 2004). Monitoring of these women is done clinically, based on symptoms such as headaches and visual disturbances, and will include visual field assessments and management jointly with an endocrinologist.

Cabergoline treatment is initiated at 500 µg/week, either as a single dose or as two divided doses on separate days. Prolactin levels must be monitored monthly and cabergoline dosage increased, if necessary, at monthly intervals. The usual therapeutic dose is 1 mg/week, ranging from 0.25 to 2 mg/week. The manufacturer advises discontinuation of the drug during pregnancy.

Quinagolide treatment should also be initiated at a low dose. Usually, it is commenced at 25 µg given at bedtime for 3 days. Then, it is increased at 3-day intervals in steps of 25 µg. The usual maintenance dose is 75–150 µg/day. The manufacturer advises discontinuation of the drug during pregnancy, unless medical reasons for continuing arise.

With the use of dopamine agonists, euprolactinaemia is achieved in approximately 60–85% of women. Regular menses are restored in 70–90% of cases and ovulation in 50–75%. If a prolactinoma is present, a decrease in tumour size is achieved in 70% of patients. Pregnancy rates of 30–70% have been reported overall.

Two large randomized controlled trials compared cabergoline with bromocriptine in women with hyperlactinaemic amenorrhoea. Cabergoline was found to be more effective than bromocriptine in achieving euprolactinaemia (83% and 93% with cabergoline vs 59% and 48% with bromocriptine). Moreover, cabergoline was more effective in restoring ovulation and increasing pregnancy rates (72% and 72% with cabergoline and 52% and 48% with bromocriptine) (Webster et al 1994, Pascal-Vigneron et al 1995). Nevertheless, consideration must be given to safety for use in pregnancy (National Institute for Health and Clinical Excellence 2004).

Side-effects

Side-effects of dopamine agonist treatment are common, but are usually mild and transient. Approximately 5% of patients will discontinue treatment for this reason. Side-effects are mainly gastrointestinal and cardiovascular. Gastrointestinal side-effects include nausea, vomiting and constipation. Other side-effects include drowsiness, orthostatic hypotension, headaches and nasal congestion. Side-effects can be minimized if treatment is initiated at a low dose and then gradually increased. Moreover, vaginal administration of the medication may help. The incidence of side-effects is lower with cabergoline than with bromocriptine.

With dopamine agonists, there is no increase in the risk of OHSS, multiple pregnancy or spontaneous miscarriage.

Insulin-Sensitizing Agents

PCOS is characterized by chronic anovulatory infertility and hyperandrogenism with clinical manifestations of oligo- or amenorrhoea, hirsutism and acne. Women with PCOS exhibit an increased prevalence of cardiovascular risk factors, including a higher incidence of obesity, insulin resistance and hyperinsulinaemia (Royal College of Obstetricians and Gynaecologists 2007). Increased insulin concentrations lead to hyperandrogenism and subsequently anovulation. In obese (BMI >30 kg/m²) PCOS women, weight reduction alone may decrease hyperinsulinaemia and hyperandrogenism, and restore ovulation. Therefore, prior to commencing drug treatment, women should be advised to lose weight, as this would improve their chance of spontaneous ovulation and improve their response to ovulation induction.

Metformin

Insulin-sensitizing agents have been tried as ovulation induction drugs in PCOS patients. The most commonly used is metformin, a biguanide oral hypoglycaemic drug used for the treatment of type 2 diabetes mellitus. Metformin has been proven to reduce serum insulin and androgen concentrations, and improve ovulation rates and hirsutism (Lord et al 2003). Metformin has been used as an ovulation induction agent in many trials; nevertheless, these trials are characterized by heterogeneity in terms of dosage, timing and duration of treatment (Al-Inany and Johnson 2006).

A Cochrane review, including women with clomiphene-resistant PCOS and a mean BMI greater than 25 kg/m², concluded that metformin as a single agent did not increase pregnancy rates compared with placebo. Treatment with both metformin and CC did increase clinical pregnancy rates compared with CC alone (OR 4.88, 95% CI 2.46–9.67). Metformin as a single agent was, however, shown to induce ovulation compared with placebo (OR 3.88, 95% CI 2.25–6.69). Moreover, metformin in combination with CC was more effective at inducing ovulation compared with CC alone (OR 4.41, 95% CI 2.37–8.22) (Lord et al 2003).

Two large randomized controlled trials (Moll et al 2006, Legro et al 2007) concluded that metformin should not be used as a primary method for ovulation induction in PCOS patients. Moll et al (2006) compared the effect of CC plus metformin and CC plus placebo on ovulation induction in PCOS women. They concluded that there were no significant differences in ovulation, ongoing pregnancy or miscarriage rates. Nevertheless, a large proportion of women in the metformin group discontinued treatment because of side-effects. Legro et al (2007) compared CC plus placebo, metformin plus placebo and CC plus metformin. The livebirth rate was 22.5% in the CC group, 7.2% in the metformin group and 26.8% in the combination group. Conception rates were lower in the metformin group (21.7%) than in the CC group (39.5%) or the combination group (46%).

Therefore, infertile women with PCOS should not be offered metformin as a first-line agent, and CC remains the drug of choice. Nonetheless, PCOS patients who have not responded to CC and who have a BMI greater than 25 kg/m² should be offered combined metformin and CC treatment (National Institute for Health and Clinical Excellence 2004).

It must be noted that metformin has not been licensed in the UK for use in women who are not diabetic. Therefore, careful counselling of women is mandatory prior to commencing treatment with metformin.

Treatment with metformin is mainly associated with gastrointestinal side-effects, including nausea, vomiting, diarrhoea and abdominal cramps. Lactic acidosis is a rare event. No evidence of fetal toxicity or teratogenicity has been reported.

