Episodic luteinizing hormone release in hyperprolactinemic women

,....

FERTILITY AND STERILITY Copyright c 1986 The American Fertility Society

~ l

Vol. 45, No. 4, April 1986· Printed in U.S A.

Episodic luteinizing hormone release in hyperprolactinemic women

R. Herbert Wiebe, M.D.* Stuart Handwerger, M.D.t Michael Soules, M.D.+ I I

University of South Alabama, Mobile, Alabama, Duke University Medical Center, Durham, North Carolina, and University of Washington, Seattle, Washington·

'

(

I

Episodic luteinizing hormone (LH) secretion was studied in 16 hyperprolactinemic women (microprolactinoma, 12; idiopathic, 4) with amenorrhea of 1.4 to 7 years' duration. Blood samples obtained through an indwelling veltous catheter at 20-minute intervals over 5 hours were assayed for LH, follicle-stimulating hormone (FSH), prolactin (PRL), and estrogen (E) (selected samples). LH pulse patterns were divided arbitrarily into high-amplitude release (LH pulse > 10 m/U!ml) (n = 7), low-amplitude release (LH pulse < 10 m/U!ml (n = 6), and no release (n = 3). Mean pulse frequencies in women with high-amplitude and low-amplitude release were 2.4 ± 0.3 (mean ± standard error) and 1.3 ± 0.2 pulses/5 hours and differed significantly (P > 0.02), whereas mean percentages of secretory increment were 155% and 62%, respectively. Mean LH concentrations in the high-amplitude (18.0 ± 0.8 m/U!ml), low-amplitude (13.2 ± 0.6 m/U!ml), and no-pulse groups (7.5 ± 0.2 mlU! ml) differed significantly (P > 0.02). Despite the different pulse patterns, mean serum FSH, PRL, and total E concentrations were similar. The lack of episodic LH release and/or low ·infrequent LH release could account for the absence of cyclic hypothalamic pituitary ovarian function, although other mechanism(s) may be operative in women with augmented LH secretory pulses. ·Fertil.Steril45:483, 1986

Several factors have been purported to contribute to the development of amenorrhea in patients with hyperprolactinemia, including a possible inhibitory effect on ovarian steroidogenesis,1 abnormalities of pituitary estrogen (E) feedback,2 and infrequent or absent:episodic luteiniz· ing hormone (LH) secretion. 3 The paucity of sponReceived July 11, 1985; revised a nd accepted J anuary 3, 1986. *Reprint requests: R. Herbert Wiebe, M.D., Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of South Alabama, Clinical Sciences Building, Room 328, Mobile, Alabama 36688. t Department of Pe diatrics, Duke University Medical Center: t Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Washington. Vol. 45, No. 4, April1986

taneous LH secretion has been thought to represent a gonadotropin-releasing hormone (GnRH) deficiency, possibly related t o alterations in hypothalamic dopaminergic4 and/or opioid tone.5 Although the pituitary response to exogenously administered GnRH·in hyperprolactinemic women is variable, the increased pituitary gonadotropin sensitivity and reserve capacity seen in a sizeable proportion of subjects studied is not con~ sistent with a GnRH deficiency.6 Alterations in the pattern of episodic LH secretion may contribute to the menstrual dysfunction in .hyperprolactinemic women.7 Recently, Sauder et al. 8 described infrequent, high-amplitude LH pulsations in seven hyperprolactinemic women. In contrast; Klibanski et al. 9 described .a more heterogenous pattern of pulsatile. gonadotropin .secretion in 20 . Wiebe et a l. PRL and episodic LH release

483

women, varying from a lack of pulsation to normal follicular phase LH pulse frequency and amplitude. This study was performed to assess the incidence of endogenous gonadotropin pulsation in a large group of hyperprolactinemic women and to determine whether alterations in hypothalamic pituitary gonadal function were associated with abnormal patterns of LH release.

