0021-972x/92/7503-0S30$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1992 by The Endocrine Society
Vol. 75, No. 3 Printed
Ovulation and Menstrual with Central Precocious Gonadotropin-Releasing NAOMI JAY, M. JOAN WILLIAM F. CROWLEY, AND PAUL A. BOEPPLE
MANSFIELD, JR., DAVID
Function of Adolescent Puberty after Therapy Hormone Agonists*
ROBERT M. BLIZZARD, SCHOENFELD, LESLIE
in U.S.A.
Girls with
RHUBIN,
Pediatric and Reproductive Endocrine Units and General Clinical Research Center, Massachusetts General Hospital, Boston, Massachusetts 02114; the Divisions of Endocrinology and Adolescent Medicine, Department of Medicine, Children’s Hospital, Boston, Massachusetts 02115; the Division of Endocrinology and Diabetes, Children’s Medical Center of the University of Virginia, Charlottesville, Virginia 22908; and Hygeia Sciences, Newton, Massachusetts 02160-1432 became increasingly regular, with cycles of 25- to 35-day duration reported by 41% of the girls in the first year postmenarche and 65% of the girls studied 3 or more years postmenarche. Ovulation was demonstrated in 50% of the girls studied within 1 yr of menarche and in 90% of the girls studied 2 yr or more postmenarche, including 5 girls who reported pregnancies. The development of regular ovulatory menstrual function in these girls with CPP is in accord with previously documented patterns in normal adolescents. While these data provide further evidence supporting the safety of long term GnRHa therapy, continued studies will be necessary to characterize fully the reproductive function in CPP patients through adolescence and adulthood. (J Clin Endocrinol Metab 75: 890-894, 1992)
ABSTRACT Chronic GmRH agonist (GnRHa) administration has been shown to suppress pituitary-gonadal function in children with central precocious puberty (CPP), but long term data after the reactivation of gonadarche posttherapy are not yet available. This study evaluated the menstrual function of 46 girls with CPP who had been treated for at least 2 yr with GnRHa (deslorelin or histrelin, SC, daily) and were up to 7 yr posttreatment, including 21 postmenarcheal girls who collected weekly overnight urine samples for 12 consecutive weeks to assess rates of ovulation by urinary pregnanediol-3@-glucuronide measurements. Menarche occurred at age 12.1 f 1.0 yr (mean + SD), on the average 1.2 + 0.8 yr posttherapy (range, 0.1-4.3 yr). Menstrual cycle lengths
L
Patient
Received July 15, 1991. Address requests for reprints to: Paul A. Boepple, M.D., Reproductive Endocrine Unit, Massachusetts General Hospital, Boston, Massachusetts 02114. * This work was supported in part by NIH Grants HD-18169, RR01066, RR-02172, and RR-08847 and FDA Grant FD-R-000388. Presented in part at the 5th Annual North American Society for Pediatric and Adolescent Gynecology, Fort Lauderdale, FL, April 11,. 1991, and the 73rd Annual Meeting of The Endocrine Society, Washington, D.C., June 22,. 1991.
Female patients who had been treated for CPP with GnRHa at the Massachusetts General Hospital, Children’s Hospital (Boston, MA), and the Children’s Medical Center at the University of Virginia (Charlottesville, VA) were recruited for this study. Diagnostic criteria for CPP and details of the treatment regimen and monitoring protocols have been described previously (1, 2). Briefly, all patients had the onset of breast development before age 8 yr, and all demonstrated a pulsatile pattern of gonadotropin secretion and pubertal response to GnRH before therapy. Pituitary-gonadal suppression was achieved with daily SC injections of either deslorelin (4-8 pg/kg.day) or histrelin (8-10 pg/kg.day). Suppression was confirmed during in-patient evaluations performed at 3- to 6-month intervals during therapy. All aspects of this project were approved by the Institutional Review Boards of the 3 participating hospitals. Forty-nine subjects had been treated with GnRHa (duration, ~1.9 yr) and were now off therapy. Of these, 3 subjects were excluded due to either the presence of other medical diagnoses that were known to interfere with menstrual functioning or noncompliance with the medical regimen, which resulted in incomplete suppression throughout the course of their therapy. An additional 4 patients who reported declining compliance in the final few months of up to 4 yr of therapy were included in this analysis, since adequate suppression had been documented in them at all prior evaluations. Thirty-eight of 46 eligible patients participated, with retrospective review of hospital records providing data on the remaining 8 girls. To evaluate the return of puberty and menstrual function in all GnRHa-treated patients, girls with neurogenic precocious puberty (NIP; n = 7) were included in this study despite the potential for this group to manifest disordered neuroendocrine function. The diagnoses represented among the NPP patients included cerebral astrocytoma treated with cranial irradiation, congenital
ONG term administration of potent agonist analogs of GnRH (GnRHa) induces persistent suppression of the pituitary-gonadal axis and has been successfully employed as a treatment for central precocious puberty (CPP) for more than a decade (1-4). While the reactivation of gonadarche after discontinuation of GnRHa treatment has been prompt in patients with CPP (5, 6), little information has been available regarding the completion of their pubertal maturation. In females, this is marked by the development of estrogen positive feedback and the ability to mount a LH surge, evidenced by attainment of regular menstrual cycles, consistent ovulation, and corpus luteum formation. The present study was undertaken to chart the development of ovulatory menstrual function in adolescent females who had completed GnRHa treatment for CPP.
Materials and Methods population
890
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
OVULATORY toxoplasmosis, myelomeningocele, tosis, supracellar cyst, and perinatal
neonatal asphyxia.
meningitis,
FUNCTION neurofibroma-
IN PRECOCIOUS TABLE therapy(n
1. Patient = 46) Clinical
Protocol This study sought to define three characteristics of our female CPP patients after GnRHa therapy: the timing of menarche, the lengths of subsequent menstrual cycles, and the rates of ovulation associated with those cycles. Even though 18 patients had experienced at least 1 episode of menstrual bleeding before the initiation of therapy, we use the term menarche here to refer to the first menstrual period following the discontinuation of GnRHa administration. The timing of menarche was established either by history at the time of a General Clinical Research Center evaluation posttherapy or by responses to a questionnaire. While the questionnaire also included items that required respondents to describe characteristics of their cycles (e.g. duration of cycle, duration of flow, and associated symptoms), only contemporaneous menstrual cycle diaries were employed to determine menstrual cycle lengths. In addition to the existing individual records that were available for analysis, 28 girls maintained prospective menstrual diaries for a 3-month period, during which 26 of the 28 patients collected first morning voided urine specimens at weekly intervals. Urine samples were immediately frozen and subsequently assayed for pregnanediol3/?-glucuronide (PDG) as an index of progesterone secretion and corpus luteum function.
