Clinical Endocrinology (1988), 28,93-107

EPISODIC PULSATILE SECRETION OF FSH,

LH, PROLACTIN, OESTRADIOL, OESTRONE,
AND LH CIRCADIAN VARIATIONS IN
POLYCYSTIC OVARY SYNDROME

S. VENTUROLI, E. PORCU, R. FABBRI, 0. M A G R I N I , L. G A M M I ,

R . PARADISI, M . FORCACCI, R. BOLZANI A N D C. F L A M I G N I

Institute of Reproductive Physiology and Pathology, University of Bologna, and

Department of Ophthalmology, University of Modena, Italy

(Received21 April 1987; returnedfor reuision 24 June 1987:finally revised30 July 1987; accepted 13 August 1987)

SUMMARY

Pulsatile secretion of LH, FSH, PRL, oestradiol and oestrone was studied in a
group of 16 patients with micropolycystic ovary syndrome (PCOS) and
compared with that of normal ovulatory women in the fifth to sixth day of the
cycle. Hormone concentrations were measured at 10 min intervals for 8 h
starting at 0930 h. In seven subjects, the study was prolonged far 24 h, with 20
min interval samples, in an attempt to evaluate the circadian rhythm of LH by
cosinor analysis. Significant fluctuations occurred in the concentration of each
hormone. Values shown are mean fSD. PCOS subjects had high LH mean
values (27.9f5.9 IU/1) (P<0.005). LH pulse amplitude was higher than
controls (1 1*6f3.7 IU/l versus 5.2 & 1.8 IU/l; P < 0.005) while no consistent
changes in frequency or interpulse interval (62.0f 10.7 rnin versus 65.8 IfI 19.2
min; P = NS) were found. A mean of 4.8 f-1.2 pulses of FSH occurred in 8 h and
the mean pulse amplitude was 2.68 k 1.1 I with no differences from controls. All
patients were normoprolactinaemic. A mean of 5.5 f 1.9 pulses occurred in 8 h,
the interpulse interval was 76.1 f 14.4 min and the amplitude was 2.87L-0.76
ng/ml and there were no significant differences from controls; 75% of PRL
pulses showed a temporal relationship with LH pulses. Oestrone mean basal
values were higher in PCOS (47.2 f 12.5pg/ml) than controls (32.0 f9.9 pg/ml;
P < 0.02), while no differences were observed as regards oestradiol. Oestradiol
pulse amplitude was nearly the same as oestrone (43.6 f 18.8 pg/ml and
37.7 f 16.1 pg/ml, respectively); 6.0 f2.2 pulses and 6.0 f 1*6pulses occurred in
8 h with an interpulse interval of 81.1&27.1 min and 71.8+11.1 min,
respectively. Sixty-five per cent of LH pulses were followed by an oestradiol and
oestrone peak. The mean time of the appearance was 17 f 15 rnin and 25 23
min, respectively. In the PCOS group a consistent 24 h rhythm in mean plasma
Correspondence: Stefano Venturoli, M.D., Servizio di Fisiopatologia della Riproduzione Clinica Ostetrica
Ginecologica, Universita di Bologna, Via Massarenti 13, 1-40138 Bologna, Italy.

93
 S. Venturoli et al.
LH levels was found with the highest hormone values at 1720 h (P<0.05)
unrelated to apparent sleep and different from that of adult women. Pulse
frequency showed a significant slowing during the night with the longest
interpulse interval at 0327 h ( P -= 0.03) while no significant periodicity was
observed in LH pulse amplitude. This study confirms the normal pulsatile
secretion of FSH, PRL, oestrone and oestradiol in PCOS, and the deranged LH
pulsatile pattern displayed in the form of enhanced pulse amplitude and
abnormal circadian periodicity.
Even though recent suggestions for altered gonadotrophin and steroid pulsatile secretion
in micropolycystic ovary syndrome (PCOS) do exist (Rebar et al., 1976; Wentz et al.,
1976; Molloy et al., 1984), only few, statistically proven data, are currently available on
gonadotrophin (Burger et aE., 1985; Waldstreicher et al., 1986) and prolactin (PRL)
(Laatikainen & Tulenheimo, 1985; Murdoch et al., 1986) pulsatility in the syndrome,
whereas no detailed information has been published on oestrogen behaviour.
Conflicting criteria on the selection of PCOS subjects and divergent experimental
methods and pulsatility analysis may account for the failure to define the endocrine
episodic and circadian pattern of patients suffering from PCOS.
The present study was undertaken to determine: (1) Episodic secretion of FSH, LH,
PRL, oestrone, oestradiol; and (2) circadian variations of LH, in a well defined group of
subjects with PCOS.

