- Email: [email protected]
Preliminary results suggesting exaggerated ovarian androgen production early in the course of polycystic ovary syndrome
JOURNALOF ADOLESCENTHEALTHCARE1990;11:480-484
ia
Preliminary Results Suggesting Androgen Production Early in the Course of Polycystic Ovary Syndrome GARY L. FRANCIS,
M.D.,
Ph.D.,
ALAN GETTS, M.D.,
Excess ovarian sndrogen production might be a cause of the polycystic ovary syndrome (PCO). Previous studies have evaluated adult women with long-standing abnormality of the hypothalamic-pituitary gqnadal axis, Abnormal ovarian function in such patients could be a primary or even a secondary finding. For that reason, this study was designed to evaluate ovarian a&ogen production in symptomatic adolescent females. Simultaneous adrenal suppression, by using dexamethasone, and ovarian stimulation, by using gonadotropin-releasing hormone (GnRH), were achieved in 12 patients. Following stimulation, blood was serially obtained over 8 hr to measure gonadotropin, esand androgen responses. Based on the mgen, androgen response, patients could be divided into two groups. Group A (five) had a significant Increase (p c 0.01) in free testostemne, whereas group B (seven) had no increase in any androgen, including free testosterone (signikantly different from group A, p = 0.01). All patients in group A had enlarged or cystic ovaries, whereas only one-quarter patients in group B had enlarged ovaries Csigniscantly different from group A, p ,: 0.03). The pituitary and estrogenic response was similar in both groups. These preliminary data suggest that some patients with PC0 (group A) have a primary abnormality in ovarian androgen production early in the course of their disease.
Fromthe Departments of Pediatrics (G.L.F., A.G.) and chid Eisenhower Army Medical Center, Fort Gordon,
Investigation WM.),
Georgia. The opinions or assertions contained herein are the privateviews Oftheauthorsand are not to be construed as ofIicia/ or as reflecting the views Of the Deparftnent of the Army or the Department of De*. eddress reprint WqWStS to: Gary L. Francis,. LTC, D~artment of Poetics,bWdte? Reed Atmy Medical Center, Washington, m 2()3075001. M@msc?ipt accepted March 6,199O. 480
J.C.
MCPHERSON,
III,
Ph.D.
KRY WORDS:
rolycystic ovary Androgen GnRH
Polycystic ovary syndrome (PCO) is a clinical disorder characterized by menstrual irregularity, hirsutism, and obesity. It often begins in the perimenarchal female adolescent, but the cause is stil uncertain (l-4). Previous studies on the pathogenesis of PC0 have evaluated adult wcmen with long-standing abnormality of the hypothalamicpituitary gonadal axis. By this age, abnormal ovarian fki:tion could be a primary or even a secondary event. In addition, most studies have used dexamethasone suppression of hyperandrogenemia to separate adrenal from ovarian causes of androgen excess. However, this may be equivocal or misleading (5). Consequently, our study was designed with two goals: First, we wanted to isolate ovarian androgen production by simultaneous adrenal suppression with dexamethasone and ovarian stimulation with gonadotropin-releasing hormone (GnRl-l); second, we chose to exclusively evaluate female adolescents with suspected PC0 rather than adult women with long-standing disease.
Methods Following institutional, parental, and patient consent, 12 female adolescents with symptoms suggestive of PC0 were enrolled in this study. All had clinical symptoms, including hirsutism (Ferriman and Galwey score > 6) (6), acne (7), irregular menses, or obesity. Patients were identified as having
Published by Eisevier science Publishiig
Co., Inc., 655 Avenue of the Americas, New York, my 1~10
November 1990
ANDROGENS IN POLYCYSTK OVARY SYNDROME