Other drugs

Other insulin-sensitizing agents tried in the management of anovulatory infertility include the thiazolidinedione hypoglycaemic drugs. Experience with these drugs in PCOS women is limited. Troglitazone was shown to be effective in inducing ovulation. Nevertheless, it was withdrawn from the market after reports of hepatotoxicity. Rosiglitazone and pioglitazone are newer drugs used for the treatment of type 2 diabetes mellitus. Regular monitoring of liver enzymes is recommended, as these drugs may also have hepatotoxic effects. These agents are ‘class C’ drugs with evidence of teratogenicity in animals (Lord et al 2003), and far more data are required in ovulation induction patients to determine their potential role and safety.

Pulsatile Gonadotrophin-Releasing Hormone

Pulsatile GnRH has been used for ovulation induction since 1980. It is the treatment of choice for women with hypogonadotrophic hypogonadism with an intact pituitary function (Braat et al 1991).

Mechanism of action

In patients with hypogonadotrophic hypogonadism, the administration of exogenous pulsatile GnRH restores normal pulsatile pituitary gonadotrophin secretion. Subsequently, follicular recruitment and growth proceed as in a normal menstrual cycle, leading to the development of a single dominant follicle (Martin et al 1993).

Regimen and monitoring

Pulsatile GnRH is administered via a portable mini-pump, either intravenously or subcutaneously. This pump must be worn continuously throughout the treatment. A 2.5–20.0 µg/bolus is administered via the pump at 60–90-min intervals. The treatment starts at a low dose (2.5–5.0 µg/pulse) and is gradually increased until ovulation is achieved. The intravenous route is preferred by some authors as it mimics the physiological spikes of the hormone and, therefore, is more physiological, requires lower doses and is more cost-effective than the subcutaneous route of administration.

Monitoring is performed with regular serum oestradiol measurements and pelvic ultrasound every 3–4 days. The luteal phase is supported either by continuing the same pulsatile GnRH regimen or with the administration of exogenous human chorionic gonadotrophin (hCG).

Results

Pulsatile GnRH is a highly effective method of ovulation induction. Ovulation rates and pregnancy rates per treatment cycle of 70–93% and 18–29% have been reported (Braat et al 1991, Martin et al 1993, Filicori et al 1994).

Nevertheless, ovulation induction with pulsatile GnRH has been less successful in women with PCOS, raised BMI and high serum LH, testosterone or insulin levels (Filicori et al 1994).

Twin pregnancy rates ranged between 3.8% and 13.5% (Braat et al 1991, Martin et al 1993, Filicori et al 1994). This risk is significantly lower than that associated with gonadotrophin treatment (approximately 15%) (Martin et al 1993). Higher order pregnancies are rarely seen with pulsatile GnRH therapy.

The risk of OHSS is very low when exogenous GnRH is used. In comparison with gonadotrophin ovulation induction, GnRH treatment is associated with a lower risk of multiple folliculogenesis and ovarian enlargement (Martin et al 1993).

Spontaneous miscarriage rates have been approximately 10–30%. These appear to be higher in PCOS patients (40%) and lower in women with hypogonadotrophic hypogonadism (20%) (Filicori et al 1994). It is believed that in PCOS women, pretreatment with GnRH analogues may improve outcomes and lower the incidence of multiple pregnancies. Nonetheless, a systematic review by Bayram et al (2004) found no evidence to support the use of pulsatile GnRH in clomiphene-resistant PCOS women.

Side-effects

Overall, exogenous pulsatile GnRH is simple to use, does not require expensive monitoring and is associated with lower risks of multiple pregnancy and OHSS in comparison with gonadotrophin treatment. Nevertheless, many women are reluctant to use this treatment as it requires maintenance of an indwelling cannula for a long period of time; moreover, cases of infection at the site of the cannula have been reported. These seem to be more serious with the intravenous route of administration, whilst the subcutaneous route is mainly associated with local inflammation at the site of needle insertion.

Gonadotrophins

Since the 1960s, exogenous gonadotrophins have been used for ovulation induction in women with WHO Group I and II anovulatory infertility. Gonadotrophins are considered highly effective in inducing ovulation in these groups of patients. Nevertheless, their use has been associated with risks of multiple pregnancy and ovarian hyperstimulation.

Preparations

A number of preparations of exogenous gonadotrophins exist but not all are still available in every country.

Urinary human menopausal gonadotrophin

Human menopausal gonadotrophin (hMG, Menotropin) derives from the urine of postmenopausal women. It contains equal amounts of FSH and LH (75 IU of each), as well as other urinary proteins (>95% of total protein content). The initial preparations were crude as they contained significant amounts of non-gonadotrophin proteins; moreover, they had to be administered intramuscularly. Newer hMG preparations are more purified, contain fewer proteins and can be administered subcutaneously.

Urinary follicle-stimulating hormone

Purified urinary FSH was first introduced in the 1980s (Urofollitropin). It was developed after removing LH by using monoclonal antibody technology (immunoaffinity chromatography). Purified urinary FSH contains 75 IU FSH and less than 1 IU LH. Still-significant amounts of non-gonadotrophin proteins were present and the preparation had to be given intramuscularly. Newer preparations are more highly purified, contain less than 0.001 IU LH and can be given subcutaneously, allowing self-administration.

Recombinant follicle-stimulating hormone

Recombinant FSH was developed by genetic engineering in 1988. Through recombinant DNA techniques, the genes responsible for expression of the α and β subunits of FSH were introduced to the genome of a Chinese hamster ovary cell line. These cells would then produce FSH. The synthesized FSH is purified using chromatography. In comparison with urinary FSH, recombinant FSH has the same structure of side chains, but contains more basic and less acidic isoforms. Its pharmacokinetic characteristics are similar to those of urinary FSH.