40

30

MATERIALS AND METHODS

Sixteen hyperprolactinemic women, 19 to 39 years of age, amenorrheic for 1.4 to 7 years, were studied. Twelve of the women had a pituitary microadenoma (6 confirmed surgically, 6 diagnosed radiologically) and 4 women had idiopathic hyperprolactinemia. Thyroid function studies (serum thyroxine and thyroid-stimulating hormone), growth hormone, and cortisol reserve, as assessed by insulin-induced hypoglycemia, were normal in all of the subjects. Subsequent to the study, all subjects were treated either surgically and/or with bromoergocryptine, resulting in cyclic menses and euprolactinemia. To assess the gonadotropin secretory pattern, blood was sampled every 20 minutes through an indwelling venous cannula over a period of 5 hours, commencing at 8:00 A.M. Serum was separated and stored at - 20°C until assayed. Serum follicle-stimulating hormone (FSH) and LH 10 were measured at 20-minute intervals, serum prolactin (PRL) 11 at 40-minute intervals, and serum E at the commencement and termination of the sampling period. The intraassay variabilities for FSH and LH were 6.8% and 6.6%, respectively. The intraassay variability for PRL was 8.4%. Several descriptive parameters were established to evaluate the pattern of pulsatile LH secretion. LH pulse analysis was done as described by Santen and Barden.12 An increment from nadir to peak > 20% associated with either a progressive increase or decrease in LH levels defined a secretory pulse. The frequency of pulsation over 5 hours, the absolute increment in LH, and the percent increment in LH per secretory pulse were also determined. The mean LH concentration, LH increment, and percent secretory increment were calculated. The mean FSH, PRL, and total.E concentrations were calculated for each subject. FSH pulse analysis was not performed, because the hyperprolactinemic subjects, like normal-cycling women, showed no definite FSH pulse pattern. 484

Wiebe et al. PRL and episodic LH release

10

0

60

120

180

240

300

Time (minutes)

Figure 1 High-amplitude LH release in subject 1. Arrows indicate LH pulse.

PRL concentrations for each subject were analyzed with the concomitant LH concentrations to determine whether there was a pulsatile pattern coincident with LH. Regression analysis between mean serum PRL, FSH, LH, and total E concentrations was performed. Regression line analysis between LH pulse frequency and amplitude, E , and PRL concentrations also was performed. LH pulse patterns were arbitrarily divided into a high-amplitude release (one or more LH pulses > 10 miU/ml), low-amplitude release (LH pulses < 10 miU/ml), and no LH release. Examples of each are depicted in Figures 1, 2, and 3. In five of seven subjects with high-amplitude release, all LH pulses approached or exceeded 10 miU/ml, representing a distinct pattern of LH release during the sampling period. In the remaining two subjects (subjects 3 and 4), only one LH pulse of lower amplitude was detected. Because the overall pattern ofLH release in these two subjects was more consistent with a high-amplitude LH release, they were included in this group (Fig. 4). The subjects were placed into three groups: highamplitude (7), low-amplitude (6), and no spontaneous LH release (3). The patients' age, duration Fertility and S terility

20

0

120

60

180

240

300

Time (minutes)

Figure 2 Low-amplitude LH release in subject 10. Arrow indicates LH pulse.

of amenorrhea, mean serum PRL, FSH, LH, total E concentrations, pulse frequency, absolute secretory increment, and percent secretory increment were compared by analysis of variance. For comparison of the LH pulse frequency between the hyperprolactinemic subjects and normal controls, the serial LH data were subjected to further analysis. The LH data from the 13 subject~s who demonstrated episodic LH release (as determined by the Santen and Barden 12 pulse analysis technique) were reanalyzed for LH pulse frequency by another method, a computerized modification of the Santen and Barden technique, as previously described. 13 Comparisons for LH pulse frequency were made with LH data collected from five normal women studied in the early follicular phase (cycle days 1 to 4). 13 In these normal women, the sampling interval was 20 minutes over the same time period (8:00A.M. to 1:00 P.M. ). Mean LH concentrations and absolute LH increments were not compared, because LH concentrations in the normal women were assayed in a different laboratory than those of the hyperprolactinemic women . Statistical comparisons between the control group and hyperprolactinemic group were assessed by analysis of variance.

