Poly cystic ovary syndrome: the spectrum of the disorder in 1741 patients

Transcription

1 Human Reproduction vol.10 no.8 pp.21o7-2111, 1995 Poly cystic ovary syndrome: the spectrum of the disorder in 1741 patients Adam H. Balen 1, Gerry S.Conway, Gregory Kaltsas, Kitirak Techatraisak, Patrick J.Manning, Christine West and Howard S.Jacobs Department of Endocrinology, University College London Medical School, The Middlesex Hospital, Mortimer Street, London WIN 8AA, UK 'To whom correspondence should be addressed at: Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK. The criteria for the diagnosis of the polycystic ovary syndrome (PCOS) have still not been agreed universally. A population of 1741 women with PCOS were studied, all of whom had polycystic ovaries seen by ultrasound scan. The frequency distributions of the serum concentrations of follicle stimulating hormone (FSH), luteinizing hormone (LH), testosterone and prolactin and the body mass index, ovarian volume, uterine cross-sectional area and endometrial thickness were determined and compared with the symptoms and signs of PCOS. Obesity was associated with hirsutism and an elevated serum testosterone concentration and was also correlated with increased rates of infertility and cycle disturbance. The rates of infertility and cycle disturbance also increased with serum LH concentrations >10 IU/1. A rising serum concentration of testosterone [mean and 95th percentiles 2.6 ( ) nmol/1] was associated with an increased risk of hirsutism, infertility and cycle disturbance. The ovarian volume was correlated with serum concentrations of testosterone, LH and the body mass index, which was also correlated with the uterine area. This descriptive data from the largest reported series of women with PCOS enables the development of a management-orientated approach to the syndrome. Women who are overweight can expect an improvement in their symptoms if they lose weight. An elevated concentration of LH (>10 IU/1) is associated with infertility and treatment should be chosen accordingly. If the serum testosterone concentration is >4.8 nmol/1, other causes of hyperandrogenism should be excluded. Key words: hyperandrogenism/infertility/luteinizing hormone/ obesity/polycystic ovary syndrome Introduction Controversy surrounds the diagnosis of polycystic ovary syndrome (PCOS), with different authors using various criteria to define the syndrome. In the UK it is now generally accepted that polycystic ovaries detected by ultrasound scan provide the unifying diagnostic criterion. The additional symptoms (oligo/amenorrhoea, obesity, hyperandrogenism) and biochemical disturbances [elevated serum concentrations of luteinizing hormone (LH), androgens and insulin] may each occur together or in isolation with the ultrasound picture and hence result in PCOS. An initial report of 556 patients who attended the endocrine clinic at the Middlesex Hospital, London described the heterogeneity of women with PCOS compared with a control group, clarified the significance of three endocrine patterns (raised concentrations of LH, testosterone and prolactin) and identified five clinical subgroups (hirsutism, infertility, obesity, alopecia and acanthosis nigricans) (Conway et al, 1989). We now extend that study and present the features of what we believe to be the largest series of patients with PCOS, all of whom had ultrasound-detected polycystic ovaries and symptoms of the syndrome, in order both to act as a reference for the spectrum of this disorder and to highlight the features of ultrasound morphology with endocrine parameters. Materials and methods All women who were referred to the reproductive endocrine clinics at the Middlesex Hospital were requested to have a pelvic ultrasound scan and a hormone profile performed before they attended for their first consultation. Ours is a tertiary referral specialist endocrine clinic and it is our routine clinical practice to obtain as much information as possible about all patients, irrespective of their diagnosis, prior to their first attendance. The indications for referral were menstrual cycle disturbances, infertility and symptoms of hyperandrogenism. Over the time of the study it was normal practice to perform the initial baseline ultrasound scan transabdominally. Ovarian morphology and volume were recorded together with uterine cross-sectional area and endometrial thickness. The criteria of 5=10, 2-8 mm in diameter, arranged around an echodense central stroma, as