Recombinant luteinizing hormone

Recombinant LH was only introduced recently. Its production is based on the same principles of genetic engineering that apply to the synthesis of recombinant FSH. Its pharmacokinetics are similar to those of human pituitary LH.

Urinary human chorionic gonadotrophin

Urinary hCG is extracted from the placenta or the urine of pregnant women. It is used in gonadotrophin-stimulated cycles to be a surrogate for the LH surge and to trigger ovulation once a follicle reaches maturation. The preparation is administered either intramuscularly or subcutaneously.

Recombinant human chorionic gonadotrophin

Recombinant hCG is also available. Studies suggest that 250 µg recombinant hCG is equivalent to 5000–10,000 IU urinary hCG.

Thus far, research indicates that different gonadotrophin preparations are equally effective. When recombinant hCG was compared with urinary hCG, no differences in clinical outcomes (including ongoing pregnancy and livebirth rates, miscarriage and OHSS rates) were demonstrated (Al-Inany et al 2005). When recombinant FSH was compared with urinary hMG, no significant differences in pregnancy rates were observed (Nugent et al 2000).

Nevertheless, recombinant gonadotrophins have certain advantages in comparison with urinary products. They have high batch-to-batch consistency, unlimited supply, high purity and are associated with less risk of allergic reaction as they do not contain any non-gonadotrophin proteins. However, their increased cost is a major disadvantage. Thus, when deciding which preparation to use, factors such as patient safety, cost and drug availability must be taken into consideration (Table 17.1)

Table 17.1 Preparations of gonadotrophins for ovulation induction

Formulation	Proprietary Name	Administered
Urinary
Human menopausal gonadotrophin FSH 75 U and LH 75 U/ampoule	Merional Menopur	IM SC
Urofollitrophin FSH 75 U and LH <1 U/ampoule	Fostimon	IM/SC
Human chorionic gonadotrophin 5000 U/ampoule	Choragon Pregnyl	IM/SC
Recombinant
Follitrophin α FSH 75 U/ampoule	Gonal-F	SC
Follitrophin β FSH 75 U or multiples in cartridges	Puregon	SC
Lutrophin α LH 75 U/ampoule	Luveris	SC
Follitrophin α with lutrophin α FSH 150 U and LH 75 U	Pergoveris	SC
Choriogonadotrophin α 6500 U ≡ 250 µg	Ovitrelle	SC

FSH, follicle-stimulating hormone; LH, luteinizing hormone; IM, intramuscular; SC, subcutaneous.

Indications

Ovulation induction with gonadotrophins is indicated for:

• women with WHO Group I anovulatory infertility (hypogonadotrophic hypogonadism);

• women with WHO Group II anovulatory infertility (including women with PCOS) who failed to ovulate on CC; and

• women with unexplained infertility or ovulatory women undergoing IUI for unexplained or mild male factor infertility.

The choice of therapeutic regimen will depend on the underlying cause of infertility. Treatment must be tailored to the individual’s needs.

Hypogonadotrophic hypogonadism

Women with hypogonadotrophic hypogonadism have extremely low levels of endogenous gonadotrophins. Therefore, both exogenous FSH and LH are required. FSH is essential for follicular growth and oocyte maturation, whereas LH is required for steroidogenesis, luteinization and ovulation. The treatment of choice for this group of patients is hMG, as it contains both FSH and LH (Shoham et al 1991). Nonetheless, the use of recombinant FSH and LH may be an alternative. The regimen used for these women is the ‘step-up protocol’. The luteal phase is usually supported with the administration of hCG, as these patients have very low levels of LH.

Polycystic ovary syndrome

Clomiphene-resistant PCOS women may be candidates for treatment with gonadotrophins. In contrast to women with hypogonadotrophic hypogonadism, PCOS patients have FSH and LH levels which are normal or may be elevated. These patients are at increased risk of multiple pregnancy and OHSS. Therefore, relatively low doses of gonadotrophins are needed for ovulation induction. The protocols used for this category of patients are the ‘chronic low-dose step-up protocol’, the ‘step-down’ protocol and the ‘sequential protocol’. Luteal phase support is not routinely required but, if needed, progesterone support is preferred.

A systematic review of 14 randomized controlled trials found that urinary hMG and urinary FSH had similar effectiveness in terms of pregnancy rates. However, the incidence of OHSS was reduced with FSH compared with hMG. No significant differences were reported between the use of subcutaneous pulsatile and intramuscular injection of gonadotrophins, daily or alternate-day administration, and ‘step-up’ or ‘standard’ regimens (Nugent et al 2000).

A Cochrane review including four randomized controlled trials concluded that when recombinant FSH was compared with urinary FSH, there were no significant differences in terms of ovulation, pregnancy, miscarriage, multiple pregnancy and OHSS rates. No significant differences were demonstrated between administering recombinant FSH as a chronic low dose or as a standard regimen (Bayram et al 2001). The reader is referred to the review by Amer (2007) for further information.

Ovulatory patients undergoing intrauterine insemination

Gonadotrophins have been used for superovulation induction in women with unexplained infertility or ovulatory women with age-related or mild male factor infertility undergoing IUI. In these women, both the step-up and step-down regimens have been applied. The purpose of gonadotrophin treatment is to induce multifollicular development (’superovulation’). Careful monitoring is essential as the risks of OHSS and multiple pregnancy are high.

Regimens and monitoring

Varying regimens of administering exogenous gonadotrophins have been developed (see Figure 17.5).

Figure 17.5 Chronic ‘low-dose step-up’ protocol vs ‘step-down’ protocol for gonadotrophin administration. FSH, follicle-stimulating hormone; hCG, human chorionic gonadotrophin.

