Mood changes and physical complaints during the normal menstrual cycle in healthy young women

Psychoneuroendocrinology, Vol. 15, No. 2, pp. 131-138,

1990

0306-4530/90 $3.00 + 0.00 ©1990 Pergamon Press plc

Printed in Great Britain

MOOD CHANGES AND PHYSICAL COMPLAINTS DURING THE NORMAL MENSTRUAL CYCLE IN HEALTHY YOUNG WOMEN REINHOLD G. LAESSLE, REINHARD J. TUSCHL, ULRICH SCHWEIGER, and KARL M. PIRKE Division of Psychoneuroendocrinology, Max-Planck-lnstitute of Psychiatry, Munich, ER.G. (Received 14 September 1988; Infinal form 24 January 1990)

SUMMARY Significant emotional and physical symptoms have been linked to hormonal changes during the menstrual cycle. A critical evaluation of the available studies questions the commonly held belief in menstrual-cycle-related complaints in the majority of normal women. The present study investigated changes in mood, somatic complaints and vegetative variables during the menstrual cycle in 30 healthy young women. Normal cycle function was evaluated and cycle phases were defined according to endocrine data. For all subjects, blood samples were taken at least three times a week to measure estmdiol and progesterone. Daily ratings of psychological variables revealed no significant changes in global mood or depression over the cycle. Somatic complaints such as abdominal pain and breast tenderness were significantly related to the luteal, premenstrual, and menstrual phases. Appetite increased in the periovulatory and premenstrual phases. There was a tendency for sexual interest to be highest in the post-menstrual period. Affect and vegetative variables showed no association with hormone levels but were significantly correlated with subjective stress ratings. We conclude that in most healthy young women, cycle-related hormone fluctuations are not accompanied by marked affective changes. Specific physical complaints, however, do occur, particularly in the luteal, premenstrual, and menstrual phases.

INTRODUCTION A VARIETYof physical and emotional changes have been linked to hormonal fluctuations during the menstrual cycle. The luteal phase and particularly the premenstrual period in many women seemed to be associated with psychological symptoms such as increased depression, irritability or tension, and also with physical symptoms such as breast tendemess, abdominal pain, headache, back pain and nausea (for review see Dennerstein & Abraham, 1982). In a subgroup of women these premenstmal changes appeared to be severe enough to constitute a syndromal disorder, the so-called "premenstrual syndrome" (PMS), which has been extensively investigated (for review see Osofsky & Blumenthal, 1985) and has been proposed as a new diagnostic category ("Late Luteal Phase Dysphoric Disorder") in DSM-III-R. Gonadal hormones were postulated as possible causative factors for menstrual-cycle-related psychological and physical changes as early as the 1930s (Frank, 1931). More recently, theories of increased levels of progesterone, as well as changes in progesterone levels or estrogen/progesterone ratio, have been proposed (Btickstr6m et al., 1983; Sanders & Reinisch, 1985). However, critical reviews of the available studies indicated serious methodological problems Correspondence to be addressed to: Dr. Karl M. Pirke, Max-Planck-lnsfitute of Psychiatry, KraepelinstraBe 10, D-8000 Miinchen 40, ER.G. 131

132

R.G. LAESSLEet al.