of the 16 patients demonstrated episodic LH activity, with > 50% of these with high-amplitude release. No correlation was found between mean LH, mean FSH, and PRL concentrations. Although FSH and mean PRL concentrations did not correlate with mean E, mean LH concentrations correlated with mean E concentrations (r = 0.5, P > 0.01). However, neither pulse frequency nor pulse amplitude correlated with mean total E concentrations. In the 13 subjects demonstrating episodic LH release, pulse frequency ranged from one to three pulses per 5-hour period with pulse amplitudes varying from 4.3 miU/ml to 23.7 miU/ml, representing a group pulse frequency of 1.92 pulses/5 hours and a mean pulse amplitude of 10.3 miU/ ml. LH pulse frequency, analyzed by the Soules et al. 13 modification of the San ten and Barden method, in these 13 hyperprolactinemic women was 2.65 ± 1.3 pulses/5 hours and was significantly slower than the LH pulse frequency of 3.9 ± 1.0 in the five normal women studied in the early follicular phase (P > 0.02). The discrepancy between pulse frequency assessed by the original Santen and Barden 12 criteria and the modified Santen and Barden method 13 is not surprising, because the latter, a computerized program, is more likely to pick up small pulses of short duration. ' · Subjects were grouped by pulse analysis into high-amplitude LH release (7), low-amplitude LH release (6), and no LH release (3). The mean age did not differ between the three groups (Table 2). The trend toward longer duration of amenorrhea in subjects with low-amplitude and no LH release was not significantly different. 20

RESULTS The clinical characteristics of the patients, along with mean PRL, FSH, LH, and total E concentration are shown in Table 1, as are the descriptive aspects of pulse analysis. Thirteen (81 %) Vol. 45, No. 4, April 1986

0

60

120

180

240

300

Time (minutes)

Figure 3 No LH release noted in subject 14. Wiebe et al. PRL and episodic LH release

485

30

Mean pulse frequency in subjects with highamplitude release was 2.4 ± 0.3 pulses/5 hours. Mean secretory increment was 13.9 miU/ml, a 155% mean increment. Pulse frequency in the low-amplitude group was slower, with 1.43 ± 0.2 pulses/5 hours (P > 0.02). The mean absolute secretory increment was 6.2 miU/ml, a 62% secretory increment. Despite the differences in mean LH concentrations, mean secretory increments, and pulse frequency, FSH concentrations and serum total E levels were the same (Table 2). No difference was noted in mean PRL concentrations between the three groups. No synchrony between LH and"PRL pulse was evident.

20

10

0

60

120

180

240

DISCUSSION

300

Time (minutes)

Episodic LH release was demonstrated in > 80% of the subjects studied. These findings are consistent with the demonstrations by Klibanski et al. 9 and Sauder et al. 8 of the presence of episodic LH release in a sizeable proportion of women with pituitary microadenoma and idiopathic hyperprolactinemia. Our findings are contrary to earlier reports, which have indicated a loss of episodic LH activity in 45%14 to 85%3 of women with hyperprolactinemia. The reasons for this discrepancy are not readily apparent, but

Figure 4 Pulse pattern consistent with high-amplitude LH release in subject 4. Arrows indicate LH pulse.

The mean LH concentration in the high-amplitude group was 18.0 ± 0.8 miU/ml, in the upper range of normal for the early follicular phase. Mean LH level in the low-amplitude group was 13.2 ± 0.6 miU/ml. The mean LH concentration in the no pulse group was significantly lower than the high-amplitude group (P > 0.02).