amenorrhea if they had absent menses for at least 6 months and oligomenorrhea if they had less than six menses per year. All ranged in age from 14 to 19 years old. Normal control subjects were not included because of potential side effects and a lack of benefit to such subjects by participation in this study. AlI patients had baseline AM determinations of thyroxine, thyrotropin, prolactin (PRL), luteinizing hormone (LH), follicle-stimulating hormone (F§H), testosterone (T), and 17-alpha-hydroxyprogesterone (17OHP). Additional measurements in the majority of patients included dehydroepiandrosterone (DHEA) or DHA sulfate (DHAS) and androstendione (A). Insulin levels were not obtained. Results were compared to the normal adolescent ranges. Pelvic examination was performed on all patients, and pelvic ultrasound was performed on nine patients. All patients had normal baseline 170HP levels, and three patients (1,5, and 8) additionally underwent 1 hr andrenocorticotropin (l-24 synthetic ACTH, 0.25 mg IV) stimulation. These data were used to exclude nonclassical forms of congenital adrenal hyperplasia in all patients. All patients were given dexamethasone (1.5 mg/ m2 per day + tid) for 6 days. At 8 AM on day 6, GnRH (100 pg IV) was given, and blood was obTable 1. Clinical and Hormonal” Observations on Patient no. (Nl . Ranges)
he (yr)
Hirsutism’ A‘Ild Mensesb (<6? --
481
tained at 0,30,60,90, and 120 min for LH and FSH determinations. Blood obtained at 0, 2, 4,. and 8 hr was assayed for cortisol, DHA, DHAS, A, T, free T, e&one (E,), and estradiol (E2). Free T levels were measured by using commercial reagents (Pantex, Santa Monica, CA), with an intra-assay coefficient of variation of 8.6%. All other steroid hormones were measured by using previously published techniques (B-10). Statistical significance was determined by linear regression analysis, unless otherwise indicated. Results
The clinical characteristics of these 12 patients are shown in Table 1. The patients were divided into two groups, based on the serum free T response to GnRH. In 5 of the 12 patients, designated group A, free T levels increased following GnRH (Fig. IA). This increase in free T was significant over 8 hr (6.2 fold increase, p < 0.01) (Fig. 2). A single patient in this group (patient 5) also had an increase in A (baseline, 8.7 nmol/L; at 8 hr, 52 nmoI/L). She was the only patient in either group to have such an increase. Group B included seven patients who did not have an increase in any serum androgen, including free
12 Adolescent
Females
with Suspected
PC0
DHAS(FmollL) A(nmol/L) Pelvic US Ideal body T(nmolfL) LH(IU/L) (2.7-7.8) (<2.5cm) mass (%) (0.79-2.60) (3-20) LHIFSH (2.22-922)
Group A If4
2’ 3 4 59 Mean f SD
16 14 18 14 17
A A DUB A 0
7 7 27 5 18 12.8~8
15.8k1.8
III NL I III I 1.6-el.2
159 96 NA 157 142
2.22 2.01 2.36 7.42 3.26
138229
3.45k1.3
6.9 6.2 29.7 27.3 20.9
0.63 0.81 2.7 2.5 2.2
18.2ir7.0 1.7kO.6
4.29 3.90 9.80 10.90 6.0
9.5 NA 15.5 4.2 NA
7.o-c3
9.7-e6
3.1 3.3h 3.2 4.2 3.5
Group B ;
17 19
A 0
27 4
III II
196 150
3.26 3.50
21.2 18.5
1.6 1.7
3.58 5.72
8.5 3.4
c2.5 4.0
&p 9 10 11 12
15 15 19 16 18
DUB 0 0 0 NL
5 6 6 8 16
I II I II II
158 88 219 143 153
4.50 1.39 4.26 2.81 1.80
10.6 <2 9.9 36” 21
1.1 0.2 1.8 2.8 2.1
9.70 6.19 5.70 NA NA
12.0 NA 7.9 NA 6.3
c2.5 NA c2.5 NA NA
1.6kO.4 -
6.2rc-2
7.6k-3
Mean 2 L;D 17.Ok1.7
10.3. 8
1.820.7
158242
3.07kO.4 19.624.4
“Not shown are normal levels of 170HP, cortisol, PRL, Tl, and TSH. 6O = oligomenorrhea; A = amenorrhea; DUB = dysfunctional uterine bleeding; NL = normal. ‘Method of Ferriman and Gallwey (6). dGraded acne score I-III (7). ‘Longest axis ovarian length in cm as determined by pelvic ultrasound. ‘Baseline free T was elevated. 8ACI’II stimulation testing was normal. ‘Cystic ovarian change also identikied on pelvic ultrasound.
482
Free
2n
jCIURNAL OF ADDLESCENT HEALTH CARE VoI. 11 h’o. 6
FRANCIS ET AL.
-
T (pmol/L)
I
Free
T (pml/L)
,.___ _,._ ,@.......l.----- ----......,,-_.,..-..---*
1
I
I
a
4
2
,
(i
I
0
8
2
TIME AFTER GnAH (Hours} -
PATIENT
i
+
PAT.ENT 2
-a-
PATIENT
4
-*-
PAllENT
Hgwe
-k-
I
I
4
6
e
TIME AFTER GPW (Hcars) “’
PATIENT 3
s
*,
PUIENT
8
PATIENT 10
,+..
WTIENT
7
,-+,
PATIENT 8
6,
PATIENT
11
“&
PATIENT
a-
PATIENT
9
1'2
1. (AI individual free T response to GnRH in flue group A
@i&s. (B) hdiuidual free T response to GnRli in seven group B patients. Patients 8,9, and 12 M ~?eeT levels beh.u the level of assay msitiuiiyatalltime points.