described by Adams et al. (1985) was used to diagnose the presence of polycystic ovaries. The ovarian volume was calculated as 4/3 Jt(l/2 diameter) 3, where the diameter was taken as the mean of the height, width and depth of the ovary, in the absence of a dominant follicle. The uterine crosssectional area was defined as the product of the maximum length and the antero-posterior diameter of the uterus, as measured in the sagittal plane (Poison et al, 1987). A full clinical history was taken and the patients were examined. The clinical data included age, body mass index (BMI, weight/ height 2 ) and the presence of acne and hirsutism, which was defined using the Ferriman and Gallway (1961) score. Acanthosis nigricans was diagnosed by clinical appearance. The menstrual cycle was described as being either regular, oligomenorrhoeic (a cycle interval of longer than 35 days but less than 6 months) or amenorrhoeic (no menstruation for more than 6 months). The fertility status was classified as 'proven fertile' (those with a previous pregnancy and no I Oxford University Press 2107

2 A.H.Balen et al. Table I. Characteristics of 1741 women with ultrasound-detected polycystic ovaries. Mean and 5-95th percentiles b. i 250 Age (years) Ovarian volume (cm 3 ) Uterine cross-sectional area (cm 2 ) Endometrium (mm) BMI (kg/m 2 ; 19-25) a FSH (IU/1; l-10) a LH (IU/1; l-10) a Testosterone (nmol/1; O.5-2.5) a Prolactin (< 350 mu/l) a 31.5 (14-50) 11.7 ( ) 27.5 ( ) 7.5 ( ) 25.4( ) 4.5 ( ) 10.9 ( ) 2.6( ) 342 (87-917) a Normal range. BMI = body mass index; FSH = follicle stimulating hormone; LH = luteinizing hormone. subsequent infertility), 'fertility untested' (those who had never tried to conceive) or 'primary/secondary' infertility of at least 1 year's duration. Serum was collected for measurement of LH, follicle stimulating hormone (FSH), testosterone, prolactin and thyroid function. Sex hormone-binding globulin was not measured routinely due to financial constraints on the laboratory. The details of the radioimmunoassays that were used are described in detail by Conway et al. (1989). In particular, the Chelsea radioimmunoassay kit for measurement of LH (which employs polyclonal antiserum F87 and NIBSC 68/40 International Reference Preparation) was used and a value of 10 IU/1 was two standard deviations above the mean for normal women in the follicular phase. The hormone measurements were usually performed in the follicular phase of the menstrual cycle and were excluded from the analysis if measurements of FSH were elevated (>10 IU/1), which suggested the presence of either a midcycle surge or a peri-menopausal state. Statistics The variables BMI, serum LH and prolactin concentrations and ovarian volume were log-transformed before parametric analysis, and geometric means are presented for these variables. Group means were compared by analysis of variance with Duncan's procedure for multiple comparisons. Grouped variables (BMI and LH) and discrete data were tested with ~f} and associations between continuous variables were sought with Pearson's correlation coefficients. Multiple regression analysis was performed using the stepwise method with a P <0.05 threshold for inclusion. Results A total of 1871 women who attended the clinic were identified as having polycystic ovaries together with symptoms and signs of PCOS. A total of 130 patients (6.9%) were excluded from the analysis because they were additionally found to be menopausal (n = 13) or also had weight-related amenorrhoea («= 46), pituitary disease (n = 27), a prolactinoma (n = 25) or congenital adrenal hyperplasia (n = 19). There were no patients with abnormal thyroid function tests or androgensecreting tumours. The characteristics of the remaining 1741 patients are recorded in Table I and Figures 1 and 2. A complete set of data was, inevitably, not available for every patient. In some cases the clinician failed to record a specific piece of information (for example the presence/absence of acanthosis nigricans) or there was insufficient blood taken to perform a full baseline endocrine profile. Whilst at least one 2108 C B FSH(nVl) ll III.!... testosterone (nmol/l) 1 Illlllh, IB i LHOufl) d. prolac&n (mu/l) Figure 1. Frequency distribution of the characteristics of 1741 women with polycystic ovary syndrome: serum concentrations of (a) follicle stimulating hormone (FSH), (b) luteinizing hormone (LH), (c) testosterone and (d) prolactin. ovary was visualized in each patient and a morphological assessment