Step-up protocol

The step-up protocol is used for women with hypogonadotrophic hypogonadism and for clomiphene-resistant PCOS patients.

Women with WHO Group I anovulation (hypogonadotrophic hypogonadism) are initially given a daily dose of 150 IU hMG. The dose is increased by 75 IU after 4–5 days depending on the response. This is assessed by serum oestradiol concentration and transvaginal ultrasonography. If an ovarian response is confirmed, the gonadotrophin dose is not altered. The frequency of monitoring may be intensified, depending on oestradiol levels and the size or number of the growing follicles. When one or two follicles reach a mean diameter of 18 mm, 5000–10,000 IU hCG is administered to trigger ovulation. Ovulation usually occurs after 36–48 h and couples are advised to have intercourse on the day of the injection and the day after. This regimen is also known as the ‘regular-dose step-up protocol’.

In women with PCOS, the ‘chronic low-dose step-up protocol’ is preferred. Treatment is initiated with a low FSH dose (50–75 IU) continued for a longer period (14 days). Small increments are allowed (25–37.5 IU every 5–7 days) until a dominant follicle emerges.

Step-down protocol

In this regimen, treatment is initiated at higher doses of FSH and then gradually reduced. The purpose of this approach is to mimic the FSH fluctuations during the normal ovulatory cycle. Therapy is commenced at 150 IU/day on day 2 or 3 of the cycle. When a dominant follicle emerges, FSH is reduced by 37.5 IU/day. After 2–3 days, FSH is further decreased to 75 IU/day. This last dose is continued until hCG is administered to trigger ovulation.

Sequential protocol

Treatment is started with a ‘step-up’ protocol, followed by a ‘step-down’ regimen once a dominant follicle (≥14 mm) has emerged.

Concomitant use of gonadotrophin-releasing hormone agonist

Combined treatment with a GnRH agonist and gonadotrophins has been used in PCOS women undergoing IVF. In ovulation induction, pretreatment with a GnRH agonist may theoretically suppress endogenous LH levels. If this treatment is continued during gonadotrophin administration, premature luteinization may be prevented.

In a randomized clinical trial when GnRH agonists were used concomitantly with gonadotrophins, no significant differences were shown in ovulation or pregnancy rates compared with using gonadotrophins alone (Vegetti 1998). Moreover, Nugent et al (2000) found an increase in the incidence of OHSS when GnRH analogues were added to gonadotrophins. Therefore, the use of GnRH agonists is not recommended for PCOS women undergoing ovulation induction with gonadotrophins.

Results

In women with WHO Group I ovulation dysfunction, pregnancy rates of 25% per cycle and cumulative pregnancy rates of 90% after six cycles of treatment have been achieved (Amer 2007). Nevertheless, success rates were lower in PCOS women (5–25% and 30–60%, respectively).

Risks

Treatment with gonadotrophins is associated with an increased risk of multiple pregnancy and OHSS. These complications will be discussed in more detail later in this chapter.

When non-purified preparations are used, their administration may be associated with an increased risk of local reaction (at the injection site) or, on rare occasions, with a severe anaphylactic reaction. In this case, consideration should be given to switch to recombinant gonadotrophin preparations.

Weight Reduction

Obesity has been described as the new worldwide health epidemic. In the UK, obesity affects 24% of the adult female population. As the number of obese women is increasing, so is the prevalence of the disease among those who seek fertility treatment.

Obesity has been associated with a number of reproductive disorders. These include menstrual disorders, infertility, miscarriage and obstetric complications, both fetal and maternal. It appears that it is not only the increased amount of fat per se, but also the fat distribution that is related with these disorders. Thus, an increased waist:hip ratio has a more important effect on fertility than weight alone.

The exact mechanism through which obesity impairs ovarian function, and thus fertility, is largely unknown. It seems that obesity causes low sex-hormone-binding globulin concentrations, hyperandrogenaemia and hyperinsulinaemia. Adipose tissue affects gonadal function via the secretion of adipokines. These include leptin, adiponectin, ghrelin and resistin. The most investigated of all is leptin. Obese women have elevated serum and follicular fluid leptin concentrations. High leptin levels cause a reduction in insulin-induced steroidogenesis in granulosa and theca cells. Leptin also inhibits oestradiol production by the granulosa cells. On the other hand, adiponectin levels decrease in obesity, leading to insulin resistance. Thus, hyperinsulinaemia may be due to the effects of low adiponectin levels and increased resistin levels. The resulting hyperandrogenaemia is caused by the inhibition of sex-hormone-binding globulin and insulin-like growth factor binding protein-1.

As far as fertility is concerned, large retrospective studies have shown a link between obesity and anovulation. Anovulation seems to be the result of hyperandrogenaemia and increased levels of leptin. Obese anovulatory women may or may not have PCOS. It is unknown whether obesity leads to PCOS or vice versa.

The impact of obesity on assisted conception techniques has also been investigated. Regarding ovulation induction, it seems that obese women can be more resistant to CC than non-obese women. As far as IVF is concerned, the evidence from the literature is inconclusive. Although obese women appear to require higher doses of drugs for ovarian stimulation, have a lower chance of pregnancy following IVF and an increased miscarriage rate, it is not clear whether there is an effect of BMI on livebirth rates, cycle cancellation, oocyte recovery or ovarian hyperstimulation incidence (Maheshwari et al 2007).

In this group of patients, weight reduction improves biochemical indices and fertility rates. A 5–10% decrease in body weight will lead to a 30% reduction in body fat, which is sufficient to restore regular menstruation and ovulation. A prospective study was conducted by Clark et al (1995). This study looked at the effect of weight loss, with diet and exercise, on women with anovulation, clomiphene resistance and a BMI greater than 30. Weight reduction led to resumption of ovulation and subsequent pregnancy, as well as a reduction of testosterone levels and increased sex-hormone-binding globulin concentrations.