Hormone
analysis
Qualitative determinations of PDG were made using a membranebased immunoassay (Hygeia Sciences, Newton, MA), employing a previously characterized anti-PDG monoclonal antibody (7). Assay sensitivity (-4 nmol/L) was set to correspond to a value 2 SD greater than the mean follicular phase urinary PDG concentration (8, 9). To further validate these results, a subset of samples (56 of the total 301, 18.6%) was randomly chosen for quantification in a previously validated PDG RIA (8, 9). In this subset, all 12 samples containing 4 nmol/L PDG or more by RIA were positive in the qualitative immunoassay, as were 5 of the 30 samples whose RIA-determined concentration was between 1.4-4 nmol/L. All of the samples whose quantitative results were less than 1.4 nmol/L PDG yielded negative results in the qualitative assay. Thus, employing an RIA determination of 4 nmol/L or more as our positive reference, the qualitative I’DG assay performed with a sensitivity level of 100% and a specificity of 89%.
Data analysis Data regarding the attainment of menarche were analyzed by using maximum likelihood estimates, assuming normal distribution, thus allowing the inclusion of girls who remained premenarcheal. To assess trends in cycle length over time, the deviation from the adult normal range (after square root transformation) was viewed relative to time, using a random patient effect model (10). For purposes of analysis, normal cycle length was considered to be 25-35 days, as previously documented in healthy adults (11). In girls with normal length cycles, intraindividual variation was further evaluated. A patient was said to have regular cycles if all were between 25-35 days and varied by no more than 5 days. Data regarding the attainment of menarche were analyzed separately for patients with idiopathic (n = 39) and neurogenic (n =. 7) precocity, but are reported together since the groups exhibited no significant differences. Onlv one patient with NPI’ (supracellar cyst) partizpated in the prospective’compbnent of this study; tius, ovulatory rates could not be compared between the two diagnostic subsets.
Results The characteristics of the 46 girls with CPP are summain Table 1. The subset whose data permitted estimates of ovulatory rates did not differ significantly from the group overall in terms of treatment and posttreatment histories. rized
PUBERTY characteristics
event (yr)
Age at initiation of therapy Age at discontinuation of therapy Duration of GnRHa therapy Age at menarche Timing of menarche posttherapy Timing of study postmenarche
891 in girls
with
CPP post-GnRHa
Mean f 7.3 10.9 3.6 12.1 1.2 1.6
f + f + f +
SD
Range
1.3 0.9 1.2 1.0 0.8 1.4
4.1-9.2 9.1-13.9 1.9-7.0 10.1-14.9 0.1-4.3 O-6.9
Menarche
Of the total 46 subjects, 44 (96%) had attained menarche by the completion of the study. The agesand intervals that are relevant to this milestone appear in Table 1. The 2 patients who had not yet reached menarche were still within 1 SD of our cohort’s mean age and time to menarche. The variance in the interval between the cessation of treatment and attainment of menarche among patients was examined for its correlation with various clinical characteristics of the population. The interval to menarche had a modest but significant (P < 0.03) negative relationship with chronological age (CA); that is, older girls tended to attain menarche more quickly after therapy stopped (r’ = 0.42). In a multiple regression model with CA as a covariate, the inclusion of duration of treatment as a second covariate accounted for an additional small increase (P < 0.04) in the interval to menarche, increasing the r* value to 0.48. In other words, having controlled for the impact of CA, girls treated for longer periods with GnRHa tended to take longer to reach menarche after stopping therapy. However, most of the variance among the patients was left unaccounted for, with other factors (e.g. age at onset of puberty, history of mensespretherapy, and bone age, weight, or Tanner stage breast development posttherapy) displaying no significant associations.In addition, there were no significant differences in timing of menarche between patients with a neurogenic etiology us. those with idiopathic CPP. Of the eight girls who reached menarche within 6 months of treatment cessation (>l SD early), declining compliance during the final months of therapy may have contributed in half of the subjects. However, the remaining four girls who reached menarche early had been consistently compliant and appeared completely suppressedwhen therapy was discontinued. The average time to menarche did not change significantly if the four patients who exhibited waning compliance in the final months of therapy were removed from analysis. Cycle length and regularity
Menstrual cycle lengths (223 cycles in 34 patients) are plotted zx years postmenarche in Fig. 1. The deviation of cycle length from that in normal adult women decreased with prolonged follow-up (P = 0.04). During the 12 weeks of prospective study of 28 girls, 20 (70%) were observed to have cycles of normal lengths, but only 13 (45%) demonstrated regular cycles, i.e. 2 or more cycles between 25-35 days, which varied less than 5 days in length. Subjective retrospective data on cycle length often differed substantially from contemporaneous reporting, in that the great majority
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
892 Q
JAY ET AL. >60
-o
l
JCE & M. 1992 Vol75.No3
100
1
A 80
P 4
60
2 0
40
8
.0 1001
-
&
1
2
n= 121
3 n = 49
Em
I
< 25 25 -35
4
5 n = 29
6
7
n = 24
Davs Da&
20
0 I
o-1
1-2
2-3
>3
YEARS POST-MENARCHE FIG. 2. Rates of ovulation in CPP girls post-GnRHa therapy plotted relative to time postmenarche. Ovulatory girls (U) were those in whom urinary PDG was positive (>4 nmol; see Materials and Methods) in 1 or more weekly samples during 12 weeks or who had reported pregnancies. Ovulatory cycles (ES@ were those in which a luteal phase ur&ary PDG test was positive. The total numbers of eirls and cvcles studied in each year appear above the relevant bar. Rates of ovulation increased significantly with increasing time postmenarche (by x2 test, P = 0.001).