PATIENTS A N D ME T H O D S

Subjects
Sixteen young women aged between 18 and 28 years, with a well-established clinical and
laboratory diagnosis of chronic anovulation and PCOS, according to criteria previously
described (Paradisi et al., I986), volunteered for this investigation. The subjects were
suffering from amenorrhoea or severe oligomenorrhoea (over 50 days) of perimenarchal
onset, and hirsutism. Eleven subjects had increased body weight; body mass index mean
values were 26.2 fSD 5.7 (normal range 18-24).
They had elevated serum LH concentrations (2 SD higher than normal adult mean
values), LH: FSH ratio above 3 and high levels of testosterone and androstenedione.
None of the subjects had clinical or laboratory signs of adrenal pathology (Venturoli et
al., 1980). All patients had enlarged and microcystic ovaries as diagnosed by ultrasound
(Orsini et al., 1985). Twelve out of 16 patients had the diagnosis confirmed by
laparoscopy .
Eight normal women aged 18-30, who had regular menstrual ovulatory cycles of 27 to
32 days duration in the previous 3 years, and without physical or psychological disease
served as a control group. Only women with a biphasic basal body temperature chart, a
luteal phase duration of at least 12 days, and mid luteal plasma progesterone levels greater
than 7 ng/ml were included.

Study design
None of the subjects had received any medications for at least 3 months before the study
which was performed on the fifth to sixth day of the menstrual cycle in normal and
 Pulsatile hormone secretion in PCOS 95
oligomenorrhoeic PCOS subjects, and on a day chosen a t random in amenorrhoeic
subjects.
An indwelling i.v. catheter was inserted into a forearm vein and kept patent with
heparinized 0.9% saline solution.
Blood samples (5 ml each blood sample) were collected every 10 min for 8 h, starting at
0930 h for the evaluation of pulsatile secretion of LH, FSH, PRL, oestrone and
oestradiol. LH and FSH pulsatile secretion was evaluated in all patients; PRL pulsatility
was studied in seven and oestrone and oestradiol fluctuations in eight.
Seven out of 16 PCOS patients continued the study for a 24 h period with blood
sampling every 20 min to evaluate the circadian LH pattern.
During the night-time sampling procedure, a sleep record was kept by a trained
observer, although polygraphic recordings were not performed.
In all subjects additional blood samples were drawn on the day of the study to
determine sex steroids.

Hormone assay
All samples from each subject were run in the same radioimmunoassay for LH, FSH,
PRL, oestrone and oestradiol.
The radioimmunoassay techniques used for hormonal measurements were: gonadotro-
phins (FSH-LH) and prolactin (PRL) by rapid double-antibody (kits purchased from
Biodata, Rome, Italy); 17a-hydroxyprogesterone (17-P), and progesterone (P) by
chromatographic separation on Sephadex LH-20 columns; dehydroepiandrosterone
sulphate (DHA-S) directly on diluted plasma, testosterone (T), 5a dihydrotestosterone
(DHT), androstenedione (A), oestrone (El), and oestradiol-17P (E2) by TLC on silica gel
60 F254 as previously described (Paradisi et al., 1986).
Intra- and inter-assay variability (coefficients of variation (CV)) and sensitivity of the
method were: 3.7%, 4.2% and 1.5 IU/I for FSH; 4.5%, 5.1% and 1.5 IU/l for LH; 3.6%,
6.0% and 1.5 ng/ml for PRL; 5.0%, 5.2% and 10 pg/tube for 17-P; 11.2%, 16.5% and 15
pg/tube for P; 3.6%, 5.6% and 15 pg/tube for DHA-S; 4.5%, 9.3% and 15 pg/tube for T;
8.3%, 13.0% and 15 pg/tube for DHT; 7.1%, 10.5% and 12 pg/tube for A; 3.3%, 5.9%
and 3 pg/tube for El; 5.1 %, 5.9% and 2.6 pg/tube for E2.

Quantitative analysis of pulsatile hormone secretion and 24 h LH periodicity

Pulsatile LH, FSH and PRL patterns were analysed by applying a modification of the
method of Santen and Bardin (1973). A LH and FSH pulse was defined as an increment in
plasma levels of more than 1 IU/I which exceeded the previous nadir by at least 20% and
characterized by more than two consecutive values. In the assays described, a 20%
increment corresponded to approximately three to four times the intraassay C.V.
A PRL pulse was defined as a rise over successive samples greater than 20% which
corresponded to four to five times the intraassay C.V.
The interpulse interval represents the time between two subsequent peaks and was
calculated individually from pulse to pulse in all subjects.
The amplitude of a pulse was defined as the difference between the highest value in the
pulse and the preceding nadir.
Pulsatile oestrogen (oestradiol and oestrone) pattern was analysed according to the
method of Backstrom et al. (1982) which defines a pulse as occurring when the hormone
96 S . Venturoli et al.
Table 1. Baseline plasma hormone values (mean+SD) of PCOS and control
subjects

PCOS ( I 6) Control (8)t

~~ - ..-~~ ~~~

FSH (IUjl) 9.1 i.2.7 9.6 4.9

LH (IUjl) 30,9*9.6*** 10-lk3.7
LH: FSH *
3.8 1.9** 1.2k0.5
P R L (ngiml) 12.5f7.1 12.6 6.1
Oestradiol (pgjml) 49.3 26.4 45.6 f 18-9
Oestrone (pgjml) 48.2 & 16* 34.2 f I2
Oestrddio1/0estrone 1 .O & 0.4 1.3 k 0 . 7
Testosterone (ngidl) 61,9* 32.6** 2 7 3 & 1 1.2
Scc-dihydrotestosteronc (ng/dl) 2 2 . 7 1 8.8 15.0t4.8
Androstenedione (ngjdl) 2x1 & l12** 114k23
DHA-S (pg/ml 3.13& 1.43 3.12f 1.04
17a-hydroxyprogesterone (ngiml) *
0.94 0.4 I 0.51 f 0 . 1 5
0.25 k 0.04
Progesteronc (ngiml) 0.35 10.03

