Clinical Endocrinology (1992) 36, 105-111
Improvement in endocrine and ovarian function during dietary treatment of obese women with polycystic ovary syndrome Deborah S. Kiddy, Diana Hamilton-Fairley, Anne Bush', Frances Short', Vtctor Anyaoku', Michael J. Reed' and Stephen Franks Department of Obstetrics and Gynaecology and 'Unit of Metabolic Medicine, Imperial College of Science, Technology and Medicine, St Mary's Hospital Medical School, London, UK (Received 11 June 1991; returned for revision 3 July 7991; finally revised 25 July 1991; accepted 15 August 1991)
Summary OBJECTIVE Obese women with poiycystic ovary syndrome have a greater frequency of menstrual disturbance and of hirsutism than lean women with the syndrome. initial studies have demonstrateda marked improvement in endocrine function following a short-term, very low calorie diet. The purpose of this study was to examine the effect of long-term calorie restriction on clinical as well as biochemical abnormalities In obese women with polycystic ovary syndrome. DESIGN We performeda wlthin-groupcomparisonof ciinical and biochemical indices before and during dietary treatment. PATIENTS Twenty-four obese women with polycystic ovary syndrome (mean weight 91.5 (SD 14.7)kg) were scheduled for treatment for 6-7 months with a 1000 kcal, low fat diet. Nineteen of the 24 had menstrual disturbances, 12 had infertility and 19 were hirsute. MEASUREMENTS AND RESULTS Thirteen subjects lost more than 5% of their starting weight (range 5.9-22%). in this group there was no significant change in gonadotrophin or total serum testosterone levels but there was a marked increase in concentrations of sex hormonebinding globulin (pretreatment: 23.6(9.5);post-treatment 36.3 (11.8)nmolll, P=0.002)and a reciprocal change in free testosterone levels (77 (26) vs 53 (21) pmol/l, P = 0.009). These changes were accompanied by a reduction in fasting serum insulin levels (median (range) 11.2 (5.2-32)vs 2.3 (0.1-13.8) mUII, P = 0.018)and the insulin response to 75 g oral glucose. There were no significant
Correspondence: Professor S. Franks, Department of Obstetrics and Gynaecology, St Mary's Hospital Medical School, London W2 lPG, UK.
changes in these Indices In the group who lost < 5 % of their inltiai body weight. Of the 13 women who lost > 5 % of their pretreatment weight, 11 had menstrual dysfunction. Amongst these women, nine of 11 showed an Improvement in reproductive function, i.e. they either conceived (five) or experienced a more regular menstrual pattern. There was a reduction In hlrsutism In 40% of the women in this group. By contrast, in the group who lost less than 5% of their Initial weight, only one of the eight with menstrual disturbances noted an improvement in reproductivefunction and none had a signtficant reduction in hirsutism. CONCLUSIONS These data indicate that moderateweight loss during long-term calorie restriction is associatedwith a marked cllnlcal improvement which reflects the reduction in insulin concentrations and reciprocal changes in SHBG. The Improvementin menstrualfunction and fertility may therefore be consequent upon an increase In insulin sensitivity which, directly or indirectly, affects ovarian function.
Polycystic ovary syndrome is the most common cause of anovulatory infertility (Adams et al., 1986; Hull, 1987). Women with polycystic ovaries (PCO) are often overweight and obese subjects have a higher prevalence of menstrual disorders and, consequently, infertility than lean women with PCO (Kiddy et al., 1990). In recent studies it has also been shown that obese women with PCO are more likely to be hirsute than non-obese subjects (Kiddy et al., 1990; Conway et al., 1990) and that although total testosterone levels in these women are similar to those in their normal weight counterparts (Kiddy et al., 1990), SHBG levels are much lower in the obese group of women. This leads to higher serum concentrations of free testosterone and increased metabolic clearance of testosterone by androgendependent tissues. It is now well recognized that there is an inverse correlation of serum SHBG with body mass index (BMI) (Plymate et al., 1981; Kiddy et al., 1989) and there is increasing evidence that insulin plays an important part in this interrelationship. Traditionally, sex steroids and thyroid hormones have been considered to be the major regulators of SHBG concentrations in serum, but dietary factors may be more important. We have recently shown that short-term 105
106
D. S. Kiddy et a / .
Table 1 Clinical details of 24 women with PCOS before (pre) and after (post) treatment with a 1000 kcal/day diet.