According to the guidelines of the British Fertility Society, it is advised that women with a BMI greater than 35 kg/m² should not receive fertility investigations or treatment until they reduce their BMI to less than 35 kg/m². Women who are trying to conceive should be advised to maintain a BMI in the range of 20–25 kg/m². Women with a BMI greater than 30 kg/m² should be encouraged to lose weight to a BMI less than 30 kg/m² before receiving fertility therapy either in the form of ovulation induction or as assisted reproduction technology.

Nevertheless, the management of obesity is difficult and requires a multidisciplinary approach. Women should be encouraged to adopt a healthy lifestyle. This would include modification of their dietary quality and change in physical activity. If this fails, pharmacotherapy ought to be considered and, ultimately, obesity surgery. If there is an underlying eating disorder, psychopathology should also be taken into consideration and appropriate referrals made.

Pharmacotherapy includes two main groups of drugs: peripherally acting and centrally acting drugs. The first category mainly infers to orlistat. This medication inhibits gastric and pancreatic lipase and, therefore, reduces fat absorption from the intestine. Orlistat has been shown to decrease testosterone levels in PCOS patients. Although it is generally well tolerated, it can cause gastrointestinal disturbances, which consequently lead to low patient compliance. Orlistat is not licensed for use in pregnancy. The second category of drugs includes sibutramine, which inhibits serotonin and noradrenaline reuptake. It can cause a greater reduction in insulin levels and insulin resistance compared with metformin. Another medication is remonabant, a selective cannabinoid-1 receptor blocker. Both sibutramine and remonabant are contraindicated in pregnancy.

As far as bariatric surgery is concerned, laparoscopic adjustable gastric banding remains the mainstream therapy. Nonetheless, further research is required in order to establish its role in improving fertility. After the procedure, contraception is required until the target weight is reached.

Surgical Management: Laparoscopic Ovarian Drilling

Surgical ovarian wedge resection, introduced in 1939, was the first established treatment for anovulatory PCOS patients. The procedure was performed via laparotomy, and 75% of each ovary was removed. It was speculated that by removing part of the hormone-producing ovarian tissue, androgen and inhibin levels would be reduced. This was followed by an increase in FSH levels and a decrease in LH, leading to spontaneous resumption of ovulation. Nevertheless, the response rate was variable and the procedure was abandoned largely because of the risk of postoperative adhesion formation. As soon as effective medical methods of ovulation became available, ovarian wedge resection became obsolete.

Nonetheless, ovulation induction with CC is not always successful and 20% of women are described as ‘CC resistant’. CC-resistant patients can receive treatment with gonadotrophins, but these are relatively expensive, require intensive monitoring and are associated with increased risk of OHSS and multiple pregnancy. Laparoscopic ovarian drilling (LOD) has therefore been introduced as an alternative therapy for this group of patients.

LOD was first described by Gjonnaess in 1984. Both laparoscopic ovarian cautery and laser vaporization (using carbon dioxide, argon or neodymium:yttrium aluminium garnet lasers) have been used since. The aim is to create approximately 10 holes per ovary in the ovarian surface and stroma. The mechanism of action is thought to be similar to that of ovarian wedge resection. With LOD, androgen-producing ovarian tissue is destroyed. This subsequently causes a decline in the serum levels of androgens, inhibin and LH and an increase in FSH levels. Therefore, disturbances in the ovarian–pituitary feedback mechanism are corrected, leading to follicular recruitment, maturation and ovulation. A long-term cohort study has also shown that up to 20 years after LOD, there was persistence of ovulation and normalization of serum androgens and sex-hormone-binding globulin levels in 60% of the participants (Gjonnaess 1998, Amer et al 2002).

With the employment of LOD, spontaneous ovulation rates of 30–90% and pregnancy rates of 13–88% have been described. In the literature, most data derive from randomized controlled trials comparing ovarian drilling with ovulation induction using exogenous gonadotrophins. There is no evidence of a difference in livebirth outcomes following either LOD (after 6–12 months of follow-up) or three to six cycles of ovulation induction with gonadotrophins in CC-resistant women. Moreover, there was no evidence of a difference in clinical pregnancy rate, miscarriage rate, ovulation rate and quality of life. Nonetheless, multiple pregnancy rates were reduced after LOD (Farquhar et al 2007). Moreover, there appears to be no difference between LOD of one ovary vs drilling of both ovaries in terms of ovulation induction.

In CC-resistant women, LOD may improve clomiphene sensitivity, since when CC or gonadotrophins were added after ovarian drilling, higher pregnancy rates were reported.

Attempts have been made to identify predictors of success in women undergoing LOD, as 43% may not ovulate spontaneously after ovarian drilling. This seems to be related to factors such as the duration of infertility, BMI and free androgen index.

In summary, as there is no evidence of a difference in efficacy between LOD and gonadotrophin ovulation induction, LOD may be the treatment of choice as it is associated with a lower risk of multiple pregnancy and OHSS. Nevertheless, LOD carries risks such as pelvic infection, postoperative adhesion formation, risks of general anaesthesia and the theoretical risk of premature ovarian failure. All these should be taken into consideration before embarking on this procedure. Careful counselling and selection of patients is therefore mandatory. The operation should only be performed by fully trained laparoscopic surgeons. At the moment, the use of LOD is restricted to anovulatory women with a normal BMI. Its use is not recommended as an attempt to decrease the risk of developing diabetes mellitus or coronary artery disease. Long-term risks for women with PCOS (Royal College of Obstetricians and Gynaecologists 2007).