T, following
GnRH (Fig. 16). This was significantly different from the change in free T levels seen over 8 hr in group A (p = 0.01) (Fig. 2). Although the androben response was different between groups, both had similar increases in LH, E,, and E2 (Table 2). Clinical observations and hormone measurements were compared between groups. Amenorrhea was more common in group A (315) than in group B (l/7), although this difference was not statistically significant. Enlarged ovaries (>2.5 cm longest axis as assessed hi pelvic ultrasound} (11) were identified in all patients in group A, one of whom had cystic changes. Of the four patients in group B who had pelvic ultrasound, only one had enlarged
Figure 2. Free T respme group B I +J. Data ure
to GnlW in patients in group A Ii@ and
expressed us mew
ovaries. This difference was also significant (p C 0.03, Fisher’s exact test). There were no differences between groups A and B in mean age, weight, acne, or hirsutism scores. Baseline hormone levels, including 17OHP, A, DHAS, T, free T, LH, and LHI FSH ratio were similar in both groups, as was the response to dexamethasone suppression--aU 12 patients had suppressed cortisol(~82.7 nmd/L, 3 pg! dl) and free T levels (~27.7 pmollL, 8 pglml). All patients were subsequently treated with a combination oral contraceptive. No other therapy for acne or hirsutism was prescribed. The results of the combined &xamethasone-GnRH stimulation test were unknown to both the patient and evaluating physician. After 2 years of treatment, nine patients (four from group A and five from group B) were available for follow-up. Clinical improvement, including decreased hirsutism, acne, and oily skin, as
T&h 2. Response of LH, E,, and E2 Following in Dexamethasone-Suppressed Patients”
+ SD.
Group A
Grerup B
Maximal LI-I (IUIL) 108.2 -c 88.4 67.9 5 1.7 E, (pm&L) at time (hr) 0 2 4 8
Er (pmol/L) at lime (hr) Q 0
a
I
I
2
4
i
6
TIME AFTER GnRH (Hours)
-
GROUP A
-
GROUP El
8
2 4 E
GnRH Probability p = 0.15
140 122 229 454
I? 50 2 40 rc_103 -c 100
172 176 218 322
-c f f it
33 54 22 66
NSb
139 155 293 667
rc 50 + 40 2 80 f 240
233 203 340 447
k f f 2
65 95 93 90
NSb
“Data are expressed as mean -r- SD. bNS = not significant.
November 1990
well as a reduction in basal androgen levels, was seen in four of four patients in grcup A and four of five patients in group B.
Discussion PC0 might be caused by abnormal androgen production in the ovary or adrenal glands or by
abnormal hypothalamic-pituitary regulation of gonadotropin secretion (l-3,5,12,13). Previous investigations of the etiology of PC0 have primarily evaluated adult women who may have had longstanding abnormalities of and perhaps secondary changes in the hypothalamic-pituitary gonadal axis. A variety of stimulation tests have been shown to increase T, free T, A, El, and E2levels in both normal women and women with PC0 (12-14). The women with PC0 had the greatest increase, suggesting an ovarian abnormality, but there was considerable overlap. A recent study, using G&I-I analog stimulation of the ovary, revealed increased 17OI-Wlevels only in women with PC0 (14). This suggests that the adult women with PC0 has an intrinsically different ovarian response to G&I-I compared to her normal counterpart. This, however, could be a primary abnormality or secondary to long-standing changes in the hypothalamic-pituitary gonadal axis. To our knowledge, our study is the first to exclusively evaluate adolescent females with suspected PCO. Abnormalities defined in this age group may be more likely to reflect the primary etiology of PC0 than in patients with long-standing disease. All 12 of our female adolescents had symptoms suggestive of PCO. Five patients (patients 3, 4, and 5 from group A and patients 11 and 12 from group B) had an LH/FSH ratio in excess of 2:l. This increased LH/ FSH ratio is sommonly but not universally seen in patients with PC0 (15). Nine patients were unequivocally hyperandrogenemis and dexamethasone-suppressible, with either a suppressed free T of c 27.7 pmol/L (8 pg/ml) or > 50% suppression of total T. This latter finding would typically be interpreted to suggest an adrenal abnormality but could be misleading due to the effects of dexamethasone on both the adrenal glands and ovary (5). The other three patients (patient 3 in group A and patients 9 and 12 in group B) may have had a subtle androgen abnormality not documented by our single measurements (4). In contrast, combined dexamethasone-GnRH stimulation allowed us to identify five patients who were distinctly different. Designated group A, they had an increase in free T but normal pituitary and