20 t-l
o-1
l-2
2-3
>3
YEARS POST-MENARCHE FIG. 1. Menstrual cycle lengths in 34 CPP girls post-GnRHa therapy plotted relative to time postmenarche. In both A (individual cycles) and B (group data), menstrual cycles between 25-35 days in length are considered normal (see Materials and Methods) and are designated by the dotted lines or solid bars in A and B, respectively. The deviation of cycle length from that in normal adult women decreased with prolonged follow-up (P = 0.04).
of patients reported via the questionnaire that their cycles were consistently 28-30 days in length while their diaries reflected far more variability. Ovulation
Of the 26 girls who contributed weekly urine samples, 3 remained premenarcheal, and 2 were excluded from analysis due to the initiation of therapy with medroxyprogesterone acetate by their primary care providers after histories of oligomenorrhea. These 21 postmenarcheal girls were considered to display ovulatory menstrual function if 1 or more urine specimenswas positive in the qualitative PDG assay. In addition, 5 patients who reported pregnancies and 1 who had a documented serum progesterone level of 37.5 nmol/L during a General Clinical ResearchCenter evaluation posttherapy were also included among the ovulatory patients for the purposesof analysis. Data for these 27 girls are depicted in Fig. 2 relative to time postmenarche. In this cross-sectional analysis, demonstration of ovulatory function increased significantly as patients were studied posttherapy. This trend was significant if the analysis was based on the percentage of patients who ovulated at least once during the 3 months of observation or on the percentage of cycles in which a urinary PDG test was positive. In the latter instance, the
lower rates of ovulation depicted in Fig. 2 reflect that fact that patients who ovulated in at least 1 cycle often did not do so in all cycles during the course of the 3-month study. Of the girls deemed to be ovulatory, only 5 had positive PDG levels in all cycles observed, and an additional 4 patients ovulated in 2 of 3 cycles. In our patients, normal cycle lengths and premenstrual/ menstrual symptoms lacked both specificity and sensitivity as clinical correlates of ovulation. Anovulatory and ovulatory cycles fell within the normal range of 25-35 days with comparable frequencies (59% and 57%, respectively), and 66% of all long cycles (>35 days) examined were associated with ovulatory values of urinary PDG. Normal, regular length menstrual cycles in our patients, thus, correlated poorly with a biochemical index of ovulation (r = 0.02). Similarly, symptoms were reported during 50% of anovulatory cycles, making their correlation with ovulation modest in our study population (r = 0.34). Discussion
In addition to reactivation of pubertal pituitary-gonadal function, completion of secondary sexual development, and attainment of adult stature, a full assessmentof the long term safety and efficacy of GnRHa therapy of CPP will hinge upon the documentation of eventual normal fertility in treated patients (5, 6). While assessmentof ultimate fertility must await a longer follow-up of this still young population, the development of ovulatory menstrual function in the present subset of adolescent girls with CPP must, for now, serve as our index of completed pubertal maturation. The mean age at menarche of 12.1 yr in the present study is within the normal range for age at menarche in North America (12). However, the timing of menarche in this patient group is determined primarily by the decision of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
OVULATORY
FUNCTION
when to stop GnRHa therapy. Therefore, examination of the interval between stopping treatment and the development of menarche facilitates comparisons among individual patients and between studies. The findings of the present study are similar to those of Manasco et al. (5) and Kauli et al. (6), in which 67% and 80% of subjects, respectively, reached menarche by 1 yr post-GnRHa treatment. While the return of pulsatile gonadotropin secretion has been uniformly documented in our patients with idiopathic CPP within 6 months of the discontinuation of suppressive therapy, the individual variation in interval to menarche has been quite substantial. While there was a significant tendency for girls to reach menarche earlier if GnRHa was discontinued at an older age and/or after treatment of shorter duration, most of the individual variation in our study population was left unexplained and was probably related to factors, such as body composition, diet, and exercise, that are known to affect the rate of sexual maturation but were not considered in this study. The natural history of CPP appears to include normal expectations for adult reproductive function (13, 14). However, it was not surprising that cycle irregularity and anovulation were frequently evident in our patient cohort who were studied an average of only 1.6 yr postmenarche, since variable cycle lengths and inconsistent ovulation typically persist during the first 5-6 yr of menstruation in normal adolescent females (15-21). The appropriate comparisons were, therefore, critical to put the data in our adolescent patients into perspective. First, it was important to base analyses of cycle length on contemporaneous diaries, since recollections reported on questionnaires frequently overestimated the regularity of cycles. Second, it was important to employ a direct biochemical index of ovulation, since regular cycle lengths and the presence of menstrual symptoms were not reliable clinical markers of ovulation in our adolescent subjects, contrary to their apparent validity in adult populations (23-25). Finally, comparisons of our findings in CPP with those reported in normal adolescents had to take methodological details into account. While the rates of ovulation in the present study were higher than those reported in many studies of normal subjects (17-22), these estimates are significantly influenced by the interval of observation and the parameter monitored (e.g. single day 21 serum sample for progesterone VS. multiple determinations of urinary PDG over several cycles). Metcalf et al. (20), who studied adolescent girls with the same serial integrative index of ovulation (weekly urinary PDG measurements) over the same extended observation period (3 months) as we employed in our CPP patients, reported similar rates of ovulation (41.7% of girls in the first year postmenarche, 80% of those in the fifth year). Thus, using the most comparable normative data, the development of ovulatory menstrual function after GnRHa therapy in our patients with CPP appears to be appropriate for their age and developmental stage. GnRH agonists are being increasingly employed in a multitude of clinical situations. The recent FDA approval of histrelin as a therapy for CPP is likely to facilitate the use of GnRHa in a growing list of pediatric indications. Consequently, long term follow-up of patients treated in childhood
IN PRECOCIOUS
PUBERTY
893
becomes all the more critical. While the results of this study provide further evidence of the safety and reversibility of long term GnRHa treatment, continued longitudinal study through adulthood will be necessary to describe fully the reproductive function of this population. In this regard, it will be particularly important to diligently follow patients with histories of neurogenic precocity. While the progression of puberty posttherapy in this subset was no different from that in patients with idiopathic CPP in the present study, menarche is a milestone that is relatively insensitive to the subtle disorders of neuroendocrine function that NPP patients may subsequently manifest. Finally, the five pregnancies in our young patients occurring as early as age 13.8 yr, while reassuring vis iz vis the demonstration of normal fertility, underscore the need to address further the psychosocial impact of early sexual maturation and its therapy. Acknowledgments The GnRHa used in these studies Rivier. The Salk Institute (deslorelin), (histrelin).
were the generous gifts of Jean and Ortho Pharmaceuticals, Inc.
