Insulin-like growth factor-1 and insulin-likegrowth factor-binding protein-3 in womenwith premenstrual syndrome

Andrea J. Rapkin, M.D.,* Marcelle Cedars, M.D.,† Melinda Morgan, Ph.D.,* andLinda Goldman, R.N.P., M.N.*

University of California, Los Angeles Medical Center, Los Angeles, California, and University of Cincinnati,Cincinnati, Ohio

Objective: To determine whether women with premenstrual syndrome (PMS) have aberrations of the GHaxis as has been demonstrated in individuals with depression.

Design: Prospective trial.

Setting: Department of Obstetrics and Gynecology, University of California, Los Angeles.

Patient(s): After prospective screening, 32 healthy women with PMS and 32 asymptomatic controls com-pleted the study.

Intervention(s): Subjects completed a daily PMS symptom diary and a Beck Depression Inventory. Theyunderwent phlebotomy 5 days and 12 days after the LH midcycle surge, which was identified with the use ofa urinary LH detection kit.

Main Outcome Measure(s): Serum levels of insulin-like growth factor-1 (IGF-1), insulin-like growthfactor-binding protein-3 (IGFBP-3), E2, and progesterone.

Result(s): Levels of IGF-1, IGFBP-3, E2, and progesterone did not differ between women with prospectivelydocumented PMS and control subjects.

Conclusion(s): Premenstrual syndrome and affective disorder share common symptoms and possibly acommon cause. Biochemical markers such as alterations in the somatotropic system often are associated withmajor depression. Levels of IGF-1 and IGFBP-3 did not differ between women with PMS and control subjects,supporting the concept that PMS and endogenous depression are biologically distinct entities. (Fertil Sterilt1998;70:1077–80. ©1998 by American Society for Reproductive Medicine.)

Key Words: Premenstrual syndrome, premenstrual dysphoric disorder, growth hormone, insulin-like growthfactor-1, insulin-like growth factor-binding protein-3

Premenstrual syndrome (PMS) is a psycho-neuroendocrine disorder consisting of disablingaffective, behavioral, and physical symptoms thatoccur for up to 2 weeks before menses with reliefsoon after the onset of the menstrual period. Pre-menstrual syndrome shares certain clinical fea-tures with affective disorder, and there is a statis-tical association between the diagnosis of PMSand the future development of affective disorder,in particular, major depressive disorder (1–4).This association has led some investigators tosuggest that PMS is a variant of affective disorder(5, 6). However, various biochemical “markers”that are highly associated with affective disorder,including dysfunction of the hypothalamic-pitu-itary-adrenal axis, have not been documented inwomen with PMS (4, 7).

Another neuroendocrine aberration noted inpatients with major depressive illness is analteration of the somatotropic system (i.e., theGH–insulin-like growth factor axis) (8–16).We therefore attempted to determine whetherthere was an abnormality of the somatotropicsystem in women with PMS compared withasymptomatic women.

MATERIALS AND METHODS

Subjects with PMS and controls were re-cruited through newspaper advertisements. Thestudy was performed in the Obstetrics and Gy-necology Consult Suite and Annex at the Uni-versity of California, Los Angeles. Each sub-ject signed a consent form, and the study was

Received June 15, 1998;revised and accepted July27, 1998.Presented at the Societyfor GynecologicInvestigation Meeting,Atlanta, Georgia, March11–14, 1998.Reprint requests: Andrea J.Rapkin, M.D., University ofCalifornia, Los AngelesMedical Center,Department of Obstetricsand Gynecology, 10833 LeConte, Room 22-177 CHS,Los Angeles, California90095-1740 (FAX: 310-206-6531).* Department of Obstetricsand Gynecology, Universityof California, Los AngelesMedical Center.† Department of Obstetricsand Gynecology, Collegeof Medicine, University ofCincinnati.

FERTILITY AND STERILITY tVOL. 70, NO. 6, DECEMBER 1998Copyright ©1998 American Society for Reproductive MedicinePublished by Elsevier Science Inc.Printed on acid-free paper in U.S.A.