ANDROGENS IN POLYCYSTIC OVARY SYNDROME
483
estrogenis responses to GnlUI. They were more likely to have amenorrhea and enlarged or cystic . __ . _ ovaries. It is our opimon that these data support the presence of a primary ovarian zbnormality in this group. AS would have been expected, these patients responded favorably when treated with oral contraceptives. Patients in group B had no increase in androgen levels and were less likely to have amenorrhea or enlarged ovaries. It is our opinion that this is most consistent with an adrenal abnormality or idiopathic hyperandrogenemia. Although it was not expected, these patients responded equally well to oral sontraseptive therapy. It is possible that by decreasing the ovarian contribution to serum androgen levels, oral contraceptives may have been effective in this group of patients regardless of the underlying cause of hyperandrogenemia. In conclusion, these data provide two important observations. First, they suggest that early in the course of PCO, some adolescent patients (group A) had evidence for a subtle alteration in ovarian androgen production, whereas their pituitary and estrogenis responses to GnFtH were still normal. This observation supports the theory of a primary ovarian abnormality in this disorder. Second, regardless of whether or not this abnormality was present in any individual patient, almost all found oral contraseptive therapy to be readily acceptable and effective. Oral contraceptive therapy would appear to be a good first choice for the treatment of adolescents with suspected PCO. An evaluation of isolated ovarian function, such as that in this report, may have given insight into the etiology of PC0 but appears to be unnecessary in selecting individual patients who will respond favorably to oral contraceptive treatment. We thank Michael j. Brennan, LTC, MC, for his helpful review, andMr. Thomas Charles for assistance with preparancn ui this manuscript.
References 1. ConeyP. ~olycysticovarian disease:
current concepts of pathophysiology and therapy. Fertil Steril 1984;42:667-82.
2. McKenna T. Pathogenesis and treatment of syndrome. N Engl J Med 1988;318:558-62.
poly~ystic
ovary
3. Louhlin T, Cunningham S, Moore A, Culliton M, Smyth E, M&enna T. Adrenal abnormalities in pdyqstic Ovary syn&ome. J Clin Endocrinol Metab 1986;62:142-147. 4. Moll G, Rosenfield R. Plasma fret testosterone in the diagnosis of polycystic ovary syndru.lt,p. J Pediatr 1983;102 461-4. 5. Ettinger B, Goldfield E, Burrill K, werckr R, Forsham E.
484
7.
8. 9.
10. 11.
12.
13.
14.
FRANCIS ET AL.
Plasma testosterone stimulation-suppression dynamics in hirsute women. Am J Med 1973;54:195-200. Ferriman D, Gallwey J. Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 1%1;21:1440-7. Treadwell PA. The pilocebaceous unit, acne vulgaris, and acneiform eruptions. In: Kelly, ed . Practice of pediatrics. Philadelphia: Harper and Row, 1987;10:1-13. Parker CR Jr, Ellegood JO, Mahesh VB. Methods for multiple steroid radioimmunoassay. J Steroid Biochem 1975;6:1. Wright K, Collins DC, Reedy IRK. Comparative specificity of antisera against estrone, estradiol-178, and estriol using 6O-carboxymethyloxine bovine serum albumin derivatives. Steroids 1973;21:755. Buster ]E, Abraham GE. Radioimmunoassay of plasma dehydroepiandrosterone sulfate. Anal Lett 1977;5:543. Nakayama RT. The female reproauctive system. In: Kelly, ed. Practice of pediatrics. Philadelphia: Harper and Row, 1987;7:1-14. Skaf R, Shelder R, Kemman E. Androgenic response in anovulatory women during menotropins stimulation. Obstet Gynecol1981;58:714-719. Heineman M, Thomas C, Doesburg W, Polland R. Pituitary and ovarian responses to LHIUI stimulation in women with clinical features of the polycystic ovary syndrome, Eur J Obstet Gynecol Reprod Biol1984;17:273-284. Barnes RB, Rosenfield RL, Burstein S, Ehrmann DA. Pitt&
JOURNAL OF ADOLESCENT HEALTH CARE Vol. 11, No. 6