References 1. Crowley Jr WF, ComiteF, Vale W, Rivier J, Loriaux DL, Cutler GB. 1981 Therapeutic use of pituitary desensitization with longacting LHRH agonist: a potential new treatment for idiopathic precocious puberty. J Clin Endocrinol Metab. 52:370-2. 2. Boepple PA, Mansfield MJ, Wierman ME, et al. 1986 Use of a potent, long-acting agonist of gonadotropin-releasing hormone in the treatment of precocious puberty. Endocr Rev. 7:24-33. 3. Manasco PK, Pescovitz OH, Suvimol C, et al. 1989 Six-year results of luteinizing hormone releasing hormone (LHRH) agonist treatment in children with LHRH-dependent precocious puberty. J Pediatr. 115:105-8. 4. Kaplan SL, Grumbach MM. 1990 Pathophysiology and treatment of sexual precocity. J Clin Endocrinol Metab. 71:785-9. 5. Manasco PK, Pescovitz OH, Feulillan PP, et al. 1988 Resumption of puberty after long-term luteinizing hormone-releasing hormone agonist treatment of central precocious puberty. J Clin Endocrinol Metab. 67:368-72. 6. Kauli R, Kornreich L, Laron Z. 1990 Pubertal development, growth and final height in girls with sexual precocity after therapy with the GnRH analogue D-TRP-6-LHRH. Horm Res. 33:11-17. 7. Bahar I. 1988 The development of over the counter assays for pregnanediol-3-glucuronide and estrone-b, d-glucuronide. Prog Clin Biol Res. 285:139-51. 8. Chatterton Jr RT, Haan JN, Jenco JM, Cheesman KL. 1982 Radioimmunoassay of pregnanediol concentrations in early morning urine specimens for assessment of luteal function in women. Fertil Steril. 37:361-6. 9. Miller MM, Hoffman DI, Creinin M, et al. 1990 Comparison of endometrial biopsy and urinary pregnanediol glucuronide concentration in the diagnosis of luteal phase defect. Fertil Stertil. 54:100811 10. Laird NA, Ware JH. 1982 Random-effects models for longitudinal data. Biometric. 38:963-74. 11. Treloar AE, Boynton RE, Behn BG, Brown BW. 1967 Variation of the human menstrual cycle through reproductive life. Int J Fertil. 12:77-126. 12. Sandler DP, Wilcox,AJ, Horney LF. 1984 Age at menarche and subsequent reproductive events, Am J Epidemiol. 119:765-74. 13. Murram D, Dewhurst J, Grant DB. 1984 Precocious puberty: a follow up study. Arch Dis Child. 59:77-B. 14. Grant DB, Murram D, Dewhurst J. 1985 Precocious puberty:
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
894
15.
16.
17.
18.
19.
JAY implications on adult reproductive function. In: Venturoli S, Flamigni C, Givens JR, eds. Adolescence in females. Chicago: YearBook; p 335. Borsos A, Lampe L, Bulough A, Csoknyay J, Ditroi F, Szckely P. 1988 Ovarian function after the menarche and hormonal contraception. Int J Gynaecol Obstet. 27:249-53. Campbell H, Edstrom K, Engstrom L, et al. 1986 World Health Organization multicenter study on menstrual and ovulatory patterns in adolescent girls. II. Longitudinal study of menstrual patterns in the early postmenarcheal period, duration of bleeding episodes and menstrual cycles. J Adolesc Health. 7:236. Lemarchand-Beraud T, Zufferey MM, Reymond M, Rey I. 1982 Maturation of the hypothalmo-pituitary-ovarian axis in adolescent girls. J Clin Endocrinol Metab. 54:241-6. Apter D. 1980 Serum steroids and pituitary hormones in female puberty: partly longitudinal study. Clin Endocrinol (Oxf). 12:10720. Apter D, Vihko R. 1985 Hormonal patterns of the first menstrual cycles. In: Venturoli S, Flaigni C, Givens JR, eds. Adolescence in
ET AL.
JCE & M. 1992 Vol75.No3
females, Chicago: YearBook; p 215 20. Metcalf MG, Skidmore DS, Lowry GF, MacKenzie JA. 1983 Incidence of ovulation in the years after the menarche. J Endocrinol. 97:213-19. 21. Venturoli S, Porcu E, Fabbri R, et al. 1986 Ovarian multifollicularity, high LH and androgen plasma levels, and anovulation are frequent-and strongly linked in adolescent irregular cycles. Acta Endocrinol (Copenh). 111:368-72. 22. Read GF, Wilson DW, Hughes IA, Griffiths K. 1984 The use of salivary progesterone assavs in the assessment of ovarian function in post;n’ena&heal girls. J Endocrinol. 102:265-8. 23. Siegberg R, Nilsson CG, Stenman U. Widholm. 1984 Sex hormone pro&es& bligomenorrheic adolescent girls and the effect of oral contraceptives. Fertil Steril. 41:888-93. 24. Prior JC, Vigna Y. Absence of molimina: the clinical or selfdiagnosis of anovulation [Abstract]. Proc of the 7th Conf of the Sot of Menstrual Cycle Res. 1988. 25. Magyar DM, Boyers SP, Marshall JR, Abraham GE. 1979 Regular menstrual cycles and premenstrual molimina as indicators of ovulation. Obstet Gynecol. 53:41 I-14.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
Vol. 75, No. 3 Printed
Ovulation and Menstrual with Central Precocious Gonadotropin-Releasing NAOMI JAY, M. JOAN WILLIAM F. CROWLEY, AND PAUL A. BOEPPLE
MANSFIELD, JR., DAVID
Function of Adolescent Puberty after Therapy Hormone Agonists*
ROBERT M. BLIZZARD, SCHOENFELD, LESLIE
in U.S.A.