Risks of Induction of Ovulation

Ovarian hyperstimulation syndrome

OHSS is a rare, serious and potentially life-threatening complication of ovulation induction. It is a systemic disease resulting from vasoactive products released from hyperstimulated ovaries. It occurs during the luteal phase of the cycle or during early pregnancy. The incidence of mild OHSS in ovulation induction with CC is 13.5%, whereas moderate and severe forms have only been described sporadically. In IVF/intracytoplasmic sperm injection (ICSI) cycles, the incidence of mild, moderate and severe OHSS may be 33%, 3–6% and 0.1–2%, respectively (Delvigne and Rozenberg 2002). Only a few cases of OHSS have occurred in spontaneous cycles. The true incidence of mortality from OHSS is unknown; nevertheless, deaths are rare.

Pathophysiology

The pathophysiology of OHSS is characterized by an increased capillary permeability, leading to a shift of fluid from the intravascular to the extravascular space, with third-space fluid accumulation and intravascular dehydration. The exact mechanism remains unknown. It is believed that the increased vascular permeability is the result of vasoactive agents released by the hyperstimulated ovaries. Several agents have been proposed, including oestradiol, LH, prolactin, histamine, prostaglandins, prorenin (renin–angiotensin cascade), serotonin, interleukins and vascular endothelial growth factor.

In the initial stages of OHSS, there is an increase in the size of the ovaries and abdominal discomfort. In more severe forms, ovaries become cystic and there is abdominal distention with pain, nausea, vomiting and diarrhoea. This may be followed by ascites, pleural effusion and, on rare occasions, pericardial effusions. There is marked intravascular volume depletion with haemoconcentration that may lead to severe manifestations, including thromboembolism, severe hypoalbuminaemia, hypovolaemia, oliguria and electrolyte disturbances. Cases of acute respiratory distress syndrome have also been reported. Occasionally, liver impairment may occur.

Diagnosis/classification

The diagnosis of OHSS is based on clinical grounds. A history of ovulation induction accompanied by symptoms of abdominal pain, nausea and vomiting should raise the possibility of OHSS. Other conditions, such as ovarian cyst accidents (e.g. torsion, rupture, haemorrhage), pelvic inflammatory disease and ectopic pregnancy should be considered.

Different classification systems of OHSS have been proposed. The latest system was suggested by Marthur in 2005 and is shown in Table 17.2 as adopted by the Royal College of Obstetricians and Gynaecologists. Four categories of OHSS are noted: mild, moderate, severe and critical. As the management of the condition will be dictated by its severity, it is mandatory to assess each case properly and classify it according to Table 17.2.

Table 17.2 Classification of severity of ovarian hyperstimulation syndrome (OHSS)

Grade	Symptoms
Mild	Abdominal bloating Mild abdominal pain Ovarian size usually <8 cm *
Moderate	Moderate abdominal pain Nausea ± vomiting Ultrasound evidence of ascites Ovarian size usually 8–12 cm *
Severe	Clinical ascites (occasionally hydrothorax) Oliguria Haemoconcentration haematocrit >45% Hypoproteinaemia Ovarian size usually >12 cm *
Critical	Tense ascites or large hydrothorax Haematocrit >55% White cell count >25,000/ml Oligo/anuria Thromboembolism Acute respiratory distress syndrome

* Ovarian size may not correlate with severity of OHSS in cases of assisted reproduction because of the effect of follicular aspiration.

Risk factors

Several risk factors of OHSS have been identified in an attempt to prevent this serious iatrogenic complication.

Age

Most studies have reported that women who developed OHSS were significantly younger than those who did not develop OHSS. Women under the age of 30 years are at higher risk. A possible explanation could be that ovaries of younger women are more responsive to gonadotrophins, as they contain a higher number of gonadotrophin receptors and a larger number of follicles responding to gonadotrophins.

Severe OHSS is only rarely associated with clomiphene ovulation induction, whereas moderate forms of the syndrome are encountered in 8% of cycles stimulated with clomiphene. The incidence is much higher when gonadotrophins are used. A study reported that the incidences of mild, moderate and severe OHSS following induction with gonadotrophins are 20%, 6–7% and 1–2%, respectively. GnRH agonists have also been associated with an increase in OHSS incidence. This may be due to inhibition of the spontaneous luteinization process linked with the LH peak.

On the other hand, the use of GnRH antagonists does not increase the risk of OHSS. Insulin-sensitizing agents do not seem to affect its incidence.

Response to ovulation stimulation

High levels of oestradiol, a large number of preovulatory follicles and the presence of large follicles have been associated with an increased risk of OHSS. Other potential predictors of OHSS may include inhibin A and B and vascular endothelial growth factor levels. Further research is required in order to evaluate their positive predictive value and possible future use.

Previous ovarian hyperstimulation syndrome

The risk of developing OHSS is increased in patients who have developed it previously, and altered stimulation protocols need to be utilized.

Prevention

Although several strategies have been developed, OHSS cannot be totally prevented. This mainly applies to late OHSS triggered by the rising serum hCG secreted in early pregnancy. Nevertheless, prevention and early recognition of OHSS are fundamental in ensuring patients’ safety. Identification of any risk factors may reduce the risk of developing OHSS, as adaptations can be made to the ovulation regimen. Careful monitoring of the stimulation with the use of ultrasonography should be implemented, as it has a good predictive value in the occurrence of OHSS (Blankstein et al 1987). For instance, starting with a lower dose of gonadotrophin, reducing the dose when OHSS is suspected or even cancelling the cycle and withholding hCG injection may prevent OHSS. In the latter case, the couple should be advised to refrain from intercourse or to use barrier methods of contraception.