Conversion factoi-s: PRL ng/ml x 20=mU/l; Oestradiol pglml x 3.67 =

pmol/l; Oestronc pg/ml x 3.69 = pmol/l; tcstostcrone ng/dl x 0.0347 = nmolil;
5ctDIIT ng/dl x 0,0344 nmol/l; androstcncdionc ng/dl x 04349 -nmol,l; 17-
OH P ng/ml x 3.03 =nniol/l; progesterone ng/ml x 3.18 = nmoljl.
* P < 0 4 5 . ** P i 0 - 0 1 . *** P<0~001.
t Fifth to sixth day of the cyclc.

concentration of two consecutive samples was greater than that of the mean of the two
previous samples (basal value) and when the value of at least one of the peak samples
exceeded the mean basal value by more than twice the coefficient of variation of the assay.
To determine the relationship between the pulses of each gonadotrophin, the number
of coinciding pulses was calculated by counting the number of pulses whose peaks
coincided. The relationship of gonadotrophin pulses to those of oestrone and oestradiol
was analysed by determining the number of responsive pulses. Considering the latent
period between the response of the gonad to gonadotrophins (Baird, 1978), a
gonadotrophin pulse was arbitrarily regarded as responsive only if a pulse of steroid
occurred within 60 min of the start of the gonadotrophin pulse. The time between the
appearance of LH and eostrogen peaks was then evaluated.
The 24 h variations in overall (mean) plasma LH concentrations, LH interpulse interval
and LH pulse amplitude were evaluated by determining the best-fitting curve by the least
square technique, specially adapted, under the term of cosinor, to the analysis of
biological rhythms (Halberg et al., 1965; Bingham et ul., 1982). To analyse and describe
periodic data, the cosinor uses the cosine function:
g ( t ) = M + A cos (cot+$)
with g(t) the value at time t of the periodic function of angular frequency o (in degrees per
time unit, with 360=a complete cycle), defined by parameters M (mesor), A (amplitude)
and # (acrophase). In circadian studies the period is assumed to be equal to 24 h.
 Pulsatile hormone secretion in PCOS 97

0 1 2 3 4 5 6 7 8

0 1 2 3 4 5 6 7 8
1 1 1

30
25
20
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
Hours Hours
Fig. 1. Examples of individual spontaneous patterns of LH secretion over 8 h, in patients with
PCOS. Arrows indicate the peak of the pulse.

The amplitude of the cosinor indicates the amount of the fluctuation. The phase of the
cosinor indicates the time of maximum fluctuation. Mesor is the mean value of the data
analysed. Single cosinor and group cosinor analysis were used to evaluate individual
variations and rhythmicity in the groups, respectively.
The null hypothesis that cosinor amplitude was zero was tested and the F value
evaluated.
Results concerning basal gonadotropin and steroid levels and LH pulse amplitude and
interpulse interval are expressed as mean iSD.
Kolmogorov test for evaluation of normal or log-normal distribution did not show
satisfactory approximation. The Mann-Whitney U test was used to determine the
significance of differences between groups (Siegel, 1956).

RESULTS

Hormone concentrations during the pulsation study (Table I )

We found significantly higher mean values of LH, testosterone, androstenedione and
oestrone in PCOS subjects with respect to controls; normal mean values were found as
regards FSH, PRL, oestradiol, DHA-S and 17 hydroxy-progesterone.
LH:FSH ratio was 3.8k 1.9 in PCOS versus 1.2kO.5 in controls (P<O.OOl) and
oestradiol: oestrone ratio was 1.O k0.4 versus 1.3 Ifr 0.7.
98 S. Venturoli et al.

- 11
I, If
I 1 1 I I I I I
7 0 1 2 3 4 5 6 7 8
-3. . . * . .. 0 1 2
1 . 1
3
1
4
1
5
' 1
6
I
7
1
8
LL
1 r . * . .
-E
8 5

1 1 1 1 L L I I I I I I I J
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5
a 25r

0 1 2 3 4 5 6 7 8
Hours
LL I I I I I I I I
0 1 2 3 4 5 6 7 8
Hours

Fig. 2. Examples of individual spontaneous patterns of PRL and FSH secretion over 8 h, in
patients with PCOS. Arrows and asterisks indicate the peak of the PRL and FSH pulse,
respectively.

Characteristics of the episodic pulsatile secretion (Table 2)

Mean LH plasma values over the 8 h LH period were 27.9 k 5.9 IU/l in PCOS versus
10.9 k2.9 IU/1 in controls (Pc 0.005). The degree of LH pulsatility varied a great deal in
different subjects as regards number of pulses: pulse amplitude was generally high, but far
from uniform within the period of study (Fig. 1). Comparing the PCOS group versus
controls, the women with PCOS showed a raised LH pulse amplitude (11-6t-3.7 IUjl
-=
versus 5.2 k 1.8 IU/l, P 0.00.5) while no consistent changes in frequency or interpulse
interval were found.