0015-0282/98/$19.00PII S0015-0282(98)00389-6

1077

approved by the Institutional Review Board at the Universityof California, Los Angeles. Eligibility criteria for all subjectsincluded age 18–40 years, history of regular menses (25–32days), and no hormonal, psychotropic, or nutritional supple-mentation or recreational drug or alcohol use for the monthbefore or during the study.

Subjects were evaluated for 2 months before acceptanceinto the study. During this time, all subjects were required tocomplete a daily symptom diary consisting of 12 of the morecommonly experienced symptoms of PMS: anxiety, depres-sion, mood swings, irritability-anger, breast tenderness, di-minished energy, edema or bloating, food craving, sleepdisturbances, avoidance of social interaction, reduced capac-ity for work, and decreased concentration. Each symptomwas rated on a scale of 1 (not at all) to 6 (extreme). Inaddition, the subjects completed the Beck Depression Inven-tory (17) during the follicular phase of the first prestudymonth. Each subject underwent a psychiatric interview andmedical examination.

Candidates accepted as PMS subjects had the followingcharacteristics: [1] a negative psychiatric history; [2] a nor-mal physical and psychiatric evaluation; [3] a follicular-phase Beck Depression Inventory score of,10; [4] at leastfive moderate to severe premenstrual symptoms severeenough to disrupt interpersonal relationships or daily activ-ities for 7–14 days before menses, with one or more of thesefive being a mood symptom; [5] at least a 50% increase inthe severity of five of the symptoms from the follicular phaseto the luteal phase; and [6] no more than minimal physical ormood symptoms during the follicular phase. Control subjectswere required to fulfill criteria 1, 2, 3, and 6, and to have noremarkable premenstrual symptoms. All women with PMSwere required to fulfill criteria of the Diagnostic and Statis-tical Manual of Mental Disorders, 4th revision, for premen-strual dysphoric disorder (18).

All subjects completed the symptom diary for 3 months.During the last study month, the subjects used a urinary LHdetection kit in the periovulatory period to time the midcycleLH peak. Fasting morning serum sampling for insulin-likegrowth factor-1 (IGF-1), insulin-like growth factor-bindingprotein-3 (IGFBP-3), E2, and progesterone was performed 5days and 12 days after the LH midcycle surge (days 19 and26 of an idealized 28-day cycle).

Serum samples were frozen and stored at270°C. Hor-mone levels were determined at the University of CincinnatiSchool of Medicine. Serum IGF-1 levels were determined bya two-site, noncompetitive immunoradiometric assay(IRMA,Diagnostic Systems Laboratories, Webster, TX) after acid-ethanol extraction. Samples were run undiluted with an assaysensitivity of 0.8 ng/mL. The intra-assay and interassaycoefficients of variation were 3% and 5%, respectively.Serum IGFBP-3 levels were determined by a two-site, non-competitive IRMA (Diagnostic Systems Laboratories) with-out extraction. All samples were run at a 1:100 dilution. The

sensitivity of the assay was 0.5 ng/mL, with intra-assay andinterassay coefficients of variation of 2% and 5%, respec-tively.

Levels of E2 were measured by RIA with the use of theCoat-a-Count system (Diagnostics Products Corporation,Los Angeles, CA). The sensitivity of the assay was 8 pg/mL.The intra-assay and interassay coefficients of variation wereboth 7%. Levels of progesterone were determined with asolid-base RIA (Diagnostic Products Corporation). The in-tra-assay and interassay coefficients of variation were 5.6%and 9%, respectively.

The within-group and between-group significance ofdaily diary mood symptoms during the follicular and lutealphases was assessed by nonparametric procedures with theuse of the Mann-WhitneyU-test for independent samplesand the Wilcoxon test for dependent samples. Between-group and across-time analyses of hormonal data were per-formed with the use of analysis of variance with repeatedmeasures. Between-group hormonal levels on days 19 and26 also were compared with the use of Student’st-test.Anovulatory subjects, as determined by progesterone levelsof ,4 ng/mL, were deleted from the data analysis.