Weight (kg) Subject 1
2 3 4 5 6 7 8 9 10
I1 12 13 14
15 16 17 18 19 20 21 22 23 24
BMI Pre 38.0 35.5 33.0 33.9 41.7 45.7 31.3 32.1 39.2 35.2 26.9 35.5 29.0 38.9 35.0 36.0 30.3 30.0 38.7 27.2 32.5 30.2 26.1 37.0
FG score
Menstrual pattern
pre
post
pre
post
Pre
post
1174 84. I
108.9' 654* 89.6' 98.8 92.6* 107.0 81.9' 77.8 96.0* 79.3* 70.2 86.6* 77.9 944* 95.2 89.1 69.8* 66.1* 106.0 72.0 81.9 75.3 69.6* 87.9*
NH 24 16 13
NH 21 13 12 17 13 16 NH 9 NH 15 12 NH 12 28
Irreg and (i) polymen Irreg and polymen Regular Regular Amenorrhoea Oligomen (i) Irregular (i) Oligomen (i) Regular Regular Menorrhagia Amenorrhoea Regular Reg. anov. (i) Oligomen. (i) Oligo-amen Amenorrhoea (i) Oligomen. Regular (i) Oligomen. (i) Oligomen. Oligomen. (i) Amenorrhoea Reg. anov. (i) Reg. anov. (i)
No change Regular No change No change Irregular No change Conceived Conceived No change Reduced blood loss Regular No change No change No change No change No change Conceived Conceived No change No change No change No change Conceived Conceived
92.3 101.3 117.1 112.5 87.7 78.9 104.6 90.0 72.6 100.0 72.2 105.6 94.5 92.1 80.5 75.1 109.0 74.0 85.7 74.3 75.6 100.7
17
13 16 NH 9 NH 15 22 NH 24 28 15 11 20 16 11 21 15 NH 15
17
11 20 -
11 -
15
NH 15
FG, Ferriman-Gallwey score for hirsutism; NH, non-hirsute; * > 5% weight loss; -, conceived in month following completion of study; i, infertility
treatment of obese women with PCO with a very low calorie diet (350450 kcal per day) leads to a twofold increase in serum SHBG levels and an accompanying fall in serum insulin (Kiddy et al., 1989). This prompted the question whet her long-term calorie restriction and weight reduction would not only improve hormone levels but also restore regular ovulatory menstrual cycles, and therefore fertility, in these subjects. This study was therefore designed to examine the endocrine and clinical consequences of weight loss in obese women with polycystic ovaries. Patients and methods
Twenty-four obese women with polycystic ovaries were recruited into the study from the gynaecologicalkndocrine and infertility clinics at the Samaritan Hospital for Women (Table 1). Each of these subjects was diagnosed by pelvic ultrasonography as having polycystic ovaries (Adams et al., 1986) and each had a raised serum LH or serum testosterone
level (or both) except four subjects, three of whom were hirsute and had a total testosterone equal to or greater than one standard deviation above the mean control value (Franks, 1989). None of the subjects had received any hormonal treatment in the 2 months prior to the study and all were premenopausal. Each subject's weight was stable for at least 4 weeks prior to the study. Before starting the study informed written consent was obtained from each subject. Approval for the study was obtained from the Ethical Committee of Parkside Health Authority. Each subject also had the following baseline investigations performed prior to the study: ultrasound ovarian scan, haemoglobin, urea and electrolytes, measurement of height and weight for calculation of body mass index (BMI) (weight (kg)/height (m)2). The Ferriman-Gallwey score (Ferriman & Gallwey, 1961) was used to document the degree of hirsutism by the same observer (DSK) before and during the study. A history was taken of the pattern of menstrual cycles, alcohol consumption, smoking habits and
Endocrine and ovarian function during dietary treatment of PCOS
Clinical Endocrinology (1992) 36
Fig. 1 Serum concentrations of a.
testosterone; b, SHBG; c, free testosterone (free T); d, fasting insulin; e, the sum of insulin concentrations during the OGIT (sumINS) and f, IGF-I B, before and D, after dietary treatment in women with PCOS who lost more or less than 5% of their initial body weight ( > 5%, < 5%). Values are mean + I SD or, in the case of insulin concentrations, median and ( - ) range.