FSH
Mean FSH plasma values over 8 h were similar in PCOS and in controls.
The exact pattern of episodic FSH secretion is difficult to estimate, due to the low
amplitude of the fluctuations of this hormone and the poorly defined outline of the pulse
 Pulsatile hormone secretion in PCOS 99
Table 2. Pulsatile characteristics of LH, FSH, PRL, oestradiol and oestrone in PCOS and control groups

Interpulse
Mean k SDt No. pulses Amplitude interval (min)

LH (IU/l) PCOS (16) 27.9 15.9 8.3f 1.4 1l.6f 3.7 62.0f 10.7
Control (8) 10.9+2.9** 7.3 f 1.7 5.2+ 1.8** 65.8 f 19.2
FSH (IU/I) PCOS (16) 8.3 2.3 4.8+ 1.2 2.68_+1.11
Control (8) 8.7 13.5 3.5k 1.9 2.3210.71
PRL (ngiml) PCOS (7) 9.3 k4.5 5 5 f 1.9 247 0.76 76.1 f 14.4
Control (8) *
9.3 3.4 6.6f 1.5
6.0 f2.2
3.7k0.9
43.6f 18.8
*
74.1 19.2
81.1 k27.1
Oestradiol (pg/ml) PCOS (8) 48.91 19.5
Control (8) 45.9 k 19.4 6.3k1.9 37.9 f 10.4 63.4+ 14.9
Oestrone (pg/ml) PCOS (8) 47.2+ 12.5 6,0+ 1.6 37.71 16.1 71.8 _+ 11.1
Control (8) 32.0 k 9.9* 7.0k 1.0 39.0+ 11.5 63.1k11.5

* Pc0.02. ** PiO.005. t Values over a period of 8 h. Conversion factors as for Table I .

I I I I I I I I I I
0 1 2 3 4 5 6 7 6
Hours
Fig. 3. Example of concomitance between LH: PRL pulses in one subject with PCOS. Arrows
indicate the peaks of each pulse: (+)indicates the coincident LH: PRL pulses in 8 h.

curve which generally spans a variable number of samples (Fig. 2). In addition no regular
pulse rhythms were observed during the period of the study making the interpulse interval
difficult to calculate. We did not find significant differences in the PCOS group and in
controls as regards amplitude and number of pulses (Table 2).

PRL
Mean PRL plasma values over 8 h were the same in PCOS and in controls.
All patients examined were normoprolactinaemic, and the secretion pattern of PRL
100 S . Venturoli et al.
was pulsatile in all cases, with some variability in frequency, amplitude and rhythm of the
pulses (Fig. 2). There were only slight differences as regard PRL pulse characteristics in
PCOS subjects with respect to controls (Table 2).
There was a temporal relationship between spontaneous LH and PRL pulses in all
subjects studied as identified by the correspondence of the peaks (Fig. 3).
In the patients we found a 75% concomitance between PRL and LH pulses and 45%
concomitance between LH and PRL, which reflects the greater number of LH pulses than
PRL ones.
Control subjects showed slight differences from PCOS, as regards the percentage of
concomitance between PRL and LH pulses (PRL versus LH 60%, LH versus PRL 38%).

Oestradiol and oestrone

Oestradiol mean plasma values over 8 h were similar in PCOS and in controls (Table 2),
while oestrone mean values were significantly different from controls (47.2 & 12.5 pg/ml
versus 32.0 k 9.9 pg/ml) ( P < 0.02).
Evident fluctuations characterized both oestradiol and oestrone behaviour in PCOS
(Fig. 4), though some differences existed between the two hormones. A very low
percentage (25.5%) of oestrone and oestradiol pulses appeared to be concomitant.
In the PCOS group, amplitude of oestradiol pulse, number and interpulse interval were
nearly the same as oestrone (Table 2).
Comparing the PCOS subjects with controls, the pulsatile behaviour of oestradiol and
oestrone did not show differences, though we observed a shorter, but not significant,
interpulse interval in normal subjects (Table 2).
A definite relationship between fluctuations of oestrogens and pulsatile secretion of LH
is very difficult to establish (Fig. 5). However, in PCOS 63.3%)of LH pulses were followed
by an oestradiol peak and 69.1% by an oestrone pulse within 60 min. The mean time of the
appearance of oestradiol peak was 16.8 _+ 15.4 and that of oestrone was 25.1 k23.1 min.
Control subjects showed the same behaviour as regards the relationship between LH and
oestrogen pulses, and no differences were recorded concerning percentage of concomi-
tance (LH versus oestradiol 62%, LH versus oestrone 64.4%) or time of appearance
(oestradiol versus LH 15.3_+ 18.9 min; oestrone versus LH 20.3 _+ 14.5 min).

Churacteristics of LH circadian rhythm (Table 3 )

Six out of the seven subjects studied had significant LH circadian changes as regards
overall LH values. Three subjects had a significant interpulse interval, while only two had
significant changes in LH pulse amplitude (Table 3). Figure 6 shows an example of a 24 h
periodicity as regards the highest LH values which occurred in the afternoon; in this
subject, the nocturnal slowing of pulse frequency is also evident.
Analysis of the group data revealed a significant degree of synchrony on the overall LH
data ( P < 0.05) and interpulse interval ( P < 043) rhythms. The highest hormone values
were found in the afternoon and the mean acrophase of the group was at 1720 h. Pulse
frequency showed a slowing during the night and the longest mean interpulse interval as
mean of the group was at 0327 h.
No evidence was found for the presence of a 24 h rhythm in LH pulse amplitude when
 Pulsatile hormone secretion in PCOS 101

Hours Hours

Fig. 4. Patterns of episodic oestrone and oestradiol secretion in three subjects with PCOS:
representative examples of an 8 h period. Arrows indicate peaks of the pulses.