responses to nafarelin testing in the polycystic ovary syndrome. N Engl J Med 1989;320:559-565. 15. ?4urdoch AP, Dunlop W, Kendall-Taylor P, Watson MJ. The acute effects of a dopamine antagonist (domperidone) on luteinising hormone, follicule stimulating hormone, prolactin and thyrotrophin secretion in polycystic ovarian syndrome. Differential effect of ovulation. Clin Endocrinol 1984;21:611619. tary-ovarian
This manuscriptwas reviewed by a total of four reviewers. The consensus was that it was an original and clever way to try to elucidatethe pathogenesis of PC0 in adolescents.On the other hand, the small samplesize, lackof uniformityof data collectionand fact that all patients responded to oral contraceptives, despite subgroup assignment, were troublesome. Recognizingits limitations,it was accepted for publicationto stimulate further research and to allow the reader to judge for him/herself. Iris F. Litt, M.D., Editor-In-Chief
ia
Preliminary Results Suggesting Androgen Production Early in the Course of Polycystic Ovary Syndrome GARY L. FRANCIS,
M.D.,
Ph.D.,
ALAN GETTS, M.D.,
Excess ovarian sndrogen production might be a cause of the polycystic ovary syndrome (PCO). Previous studies have evaluated adult women with long-standing abnormality of the hypothalamic-pituitary gqnadal axis, Abnormal ovarian function in such patients could be a primary or even a secondary finding. For that reason, this study was designed to evaluate ovarian a&ogen production in symptomatic adolescent females. Simultaneous adrenal suppression, by using dexamethasone, and ovarian stimulation, by using gonadotropin-releasing hormone (GnRH), were achieved in 12 patients. Following stimulation, blood was serially obtained over 8 hr to measure gonadotropin, esand androgen responses. Based on the mgen, androgen response, patients could be divided into two groups. Group A (five) had a significant Increase (p c 0.01) in free testostemne, whereas group B (seven) had no increase in any androgen, including free testosterone (signikantly different from group A, p = 0.01). All patients in group A had enlarged or cystic ovaries, whereas only one-quarter patients in group B had enlarged ovaries Csigniscantly different from group A, p ,: 0.03). The pituitary and estrogenic response was similar in both groups. These preliminary data suggest that some patients with PC0 (group A) have a primary abnormality in ovarian androgen production early in the course of their disease.
Fromthe Departments of Pediatrics (G.L.F., A.G.) and chid Eisenhower Army Medical Center, Fort Gordon,
Investigation WM.),
Georgia. The opinions or assertions contained herein are the privateviews Oftheauthorsand are not to be construed as ofIicia/ or as reflecting the views Of the Deparftnent of the Army or the Department of De*. eddress reprint WqWStS to: Gary L. Francis,. LTC, D~artment of Poetics,bWdte? Reed Atmy Medical Center, Washington, m 2()3075001. M@msc?ipt accepted March 6,199O. 480
J.C.
MCPHERSON,
III,
Ph.D.
KRY WORDS:
rolycystic ovary Androgen GnRH
Polycystic ovary syndrome (PCO) is a clinical disorder characterized by menstrual irregularity, hirsutism, and obesity. It often begins in the perimenarchal female adolescent, but the cause is stil uncertain (l-4). Previous studies on the pathogenesis of PC0 have evaluated adult wcmen with long-standing abnormality of the hypothalamicpituitary gonadal axis. By this age, abnormal ovarian fki:tion could be a primary or even a secondary event. In addition, most studies have used dexamethasone suppression of hyperandrogenemia to separate adrenal from ovarian causes of androgen excess. However, this may be equivocal or misleading (5). Consequently, our study was designed with two goals: First, we wanted to isolate ovarian androgen production by simultaneous adrenal suppression with dexamethasone and ovarian stimulation with gonadotropin-releasing hormone (GnRl-l); second, we chose to exclusively evaluate female adolescents with suspected PC0 rather than adult women with long-standing disease.
Methods Following institutional, parental, and patient consent, 12 female adolescents with symptoms suggestive of PC0 were enrolled in this study. All had clinical symptoms, including hirsutism (Ferriman and Galwey score > 6) (6), acne (7), irregular menses, or obesity. Patients were identified as having
Published by Eisevier science Publishiig
Co., Inc., 655 Avenue of the Americas, New York, my 1~10
November 1990
ANDROGENS IN POLYCYSTK OVARY SYNDROME