Management

OHSS is a rare disease and many healthcare professionals may be unfamiliar with the syndrome. Protocols for the management of OHSS should therefore be in place. Treatment is predominantly supportive, as OHSS resolves spontaneously. Usually, this takes 7 days in non-pregnant women and 10–20 days in pregnant women.

A physical examination, including body weight, abdominal girth measurements, abdominal and cardiorespiratory examination, and assessment of hydration, must be performed. A pelvic ultrasound will help to assess the size of the ovaries and check for ascites. Blood tests should include full blood count, urea and electrolytes, liver function tests and clotting screen. In case of dyspnoea, a chest X-ray or chest ultrasonography should be requested. If pericardial effusion is suspected, an electrocardiogram and echocardiogram will be essential.

Outpatient management

Women with mild OHSS and many with moderate OHSS may be managed as outpatients. Paracetamol and codeine are the analgesics of choice. Non-steroidal anti-inflammatory drugs are contraindicated, as they may compromise renal function. Women should be advised to drink to thirst and not to excess. A review is required every 2–3 days. Nevertheless, urgent reassessment is warranted if the woman develops severe abdominal pain, increasing abdominal distention, dyspnoea or has a subjective impression of decreasing urine output (Royal College of Obstetricians and Gynaecologists 2006).

Inpatient management

Women with severe OHSS and those with moderate OHSS whose symptoms cannot be controlled with oral medication should be admitted to hospital. Critical cases of OHSS may require admission to an intensive care unit for invasive monitoring (see Figure 17.6).

Figure 17.6 Vaginal scan of a patient with moderate ovarian hyperstimulation syndrome, showing enlarged ovary, luteal cysts and ascitic fluid.

Monitoring of these patients should include:

• heart rate and blood pressure;

• abdominal girth and weight (measured daily);

• intake and output chart;

• daily haemoglobin, haematocrit and platelet count; and

• daily serum creatinine and electrolytes.

Measures of worsening OHSS include: increasing abdominal pain, oliguria (in particular, a persistent positive fluid balance or daily urine output <1000 ml/day), weight gain, increased girth circumference, worsening dyspnoea and haemoconcentration (as measured by raised haemoglobin, haematocrit or white cell count).

If pain is not adequately controlled with paracetamol, oral or parenteral opiates ought to be considered. Nausea is usually caused by the presence of ascites, and measures taken to reduce ascites also contribute to relief from nausea. The recommended antiemetics are those appropriate for early pregnancy, such as metoclopramide, cyclizine and prochlorperazine. The intravascular volume is maintained by encouraging women to drink to thirst. In cases of severe vomiting, intravenous crystalloids, such as normal saline, must be administered, aiming for a fluid intake of 2–3 l/day. In the case of haemoconcentration, more intensive hydration is needed; if the problem persists, colloid therapy needs to be considered. Colloids that have been used for this purpose include human albumin, dextran, Haemaccel, 6% hydroxyethylstarch and mannitol. Nevertheless, there is no evidence to recommend any specific fluid regimen. Diuretics should be avoided for fear of intravascular volume depletion. In cases of increasing abdominal distention or persistent oliguria, an ultrasound-guided paracentesis should be considered. Drainage of ascitic fluid provides rapid symptomatic relief from pain and dyspnoea, and is followed by a significant improvement in urine output. Intravenous colloid replacement should be considered in order to avoid cardiovascular collapse. Thromboprophylaxis should be given to all women with OHSS who are admitted. This will include full-length venous support stockings and subcutaneous prophylactic heparin. Surgical treatment is reserved for cases of suspected adnexal torsion.

Counselling

Women who undergo ovulation induction should be counselled carefully about the risks and possible symptoms of OHSS. A 24-h contact number and access to a clinician should be offered to these women.

In the case of OHSS, further counselling is needed to provide information and reassurance. For an overview of management of OHSS, the reader is referred to the guidelines of the Royal College of Obstetricians and Gynaecologists (2006).

Multiple pregnancy

In many developed countries, there has been a 30–50% increase in twin births since 1980. Triplet deliveries have increased three- to five-fold during this period. The incidence of multiple births in the UK has risen from 10–15/100,000 maternities before 1981 to 12.1/1000 maternities in 1991 and 14.9/1000 maternities in 2004 (Office for National Statistics 2006). The number of multiple pregnancies conceived is even higher, as spontaneous and iatrogenic fetoreductions are not included in birth statistics.

Multiple pregnancies are high risk as they carry risks for both the mother and the fetus. They are often complicated by preterm delivery, intrauterine growth restriction or a small-for-gestational-age fetus and their sequelae, pre-eclampsia and eclampsia, gestational diabetes, antepartum haemorrhage and assisted delivery. Multiple pregnancies are associated with high infant mortality and morbidity, and carry major long-term consequences for childhood and adult life, especially in terms of neurodevelopmental impairments. Moreover, twin and higher order pregnancies carry significant socioeconomic and psychological consequences for family life.

Approximately 20% of the increase in multiple births can be attributed to advanced maternal age, as it is known that older women are more likely to have a multiple pregnancy. The remainder is associated with ovulation induction and assisted reproductive technologies, such as IVF and ICSI. The exact numbers of higher order pregnancies resulting from ovarian stimulation, with or without IUI, are unknown as there are no national registers that record the outcome of controlled ovarian simulation (ESHRE Task Force on Ethics and Law 2003).

Nonetheless, ovarian simulation, with or without IUI, contributes significantly to the occurrence of multiple births. A study by Lynch et al (2001) showed that 20% of all multiple pregnancies were attributable to ovulation induction agents, which is considerably higher than the 14% attributable to IVF. Similarly, 34% of higher order pregnancies in the UK in 1989 resulted from ovulation induction (Levene et al 1992). Multiple pregnancies occurred in 2–13% of women with all causes of infertility taking CC, whereas women with clomiphene-resistant PCOS treated with gonadotrophins have a multiple pregnancy rate of 36%.