made, it was not always possible to measure accurately all of the ovarian diameters or uterine parameters. For the clinical and endocrine variables, a mean of 21.4% (range ) of data points were missing and for the ultrasound data 32.8% (range ). Of patients examined, 38.4% were overweight (BMI >25 kg/m 2 ), 39.8% had an elevated serum concentration of LH (> 10 IU/1) and 28.9% had an elevated serum testosterone concentration (>2.5 nmol/1). With respect to menstrual history, 47.0% had oligomenorrhoea, 29.7% had a normal menstrual cycle, 19.2% had amenorrhoea, 2.7% had polymenorrhoea and 1.4% had menorrhagia; 66.2% of the patients had mild (20.6%), moderate (40.7%) and severe (4.9%) hirsutism; 34.7% of patients had acne and 2.5% had acanthosis nigricans. Using Pearson's correlation coefficient the patients' BMI was significantly correlated with ovarian volume (r = 0.11; P < ) and uterine cross-sectional area (r = 0.15, P < ). A higher BMI was associated with a rise in serum testosterone concentration (r = 0.254, P < ) and the prevalence of hirsutism (f = ) (Figure 3). Obesity was also associated with an increased rate of infertility and cycle disturbances: the rates of primary (15%) and secondary (8%) infertility were fairly constant with BMIs of kg/m 2

3 PCOS - spectrum of the disorder a. 350 b nmol/l lift..;.".;: IS IB Z BM1.1 II B S ovarian volume (ml) c, 200 ' d. 350,1.ll otcnne area (cm?) 1 I. III!... cndomelrium (mm) Figure 2. Frequency distribution of the characteristics of 1741 women with polycystic ovary syndrome: (a) body mass index (BMI) (kg/m 2 ), (b) ovarian volume, (c) uterine cross-sectional area (UXA) and (d) endometrial thickness (mm). but rose to 26 and 14% respectively, when the BMI was >30 kg/m 2 (P < ). Similarly the percentage of women with a regular menstrual cycle fell from - 32% to 22% when the BMI rose above 30 kg/m 2 (P = 0.032). With respect to fertility, 804 of a total of 1269 (63.4%) women had not yet tested their fertility. Of the remaining 465 women, 228 (49%) had primary infertility, 121 (26%) had secondary infertility and 116 (25%) had proven fertility. The LH concentrations related to fertility are recorded in Table II. The serum LH concentration of those with primary infertility was significantly higher than that of women with secondary infertility and both were higher than the LH concentration of those with proven infertility (P < ). The rate of infertility increased as the serum LH concentration rose (Figure 4, x , DF 12, P < ). There was also a significant increase in the rate of cycle disturbance with increasing LH concentrations (Figure 4, x , DF 24, P < ). The ovarian volume was significantly correlated with serum concentrations of LH (r = 0.24, P < ), testosterone (/- = 0.16, P < ) and the BMI (r = 0.11, P < ). These variables were independent of each other when tested with multiple regression analysis. A rising serum concentration of testosterone was associated with an increased risk of BMI -+-infertility hirsutism -x-testosterone Figure 3. The relationship between body mass index (BMI) and the rates of hirsutism and serum testosterone concentration. Table II. Serum luteinizing hormone (LH) concentrations (IU/1) with respect to fertility status Proven fertility Untested fertility Primary infertility Secondary infertility 7.2 ± ± ± 2.2" 9.0 ± 2.0 b a Different from proven fertile and secondary infertile groups. b Different from proven fertile group. hirsutism {P < ), infertility (P = ) and cycle disturbance (P = ). Discussion Since the advent of high resolution ultrasound scanning an accurate estimate of the prevalence of polycystic ovaries has been made possible. Ovarian morphology appears to be the most sensitive marker for PCOS compared with the classical endocrine features of a raised serum LH and/or testosterone concentration which were found in only 39.8 and 28.9% of our patients respectively, whilst the symptoms of obesity, hyperandrogenism and menstrual cycle disturbances occurred in 38.4, 70.3 and 66.2% of patients respectively. We have therefore preferred to make the diagnosis of the PCOS when there are, in addition to the ultrasound finding of polycystic ovaries, the associated symptoms (menstrual irregularity, hyper-androgenization, obesity) or endocrine abnormalities (raised serum LH and testosterone concentrations) (Jacobs, 1987). Most European clinicians now use ultrasonography to make the diagnosis of polycystic ovaries. In the USA there are, however, those who consider that the polycystic ovary 2109