Girls with
RHUBIN,
Pediatric and Reproductive Endocrine Units and General Clinical Research Center, Massachusetts General Hospital, Boston, Massachusetts 02114; the Divisions of Endocrinology and Adolescent Medicine, Department of Medicine, Children’s Hospital, Boston, Massachusetts 02115; the Division of Endocrinology and Diabetes, Children’s Medical Center of the University of Virginia, Charlottesville, Virginia 22908; and Hygeia Sciences, Newton, Massachusetts 02160-1432 became increasingly regular, with cycles of 25- to 35-day duration reported by 41% of the girls in the first year postmenarche and 65% of the girls studied 3 or more years postmenarche. Ovulation was demonstrated in 50% of the girls studied within 1 yr of menarche and in 90% of the girls studied 2 yr or more postmenarche, including 5 girls who reported pregnancies. The development of regular ovulatory menstrual function in these girls with CPP is in accord with previously documented patterns in normal adolescents. While these data provide further evidence supporting the safety of long term GnRHa therapy, continued studies will be necessary to characterize fully the reproductive function in CPP patients through adolescence and adulthood. (J Clin Endocrinol Metab 75: 890-894, 1992)
ABSTRACT Chronic GmRH agonist (GnRHa) administration has been shown to suppress pituitary-gonadal function in children with central precocious puberty (CPP), but long term data after the reactivation of gonadarche posttherapy are not yet available. This study evaluated the menstrual function of 46 girls with CPP who had been treated for at least 2 yr with GnRHa (deslorelin or histrelin, SC, daily) and were up to 7 yr posttreatment, including 21 postmenarcheal girls who collected weekly overnight urine samples for 12 consecutive weeks to assess rates of ovulation by urinary pregnanediol-3@-glucuronide measurements. Menarche occurred at age 12.1 f 1.0 yr (mean + SD), on the average 1.2 + 0.8 yr posttherapy (range, 0.1-4.3 yr). Menstrual cycle lengths
L
Patient
Received July 15, 1991. Address requests for reprints to: Paul A. Boepple, M.D., Reproductive Endocrine Unit, Massachusetts General Hospital, Boston, Massachusetts 02114. * This work was supported in part by NIH Grants HD-18169, RR01066, RR-02172, and RR-08847 and FDA Grant FD-R-000388. Presented in part at the 5th Annual North American Society for Pediatric and Adolescent Gynecology, Fort Lauderdale, FL, April 11,. 1991, and the 73rd Annual Meeting of The Endocrine Society, Washington, D.C., June 22,. 1991.
Female patients who had been treated for CPP with GnRHa at the Massachusetts General Hospital, Children’s Hospital (Boston, MA), and the Children’s Medical Center at the University of Virginia (Charlottesville, VA) were recruited for this study. Diagnostic criteria for CPP and details of the treatment regimen and monitoring protocols have been described previously (1, 2). Briefly, all patients had the onset of breast development before age 8 yr, and all demonstrated a pulsatile pattern of gonadotropin secretion and pubertal response to GnRH before therapy. Pituitary-gonadal suppression was achieved with daily SC injections of either deslorelin (4-8 pg/kg.day) or histrelin (8-10 pg/kg.day). Suppression was confirmed during in-patient evaluations performed at 3- to 6-month intervals during therapy. All aspects of this project were approved by the Institutional Review Boards of the 3 participating hospitals. Forty-nine subjects had been treated with GnRHa (duration, ~1.9 yr) and were now off therapy. Of these, 3 subjects were excluded due to either the presence of other medical diagnoses that were known to interfere with menstrual functioning or noncompliance with the medical regimen, which resulted in incomplete suppression throughout the course of their therapy. An additional 4 patients who reported declining compliance in the final few months of up to 4 yr of therapy were included in this analysis, since adequate suppression had been documented in them at all prior evaluations. Thirty-eight of 46 eligible patients participated, with retrospective review of hospital records providing data on the remaining 8 girls. To evaluate the return of puberty and menstrual function in all GnRHa-treated patients, girls with neurogenic precocious puberty (NIP; n = 7) were included in this study despite the potential for this group to manifest disordered neuroendocrine function. The diagnoses represented among the NPP patients included cerebral astrocytoma treated with cranial irradiation, congenital
ONG term administration of potent agonist analogs of GnRH (GnRHa) induces persistent suppression of the pituitary-gonadal axis and has been successfully employed as a treatment for central precocious puberty (CPP) for more than a decade (1-4). While the reactivation of gonadarche after discontinuation of GnRHa treatment has been prompt in patients with CPP (5, 6), little information has been available regarding the completion of their pubertal maturation. In females, this is marked by the development of estrogen positive feedback and the ability to mount a LH surge, evidenced by attainment of regular menstrual cycles, consistent ovulation, and corpus luteum formation. The present study was undertaken to chart the development of ovulatory menstrual function in adolescent females who had completed GnRHa treatment for CPP.
Materials and Methods population
890
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
OVULATORY toxoplasmosis, myelomeningocele, tosis, supracellar cyst, and perinatal
neonatal asphyxia.
meningitis,
FUNCTION neurofibroma-
IN PRECOCIOUS TABLE therapy(n
1. Patient = 46) Clinical
Protocol This study sought to define three characteristics of our female CPP patients after GnRHa therapy: the timing of menarche, the lengths of subsequent menstrual cycles, and the rates of ovulation associated with those cycles. Even though 18 patients had experienced at least 1 episode of menstrual bleeding before the initiation of therapy, we use the term menarche here to refer to the first menstrual period following the discontinuation of GnRHa administration. The timing of menarche was established either by history at the time of a General Clinical Research Center evaluation posttherapy or by responses to a questionnaire. While the questionnaire also included items that required respondents to describe characteristics of their cycles (e.g. duration of cycle, duration of flow, and associated symptoms), only contemporaneous menstrual cycle diaries were employed to determine menstrual cycle lengths. In addition to the existing individual records that were available for analysis, 28 girls maintained prospective menstrual diaries for a 3-month period, during which 26 of the 28 patients collected first morning voided urine specimens at weekly intervals. Urine samples were immediately frozen and subsequently assayed for pregnanediol3/?-glucuronide (PDG) as an index of progesterone secretion and corpus luteum function.