amenorrhea if they had absent menses for at least 6 months and oligomenorrhea if they had less than six menses per year. All ranged in age from 14 to 19 years old. Normal control subjects were not included because of potential side effects and a lack of benefit to such subjects by participation in this study. AlI patients had baseline AM determinations of thyroxine, thyrotropin, prolactin (PRL), luteinizing hormone (LH), follicle-stimulating hormone (F§H), testosterone (T), and 17-alpha-hydroxyprogesterone (17OHP). Additional measurements in the majority of patients included dehydroepiandrosterone (DHEA) or DHA sulfate (DHAS) and androstendione (A). Insulin levels were not obtained. Results were compared to the normal adolescent ranges. Pelvic examination was performed on all patients, and pelvic ultrasound was performed on nine patients. All patients had normal baseline 170HP levels, and three patients (1,5, and 8) additionally underwent 1 hr andrenocorticotropin (l-24 synthetic ACTH, 0.25 mg IV) stimulation. These data were used to exclude nonclassical forms of congenital adrenal hyperplasia in all patients. All patients were given dexamethasone (1.5 mg/ m2 per day + tid) for 6 days. At 8 AM on day 6, GnRH (100 pg IV) was given, and blood was obTable 1. Clinical and Hormonal” Observations on Patient no. (Nl . Ranges)
he (yr)
Hirsutism’ A‘Ild Mensesb (<6? --
481
tained at 0,30,60,90, and 120 min for LH and FSH determinations. Blood obtained at 0, 2, 4,. and 8 hr was assayed for cortisol, DHA, DHAS, A, T, free T, e&one (E,), and estradiol (E2). Free T levels were measured by using commercial reagents (Pantex, Santa Monica, CA), with an intra-assay coefficient of variation of 8.6%. All other steroid hormones were measured by using previously published techniques (B-10). Statistical significance was determined by linear regression analysis, unless otherwise indicated. Results
The clinical characteristics of these 12 patients are shown in Table 1. The patients were divided into two groups, based on the serum free T response to GnRH. In 5 of the 12 patients, designated group A, free T levels increased following GnRH (Fig. IA). This increase in free T was significant over 8 hr (6.2 fold increase, p < 0.01) (Fig. 2). A single patient in this group (patient 5) also had an increase in A (baseline, 8.7 nmol/L; at 8 hr, 52 nmoI/L). She was the only patient in either group to have such an increase. Group B included seven patients who did not have an increase in any serum androgen, including free
12 Adolescent
Females
with Suspected
PC0
DHAS(FmollL) A(nmol/L) Pelvic US Ideal body T(nmolfL) LH(IU/L) (2.7-7.8) (<2.5cm) mass (%) (0.79-2.60) (3-20) LHIFSH (2.22-922)
Group A If4
2’ 3 4 59 Mean f SD
16 14 18 14 17
A A DUB A 0
7 7 27 5 18 12.8~8
15.8k1.8
III NL I III I 1.6-el.2
159 96 NA 157 142
2.22 2.01 2.36 7.42 3.26
138229
3.45k1.3
6.9 6.2 29.7 27.3 20.9
0.63 0.81 2.7 2.5 2.2
18.2ir7.0 1.7kO.6
4.29 3.90 9.80 10.90 6.0
9.5 NA 15.5 4.2 NA
7.o-c3
9.7-e6
3.1 3.3h 3.2 4.2 3.5
Group B ;
17 19
A 0
27 4
III II
196 150
3.26 3.50
21.2 18.5
1.6 1.7
3.58 5.72
8.5 3.4
c2.5 4.0
&p 9 10 11 12
15 15 19 16 18
DUB 0 0 0 NL
5 6 6 8 16
I II I II II
158 88 219 143 153
4.50 1.39 4.26 2.81 1.80
10.6 <2 9.9 36” 21
1.1 0.2 1.8 2.8 2.1
9.70 6.19 5.70 NA NA
12.0 NA 7.9 NA 6.3
c2.5 NA c2.5 NA NA
1.6kO.4 -
6.2rc-2
7.6k-3
Mean 2 L;D 17.Ok1.7
10.3. 8
1.820.7
158242
3.07kO.4 19.624.4
“Not shown are normal levels of 170HP, cortisol, PRL, Tl, and TSH. 6O = oligomenorrhea; A = amenorrhea; DUB = dysfunctional uterine bleeding; NL = normal. ‘Method of Ferriman and Gallwey (6). dGraded acne score I-III (7). ‘Longest axis ovarian length in cm as determined by pelvic ultrasound. ‘Baseline free T was elevated. 8ACI’II stimulation testing was normal. ‘Cystic ovarian change also identikied on pelvic ultrasound.
482
Free
2n
jCIURNAL OF ADDLESCENT HEALTH CARE VoI. 11 h’o. 6
FRANCIS ET AL.
-
T (pmol/L)
I
Free
T (pml/L)
,.___ _,._ ,@.......l.----- ----......,,-_.,..-..---*
1
I
I
a
4
2
,
(i
I
0
8
2
TIME AFTER GnAH (Hours} -
PATIENT
i
+
PAT.ENT 2
-a-
PATIENT
4
-*-
PAllENT
Hgwe
-k-
I
I
4
6
e
TIME AFTER GPW (Hcars) “’
PATIENT 3
s
*,
PUIENT
8
PATIENT 10
,+..
WTIENT
7
,-+,
PATIENT 8
6,
PATIENT
11
“&
PATIENT
a-
PATIENT
9
1'2
1. (AI individual free T response to GnRH in flue group A
@i&s. (B) hdiuidual free T response to GnRli in seven group B patients. Patients 8,9, and 12 M ~?eeT levels beh.u the level of assay msitiuiiyatalltime points.