I:,
2
v

-
25

-5 120
-
4

40

0 1 2 3 4 5 6 7 8
Hours
Fig. 5. Relationship between LH, oestrone and oestradiol pulses in one subject with PCOS.
Concomitance among pulses is very difficult to establish.
 L
0
h)

Table 3. Cosinor analysis of LH secretory pattern (24 h) in patients with polycystic ovary syndrome

LH pulse amplitude LH interpulse interval Overall mean plasma LH data

Cosinor Cosinor Mean Cosinor Cosinor Cosinor Mean interpulse Cosinor Cosinor Cosinor Cosinor
Amplitude phase amplitude signif. amplitude phase interval signif. amplitude phase Mean LH signif.
Patients (IUjl) (time of day) (IUjl) (F) (min.) (time of day) (min.) (F) (IU/I) (timeofday) (IUjl) (F)
9
A 1.64 2032 6.25 1.27 0617 87 1.17 1547 17.14
B 2.08 0940 9.25 0.45 2331 63 2.35 1334 33.98 3
C 2.38 0334 9.19 2.56 0255 102 t 2.62 1640 18.33
$ 2
7 %
D 5.63 0030 12.49 7 0.94 0332 78 3.00 2119 22.8 1
E 2.86 0120 8.22 f 0.94 0345 78 4 15 2 2231 16.36 I %
F 0.20 1531 5.85 0.76 1813 80 1.24 1421 25.02 * 9
G 2.85 1320 1464 0.63 0522 63 * 5-42 1622 31.17 7 %
Group mean 0.94 0124 9.41 0.79 0327 79 t 1.85 1720 23.54 *

* P < 0.05.
P < 0-03.
$ P < 0.02.
5 P<O.OI.
y P<O.OOl.
 Pulsatile hormone secretion in PCOS 103
Sleep

50-

40-

-- 30 -
2
-
)

I
.
.I

-1
20 -

10 -

OLO930h 1i30h V30h d30h OliOh 0530h 0930h

Fig. 6. Twenty-fourhour profile of LH secretionin one subject with PCOS.The highest LH values
are apparent at 1600h: the cosine curve has been drawn. A nocturnal slowing of pulse frequency is
also evident with the largest interpulse interval at 0500 h.

the grouped data were analysed (Table 3). There was no clear relationship between
apparent sleep and changes in LH secretory profiles.