Table 1. Clinical Characteristics and LH Pulse Patterns in Hyperprolactinemic Women Patient

Duration of amenorrhea

PRL

FSH

LH

mos

nglml

mlU!ml (i ± SDJ"

mJUiml (i ± SDJ

18 45 27 22 36 33 24

35 64 67 83 92 116 129

11 11.2 12.5 12.3 11.2 10.7 11.3

± ± ± ± ± ± ±

1.4 1 1.1 3.4 2.3 2.6 0.8

25.9 18.8 15 15.85 14.8 18.3 17.26

± ± ± ± ± ± ±

5.3 8 4.5 4.4 7.5 6.5 5.8

60 23 24 48 84 23

33 52 55 187 234 342

14.1 12.4 11 13.8 10.5 9.3

± ± ± ± ± ±

1.2 1.0 0.5 0.08 0.6 2.0

27.9 14 8.7 14.9 10.1 9.4

± ± ± ± ± ±

2.5 2.0 1.5 1.7 1.8 2.5

50 34 35

105 106 133

13.9 ± 1.1 15.3 ± 1.2 7.9 ± 0.5

High amplitude lb 2b sc 4c 5c 6c 7c Low amplitude 8b 9c lOb 11c 12c 13c No release 14c 15c 16c

7.35 ± 0.8 7.9 ± 1.2 7.3 ± 0.6

Pulse fre'tuency/ 5 ours

Arn~tude

Total E

mlU!ml (i ± SD)

pglml

80 135 112 83 275 192 208

12.4 ± 2.5 12 ± 2.3 9.3 ± 2.6 12.3 ± 3.7 23.7 16.3 ± 1.8 15.3 ± 3.9

69 75 43 42 53 64 20

3.0 3.0 2.0 2.0 1.0 2.0 3.0

36 47 75 38 66 110

7.8 5.8 ± 1.0 5.4 4.5 ± 0.5 5.0 7.0

75 42

1.0 2.0 1.0 2.0 1.0 1.0

%
42 20 55 34 26 45

asn, standard deviation. bldiopathic.
· Wiebe et al. PRL and episodic LH release

Fertility and Sterility

Table 2. Episodic LH Release Patterns High amplitude Low amplitude

No. Age (yrs) Duration of amenorrhea (mos) PRL (ng/ml) LH (miU/ml) LH pulse frequency/5 hours LH (miU/ml) % LH secretory increment Total E (pg/ml)

7 28 ± 1.96a 30.1 ± 3.1 83.3 ± 12.3 18.0 ± 0.8b 2.4 ± 0.3c 13.9 ± 1.3c 155 ± 19.8c 52.3 ± 7.2

6 27 ± 2.2 43.6 ± 10.3

No release

3 29.5 ± 3.4 42 ± 7.9

150.05 ± 5.9 114.6 ± 11.0 7.5 ± 0.7 13.2 ± 0.6 1.33 ± 0.2 6.2 ± 0.6 62 ± 9.1 46.8 ± 9.1

35 ± 3.9

aMean ± standard error of the mean. bHigher than no release group (P > 0.02). cHigher than low-amplitude group (P > 0.02).

heterogenicity of subjects studied and/or higher concentrations of PRL in previous studies may have contributed to the lower incidence of episodic LH activity. Klibanski et al. 9 and Sauder et al. 8 demonstrated variable patterns of LH pulsations. However, we were able to identify three distinct patterns of episodic LH activity: no episodic LH release, low, infrequent LH release, and highamplitude release. Although studies in normal women have failed to show any significant difference in pulse patterns over a 24-hour period, 12 Sauder et al. 8 reported considerable individual variability in pulse pattern within a 24-hour period in seven hyperprolactinemic women. However, Klibanski et al., 9 in a larger series, demonstrated more consistent patterns within hyperprolactinemic individuals studied over a 24-hour period. Therefore, the time span of 5 hours in our study should be a representative segment of LH secretion. Despite differing patterns of LH release, mean plasma PRL concentrations, the duration of amenorrhea, and the presence or absence of a pituitary microadenoma were not statistically different in the three groups. Although PRL concentrations tended to be higher in subjects with low-amplitude LH release and no LH release, compared with those with high-amplitude release, this was not statistically significant. With analysis of a larger sample size, this trend might achieve statistical significance. The retention of episodic release, then, in subjects with lower PRL concentrations and the loss of episodic release at higher PRL concentrations would suggest a proVol. 45, No.