T, following
GnRH (Fig. 16). This was significantly different from the change in free T levels seen over 8 hr in group A (p = 0.01) (Fig. 2). Although the androben response was different between groups, both had similar increases in LH, E,, and E2 (Table 2). Clinical observations and hormone measurements were compared between groups. Amenorrhea was more common in group A (315) than in group B (l/7), although this difference was not statistically significant. Enlarged ovaries (>2.5 cm longest axis as assessed hi pelvic ultrasound} (11) were identified in all patients in group A, one of whom had cystic changes. Of the four patients in group B who had pelvic ultrasound, only one had enlarged
Figure 2. Free T respme group B I +J. Data ure
to GnlW in patients in group A Ii@ and
expressed us mew
ovaries. This difference was also significant (p C 0.03, Fisher’s exact test). There were no differences between groups A and B in mean age, weight, acne, or hirsutism scores. Baseline hormone levels, including 17OHP, A, DHAS, T, free T, LH, and LHI FSH ratio were similar in both groups, as was the response to dexamethasone suppression--aU 12 patients had suppressed cortisol(~82.7 nmd/L, 3 pg! dl) and free T levels (~27.7 pmollL, 8 pglml). All patients were subsequently treated with a combination oral contraceptive. No other therapy for acne or hirsutism was prescribed. The results of the combined &xamethasone-GnRH stimulation test were unknown to both the patient and evaluating physician. After 2 years of treatment, nine patients (four from group A and five from group B) were available for follow-up. Clinical improvement, including decreased hirsutism, acne, and oily skin, as
T&h 2. Response of LH, E,, and E2 Following in Dexamethasone-Suppressed Patients”
+ SD.
Group A
Grerup B
Maximal LI-I (IUIL) 108.2 -c 88.4 67.9 5 1.7 E, (pm&L) at time (hr) 0 2 4 8
Er (pmol/L) at lime (hr) Q 0
a
I
I
2
4
i
6
TIME AFTER GnRH (Hours)
-
GROUP A
-
GROUP El
8
2 4 E
GnRH Probability p = 0.15
140 122 229 454
I? 50 2 40 rc_103 -c 100
172 176 218 322
-c f f it
33 54 22 66
NSb
139 155 293 667
rc 50 + 40 2 80 f 240
233 203 340 447
k f f 2
65 95 93 90
NSb
“Data are expressed as mean -r- SD. bNS = not significant.
November 1990
well as a reduction in basal androgen levels, was seen in four of four patients in grcup A and four of five patients in group B.
Discussion PC0 might be caused by abnormal androgen production in the ovary or adrenal glands or by
abnormal hypothalamic-pituitary regulation of gonadotropin secretion (l-3,5,12,13). Previous investigations of the etiology of PC0 have primarily evaluated adult women who may have had longstanding abnormalities of and perhaps secondary changes in the hypothalamic-pituitary gonadal axis. A variety of stimulation tests have been shown to increase T, free T, A, El, and E2levels in both normal women and women with PC0 (12-14). The women with PC0 had the greatest increase, suggesting an ovarian abnormality, but there was considerable overlap. A recent study, using G&I-I analog stimulation of the ovary, revealed increased 17OI-Wlevels only in women with PC0 (14). This suggests that the adult women with PC0 has an intrinsically different ovarian response to G&I-I compared to her normal counterpart. This, however, could be a primary abnormality or secondary to long-standing changes in the hypothalamic-pituitary gonadal axis. To our knowledge, our study is the first to exclusively evaluate adolescent females with suspected PCO. Abnormalities defined in this age group may be more likely to reflect the primary etiology of PC0 than in patients with long-standing disease. All 12 of our female adolescents had symptoms suggestive of PCO. Five patients (patients 3, 4, and 5 from group A and patients 11 and 12 from group B) had an LH/FSH ratio in excess of 2:l. This increased LH/ FSH ratio is sommonly but not universally seen in patients with PC0 (15). Nine patients were unequivocally hyperandrogenemis and dexamethasone-suppressible, with either a suppressed free T of c 27.7 pmol/L (8 pg/ml) or > 50% suppression of total T. This latter finding would typically be interpreted to suggest an adrenal abnormality but could be misleading due to the effects of dexamethasone on both the adrenal glands and ovary (5). The other three patients (patient 3 in group A and patients 9 and 12 in group B) may have had a subtle androgen abnormality not documented by our single measurements (4). In contrast, combined dexamethasone-GnRH stimulation allowed us to identify five patients who were distinctly different. Designated group A, they had an increase in free T but normal pituitary and
ANDROGENS IN POLYCYSTIC OVARY SYNDROME
483
estrogenis responses to GnlUI. They were more likely to have amenorrhea and enlarged or cystic . __ . _ ovaries. It is our opimon that these data support the presence of a primary ovarian zbnormality in this group. AS would have been expected, these patients responded favorably when treated with oral contraceptives. Patients in group B had no increase in androgen levels and were less likely to have amenorrhea or enlarged ovaries. It is our opinion that this is most consistent with an adrenal abnormality or idiopathic hyperandrogenemia. Although it was not expected, these patients responded equally well to oral sontraseptive therapy. It is possible that by decreasing the ovarian contribution to serum androgen levels, oral contraceptives may have been effective in this group of patients regardless of the underlying cause of hyperandrogenemia. In conclusion, these data provide two important observations. First, they suggest that early in the course of PCO, some adolescent patients (group A) had evidence for a subtle alteration in ovarian androgen production, whereas their pituitary and estrogenis responses to GnFtH were still normal. This observation supports the theory of a primary ovarian abnormality in this disorder. Second, regardless of whether or not this abnormality was present in any individual patient, almost all found oral contraseptive therapy to be readily acceptable and effective. Oral contraceptive therapy would appear to be a good first choice for the treatment of adolescents with suspected PCO. An evaluation of isolated ovarian function, such as that in this report, may have given insight into the etiology of PC0 but appears to be unnecessary in selecting individual patients who will respond favorably to oral contraceptive treatment. We thank Michael j. Brennan, LTC, MC, for his helpful review, andMr. Thomas Charles for assistance with preparancn ui this manuscript.