DISCUSSION
Our data confirm the reports of other authors (Burger et al., 1985; Waldstreicher et al.,
1986), indicating that the high circulating LH levels, typical of PCOS are largely
maintained by an exaggerated pulsatile discharge of LH, displayed in the form of
enhanced amplitude, with respect to gonadotrophin pulsatile secretion in normal women.
A generally increased frequency of LH pulse, has also been reported by several authors in
studies referring to a few hours and without statistical analysis (Rebar et al., 1976; Wentz
et al., 1976; Molloy et al., 1984) or in comparison with a random day of the follicular
phase of a normal cycle (Burger et al., 1985). We did not find increased LH frequency
(Table 2); however, the differences are much less striking when the LH pulse frequency of
PCOS subjects is compared with that of normal adults in the mid- and late follicular phase
of the cycle (Waldstreicher et al., (1986).
Recent studies by Crowely et al. (1983) showed that the appropriate sequence of
gonadotrophin amplitude and frequency changes are essential for follicular development
and any deviation from the usual progression may lead to anovulation. Data on the day to
day variations of the pulsatile pattern in PCOS, are not yet available, but these patients
104 S . Venturoli et al.
may be unable to progressively increase gonadotrophin pulse frequency to drive normal
follicular development.
FSH secretory pattern did not substantially differ from that of controls. Although
pulsatile FSH release may occur in PCOS (Fig. 2) we were unable to prove it, thus
confirming the difficulty in evaluating FSH pulsatility.
Discordant data have been reported on PRL levels and the incidence and the role of
hyperprolactinaemia in PCOS (Buvant et al., 1982; Corenblum & Taylor, 1982; Carmina
ef a]., 1984, Shoupe & Lobo, 1985).
All our patients had normal prolactin concentrations. The fact that the prolactin
secretion is susceptible to a number of factors such as stress and diet (Quigley et al.,
1981a), probably accounts for the difficulty in identifying the pulses correctly. However,
the secretion pattern in our patients was pulsatile in all subjects with a certain degree of
variability in frequency and rhythm (Fig. 2).
No differences exist in our PCOS group with respect to normal females in terms of mean
PRL values and pulsatile behaviour, confirming normal prolactin function in normopro-
lactinaemic PCOS subjects and excluding a specific impairment of lactotroph function
(Laatikainen & Tulenheimo, 1985; Murdoch et al., 1986).
A relationship between lactotroph function and gonadotropin function is frequently
postulated in PCOS: some synchronizations among neuroendocrine mechanisms con-
trolling LH and PRL release exist in females (Cetel & Yen, 1983; Braund et al., 1984) and
the temporal relationship we found between spontaneous PRL and LH pulses, with a high
degree of concomitance between hormone peaks, seem to confirm this hypothesis in
PCOS (Laatikainen & Tulenheimo, 1985; Murdoch et al., 1986).
Despite the synchronization of mechanisms controlling LH and PRL release in PCOS
which both involve the dopaminergic system (Quigley et al., 1981b; Paradisi et al., 1985)
and/or the chronic unopposed high oestrogen levels of this condition (Rebar et al., 1976;
Baird et al., 1977), the impaired LH release typical of the syndrome is probably due to
mechanisms other than those governing normal prolactin release (Falaschi et al., 1986).
It is generally accepted that the oestrogen concentrations in normal women fluctuate
markedly from hour to hour (Korenman & Sherman, 1973; Lenton et al., 1978). A
relationship has been demonstrated in normal women (Backstrom et al., 1982) between
those fluctuations and pulsatile LH secretion. On the contrary, the increase in oestrogens,
particularly of oestrone levels in PCOS subjects, is commonly accepted to derive mainly
from extraglandular production as a consequence of increased peripheral conversion of
androgens (Siiteri & McDonald, 1973; Longcope et al., 1978). There are many technical
difficulties in identifying an eostrogen pulse, close to the precision of the assay. However,
in PCOS quite evident fluctuations characterize oestradiol and oestrone which generally
have the same amplitude (Table 2). Oestrone and oestradiol pulses do not coincide,
suggesting partly different mechanisms of regulation and secretion of the two hormones
in PCOS.
The pulsatile behaviour of oestradiol and oestrone in PCOS does not show differences
from controls in the early-mid follicular phase even though oestradiol pulses, were more
frequent in normal females. In particular, we observed the same pulse amplitude and
frequency of oestrone in normal and PCOS subjects, even though the mean plasma values
of oestrone were higher in the latter group. These data emphasize that the production of
oestrogens is pulsatile in PCOS and, considering the temporal relationship with LH
pulses, probably functionally related to LH pulsatile behaviour (no data exist for
 Pulsatile hormone secretion in PCOS 105
comparison). It is likely that the extraglandular non-pulsatile production of oestrone
partly contributes to the higher plasma levels of this hormone typical of the syndrome.
The circadian study shows a 24 h rhythm of the LH secretory pattern in patients with
PCOS. The frequency of LH secretory episodes, which is not significantly different from
that of controls during the day (Table 2), shows the same nocturnal slowing described by
Soules et al. (1985) in normal adults in the early follicular phase.
Enhanced pulse amplitude, which seems the main factor responsible for the high LH
levels, does not show any appreciable variation in the 24 h, being elevated to the same
extent during both day and night (Table 3).
The remarkable finding is that the overall LH data present a significant rhythm in the
24 h with the highest levels in the afternoon (Fig. 6) and an LH profile different from that
of normal adults (Alford et al., 1973; Kapen et al., 1983; Soules et al., 1985). This confirms
and adds weight to the previous observations of Zumoff et al. (1983) in young subjects
with PCOS.
Although the biological significanceof circadian rhythms remains unknown, in the case
of LH, circadian pulsatility together with episodic LH pulsatility, could conceivably
promote a rhythmic ovarian activity. Disturbed circadian rhythms probably account for
pathological conditions and may contribute towards chronic anovulation.
Since the normal mid-cycle rise in LH initiating ovulation usually begins during sleep
(Kapen et al., 1983), Zumoff et al. (1983) hypothesized that desynchronization of the LH
clock from the sleep period may result in failure of the surge of LH and prevent
ovulation. However, it is difficult to believe that this is the mechanism of anovulation in
PCOS. High nocturnal levels do not seem to be a prerequisite for ovulation. The
beginning of the LH surge during the night does not require high nocturnal LH levels to
help this event, considering that high nocturnal LH values do not exist in normal adults
during the mid- and late-follicular phase of the menstrual cycle. Yet it cannot be excluded
that an abnormal pattern with higher diurnal LH levels could in some way upset the LH
driven follicle maturation and steroidogenesis.
Recent studies (Quigley et al., 1981b; Paradisi et al., 1985)indicate the presence of a low
dopaminergic inhibitory tone responsible for the enhanced pulsatile LH secretion in
PCOS: our data support the hypothesis that, in addition to a low dopaminergic inhibitory
tone, an abnormal circadian control of LH also exists. The association of abnormalities in
both the ultradian and circadian signal strongly points to the central nervous system as
the main site responsible for the onset and the persistence of the syndrome.
In our knowledge, similar episodic (Venturoli et al., 1986a) and circadian (Porcu et al.,
1986) LH patterns have only been observed in anovulatory cycles of adolescents with high
LH values, who are in fact a risk group for the development of PCOS in adolescence
(Venturoli et al., 1986b,c).