4, April1986

gressive effect. Larger sample size and longitudinal studies would be required to either confirm or refute this hypothesis. In subjects with episodic LH release, the frequency of LH release was slower than in the controls. However, in the high-amplitude group, the mean pulse frequency was 2.4 pulses/5 hours, comparable to a frequency of 2. 7 pulses/5 hours in the normal follicular phase reported by Santen and Barden.12 Further, the mean secretory increment of 155% in our subjects was greater than the 56% secretory increment noted by Santen and Barden12 in the early follicular phase. The mechanism(s) of the high-amplitude release is unknown, but a slight slowing of the frequency of hypothalamic GnRH release and/or an increase in the refractory period of the GnRH receptor could result in greater time available for hormone accumulation within the gonadotrope, and thus high-amplitude release. This pattern of gonadotropin release is not compatible with any significant GnRH compromise. In subjects with low-amplitude release, the frequency of LH pulses was considerably less than the normal follicular phase and the high-amplitude release group. The mean secretory increment of 62% in the group was similar to the 56% secretory increment noted by Santen and Barden in the early follicular phase. 12 A marked slowing of GnRH release, presumably resulting in infrequent LH release, could account for the absence of cyclic hypothalamic pituitary ovarian activity noted in these subjects. This pattern, along with the lack of LH activity noted in three of our subjects, is compatible with a relative GnRH deficiency. Mean LH concentrations in the high-amplitude group were in the upper range of normal for the early follicular phase; mean LH concentrations in the low-amplitude or no release groups were in the lower range of normal. Mean FSH concentrations were normal in all three groups. Despite the difference in pulse patterns and mean LH concentrations, all subjects were amenorrheic. Further, serum total E concentrations were low in all three groups. Because infrequent episodic LH release in the rhesus monkey has been shown to be associated with low serum E concentration and a lack of follicular development, 15 a lack of episodic or the presence of low-amplitude, infrequent LH release could account for the low E concentration and the absence of cyclic hypothalamic pituitary ovarian function in some women with hyperprolactinWiebe et al. PRL and episodic LH release

487

emia. It is difficult to ascribe the relative ovarian quiescence in the subjects with high-amplitude release to insufficient gonadotropin stimulation. It is unlikely that the slight slowing of LH activity observed in the high-amplitude group could account for the low E concentrations, because pulsatile GnRH administration to euprolactinemic hypoestrogenic amenorrheic women at a frequency slower than that demonstrated in our study has resulted in ovulation. 16 In normal men and women, intensified rates of venous sampling at 1- to 5-minute intervals may unmask frequent LH pulses of low amplitude not discernible at the usual sampling frequency (i.e., 15 to 20 minutes)P· 18 Although the physiologic significance of these rapid fluctuations in LH concentrations is unknown, frequent venous sampling may define significant differences in pulse patterns in normal and hyperprolactinemic women. There is controversy regarding a possible direct PRL effect on ovarian function, with many conflicting reports. Ev(;!n so, an alteration in ovarian responsivity in hyperprolactinemic women, possibly secondary to PRL, should be considered. PRL receptors have been identified within the ovary, and McNatty1 has shown an inhibiting effect of PRL on ovarian steroidogenesis in vitro. The different patterns of gonadotropin release in amenorrheic hyperprolactinemic women may be indicative of group heterogeneity with a diverse pathophysiology for hyperprolactinemia and amenorrhea. Acknowledgment. We would like to acknowledge Donald K. Clifton, Ph.D., for his aid in the evaluation and a,nalysis of the data. REFERENCES 1. McNatty KP: Relationship between plasma prolactin and the endocrine microenvironment of the developing human antral follicle. Fertil Steril 32:433, 1979 2. Glass MR, Shaw RW, Butler WR, Logan-Edwards R, London DR: An abnormality of estrogen feedback in amenorrhea-galactorrhea. Br Med J 3:274, 1975 3. Bohnet HG, Dahlen HG, Wuttke W, Schneider HPG: Hyperprolactinemic anovulatory syndrome. J Clin Endocrinol Metab 42:132, 1976 4. Bression D, Brandi AM, Martres MP, Nausbaum A, Cesselin F, Racadot J, Peillon F : Dopaminergic receptors in human prolactin secreting adenomas: a quantitative study. J Clin Endocrinol Metab 51:1037, 1980