Hormone
analysis
Qualitative determinations of PDG were made using a membranebased immunoassay (Hygeia Sciences, Newton, MA), employing a previously characterized anti-PDG monoclonal antibody (7). Assay sensitivity (-4 nmol/L) was set to correspond to a value 2 SD greater than the mean follicular phase urinary PDG concentration (8, 9). To further validate these results, a subset of samples (56 of the total 301, 18.6%) was randomly chosen for quantification in a previously validated PDG RIA (8, 9). In this subset, all 12 samples containing 4 nmol/L PDG or more by RIA were positive in the qualitative immunoassay, as were 5 of the 30 samples whose RIA-determined concentration was between 1.4-4 nmol/L. All of the samples whose quantitative results were less than 1.4 nmol/L PDG yielded negative results in the qualitative assay. Thus, employing an RIA determination of 4 nmol/L or more as our positive reference, the qualitative I’DG assay performed with a sensitivity level of 100% and a specificity of 89%.
Data analysis Data regarding the attainment of menarche were analyzed by using maximum likelihood estimates, assuming normal distribution, thus allowing the inclusion of girls who remained premenarcheal. To assess trends in cycle length over time, the deviation from the adult normal range (after square root transformation) was viewed relative to time, using a random patient effect model (10). For purposes of analysis, normal cycle length was considered to be 25-35 days, as previously documented in healthy adults (11). In girls with normal length cycles, intraindividual variation was further evaluated. A patient was said to have regular cycles if all were between 25-35 days and varied by no more than 5 days. Data regarding the attainment of menarche were analyzed separately for patients with idiopathic (n = 39) and neurogenic (n =. 7) precocity, but are reported together since the groups exhibited no significant differences. Onlv one patient with NPI’ (supracellar cyst) partizpated in the prospective’compbnent of this study; tius, ovulatory rates could not be compared between the two diagnostic subsets.
Results The characteristics of the 46 girls with CPP are summain Table 1. The subset whose data permitted estimates of ovulatory rates did not differ significantly from the group overall in terms of treatment and posttreatment histories. rized
PUBERTY characteristics
event (yr)
Age at initiation of therapy Age at discontinuation of therapy Duration of GnRHa therapy Age at menarche Timing of menarche posttherapy Timing of study postmenarche
891 in girls
with
CPP post-GnRHa
Mean f 7.3 10.9 3.6 12.1 1.2 1.6
f + f + f +
SD
Range
1.3 0.9 1.2 1.0 0.8 1.4
4.1-9.2 9.1-13.9 1.9-7.0 10.1-14.9 0.1-4.3 O-6.9
Menarche
Of the total 46 subjects, 44 (96%) had attained menarche by the completion of the study. The agesand intervals that are relevant to this milestone appear in Table 1. The 2 patients who had not yet reached menarche were still within 1 SD of our cohort’s mean age and time to menarche. The variance in the interval between the cessation of treatment and attainment of menarche among patients was examined for its correlation with various clinical characteristics of the population. The interval to menarche had a modest but significant (P < 0.03) negative relationship with chronological age (CA); that is, older girls tended to attain menarche more quickly after therapy stopped (r’ = 0.42). In a multiple regression model with CA as a covariate, the inclusion of duration of treatment as a second covariate accounted for an additional small increase (P < 0.04) in the interval to menarche, increasing the r* value to 0.48. In other words, having controlled for the impact of CA, girls treated for longer periods with GnRHa tended to take longer to reach menarche after stopping therapy. However, most of the variance among the patients was left unaccounted for, with other factors (e.g. age at onset of puberty, history of mensespretherapy, and bone age, weight, or Tanner stage breast development posttherapy) displaying no significant associations.In addition, there were no significant differences in timing of menarche between patients with a neurogenic etiology us. those with idiopathic CPP. Of the eight girls who reached menarche within 6 months of treatment cessation (>l SD early), declining compliance during the final months of therapy may have contributed in half of the subjects. However, the remaining four girls who reached menarche early had been consistently compliant and appeared completely suppressedwhen therapy was discontinued. The average time to menarche did not change significantly if the four patients who exhibited waning compliance in the final months of therapy were removed from analysis. Cycle length and regularity
Menstrual cycle lengths (223 cycles in 34 patients) are plotted zx years postmenarche in Fig. 1. The deviation of cycle length from that in normal adult women decreased with prolonged follow-up (P = 0.04). During the 12 weeks of prospective study of 28 girls, 20 (70%) were observed to have cycles of normal lengths, but only 13 (45%) demonstrated regular cycles, i.e. 2 or more cycles between 25-35 days, which varied less than 5 days in length. Subjective retrospective data on cycle length often differed substantially from contemporaneous reporting, in that the great majority
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
892 Q
JAY ET AL. >60
-o
l
JCE & M. 1992 Vol75.No3
100
1
A 80
P 4
60
2 0
40
8
.0 1001
-
&
1
2
n= 121
3 n = 49
Em
I
< 25 25 -35
4
5 n = 29
6
7
n = 24
Davs Da&
20
0 I
o-1
1-2
2-3
>3
YEARS POST-MENARCHE FIG. 2. Rates of ovulation in CPP girls post-GnRHa therapy plotted relative to time postmenarche. Ovulatory girls (U) were those in whom urinary PDG was positive (>4 nmol; see Materials and Methods) in 1 or more weekly samples during 12 weeks or who had reported pregnancies. Ovulatory cycles (ES@ were those in which a luteal phase ur&ary PDG test was positive. The total numbers of eirls and cvcles studied in each year appear above the relevant bar. Rates of ovulation increased significantly with increasing time postmenarche (by x2 test, P = 0.001).