(Ruble & Brooks-Durra, 1979; Halbreich & Endicott, 1985). A major difficulty is the lack of endocrinological data to specify cycle phases. Without such measurement it is impossible to evaluate menstrual cycles for normality, which is a necessary precondition for the existence of meaningful cycle phases. Anovulatory cycles seem to occur relatively frequently among young women (Metcalf, 1983), but they usually remain undetected. In such cases, there is no significant rise and fall of progesterone, and therefore effects from this hormone seem unlikely. Furthermore, the variability of normal menstrual cycles in total length, duration of phases, and levels of reproductive hormones is large. Inferring cycle phases merely from onset of menses and assuming equal phase length for all subjects under study is inadequate and may yield dubious results when correlating cycle phases to physical or psychological symptoms. Another problem of many studies has been the lack of prospective measurement of behavioral changes during the cycle. Generally, there is a great discrepancy between the results gained from retrospective assessments and those based on daily reports. Since this discrepancy becomes even greater when women do not know the aim of the investigation, it has been concluded that retrospective data on the relationship between cycle phase and symptoms mainly reflect socioculturally transmitted expectations and beliefs about menstruation. A third problem of most of the previous investigations is the lack of samples representative of normal women. In a number of studies the subjects were explicitly recruited on the basis of having experienced (pre)menstrual problems. The physical and psychological health of research subjects also often remains unproven. Whether results based on such samples can be generalized and applied to the majority of normal women seems questionable. Considering the above-mentioned and other methodological problems, no definite conclusions from the available data can be drawn, and the common belief of menstrual-cycle-related affective and physical symptoms in normal women has to be seriously questioned. Surprisingly enough, during the last decade there have been only few attempts to further investigate cycle-related changes in normal women. The studies using adequate methodology (Abplanalp et al., 1979; Sanders et al., 1983) found no systematic variation of symptoms across endocrinologically defined cycle phases, nor was there a significant relationship between levels of ovarian hormones and symptoms. A further clarification of those mechanisms seems necessary and also of particular relevance for a better understanding of PMS. Of great importance is whether symptoms constituting PMS are merely a quantitative exaggeration of cyclical changes occurring in most normal women or whether PMS is a qualitatively distinct syndrome, possibly linked to other psychiatric disturbances such as affective disorders. The present study is an attempt to retest the hypothesis of menstrual-cycle-related affective and physical changes in normal women, healthy both physically and psychologically. The major methodological problems of previous studies were addressed by (1) endocrinological monitoring of the cycle, (2) prospective daily assessment of symptoms, and (3) keeping subjects blind to the purpose of the analysis. SUBJECTS AND METHODS

Subjects

The present study was part of a larger investigation dealing with effects of dieting on the menstrual cycle (for details see Pirke et al., 1986; Schweiger et al., 1987). All data reported here were collected during the control phase of this study, where subjects lived under normal everyday conditions and were not dieting. It was also assured that subjects who reported repeated dieting in the past were not included in the study. Originally, 38 women were recruited and gave their informed consent to participate in the study. After evaluation of endocrine data it was determined that only 30 subjects had normal menstrual cycles when the following criteria were applied (Schweiger et al., 1987): (1) Maximum value of progesterone at least 6 ng/ml and (2) length of luteal phase at least eight days. These remaining 30 women formed the sample for the present analysis. Their mean age was

133

PSYCHOSOMATICCHANGESANDTHE MENSTRUALCYCLE

24.4 yr (SD=2.5 yr), and their mean body mass index was 22.1 kg/m 2 (SD=I.3 kg/m2). All subjects were students. Laboratory and medical examination and psychiatric assessment by a semistructured clinical interview detected no presence or history of any major medical or psychiatric disorder. None of the subjects took oral contraceptives or had taken medication during the previous four months. All were told that the aim of the study was to determine the effect of dieting on menstrual function. It was explained that this makes necessary the investigation of a control cycle under non-dieting conditions and the daily assessment of a wide variety of variables. All subjects were paid for their participation.

Endocrine methods Data collection started on the first day of a new menstrual cycle. Blood was sampled by venipuncture between 0800h and 1000h every Monday, Wednesday and Friday. Additional daily samples were taken from 26 subjects around the midcycle peak. Estradiol (E2) and progesterone (P4) were measured by radioimmunoassay after ether extraction (Radioisotopen Service, Wuerlingen, Switzerland; Travenol, Munich, Germany). Intra-assay variability was 7.3%, and interassay variability was 12.5% at an average P4 concentration of 2.6 ng/mi. Intra-assay variability was 6.9%, and interassay variability was 11.3% at an average E2 concentration of 83 pg/ml. All samples were analysed under continuous quality control.