4 Figure 4. The relationship between serum luteinizing hormone (LH) concentration and the rates of infertility and cycle disturbance. syndrome can occur in the absence of polycystic ovaries (Kim et al, 1979). Swanson et al. (1981) were the first to provide an ultrasound description of polycystic ovaries and we use a modification of this, with a requirement of 5 s 10 cysts in a single plane, as defined by Adams et al. (1985) (see above). With the ability of transvaginal ultrasonography to provide higher resolution and a clearer picture of the ovary than a transabdominal scan, some groups have chosen to define the polycystic ovary as an ovary that contains at least 15 cysts and usually more than 20 (Fox et al., 1991). Furthermore, Fox et al. (1991) state that transabdominal ultrasound failed to detect 30% of polycystic ovaries compared with an almost 100% detection rate with a transvaginal scan. Three-dimensional ultrasonography might soon provide a more precise definition of the polycystic ovary. It has also been suggested that magnetic resonance imaging might enhance further the ability to detect polycystic ovaries (Faure et al., 1989), although this costly procedure is unlikely to replace the quick and simple ultrasound scan. Several studies have tried to estimate the prevalence of polycystic ovaries in 'normal women' and have all found rates of ~ 22% (Poison et al, 1988; Tayob et al, 1990; Gadir et al., 1992; Clayton et al, 1992; Farquhar et al, 1994). Each of these studies is potentially flawed by selection bias (Jacobs, 1994) and a truly representative population-based study of the prevalence of ultrasound detected polycystic ovaries is still awaited. The polycystic ovary, at any rate, appears to have a significant prevalence in the normal population. The original descriptive triad of amenorrhoea, obesity and hirsutism (Stein and Leventhal, 1935), however, appears to be the extreme end of the spectrum of the disorder. Indeed, many women with polycystic ovaries detected by ultrasound do not have symptoms of PCOS, although symptoms may develop later, after a gain in weight, for example When polycystic ovaries are detected by ultrasound there are a variety of possible associated clinical and endocrine disturbances and a number of studies have been performed to determine which diagnostic criteria can best enhance the accuracy of diagnosing the syndrome. Fox et al. (1991) found that isolated measurements of serum concentrations of androgens, oestradiol, gonadotrophins and the LH:FSH ratio provided a diagnostic accuracy of <75% in women with polycystic ovaries and oligo/amenorrhoea, whilst a positive progestogen challenge test and the free androgen index provided diagnostic accuracies of 89 and 94% respectively. Robinson et al (1992) described the endocrine profiles of 63 women with clinical and ultrasound features of PCOS and found that serum testosterone, androstenedione or LH concentrations were elevated either alone, or in combination, in 86% of women and the LH:FSH ratio was of little value. Sampling LH every 20 min over a 6 h period gives a variability of 38% in the follicular phase and 92% in the luteal phase of normal women (Santen and Bardin, 1973), so single hormone estimations are unreliable in representing the hormone status of a subject. In a study of 40 women with ultrasonically detected polycystic ovaries, oligo/amenorrhoea, hirsutism and at least one endocrine abnormality (testosterone >3.0 nmol/1, LH >10 IU/1, LH:FSH >2.0), Gadir et al (1991) performed 15 min sampling over 6 h in the follicular phase of the cycle. Apart from a correlation between the degree of hirsutism and the serum testosterone concentrations, there was no correlation between the ovarian volume, BMI, duration of symptoms, age and the endocrine measurements (FSH, LH, oestradiol, testosterone, insulin). It was suggested that ovarian size does not indicate the severity or duration of the syndrome. The literature contains studies which both support (Berger et al, 1975) and refute (Givens et al, 1986; Franks, 1989) an association between ovarian size and serum LH concentrations. We have found that the serum concentrations of LH and testosterone did correlate significantly with ovarian volume. The important questions to ask are: what is the relevance of these correlations and which tests are most useful in assessing women with PCOS? The body mass index correlated with an increased rate of hirsutism (and serum testosterone