REFERENCES
ALFORD,F.P., BAKER, H.W.G., PATEL,Y.C., RENNIE, G.C., YOVALT,G . ,BURGER, M.G. & HUDSON,B. (1973)
Temporal patterns of circulatinghormones as assessed by continuous blood sampling. Journal of Clinical
Endocrinology and Metabolism, 36, 108-1 16.
BACKSTROM, C.T., MCNEILLY, A S . , LEASK,R.M. & BAIRD,D.T. (1982) Pulsatile secretion of LH, FSH,
prolactin, oestradiol and progesterone during the human menstrual cycle. Clinical Endocrinology, 17,
2942.
BAIRD,D.T., CORKER, C.S.,DAVISON, D.W.,HUNTER, W.M., MICHIE, E.A. & VAN LOOK,P.F.A. (1977) Pituitary
106 S. Venturoli et al.
ovarian relationships in polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism, 45,
798-809.
BAIRD,D.T. (1978) Pulsatile secretion of LH and ovarian estradiol during the follicle phase of the sheep estrous
cycle. Biology of Reproduction, 18, 359--364.
BINGHAM, L., ARBOGAST, B., GUILLAUME, G.C., LEE,J.K., HALBERG, F. (1982) Inferential statistical methods for
estimating and comparing cosinor parameters. Chronobiologia, 9, 397439.
BRAUND, R., RODGER, D.C. & JUDD,S.J.(1984) Synchronous secretion of luteinizing hormone and prolactin in
the human Iuteal phase: neuroendocrine mechanisms. Journal of Clinical Endocrinology and Metabolism,
58,293-297.
BURGER, C.W., KORSEN, T., VAN KESSEL, H., VAN DOP,P.A., CARON, F.J.M. & SCHOEMAKER, J. (1985) Pulsatile
leuteinizing hormone patterns in the follicular phase of the menstrual cycle, polycystic ovarian disease
(PCOD) and non PCOD Secondary amenorrhea. Journal of Clinical Endocrinology and Metabolism, 61,
1126-1132.
BUVAT, J., SIAME-MOUROT, C., FOURLINNIE, J.C., LEMAIRE, A., BUVAT-HERBAUT, M. & HERMAND, E. (1982)
Androgens and prolactin levels in hirsute women with either polycystic ovaries or borderline ovaries.
Fertility and Sterility, 38, 695-700.
CARMINA, E., ROSATO,F., MAGGIORE, M., GAGLIANO, A.M., INDOVINA, D. & JANNI,A. (1984) Prolactin
secretion in polycystic ovary syndrome: correlation with the steroid pattern. Acta Endocrinologica, 105,
99-104.
CETEL,N.S. & YEN,S.S.C. (1983) Concomitant pulsatile release of prolactin and leuteinizing hormone in
hypogonad women. Journal of Clinical Endocrinology and Metabolism, 56, 1313-1 3 15.
CORENBLUM, B. &TAYLOR, P.J. (1982) The hyperprolactinemic polycystic ovary syndrome may not be a distinct
entity. Fertility and Sterility, 38, 549-552.
CROWLEY, W.F., FILICORI,M., SPRATT,D.I. & SANTORO, M.F. (1985) The physiology of gonadotropin-
releasing hormone (GnRH) secretion in men and women. Recent Progress in Hormone Research, 41,
471-526.
HALBERG, F., ENGELI, M., HAMBURGER, C. & HILLMAN, D. (1965) Spectral resolution of low-frequency, small-
amplitude rhythms in excreted 17-ketosteroids; probable androgen induced circaseption desynchroniza-
tion. Acta Endocrinologica, (Suppl. 103), 5-54.
FALASCHI, P., ROCCO,A. & DEL POZO,E. (1986) Inhibitory effect of bromocryptine treatment on luteinizing
hormone secretion in polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism, 62,
348-351.
KAPEN,S., BOYAR, R., HELLMAN, L. & WEITZMAN, E.D. (1973) Episodic release ofluteinizing hormone and mid-
menstrual cycle in normal adult women. Journal of Clinical Endocrinology and Metabolism, 36,72&729.
KORENMAN, S.G. & SHERMAN, B.M. (1973) Further studies of gonadotropin and estradiol secretion during the
pre-ovulatory phase of the human menstrual cycle. Journal of Clinical Endocrinology and Metabolism, 36,
1205-1 209.
LAATIKAINEN, T. & TULENHEIMO, A. (1985) Prolactin pulsatility in polycystic ovarian disease. Journal of
Endocrinological Investigation, 8, 151-1 61.
LENTON,E.A., COOKE,I.D., SAMPSON, G.A. & SEXTON,L. (1978) Episodic secretion of oestradiol in
premenopausal women, men and post-menopausal women. Clinical Endocrinology, 9, 37-47.
LONGCOPE, C., PRATT,J.H., SCHNEIDER, S.H. & FINEBERG, S.E. (1978) Aromatization of androgens by muscle
and adipose tissue in vivo. Journal of Clinical Endocrinology and Metabolism, 45, 146-152.
MOLLOY, B.G., EL SCHEIKH,M.A.A., CHAPMAN, C., OAKEY,R.E., HANCOCK, K.W. & GLASS,M.R. (1984)
Pathological mechanisms in polycystic ovary syndrome: modulation of LH pulsatility by progesterone.
British Journal of Obstetrics and Gynaecology, 91,457465.