References 1. ConeyP. ~olycysticovarian disease:
current concepts of pathophysiology and therapy. Fertil Steril 1984;42:667-82.
2. McKenna T. Pathogenesis and treatment of syndrome. N Engl J Med 1988;318:558-62.
poly~ystic
ovary
3. Louhlin T, Cunningham S, Moore A, Culliton M, Smyth E, M&enna T. Adrenal abnormalities in pdyqstic Ovary syn&ome. J Clin Endocrinol Metab 1986;62:142-147. 4. Moll G, Rosenfield R. Plasma fret testosterone in the diagnosis of polycystic ovary syndru.lt,p. J Pediatr 1983;102 461-4. 5. Ettinger B, Goldfield E, Burrill K, werckr R, Forsham E.
484
7.
8. 9.
10. 11.
12.
13.
14.
FRANCIS ET AL.
Plasma testosterone stimulation-suppression dynamics in hirsute women. Am J Med 1973;54:195-200. Ferriman D, Gallwey J. Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 1%1;21:1440-7. Treadwell PA. The pilocebaceous unit, acne vulgaris, and acneiform eruptions. In: Kelly, ed . Practice of pediatrics. Philadelphia: Harper and Row, 1987;10:1-13. Parker CR Jr, Ellegood JO, Mahesh VB. Methods for multiple steroid radioimmunoassay. J Steroid Biochem 1975;6:1. Wright K, Collins DC, Reedy IRK. Comparative specificity of antisera against estrone, estradiol-178, and estriol using 6O-carboxymethyloxine bovine serum albumin derivatives. Steroids 1973;21:755. Buster ]E, Abraham GE. Radioimmunoassay of plasma dehydroepiandrosterone sulfate. Anal Lett 1977;5:543. Nakayama RT. The female reproauctive system. In: Kelly, ed. Practice of pediatrics. Philadelphia: Harper and Row, 1987;7:1-14. Skaf R, Shelder R, Kemman E. Androgenic response in anovulatory women during menotropins stimulation. Obstet Gynecol1981;58:714-719. Heineman M, Thomas C, Doesburg W, Polland R. Pituitary and ovarian responses to LHIUI stimulation in women with clinical features of the polycystic ovary syndrome, Eur J Obstet Gynecol Reprod Biol1984;17:273-284. Barnes RB, Rosenfield RL, Burstein S, Ehrmann DA. Pitt&
JOURNAL OF ADOLESCENT HEALTH CARE Vol. 11, No. 6
responses to nafarelin testing in the polycystic ovary syndrome. N Engl J Med 1989;320:559-565. 15. ?4urdoch AP, Dunlop W, Kendall-Taylor P, Watson MJ. The acute effects of a dopamine antagonist (domperidone) on luteinising hormone, follicule stimulating hormone, prolactin and thyrotrophin secretion in polycystic ovarian syndrome. Differential effect of ovulation. Clin Endocrinol 1984;21:611619. tary-ovarian
This manuscriptwas reviewed by a total of four reviewers. The consensus was that it was an original and clever way to try to elucidatethe pathogenesis of PC0 in adolescents.On the other hand, the small samplesize, lackof uniformityof data collectionand fact that all patients responded to oral contraceptives, despite subgroup assignment, were troublesome. Recognizingits limitations,it was accepted for publicationto stimulate further research and to allow the reader to judge for him/herself. Iris F. Litt, M.D., Editor-In-Chief