RESULTS

Thirty-two women with PMS and 32 controls completedthe study. There were no statistically significant differencesin the mean (6SD) age of the women with PMS (316 6 y)and the controls (286 8 y). All subjects were of normalweight and none had any physical illnesses. Symptom scoresfor the two groups assessed during 5 days of the follicularphase (days 8–12) and 5 days of the luteal phase (days24–28) of the menstrual cycle are presented in Table 1.

Subjects with PMS were significantly more symptomaticwith respect to all diary symptoms in the luteal phase. Asexpected, controls did not manifest this difference. The twogroups did not differ significantly during the follicular phase.

Mean (6SD) IGF-1, IGFBP-3, and E2, and progesteroneconcentrations are presented in Table 2. Mean (6SD) IGF-1,IGFBP-3, E2, and progesterone concentrations did not differsignificantly between the two groups, and only progesteronewas significantly different across time (Table 2).

DISCUSSION

There are similarities as well as differences between thepremenstrual appearance of affective symptoms as part ofPMS and affective symptoms associated with major depres-sive disorder. Common features include clinical symptomssuch as depression, irritability, anxiety, and mood swings(6, 19). There also is an increased lifetime prevalence ofaffective disorder in women with PMS (1–3, 20). In addi-tion, PMS and endogenous depression may involve similaralterations in the activity of the neurotransmitter, serotonin

1078 Rapkin et al. IGF-1 and IGFBP-3 in PMS Vol. 70, No. 6, December 1998

(21–23), and both PMS and major depression respond fa-vorably to treatment with selective serotonin reuptake inhib-itors (24).

However, there are differences between PMS and affec-tive disorder as well. The biochemical indices that often areaberrant in depression, such as decreased levels of mono-amine oxidase-b (25) and abnormalities of the hypothalam-ic-pituitary-adrenal axis (26, 27), have not been observed inwell-screened women with PMS (4, 7, 28). The primaryobjective of this study, therefore, was to determine whetherthe somatotropic system was altered in women with PMS

compared with controls. We elected to assess levels ofIGF-1, the major effector protein for GH, and IGFBP-3, themajor binding protein for IGF-1, rather than serum GHlevels, which show a marked diurnal and pulsatile variation(29–32).

Individuals with affective disorder have altered GH se-cretory activity (8–16). The basal GH secretory activitypatterns of individuals with affective disorder reported byother investigators have been inconsistent; for example, the24-hour secretory patterns revealed daytime GH hypersecre-tion (8, 10), but diminished sleep-related GH secretion alsohas been described (9, 11). On the other hand, the stimulatedGH response to GH-releasing hormone, clonidine, and dexa-methasone consistently has been found to be attenuated inindividuals with affective disorder (12, 13). Levels of IGF-1were found to be uniformly elevated in three separate studiesin patients with endogenous depression (14, 15, 33).

The lack of concordance in the results of GH studies inpatients with depression may be due at least in part to thepulsatility and circadian rhythm inherent in GH secretion.However, IGF-1, which lacks pulsatile variation, is the pri-mary predictor of GH action, and its activity depends on GHsecretion (29–31). Levels of IGF-1 and IGFBP-3 remainrelatively constant and are less affected by acute changes inGH secretion related to the time of day or the stress of testing(29, 32).

In our studies, we did not observe any differences in basalIGF-1 and IGFBP-3 levels in women with PMS comparedwith controls. Our findings, taken together with the findingsof other investigators who failed to observe differences in

T A B L E 1

Daily diary symptoms in women with PMS and in controls.