Definition of cycle phases Each cycle was divided into five phases - - menstrual, mid-follicular, peri-ovulatory, early/mid-luteal, and premenstrual. The day of the E 2 maximum was defined as the day of the midcycle peak. The menstrual phase was the time from the first day to the last day of menstrual bleeding. The mid-follicular (post-menstrual) phase was defined as the time from the end of the menstrual phase until two days before midcycle. The periovulatory phase was the five-day period centered around the midcycle. The early/mid-luteal phase was defined from three days after midcycle until three days before the onset of the next menstrual bleeding. These last three days were regarded as the late luteal or premenstrual phase. The endocrine cycle data for the sample, based on individually calculated longitudinal means per cycle phase, are presented in Table I.

TABLE I. ENDOCRINE CYCLE CHARACTERISTICS OF THE SAMPLE {N = 30) Mean

SD

Cycle len tg~ (days)

27.80

5.80

Follicular.phase* Estradiol ( p g / m l ) Progesterone (ng/ml)

78.50 0.68

40.30 0.21

Periovulator.y_phase Estradiol m a x i m u m ( p g / m l )

356.00

138.60

Early/mid-luteal phase Estradiol ( p g / m l ) Progesterone ( n g / m l )

190.40 11.70

83.50 3.40

Premenstrual phase Estradiol ( p g / m l ) Progesterone ( n g / m l )

150.10 6.90

91.40 4.50

*Average concentrations calculated from individual mean valuesper phase.

Assessment of affective and physical changes During the entire cycle, subjects kept a standardized diary of proven reliability and validity (StrauB & Appelt, 1983). Global mood and depression were rated on visual analogue scales (Celia & Perry, 1986); higher values indicate more negative affect. Somatic complaints included headache, nausea, abdominal pain, breast tenderness and back pain and were rated on a three-point scale (1: no, 2: moderate, 3: severe). Appetite and sexual interest were also rated on three-point scales. Sleep quality, stress at work/study and interpersonal stress were rated on five-point scales; higher values indicate more negative experiences. For each individual, mean values for each

R.G. LAESSLEetal.

134

cycle phase were calculated for global mood, depression, sleep duration, sleep quality and stress. For somatic complaints, the number of days with moderate or severe complaints within each cycle phase was expressed as a percentage of total phase length. The same procedure was applied for appetite and sexual interest. A total score of somatic complaints also was calculated.

Statistical analysis The effects of cycle phase on the variables described above were tested by multivariate analysis of variance for repeated measurements (MANOVA: O'Brien & Kaiser, 1985). If the overall test was significant at the 0.05 level or less, pairwise post-hoc comparisons between cycle phases were carded out. These comparisons were adjusted for multiple testing according to the Bonferroni-Holm method. In the present case of 10 pairwise post-hoc tests, a mean difference was considered significant at an adjusted 0.05 level if the single test reached a probability level of at least 0.005. RESULTS Table II shows the m e a n changes in affect variables over the menstrual cycle phases. Neither global m o o d nor depression varied significantly over the cycle.

TABLE II. MOOD AS A FUNCTION OF CYCLE PHASE [MEAN (SD}] Early Follicular Mid-Follicular Peri(Menstrual) (Post-Menstrual) Ovulatory Global Mood Depression

35.3 (13.2) 29.9 (14.0)

33.3 (12.8) 30.0 (17.0)

35.1 (15.0) 33.3 (18.6)

Early/MidPreLuteal Menstrual 35.5 (14.5) 34.4 (19.0)

32.7 (16.3) 31.0 (21.7)

MANOVA df F p< 4,25 4,25

.45 .80 .66 .70

Mean changes in physical complaints are presented in Table IlL The overall comparisons for effects o f cycle phase revealed significant results for the the total complaint score and for abdominal pain and breast tenderness. Pairwise post-hoc tests showed a significant increase of complaints in the p r e m e n s t r u a l and menstrual phases c o m p a r e d to the three other phases. Breast tenderness was already significantly greater in the early/mid-luteal phase than in the mid-follicular or p e r i o v u l a t o r y phase. A b d o m i n a l pain was highest during the p r e m e n s t r u a l and menstrual phases, but showed an increase already in the early/mid-luteal phase.