20 t-l
o-1
l-2
2-3
>3
YEARS POST-MENARCHE FIG. 1. Menstrual cycle lengths in 34 CPP girls post-GnRHa therapy plotted relative to time postmenarche. In both A (individual cycles) and B (group data), menstrual cycles between 25-35 days in length are considered normal (see Materials and Methods) and are designated by the dotted lines or solid bars in A and B, respectively. The deviation of cycle length from that in normal adult women decreased with prolonged follow-up (P = 0.04).
of patients reported via the questionnaire that their cycles were consistently 28-30 days in length while their diaries reflected far more variability. Ovulation
Of the 26 girls who contributed weekly urine samples, 3 remained premenarcheal, and 2 were excluded from analysis due to the initiation of therapy with medroxyprogesterone acetate by their primary care providers after histories of oligomenorrhea. These 21 postmenarcheal girls were considered to display ovulatory menstrual function if 1 or more urine specimenswas positive in the qualitative PDG assay. In addition, 5 patients who reported pregnancies and 1 who had a documented serum progesterone level of 37.5 nmol/L during a General Clinical ResearchCenter evaluation posttherapy were also included among the ovulatory patients for the purposesof analysis. Data for these 27 girls are depicted in Fig. 2 relative to time postmenarche. In this cross-sectional analysis, demonstration of ovulatory function increased significantly as patients were studied posttherapy. This trend was significant if the analysis was based on the percentage of patients who ovulated at least once during the 3 months of observation or on the percentage of cycles in which a urinary PDG test was positive. In the latter instance, the
lower rates of ovulation depicted in Fig. 2 reflect that fact that patients who ovulated in at least 1 cycle often did not do so in all cycles during the course of the 3-month study. Of the girls deemed to be ovulatory, only 5 had positive PDG levels in all cycles observed, and an additional 4 patients ovulated in 2 of 3 cycles. In our patients, normal cycle lengths and premenstrual/ menstrual symptoms lacked both specificity and sensitivity as clinical correlates of ovulation. Anovulatory and ovulatory cycles fell within the normal range of 25-35 days with comparable frequencies (59% and 57%, respectively), and 66% of all long cycles (>35 days) examined were associated with ovulatory values of urinary PDG. Normal, regular length menstrual cycles in our patients, thus, correlated poorly with a biochemical index of ovulation (r = 0.02). Similarly, symptoms were reported during 50% of anovulatory cycles, making their correlation with ovulation modest in our study population (r = 0.34). Discussion
In addition to reactivation of pubertal pituitary-gonadal function, completion of secondary sexual development, and attainment of adult stature, a full assessmentof the long term safety and efficacy of GnRHa therapy of CPP will hinge upon the documentation of eventual normal fertility in treated patients (5, 6). While assessmentof ultimate fertility must await a longer follow-up of this still young population, the development of ovulatory menstrual function in the present subset of adolescent girls with CPP must, for now, serve as our index of completed pubertal maturation. The mean age at menarche of 12.1 yr in the present study is within the normal range for age at menarche in North America (12). However, the timing of menarche in this patient group is determined primarily by the decision of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
OVULATORY
FUNCTION
when to stop GnRHa therapy. Therefore, examination of the interval between stopping treatment and the development of menarche facilitates comparisons among individual patients and between studies. The findings of the present study are similar to those of Manasco et al. (5) and Kauli et al. (6), in which 67% and 80% of subjects, respectively, reached menarche by 1 yr post-GnRHa treatment. While the return of pulsatile gonadotropin secretion has been uniformly documented in our patients with idiopathic CPP within 6 months of the discontinuation of suppressive therapy, the individual variation in interval to menarche has been quite substantial. While there was a significant tendency for girls to reach menarche earlier if GnRHa was discontinued at an older age and/or after treatment of shorter duration, most of the individual variation in our study population was left unexplained and was probably related to factors, such as body composition, diet, and exercise, that are known to affect the rate of sexual maturation but were not considered in this study. The natural history of CPP appears to include normal expectations for adult reproductive function (13, 14). However, it was not surprising that cycle irregularity and anovulation were frequently evident in our patient cohort who were studied an average of only 1.6 yr postmenarche, since variable cycle lengths and inconsistent ovulation typically persist during the first 5-6 yr of menstruation in normal adolescent females (15-21). The appropriate comparisons were, therefore, critical to put the data in our adolescent patients into perspective. First, it was important to base analyses of cycle length on contemporaneous diaries, since recollections reported on questionnaires frequently overestimated the regularity of cycles. Second, it was important to employ a direct biochemical index of ovulation, since regular cycle lengths and the presence of menstrual symptoms were not reliable clinical markers of ovulation in our adolescent subjects, contrary to their apparent validity in adult populations (23-25). Finally, comparisons of our findings in CPP with those reported in normal adolescents had to take methodological details into account. While the rates of ovulation in the present study were higher than those reported in many studies of normal subjects (17-22), these estimates are significantly influenced by the interval of observation and the parameter monitored (e.g. single day 21 serum sample for progesterone VS. multiple determinations of urinary PDG over several cycles). Metcalf et al. (20), who studied adolescent girls with the same serial integrative index of ovulation (weekly urinary PDG measurements) over the same extended observation period (3 months) as we employed in our CPP patients, reported similar rates of ovulation (41.7% of girls in the first year postmenarche, 80% of those in the fifth year). Thus, using the most comparable normative data, the development of ovulatory menstrual function after GnRHa therapy in our patients with CPP appears to be appropriate for their age and developmental stage. GnRH agonists are being increasingly employed in a multitude of clinical situations. The recent FDA approval of histrelin as a therapy for CPP is likely to facilitate the use of GnRHa in a growing list of pediatric indications. Consequently, long term follow-up of patients treated in childhood
IN PRECOCIOUS
PUBERTY
893
becomes all the more critical. While the results of this study provide further evidence of the safety and reversibility of long term GnRHa treatment, continued longitudinal study through adulthood will be necessary to describe fully the reproductive function of this population. In this regard, it will be particularly important to diligently follow patients with histories of neurogenic precocity. While the progression of puberty posttherapy in this subset was no different from that in patients with idiopathic CPP in the present study, menarche is a milestone that is relatively insensitive to the subtle disorders of neuroendocrine function that NPP patients may subsequently manifest. Finally, the five pregnancies in our young patients occurring as early as age 13.8 yr, while reassuring vis iz vis the demonstration of normal fertility, underscore the need to address further the psychosocial impact of early sexual maturation and its therapy. Acknowledgments The GnRHa used in these studies Rivier. The Salk Institute (deslorelin), (histrelin).
were the generous gifts of Jean and Ortho Pharmaceuticals, Inc.