Symptom

Median (range) value of indicated symptom

Follicular phase Luteal phase

PMS Controls PMS Controls

Avoid social activity 1.25 (1.0–3.0) 1.07 (1.0–2.0) 2.89 (1.0–6.0)* 1.30 (1.0–6.0)Decreased interest in work 1.59 (1.0–6.0) 1.09 (1.0–2.0) 3.17 (1.0–6.0)* 1.27 (1.0–5.0)Edema/weight gain 1.37 (1.0–4.0) 1.05 (1.0–2.0) 3.69 (1.0–6.0)* 1.54 (1.0–4.0)†Depression 1.31 (1.0–3.0) 1.12 (1.0–2.0) 3.42 (1.0–5.0)* 1.22 (1.0–4.0)Anxiety 1.31 (1.0–3.0) 1.13 (1.0–2.0) 3.16 (1.0–6.0)* 1.31 (1.0–3.0)Mood swings 1.26 (1.0–2.0) 1.04 (1.0–3.0) 3.41 (1.0–6.0)* 1.25 (1.0–6.0)Irritability 1.48 (1.0–6.0) 1.12 (1.0–2.0) 3.24 (1.0–6.0)* 1.36 (1.0–6.0)Increased appetite/cravings 1.38 (1.0–3.0) 1.12 (1.0–3.0) 3.61 (1.0–6.0)* 1.59 (1.0–6.0)†Low energy/fatigue 1.47 (1.0–3.0) 1.25 (1.0–3.0) 3.55 (1.0–5.0)* 1.47 (1.0–4.0)Headaches 1.47 (1.0–4.0) 1.23 (1.0–3.0) 2.24 (1.0–5.0)* 1.22 (1.0–3.0)Breast pain 1.08 (1.0–3.0) 1.02 (1.0–2.0) 3.16 (1.0–5.0)* 1.34 (1.0–5.0)†Sleep disturbance 1.06 (1.0–2.0) 1.09 (1.0–3.0) 1.7 (1.0–5.0)* 1.08 (1.0–2.0)Difficulty concentrating 1.18 (1.0–2.0) 1.13 (1.0–2.0) 2.97 (1.0–6.0)* 1.22 (1.0–4.0)

Note: PMS 5 premenstrual syndrome.* P ,.001 (follicular vs. luteal phase).† P ,.01 (luteal phase PMS vs. control).

T A B L E 2

Hormonal parameters in women with PMS and in controls.

Day of measurement

Mean6 SD level of indicated hormone

Women with PMS Control group

Day 19IGF-1 (ng/mL) 304.536 104.38 342.596 149.54IGFBP-3 (ng/mL) 2,972.436 421.76 3,029.386 481.70E2 (pg/mL) 115.856 51.16 109.746 46.14Progesterone (ng/mL) 10.666 3.79* 8.886 4.66

Day 26IGF-1 (ng/mL) 315.236 102.21 337.666 151.68IGFBP-3 (ng/mL) 3,088.156 424.39 2,998.676 556.69E2 (ng/mL) 101.046 59.35 99.056 61.41Progesterone (ng/mL) 7.646 5.41* 6.396 4.96

Note:IGF-15 insulin-like growth factor-1; IGFBP-35 insulin-like growthfactor binding protein-3; PMS5 premenstrual syndrome.* P ,.01 (day 19 vs. day 26).

FERTILITY & STERILITY t 1079

the hypothalamic-pituitary-adrenal axis, indicate that the af-fective symptoms that occur in women with PMS are un-likely to be biologically similar to the symptoms of majordepression. This may be due either to differences in thepathophysiology of the disorders or to the fact that thehypothalamic-pituitary-adrenal axis or GH axis changes thatcharacterize depression require prolonged affective symp-tomatology as opposed to the brief cyclic occurrence ofmonthly premenstrual symptoms (4).

Our secondary objective was to assess whether thechanges in E2 and progesterone levels between the midlutealand late luteal phase also were accompanied by changes inIGF-1 and IGFBP-3 levels. We did not find any differencesin IGF-1 and IGFBP-3 levels, suggesting that fluctuations inE2 levels and, in particular, the rise and fall of progesteronelevels, did not affect IGF-1 or IGFBP-3 levels. These resultsconfirmed previously described findings of a lack of cycle-related differences in IGF-1 and IGFBP-3 levels (34).