= 15
60
-g 4 z:23 .-
50
1
40 30 W
g
Y)
5
v,
I
I-"
g e
'
12 8 4
E
5
20 10 0
107
I20
= too 1 g
00
2 60 F g 40
t 20 0
600 400
200
"
n
>5%
amount of exercise taken. A detailed dietary assessment was made on each subject by one of two experienced dietitians and, following this, a food diary was kept for one week so that calculations of calorie intake and composition of the diet could be made. The subjects were scheduled for a low calorie, low fat diet. Those with a BMI greater than 30 kg/m2were initially given the option of a very low calorie diet (the 'Cambridge Diet', comprising 330 kcal per day (Kiddy et al., 1989) for 4 weeks, followed by a 1000 kcal per day low fat diet (approximately 20 g fat per day) for a further 6 months. Those subjects with a BMI between 25 and 30 kg/m2and those who opted against the very low calorie diet were started on the 1000 calorie low fat diet and continued for the full 7-month period. Blood was taken via an intravenous cannula between 0800 and 1000 hours, after an overnight fast, to measure levels of serum LH and FSH, total testosterone, free testosterone, SHBG, glucose, insulin, IGF-I and IGFBP-I. In nine subjects, serial blood samples were obtained at 15-minute intervals for 8 hours, before and after dieting, for measurement of LH and FSH (Mason et d., 1988). In those women who had regular menses, blood samples were taken in the early follicular phase (days 1-5) of the cycle. Following the initial fasting sample, a 75 g oral glucose tolerance test (OGTT) was carried out; blood samples were obtained at 30-minute intervals for 3 hours after glucose ingestion for determination of blood glucose and plasma insulin concentrations. Normal saline was used to maintain the patency of the intravenous cannula. After starting the diet each woman was assessed at monthly intervals. At each visit a fasting blood sample was obtained and the woman was seen by the dietitian, weighed, and the dietary history was reviewed. Each subject was asked to record the date of menses during the study period. After 7 months all the participants underwent a complete review of the clinical and biochemical examination and a further OGTT was performed.
5%
5% of initial weight) at the end of the study period. Of the 24 women studied, 13 lost more than 5% of their starting weight (mean (SD) 11.55 (4.8)% range 5.9-22%). In three subjects body weight at 7 months was not significantly lower than the pretreatment weight although they completed the full period of assessment. In addition, four subjects withdrew from the study for social reasons or non-compliance after 4-6 months of study. A further six women conceived during the study (see below) and did not, therefore, complete the full 7 months of treatment. In those cases of withdrawal from the study for whatever reason, the blood samples collected immediately before withdrawal (preceding pregnancy in those subjects who conceived) were analysed and the clinical data recorded. In the group as a whole, the mean (SD) body weight decreased from 91.9 (14.5) to 85.0 (13-2) kg (P
Improvement in endocrine and ovarian function during dietary treatment of obese women with polycystic ovary syndrome Deborah S. Kiddy, Diana Hamilton-Fairley, Anne Bush', Frances Short', Vtctor Anyaoku', Michael J. Reed' and Stephen Franks Department of Obstetrics and Gynaecology and 'Unit of Metabolic Medicine, Imperial College of Science, Technology and Medicine, St Mary's Hospital Medical School, London, UK (Received 11 June 1991; returned for revision 3 July 7991; finally revised 25 July 1991; accepted 15 August 1991)
Summary OBJECTIVE Obese women with poiycystic ovary syndrome have a greater frequency of menstrual disturbance and of hirsutism than lean women with the syndrome. initial studies have demonstrateda marked improvement in endocrine function following a short-term, very low calorie diet. The purpose of this study was to examine the effect of long-term calorie restriction on clinical as well as biochemical abnormalities In obese women with polycystic ovary syndrome. DESIGN We performeda wlthin-groupcomparisonof ciinical and biochemical indices before and during dietary treatment. PATIENTS Twenty-four obese women with polycystic ovary syndrome (mean weight 91.5 (SD 14.7)kg) were scheduled for treatment for 6-7 months with a 1000 kcal, low fat diet. Nineteen of the 24 had menstrual disturbances, 12 had infertility and 19 were hirsute. MEASUREMENTS AND RESULTS Thirteen subjects lost more than 5% of their starting weight (range 5.9-22%). in this group there was no significant change in gonadotrophin or total serum testosterone levels but there was a marked increase in concentrations of sex hormonebinding globulin (pretreatment: 23.6(9.5);post-treatment 36.3 (11.8)nmolll, P=0.002)and a reciprocal change in free testosterone levels (77 (26) vs 53 (21) pmol/l, P = 0.009). These changes were accompanied by a reduction in fasting serum insulin levels (median (range) 11.2 (5.2-32)vs 2.3 (0.1-13.8) mUII, P = 0.018)and the insulin response to 75 g oral glucose. There were no significant
Correspondence: Professor S. Franks, Department of Obstetrics and Gynaecology, St Mary's Hospital Medical School, London W2 lPG, UK.