concentration), cycle disturbance and infertility. Obese women (BMI >30 kg/m 2 ) should therefore be encouraged to lose weight. We have found that weight loss improves the symptoms of PCOS and others have shown that diet improves the endocrine profile of women with PCOS (Kiddy et al, 1989). It is thought that obese women with PCOS hypersecrete insulin, which stimulates ovarian secretion of androgens, leading to hirsutism, menstrual disturbance and infertility (Conway and Jacobs, 1993). It is seldom necessary to measure the serum insulin concentration, as this will not affect the management of the patient. We have found, however, that the prevalence of diabetes in obese women with PCOS is 11% (Conway et al, 1992) and so assessment of glucose tolerance is important in these women. Hypersecretion of LH is particularly associated with menstrual disturbances and infertility. Indeed, it is this endocrine feature that results in reduced conception rates and increased rates of miscarriage in both natural and assisted conception

5 PCOS - spectrum of the disorder (recently reviewed, Balen et al., 1993). The finding of a persistently elevated early to mid-follicular phase LH concentration (>10 IU/1 with our assay) in a woman who is trying to conceive indicates the need to suppress LH levels by either pituitary desensitization with a gonadotrophin-releasing hormone agonist, or laparoscopic ovarian diathermy (Balen et al., 1993). However, prospective randomized studies still have to be performed of large numbers of women in whom serum LH concentrations have been suppressed. Whilst an elevated serum testosterone concentration is associated with an increased rate of hirsutism, and a higher BMI, ovarian volume and serum LH concentration, it is important to exclude other causes of hyperandrogenism. If the testosterone concentration is > 4.8 nmol/1 the patient should be investigated further to exclude androgen-secreting tumours of the ovary or adrenal gland, Cushing's syndrome and congenital adrenal hyperplasia. Women with type A insulin resistance secondary to mutations in the insulin receptor gene may have serum testosterone concentrations in the male range but are also likely to have acanthosis nigricans (Conway and Jacobs, 1993). These patients will respond to gonadotrophin suppression, for example using gonadotrophin-releasing hormone agonists, with a fall in testosterone concentration. An ultrasound assessment of endometrial thickness provides a bioassay for serum oestradiol concentration and is more relevant than the measurement of oestradiol, which can fluctuate considerably. An endometrial thickness of >5 mm indicates oestrogenization and negates the necessity to perform a progestogen withdrawal test, which is both time consuming and insensitive (Shulman et al., 1989). If the endometrium is thicker than 15 mm a withdrawal bleed should be induced, and if the endometrium fails to be shed then endometrial sampling is required to exclude endometrial hyperplasia. We have described the clinical and endocrine spectrum in a large population of women with PCOS. There is inevitably an element of referral bias in a group of patients who are seen in an endocrine clinic; a gynaecological clinic, for example, would have more patients referred with subfertility. Nonetheless, with these data we have developed a management approach to the condition: women who are overweight should be encouraged to lose weight and can then be expected to experience an improvement in the symptoms of menstrual disturbance, infertility and hyperandrogenism. A raised serum concentration of testosterone should be investigated to exclude other causes of hyperandrogenaemia. If the serum LH concentration is elevated, there is a strong association with infertility, and treatment strategies include pituitary desensitization and laparoscopic ovarian diathermy (see Balen et al., 1993). References Adams, J., Franks, S., Poison, D.W., Mason, H.D., Abdulwahid, N., Tucker, M., Morris, D.V., Price, J. and Jacobs H.S. (1985) Multifollicular ovaries: clinical and endocrine features and response to pulsatile gonadotrophinreleasing hormone. Lancet, ii, Adams, J., Poison, D.W. and Franks, S. (1986) Prevalence of polycystic ovaries in women with anovulation and idiopathic hirsutism. Br. Med. J., 293, Balen, A.R. Tan, S.L. and Jacobs, H.S. (1993) Hypersecretion of luteinising hormone: a significant cause of infertility and miscarriage. Br. J. Obslel. CynaecoL, 100, Berger, M.J., Taymor, M.L. and Patton, W.C. (1975) Gonadotropin levels and secretory patterns in patients with typical and atypical polycystic ovarian disease. Fertil. Sierii. 