In view of these facts, it is mandatory to design prevention strategies to avoid multiple gestations. In the case of ovarian stimulation, judicious use of ovulation induction drugs and monitoring of the follicular size and number with ultrasonography are essential. In addition to this, ovulation induction ought to be carried out in specialist clinics so that careful monitoring is achieved. Furthermore, specific interventions may reduce the risk of multiple pregnancy. These include withholding the hCG injection and cancelling the cycle, selective aspiration of supernumerary follicles and conversion to IVF cycles. Women who undergo ovulation induction must be counselled regarding the risk of multiple pregnancy and its consequences.

In the event of multiple pregnancy, selective fetal reduction ought to be considered. Multifetal pregnancy reduction (MFPR) refers to ‘the termination of one or more normal fetuses in a multifetal pregnancy in order to improve survival rates for the remaining fetuses and decrease maternal morbidity’. For higher order pregnancies, not performing a reduction will increase the risk of losing the pregnancy.

Three different techniques of MFPR have been described. The most commonly used is the transabdominal needle insertion of potassium chloride to the fetal thorax above the diaphragm. This procedure is usually performed at 10–12 weeks of gestation. The selection of the embryo is based on which one is easiest to reach. The second method is the transcervical mini-suction, done between 8 and 11 weeks. Loss of the entire pregnancy has been reported in 50% of cases managed with this technique. The third method consists of transvaginal aspiration of the embryo, usually done at 10–12 weeks (Evans et al 1998).

It is widely accepted that for any higher order multiple pregnancy, reduction to twins has the highest survival rate. Reduction to a singleton pregnancy is not accepted because of the risk of losing that pregnancy if there is a problem later. Women with significant medical disease, such as cardiac disease, or uterine malformations are exceptions to this rule, and reduction to a singleton pregnancy may be considered in such cases (Evans et al 1998).

Prevention of multiple pregnancies should be preferred to MFPR, since it is associated with the ethical dilemmas of abortion. Moreover, it is acknowledged that the original higher order pregnancy may have detrimental effects on the development of the remaining fetuses, in terms of risk of preterm delivery, even after the event of fetocide (ESHRE Task Force on Ethics and Law 2003).

Ovarian cancer

Ovarian cancer is the fourth most common cancer among women in England and Wales, and is the most common cause of gynaecological cancer death. The possibility of a link between ovulation and ovarian oncogenesis led to concerns that fertility treatments may increase the risk of developing ovarian cancer. In particular, the possible association between drugs used for ovulation induction and the risk of ovarian cancer has been the subject of much debate. Concerns have been raised about the effects of multiple ovulations and trauma to the ovarian epithelium (Fathalla 1971), as well as the high levels of gonadotrophins reached during fertility treatment. It has been speculated that the latter may lead to the production of intraovarian carcinogens and malignant transformation (Daly 1992).

A causal relationship between fertility treatment and ovarian cancer was supported by some anecdotal case reports and epidemiological studies. However, other studies showed conflicting results.

A collaborative analysis on data from 12 case–control studies of ovarian cancer conducted in the USA (Whittemore et al 1992) showed that the risk was increased among women who had used fertility drugs (OR 2.8, 95% CI 1.3–6.1) compared with women without a clinical history of infertility. On the other hand, infertile women who had not used fertility drugs experienced no increase in risk (OR 0.91, 95% CI 0.66–1.3). The above risk was higher among nulligravid women than among gravid women. Nevertheless, the information available on specific fertility medication was too incomplete to draw any conclusions.

A case–control study found that women taking clomiphene had a higher risk of developing the disease compared with women who were not taking clomiphene (RR 2.3, 95% CI 0.5–11.4). Prolonged use of clomiphene (12 months or more) was associated with a higher risk of ovarian cancer (RR 11.1, 95% CI 1.5–82.3). Nonetheless, treatment with the drug for less than 1 year was not associated with increased risk (Rossing et al 1994).

On the other hand, several reviews have reported insufficient evidence to support a direct causal relationship between fertility drug treatment and ovarian cancer (Nugent et al 1998, Klip et al 2000).

A UK epidemiological report of cancer incidence among women who had ovarian stimulation treatment found no evidence of a link between infertility treatment and ovarian cancer. This cohort study included 5556 women of whom 75% received ovulation induction drug treatment. The incidence rates of ovarian carcinoma were not significantly different from expectation based on national cancer rates (Doyle et al 2002).

It is acknowledged that nulliparous women who have not received any fertility treatment have almost double the risk of ovarian malignancy. Therefore, the association between nulliparity, infertility and ovarian cancer needs to be considered, as infertility appears to be an independent risk factor for ovarian malignancy. This link may be the reason behind the conflicting results of research.

In view of the above controversial evidence, women who receive ovulation induction therapy should be monitored closely and the number of treatment cycles should be shortened. It is accepted that ovulation induction with clomiphene is not associated with increased risk of ovarian malignancy, provided it is not used for more than 12 cycles. Its administration should be confined to the lowest effective dose. Gonadotrophins should be used for the lowest number of cycles and at the lowest effective doses possible.

Women who undergo fertility treatment ought to be counselled regarding the possible association between medical ovulation induction and ovarian cancer, and their informed consent should be sought (National Institute for Health and Clinical Excellence 2004). A survey of women attending a fertility clinic reported that the majority (67%) of women were aware of the potential link between fertility treatment and ovarian cancer. Moreover, 79% of participants were willing to accept this potential increased risk (Rosen et al 1997).