488

Wiebe et al. PRL and episodic LH release

5. Quigley ME, Sheehan KL, Casper RF, Yen SSC: Evidence of increased opioid inhibition of luteinizing hormone secretion in hyperprolactinemic patients with pituitary microadenoma. J Clin Endocrinol Metab 50:427, 1980 6. Wiebe RH: Endocrine evaluation of hyperprolactinemia. Clin Obstet Gynecol 23:349, 1980 7: Mault PJA, Rees LH, Besser CM: Pulsatile gonadotropin secretion in hyperprolactinemic amenorrhea and the response to bromocriptine therapy. Clin Endocrinol (OxO 16:153, 1982 8. Sauder SE, Froger M, Case GD, Kelch RP, Marshall JC: Abnormal patterns of pulsatile luteinizing hormone secretion in women with hyperprolactinemia and amenorrhea: responses to bromocriptine. J Clin Endocrinol Metab 59:941, 1984 9. Klibanski A, Beitens JZ, Merriam GR, McArthur JW, Zervas NT, Ridgway EC: Gonadotropin and prolactin pulsations in hyperprolactinemic women before and during bromocriptine therapy. J Clin Endocrinol Metab 58:1141, 1984 10. Odell WD, Rayford PL, Ross GT: Simplified, partially automated method for radioimmunoassay of human thyroid-stimulating, growth, luteinizing, and folliclestimulating hormones. J Clin Med 70:973, 1967 11. Sinha YN, Selby FW, Lewis VJ, Vanderlaan WP: A homologous radioimmunoassay for human prolactin. J Clin Endocrinol Metab 36:509, 1973 12. Santen RJ, Barden CW: Episodic luteinizing hormone secretion in man: pulse analysis, clinical interpretation, physiological mechanism. J Clin Invest 52:2617, 1973 13. Soules MR, Steiner RA, Clifton DK, Cohen NL, Aksel S, Bremmer WJ: Progesterone modulation of pulsatile luteinizing hormone secretion in normal women. J Clin Endocrinol Metab 58:378, 1984 14. Buckman MT, Peake GT, Srivastava L: Pattern of spontaneous LH release in normal and hyperprolactinemic women. Acta Endocrinol (Copenh) 97:305, 1981 15. Wildt L, Hausler A, Marshall G, Hutchinson JS, Plant TM, Belchetz PE, Knobil E: Frequency and amplitude of gonadotropin secretion in Rhesus monkey. Endocrinology 109:376, 1981 16. Reid RL, Leopold GR, Yen SSC: Induction of ovulation and pregnancy with pulsatile luteinizing hormone releasing factor: dosage and mode of delivery. Fertil Steril 36:553, 1981 17. Veldhuis JD, Evans WS, Rogol AD, Drake CR, Thorner MO, Merriam GR, Johnson ML: Intensified rates of venous sampling unmask the presence of spontaneous, high-frequency pulsations ofluteinizing hormone in man. J Clin Endocrinol Metab 59:96, 1984 18. Filicori M, Hoffman A, Mansfield M, Dunaif A, Beardsworth D, Trigilio S, Donnelly J , Crowley W: Discernible FSH pulsations in the human menstrual cycle: their concordance with LH pulsations and the critical nature of sampling frequency in their demonstration. Clin Res 30:270A, 1982

Fertility and Sterility