MURDOCH, A.P., DUNLOP, W. & KENDALL-TAYLOR, P. (1986) Studies of prolactin secretion in polycystic ovary
syndrome. Clinical Endocrinology, 24, 165-175.
ORSINI,L.F., VENTUROLI, S., LORUSSO,R., PLUCHINOTTA, V., PARADISI, R.& BOVICELLI, L. (1985) Ultrasonic
findings in polycystic ovarian disease. Fertility and Sterility, 43,709-714.
PARADISI, R., MAGRINI, O., VENTUROLI, S., FABBRI,R., PORCU,E., STANZANI, L. & FLAMIGNI, C. (1985)
Naloxone does not interfere with the dopamine induced decrease in gonadotropin secretion in women with
polycystic ovarian disease. Hormone and Metabolic Research, 17,29-3 1.
PARADISI, R.,VENTUROLI, S., PASQUALI, R., CAPELLI, M., PORCU, E., FABBRI, R. & FLAMIGNI, C. (1986) Effects of
obesity on gonadotropin secretion in patients with polycystic ovarian disease. Journal of Endocrinological
Investigation, 9, 139-144.
 Pulsatile hormone secretion in PCOS 107
PORCU,E., VENTUROLI, S., MAGRINI, O., BOLZANI, R., GABBI,D., PASSARINI, M., FABBRI, R., PARADISI,R. &
FLAMIGNI, C. (1986) Circadian rhythms of luteinizing hormone in adolescent anovulation. Human
Reproduction, 1 (Suppl. 2), 76-77, Abstract No. 258.
QUIGLEY, M.E., ROPERT,J.F. & YEN,S.S.C. (1981a) Acute prolactin release triggered by feeding. Journal of
Clinical Endocrinology and Metabolism, 52, 1043-1045.
QUIGLEY, M.E., RAKOFF,J.S. & YEN,S.S.C. (1981b) Increased luteinizing hormone sensitivity to dopamine
inhibition in polycystic ovarian syndrome. Journal of Clinical Endocrinology and Metabolism, 52,23 1-234.
REBAR,R., JUDD,H.L., YEN,S.S.C., RAKOFF, J., VANDEMBERG, G. & NAFTOLIN, E. (1976) Characterization of
the inappropriate gonadotropin secretion in polycystic ovary syndrome. Journal of Clinical Inuestigution,
57, 1320-1329.
SANTEN, R.J. & BARDIN,C.W. (1973) Episodic luteinizing hormone secretion in man. Journal of Clinical
Investigation, 52,26 17-2628.
SHOUPE, D. & LOBO,R.A. (1985) Prolactin response after gonadotropin-releasing hormone in the polycystic
ovary syndrome. Fertility and Sterililty, 43, 549-553.
SIEGEL, S. (1956) Non-parametric Statistics. McGraw Hill Book Company, New York.
SIITERI,P.K. & MACDONALD, P.C. (1973) Role of extraglandular estrogen in human endocrinology. In
Handbook of Physiology; Endocrinology, vol. I1 (ed. R.O. Creep & E. Astwood), pp. 615429, American
Physiological Society, Washington, D.C.
SOULES,M.R., STEINER,R.A., COHEN,N.L., BREMNER, W.J. & CLIFTON,D.K. (1985) Nocturnal slowing of
pulsatile luteinizing hormone secretion in women during the follicular phase of the menstrual cycle. Journal
of Clinical Endocrinology and Metabolism, 6 1 , 4 3 4 9 .
VENTUROLI,S., FABBRI, R., PARADISI, R., GAMMI, L., JASONNI, V.M. & FLAMIGNI, C. (1980) Diagnostica degli
irsutismi: tests dinamici surrenalici. Giornale Italian0 di Ostetricia e Gineocologia, 214, 15-23.
VENTUROLI, S., PORCU,E., GAMMI, L., MAGRINI, o.,FABBRI, R., PARADISI,R., FOCACCI, M., DALPRATO, L. &
FLAMIGNI, C. (1986a) Luteinizing hormone pulsatility in adolescent anovulation. Human Reproduction, 1
(suppl. 2), 78, Abstract No. 263.
VENTUROLI, S., PORCU,E., FABBRI,R., PARADISI, R., RUGGERI,S., BOLELLI, G., ORSINI,L.F., GABBI,D. &
FLAMIGNI, C. (l986b) Menstrual irregularities in adolescents: hormonal pattern and ovarian morphology.
Hormone Research, 24,269-219.
VENTUROLI, S . , PORCU,E., FABBRI, R., PARADISI, R., GAMMI, L., PASSARINI,M., ORSINI,L.F. & FLAMIGNI, C.
(19864 Ovarian multifollicularity, high LH and androgen plasma levels, and anovulation are frequent and
strongly linked in adolescent irregular cycles. Acta Endocrinologica, 111, 368-372.
WALDSTREICHER, J., SANTORO, N., FILICORI, M. & CROWLEY, W.F. (1986) Hyperfunction of the hypothalamic-
pituitary axis in women with polycystic ovary disease (PCOD). Endocrinology, 118 (Suppl.), 1SO,
Abstract 417.
WENTZ,A.C., JONES,G.S. & SAPP, K. (1976) Pulsatile gonadotrophin output in menstrual dysfunction.
Obstetrics and Gynaecology, 41, 309-3 18.
ZUMOFF,B., FREEMAN, R., COUPEY,S., SAENGER, P., MARKOWITZ, & KREAM,J. (1983) A chronobiological
abnormality in luteinizing hormone secretion in teenage girls with the polycystic-ovary syndrome. New
England Journal of Medicine, 309, 120&1209.