changes in these Indices In the group who lost < 5 % of their inltiai body weight. Of the 13 women who lost > 5 % of their pretreatment weight, 11 had menstrual dysfunction. Amongst these women, nine of 11 showed an Improvement in reproductive function, i.e. they either conceived (five) or experienced a more regular menstrual pattern. There was a reduction In hlrsutism In 40% of the women in this group. By contrast, in the group who lost less than 5% of their Initial weight, only one of the eight with menstrual disturbances noted an improvement in reproductivefunction and none had a signtficant reduction in hirsutism. CONCLUSIONS These data indicate that moderateweight loss during long-term calorie restriction is associatedwith a marked cllnlcal improvement which reflects the reduction in insulin concentrations and reciprocal changes in SHBG. The Improvementin menstrualfunction and fertility may therefore be consequent upon an increase In insulin sensitivity which, directly or indirectly, affects ovarian function.
Polycystic ovary syndrome is the most common cause of anovulatory infertility (Adams et al., 1986; Hull, 1987). Women with polycystic ovaries (PCO) are often overweight and obese subjects have a higher prevalence of menstrual disorders and, consequently, infertility than lean women with PCO (Kiddy et al., 1990). In recent studies it has also been shown that obese women with PCO are more likely to be hirsute than non-obese subjects (Kiddy et al., 1990; Conway et al., 1990) and that although total testosterone levels in these women are similar to those in their normal weight counterparts (Kiddy et al., 1990), SHBG levels are much lower in the obese group of women. This leads to higher serum concentrations of free testosterone and increased metabolic clearance of testosterone by androgendependent tissues. It is now well recognized that there is an inverse correlation of serum SHBG with body mass index (BMI) (Plymate et al., 1981; Kiddy et al., 1989) and there is increasing evidence that insulin plays an important part in this interrelationship. Traditionally, sex steroids and thyroid hormones have been considered to be the major regulators of SHBG concentrations in serum, but dietary factors may be more important. We have recently shown that short-term 105
106
D. S. Kiddy et a / .
Table 1 Clinical details of 24 women with PCOS before (pre) and after (post) treatment with a 1000 kcal/day diet.
Weight (kg) Subject 1
2 3 4 5 6 7 8 9 10
I1 12 13 14
15 16 17 18 19 20 21 22 23 24
BMI Pre 38.0 35.5 33.0 33.9 41.7 45.7 31.3 32.1 39.2 35.2 26.9 35.5 29.0 38.9 35.0 36.0 30.3 30.0 38.7 27.2 32.5 30.2 26.1 37.0
FG score
Menstrual pattern
pre
post
pre
post
Pre
post
1174 84. I
108.9' 654* 89.6' 98.8 92.6* 107.0 81.9' 77.8 96.0* 79.3* 70.2 86.6* 77.9 944* 95.2 89.1 69.8* 66.1* 106.0 72.0 81.9 75.3 69.6* 87.9*
NH 24 16 13
NH 21 13 12 17 13 16 NH 9 NH 15 12 NH 12 28
Irreg and (i) polymen Irreg and polymen Regular Regular Amenorrhoea Oligomen (i) Irregular (i) Oligomen (i) Regular Regular Menorrhagia Amenorrhoea Regular Reg. anov. (i) Oligomen. (i) Oligo-amen Amenorrhoea (i) Oligomen. Regular (i) Oligomen. (i) Oligomen. Oligomen. (i) Amenorrhoea Reg. anov. (i) Reg. anov. (i)
No change Regular No change No change Irregular No change Conceived Conceived No change Reduced blood loss Regular No change No change No change No change No change Conceived Conceived No change No change No change No change Conceived Conceived
92.3 101.3 117.1 112.5 87.7 78.9 104.6 90.0 72.6 100.0 72.2 105.6 94.5 92.1 80.5 75.1 109.0 74.0 85.7 74.3 75.6 100.7
17
13 16 NH 9 NH 15 22 NH 24 28 15 11 20 16 11 21 15 NH 15
17
11 20 -
11 -
15
NH 15
FG, Ferriman-Gallwey score for hirsutism; NH, non-hirsute; * > 5% weight loss; -, conceived in month following completion of study; i, infertility
treatment of obese women with PCO with a very low calorie diet (350450 kcal per day) leads to a twofold increase in serum SHBG levels and an accompanying fall in serum insulin (Kiddy et al., 1989). This prompted the question whet her long-term calorie restriction and weight reduction would not only improve hormone levels but also restore regular ovulatory menstrual cycles, and therefore fertility, in these subjects. This study was therefore designed to examine the endocrine and clinical consequences of weight loss in obese women with polycystic ovaries. Patients and methods
Twenty-four obese women with polycystic ovaries were recruited into the study from the gynaecologicalkndocrine and infertility clinics at the Samaritan Hospital for Women (Table 1). Each of these subjects was diagnosed by pelvic ultrasonography as having polycystic ovaries (Adams et al., 1986) and each had a raised serum LH or serum testosterone