This study assessing IGF-1 and IGFBP-3 as markers ofGH activity did not find an alteration in the GH–insulin-likegrowth factor axis in the luteal phase in well-screened sub-jects with PMS. The lack of variation in IGF-1 and IGFBP-3levels between groups of women with and without PMSonce again supports the concept that PMS and endogenousdepression are most likely biologically distinct entities. Fu-ture studies evaluating these parameters and GH-releasinghormone–or dexamethasone-stimulated growth hormone se-cretory responses in groups of women with PMS and majordepressive disorder would be of interest.

Acknowledgments:The authors thank Ms. Lily Kao for her technical ex-pertise.

References1. Endicott J, Halbreich U, Schacht S, Nee J. Premenstrual changes and

affective disorders. Psychosom Med 1981;43:519–29.2. Kashiwagi T, McClure JH, Wetzel RD. Premenstrual affective syn-

drome and psychiatric disorder. Dis Nerv Syst 1976;37:116–9.3. DeJong R, Rubinow DR, Roy-Byrne P, Hoban MC, Grover GN, Post

RM. Premenstrual mood disorder and psychiatric illness. Am J Psychi-atry 1985;142:1359–61.

4. Mortola JF, Girton L, Yen SC. Depressive episodes in premenstrualsyndrome. Am J Obstet Gynecol 1989;161:1682–7.

5. Endicott J. The menstrual cycle and mood disorders. J Affect Disord1993;29:193–200.

6. Rubinow DR, Schmidt PJ. Models for the development and expressionof symptoms in premenstrual syndrome. Psychiatr Clin North Am1989;12:53–68.

7. Roy-Byrne PP, Rubinow DR, Gwirtsman H, Hoban MC, Grover GN.Cortisol response to dexamethasone in women with premenstrual syn-drome. Neuropsychobiology 1986;16:61–3.

8. Linkowski P, Mendlewicz J, Kerkhofs M, Leclercq R, Golstein J,Brasseur M, et al. 24-Hour profiles of adrenocorticotropin, cortisol andgrowth hormone in major depressive illness: effect of antidepressanttreatment. J Clin Endocrinol Metab 1987;65:141–52.

9. Jarrett DB, Miewald JM, Kupfer DJ. Recurrent depression is associatedwith a persistent reduction in sleep-related growth hormone secretion.Arch Gen Psychiatry 1990;47:113–8.

10. Mendlewicz J, Linkowski P, Kerkhofs M, Desmedt D, Golstein J,Copinschi G, et al. Diurnal hypersecretion of growth hormone indepression. J Clin Endocrinol Metab 1985;60:505–12.

11. Voderholzer U, Laakmann G, Wittmann R, Daffner-Bujia C, Hinz A,Haag C, et al. Profiles of spontaneous 24-hour and stimulated growthhormone secretion in male patients with endogenous depression. Psy-chiatry Res 1993;47:215–27.

12. Laakmann G, Neuhauser H, Hinz A, Neulinger E, Wittmann M. Stim-ulation of growth hormone, prolactin and cortisol by GNRH and DMIin depressed patients. Pharmacopsychiatry 1986;19:235–6.

13. Lesch K-P, Laux G, Pfu¨ller H, Erb A, Beckmann H. Growth hormone(GH) response to GH-releasing hormone in depression. J Clin Endo-crinol Metab 1987;6:1278–81.

14. Lesch K-P, Laux G, Erb A, Pfu¨ller H, Beckmann H. Growth hormone(GH) responses to GH-releasing hormone in depression: correlationwith GH release following clonidine. Psychiatry Res 1988;25:301–10.

15. Deuschle M, Blum WF, Strasburger CJ, Schweiger U, Weber B, Ko¨rnerA, et al. Insulin-like growth factor-I (IGF-I) plasma concentrations areincreased in depressed patients. Psychoneuroendocrinology 1997;22:493–503.