References 1. Crowley Jr WF, ComiteF, Vale W, Rivier J, Loriaux DL, Cutler GB. 1981 Therapeutic use of pituitary desensitization with longacting LHRH agonist: a potential new treatment for idiopathic precocious puberty. J Clin Endocrinol Metab. 52:370-2. 2. Boepple PA, Mansfield MJ, Wierman ME, et al. 1986 Use of a potent, long-acting agonist of gonadotropin-releasing hormone in the treatment of precocious puberty. Endocr Rev. 7:24-33. 3. Manasco PK, Pescovitz OH, Suvimol C, et al. 1989 Six-year results of luteinizing hormone releasing hormone (LHRH) agonist treatment in children with LHRH-dependent precocious puberty. J Pediatr. 115:105-8. 4. Kaplan SL, Grumbach MM. 1990 Pathophysiology and treatment of sexual precocity. J Clin Endocrinol Metab. 71:785-9. 5. Manasco PK, Pescovitz OH, Feulillan PP, et al. 1988 Resumption of puberty after long-term luteinizing hormone-releasing hormone agonist treatment of central precocious puberty. J Clin Endocrinol Metab. 67:368-72. 6. Kauli R, Kornreich L, Laron Z. 1990 Pubertal development, growth and final height in girls with sexual precocity after therapy with the GnRH analogue D-TRP-6-LHRH. Horm Res. 33:11-17. 7. Bahar I. 1988 The development of over the counter assays for pregnanediol-3-glucuronide and estrone-b, d-glucuronide. Prog Clin Biol Res. 285:139-51. 8. Chatterton Jr RT, Haan JN, Jenco JM, Cheesman KL. 1982 Radioimmunoassay of pregnanediol concentrations in early morning urine specimens for assessment of luteal function in women. Fertil Steril. 37:361-6. 9. Miller MM, Hoffman DI, Creinin M, et al. 1990 Comparison of endometrial biopsy and urinary pregnanediol glucuronide concentration in the diagnosis of luteal phase defect. Fertil Stertil. 54:100811 10. Laird NA, Ware JH. 1982 Random-effects models for longitudinal data. Biometric. 38:963-74. 11. Treloar AE, Boynton RE, Behn BG, Brown BW. 1967 Variation of the human menstrual cycle through reproductive life. Int J Fertil. 12:77-126. 12. Sandler DP, Wilcox,AJ, Horney LF. 1984 Age at menarche and subsequent reproductive events, Am J Epidemiol. 119:765-74. 13. Murram D, Dewhurst J, Grant DB. 1984 Precocious puberty: a follow up study. Arch Dis Child. 59:77-B. 14. Grant DB, Murram D, Dewhurst J. 1985 Precocious puberty:
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
894
15.
16.
17.
18.
19.
JAY implications on adult reproductive function. In: Venturoli S, Flamigni C, Givens JR, eds. Adolescence in females. Chicago: YearBook; p 335. Borsos A, Lampe L, Bulough A, Csoknyay J, Ditroi F, Szckely P. 1988 Ovarian function after the menarche and hormonal contraception. Int J Gynaecol Obstet. 27:249-53. Campbell H, Edstrom K, Engstrom L, et al. 1986 World Health Organization multicenter study on menstrual and ovulatory patterns in adolescent girls. II. Longitudinal study of menstrual patterns in the early postmenarcheal period, duration of bleeding episodes and menstrual cycles. J Adolesc Health. 7:236. Lemarchand-Beraud T, Zufferey MM, Reymond M, Rey I. 1982 Maturation of the hypothalmo-pituitary-ovarian axis in adolescent girls. J Clin Endocrinol Metab. 54:241-6. Apter D. 1980 Serum steroids and pituitary hormones in female puberty: partly longitudinal study. Clin Endocrinol (Oxf). 12:10720. Apter D, Vihko R. 1985 Hormonal patterns of the first menstrual cycles. In: Venturoli S, Flaigni C, Givens JR, eds. Adolescence in
ET AL.
JCE & M. 1992 Vol75.No3
females, Chicago: YearBook; p 215 20. Metcalf MG, Skidmore DS, Lowry GF, MacKenzie JA. 1983 Incidence of ovulation in the years after the menarche. J Endocrinol. 97:213-19. 21. Venturoli S, Porcu E, Fabbri R, et al. 1986 Ovarian multifollicularity, high LH and androgen plasma levels, and anovulation are frequent-and strongly linked in adolescent irregular cycles. Acta Endocrinol (Copenh). 111:368-72. 22. Read GF, Wilson DW, Hughes IA, Griffiths K. 1984 The use of salivary progesterone assavs in the assessment of ovarian function in post;n’ena&heal girls. J Endocrinol. 102:265-8. 23. Siegberg R, Nilsson CG, Stenman U. Widholm. 1984 Sex hormone pro&es& bligomenorrheic adolescent girls and the effect of oral contraceptives. Fertil Steril. 41:888-93. 24. Prior JC, Vigna Y. Absence of molimina: the clinical or selfdiagnosis of anovulation [Abstract]. Proc of the 7th Conf of the Sot of Menstrual Cycle Res. 1988. 25. Magyar DM, Boyers SP, Marshall JR, Abraham GE. 1979 Regular menstrual cycles and premenstrual molimina as indicators of ovulation. Obstet Gynecol. 53:41 I-14.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 06:18 For personal use only. No other uses without permission. . All rights reserved.