level (or both) except four subjects, three of whom were hirsute and had a total testosterone equal to or greater than one standard deviation above the mean control value (Franks, 1989). None of the subjects had received any hormonal treatment in the 2 months prior to the study and all were premenopausal. Each subject's weight was stable for at least 4 weeks prior to the study. Before starting the study informed written consent was obtained from each subject. Approval for the study was obtained from the Ethical Committee of Parkside Health Authority. Each subject also had the following baseline investigations performed prior to the study: ultrasound ovarian scan, haemoglobin, urea and electrolytes, measurement of height and weight for calculation of body mass index (BMI) (weight (kg)/height (m)2). The Ferriman-Gallwey score (Ferriman & Gallwey, 1961) was used to document the degree of hirsutism by the same observer (DSK) before and during the study. A history was taken of the pattern of menstrual cycles, alcohol consumption, smoking habits and
Endocrine and ovarian function during dietary treatment of PCOS
Clinical Endocrinology (1992) 36
Fig. 1 Serum concentrations of a.
testosterone; b, SHBG; c, free testosterone (free T); d, fasting insulin; e, the sum of insulin concentrations during the OGIT (sumINS) and f, IGF-I B, before and D, after dietary treatment in women with PCOS who lost more or less than 5% of their initial body weight ( > 5%, < 5%). Values are mean + I SD or, in the case of insulin concentrations, median and ( - ) range.
= 15
60
-g 4 z:23 .-
50
1
40 30 W
g
Y)
5
v,
I
I-"
g e
'
12 8 4
E
5
20 10 0
107
I20
= too 1 g
00
2 60 F g 40
t 20 0
600 400
200
"
n
>5%
amount of exercise taken. A detailed dietary assessment was made on each subject by one of two experienced dietitians and, following this, a food diary was kept for one week so that calculations of calorie intake and composition of the diet could be made. The subjects were scheduled for a low calorie, low fat diet. Those with a BMI greater than 30 kg/m2were initially given the option of a very low calorie diet (the 'Cambridge Diet', comprising 330 kcal per day (Kiddy et al., 1989) for 4 weeks, followed by a 1000 kcal per day low fat diet (approximately 20 g fat per day) for a further 6 months. Those subjects with a BMI between 25 and 30 kg/m2and those who opted against the very low calorie diet were started on the 1000 calorie low fat diet and continued for the full 7-month period. Blood was taken via an intravenous cannula between 0800 and 1000 hours, after an overnight fast, to measure levels of serum LH and FSH, total testosterone, free testosterone, SHBG, glucose, insulin, IGF-I and IGFBP-I. In nine subjects, serial blood samples were obtained at 15-minute intervals for 8 hours, before and after dieting, for measurement of LH and FSH (Mason et d., 1988). In those women who had regular menses, blood samples were taken in the early follicular phase (days 1-5) of the cycle. Following the initial fasting sample, a 75 g oral glucose tolerance test (OGTT) was carried out; blood samples were obtained at 30-minute intervals for 3 hours after glucose ingestion for determination of blood glucose and plasma insulin concentrations. Normal saline was used to maintain the patency of the intravenous cannula. After starting the diet each woman was assessed at monthly intervals. At each visit a fasting blood sample was obtained and the woman was seen by the dietitian, weighed, and the dietary history was reviewed. Each subject was asked to record the date of menses during the study period. After 7 months all the participants underwent a complete review of the clinical and biochemical examination and a further OGTT was performed.
5%
5% of initial weight) at the end of the study period. Of the 24 women studied, 13 lost more than 5% of their starting weight (mean (SD) 11.55 (4.8)% range 5.9-22%). In three subjects body weight at 7 months was not significantly lower than the pretreatment weight although they completed the full period of assessment. In addition, four subjects withdrew from the study for social reasons or non-compliance after 4-6 months of study. A further six women conceived during the study (see below) and did not, therefore, complete the full 7 months of treatment. In those cases of withdrawal from the study for whatever reason, the blood samples collected immediately before withdrawal (preceding pregnancy in those subjects who conceived) were analysed and the clinical data recorded. In the group as a whole, the mean (SD) body weight decreased from 91.9 (14.5) to 85.0 (13-2) kg (P