16. Lesch K-P, Mu¨ller U, Rupprecht R, Kruse K, Schulte HM. Endocrineresponses to growth-releasing hormone, thyrotropin-releasing hormoneand corticotropin-releasing hormone in depression. Acta PsychiatrScand 1989b;79:597–602.

17. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventoryfor measuring depression. Arch Gen Psychiatry 1961;4:561–71.

18. American Psychiatric Association. Diagnostic and statistical manual ofmental disorders–IV. 4th ed. Washington, DC: American PsychiatricAssociation, 1994.

19. Halbreich U. Menstrually related disorders: what we do know, what weonly believe that we know, and what we know that we do not know. CritRev Neurobiol 1995;9:163–75.

20. Steiner M. Female-specific mood disorders. Clin Obstet Gynecol 1992;35:599–611.

21. Rapkin AJ, Edelmuth E, Chang LC, Reading AE, McGuire MT, Su T.Whole-blood serotonin in premenstrual syndrome. Obstet Gynecol1987;70:533.

22. Rapkin A. The role of serotonin in PMS. Clin Obstet Gynecol 1992;35:629–35.

23. Coppen AJ, Doogan DP. Serotonin and its place in the pathogenesis ofdepression. J Clin Psychiatry 1988;49:4–11.

24. Steiner M, Steinberg S, Stewart D, Carter D, Berger C, Reid R.Fluoxetine in the treatment of premenstrual dysphoria. N Engl J Med1995;332:1529–34.

25. Murphy DL, Weiss R. Reduced monoamine oxidase activity in bloodplatelets from bipolar depressed patients. Am J Psychiatry 1972;128:1351–7.

26. Linkowski P, Mendelewicz J, Leclercq R, Brasseur M, Huboin P,Golstein J. The 24-hour profile of adrenocorticotropin and cortisol inmajor depressive illness. J Clin Endocrinol Metab 1985;61:429–38.

27. Rubin AL, Price LH, Charney S, Heninger GR. Noradrenergic functionand the cortisol response to dexamethasone in depression. PsychiatryRes 1985;15:5–15.

28. Rapkin AJ, Buckman TD, Sutphin MS, Chang LC, Reading AE. Plate-let monoamine oxidase B activity in women with premenstrual syn-drome. Am J Obstet Gynecol 1988;159:1536–40.

29. Blum WF, Albertsson-Wikland K, Rosberg S, Ranke MB. Serum levelsof insulin-like growth factor I (IGF-I) and IGF binding protein 3 reflectspontaneous growth hormone secretion. J Clin Endocrinol Metab 1993;76:1610–6.

30. Blum WF, Ranke MB, Jenne EW, Kietzmann K, Bierich JR. Thephysiological role of IGF binding protein. In: Frisch H, Thorner MO,editors. Hormonal regulation of growth. New York: Raven Press,1989:153–70.

31. Blum WF, Ranke MB, Kietzmann K, Gauggel E, Zeisel HJ, Bierich JR.A specific radioimmunoassay for the growth hormone (GH)-dependentsomatomedin-binding protein: its use for diagnosis of GH deficiency.J Clin Endocrinol Metab 1990;70:1292–8.

32. Brier BH, Bass JJ, Butler JH, Gluckman PD. The somatotrophic axis inyoung steers: influence of nutritional status on pulsatile release ofgrowth hormone and circulating concentrations of insulin-like growthfactor 1. J Endocrinol 1986;111:209–15.

33. Lesch K-P, Rupprecht R, Mu¨ller U, Pfuller H. Comparison of GHresponses after human GHRH-44 amide administration and TRH-in-duced TSH release in depressed patients. Biol Psychiatry 1989a;25:235–8.

34. Wang H-S, Lee J-D, Soong Y-K. Serum levels of insulin-like growthfactor I and insulin-like growth factor-binding protein-1 and -3 inwomen with regular menstrual cycles. Fertil Steril 1995;63:1204–9.

1080 Rapkin et al. IGF-1 and IGFBP-3 in PMS Vol. 70, No. 6, December 1998