26, Clayton, R.N., Ogden, V., Hodgkinson, J., Worswick, L., Rodin, D.A., Dyer, S. and Meade, T.W. (1992) How common are polycystic ovaries in normal women and what is their significance for the fertility of the population? Clin. Endocrinol.. 37, Conway, G.S. and Jacobs, H.S. (1993) Clinical implications of hyperinsulinaemia in women. Clin. Endocrinol., 39, Conway, G.S., Honour, J.W. and Jacobs, H.S. (1989) Heterogeneity of the polycystic ovary syndrome: clinical, endocrine and ultrasound features in 556 patients. Clin. Endocrinol, 30, Conway, G.S., Agrawal, R., Betteridge, D.J. and Jacobs, H.S. (1992) Risk factors for coronary artery disease in lean and obese women with the polycystic ovary syndrome. Clin. Endocrinol, 37, Eden, J.A. (1989) The polycystic ovary syndrome. Aust. N.Z. J. Obslel. CynaecoL 29, Farquhar, CM., Birdsall, M., Manning, P. and Mitchell, J.M. (1994) Transabdominal versus transvaginal ultrasound in the diagnosis of polycystic ovaries on ultrasound scanning in a population of randomly selected women. Ultrasound Obstet. CynaecoL 4, Faure, N., Prat, X. Bastide, A. and Lemay, A. (1989) Assessment of ovaries by magnetic resonance imaging in patients with polycystic ovarian syndrome. Hum. Reprod., 4, Ferriman, D. and Gallway, J.D. (1961) Clinical assessment of body hair growth in women. J. Clin. Endocrinol. Metab., 21, Fox, R., Corrigan, E., Thomas, P.A. and Hull, M.G.R. (1991) The diagnosis of polycystic ovaries in women with oligo-amenorrhoea: predictive power of endocrine tests. Clin. Endocrinol, 34, Franks, S. (1989) Polycystic ovary syndrome: A changing perspective. Clin. Endocrinol, 31, Gadir, A.A., Khatim, M.S., Mowafi, R.S., Alnaser, H.M.I., Alzaid, H.G.N. and Shaw, R.W. (1991) Polycystic ovaries: do these represent a specific endocrinopathy. Br. J. Obstet. Gynaecol., 98, Gadir, A.A., Khatim, M.S., Mowafi, R.S., Alnaser, H.M.I., Muharib, N.S. and Shaw, R.W. (1992) Implications of ultrasonically detected polycystic ovaries. I. Correlations with basal hormone profiles. Hum. Reprod., 7, Givens, J.R., Andersen, R.N., Umstot, E.S. and Wiser, W.L. (1986) Clinical findings and hormonal responses in patients with polycystic ovarian disease with normal versus elevated LH levels. Obstei. GynecoL 47, Jacobs, H.S. (1987) Polycystic ovaries and polycystic ovary syndrome. GynecoL Endocrinol, 1, Jacobs, H.S. (1994) Prevalence and significance of polycystic ovaries: Opinion. Ultrasound Obstet. Gynaecol, 4, 3 4. Kiddy, D.S., Hamilton-Fairley, D., Seppala, M., Koistinen, R., James, V.H.T., Reed, M.J. and Franks, S. (1989) Diet-induced changes in sex hormone binding globulin and free testosterone in women with normal or polycystic ovaries: correlation with serum insulin and insulin-like growth factor-1. Clin. Endocrinol, 31, Kim, M.H., Rosenfield, R.L., Hosseinian, A.R. and Schneir, H.G. (1979) Ovarian hyperandrogenism with normal and abnormal histological findings of the ovaries. Am. J. Obstet. GynecoL 134, Poison, D.W., Mason, H.D., Saldahna, B.Y. and Franks, S. (1987) Ovulation of a single dominant follicle during treatment with low dose pulsatile follicle stimulating hormone in women with polycystic ovary syndrome. Clin. Endocrinol, 26, Poison, D.W., Wadsworth, J., Adams, J. and Franks, S. (1988) Polycystic ovaries: a common finding in normal women. Lancet, ii, Robinson, S., Rodin, D.A., Deacon, A., Wheeler, M.J. and Clayton, R.N. (1992) Which hormone tests for the diagnosis of polycystic ovary syndrome? Br. J. Obstet. Gynaecol, 99, Santen, R.J. and Bardin, C.W. (1973) Episodic LH secretion in man. Pulse analysis, clinical interpretation, physiologic mechanisms. J. Clin. Invest., 52, Shulman, A., Shulman, N., Weissenglass, L. and Bahary, C. (1989). Ultrasonic assessment of the endometrium as a predictor of oestrogen status in amenorrhoeic patients. Hum. Reprod., 4, Stein, I.F. and Leventhal, M.L. (1935). Amenorrhoea associated with bilateral polycystic ovary syndrome. Am. J. Obstei. Gynaecol, 29, Swanson, M., Sauerbrei, E.E. and Cooperberg, PL. (1981) Medical implications of ultrasonically detected polycystic ovaries. J. Clin. Ultrasound, 9, Tayob, Y., Robinson, G., Adams, J., Nye, M., Whitelaw, N., Shaw, R.W., Jacobs, H.S. and Guillebaud, J. (1990) Ultrasound appearance of the ovaries during the pill-free interval. Br. J. Family Planning, 16, Received on January 24, 1995; accepted on May 18,