TABLE III. SOMATIC COMPLAINTS AS A FUNCTION OF CYCLE PHASE [MEAN {SD}] Early Follicular Mid-Follicular Peri(Menstrual) (Post-Menstrual) Ovulatory Total Score Headache Nausea Back pain Abdominal pain Breast tenderness

15.1 (15.5) A8" 6.0 (7.4) Ac 14.1 (26.3) 18.7 (24.6) 6.2 (11.3) 4.2 (9.2) 19.5 (31.6) 6.9 (11.1) 28.9 (39.0) A8 0.0 (0.0) AcD 12.4 (26.9) 0.9 (4.6) A8

6.4 (6.9) 8D 11.7 (16.4) 10.7 (18.0) 9.7 (19.7) 2.0 (8.0) BE 0.7 (3.7) cD

Early/MidPreLuteal Menstrual 11.2 (12.9) 19.8 (15.3) 17.4 (22.4) 26.4 (33.8) 7.4 (13.1) 8.9 (13.0) 13.3 (23.2) 14.9 (27.6) 8.3 (16.8) c 18.9 (24.3) DE 14.8 (23.9) AC 22.2 (36.4) 8D

MANOVA df F p< 4,25 4,25 4,26 4,25 4,25 4,26

5.6 1.2 1.0 1.5 7.6 3.7

.002 .40 .50 .30 .001 .02

* If two means are followed by the same letter, they are significantly different with an adjusted p<0.05 for 10 pairwise comparisons.

135

PSYCHOSOMATICCHANGESANDTHE MENSTRUALCYCLE

Vegetative function variables are shown in Table IV. A significant effect of cycle phase was found for appetite, with peaks in the periovulatory and in the premenstrual phase. The post-hoc comparisons revealed a significant difference between the mid-follicular and the peri-ovulatory phases. Sexual interest showed a tendency to be highest in the postmenstrual period (MANOVA: p<0.06). Post-hoc tests were significant for the comparison of postmenstrual with early/midluteal phase. No significant effects were found for either sleep duration or sleep quality.

TABLE IV. VEGETATIVEVARIABLESAS A FUNCTION OF CYCLE PHASE [MEAN (SD}] Early Follicular Mid-Follicular Peri(Menstrual) (Post-Menstrual) Ovulatory

Early/MidLuteal

PreMenstrual

MANOVA df F p<

Appetite 20.0 (36.0) 19.2 (31.3)a " 34.0(33.7)a 24.0 (32.6) 33.3(39.1) 4,24 2.8 .05 Sexual interest 47.9 (38.0) 65.1 (33.0)A 50.7 (33.8) 45.9(31.4)a 54.8 (40.8) 4,24 2.7 .06 Sleep duration 7.1 (1.7) 7.4 (0.7) 7.3 (0.9) 7.3 (0.7) 7.2 (1.0) 4,25 0.2 .92 (hours) Sleep quality 2.2 (0.7) 1.7 (0.6) 1.9 (0.7) 2.1 (0.6) 2.1 (0.8) 4,26 1.9 .14 *if two means are followedby the same letter, they are significantly different with an adjusted p
To investigate possible relationships of hormone levels to psychological and physical symptoms, correlations of these variables to progesterone and estradiol levels were calculated within the early/mid-luteal and the premenstrual phases. Progesterone levels during the premenstrual phase were negatively related to headache (p =-.41, p <0.01) and back pain (p =-.42, p <0.01). No relationships between levels of gonadal hormones and affect variables could be shown. However, subjective ratings of interpersonal stress were significantly correlated with both global mood (p = .38, p < 0.05) and depression (p = .53, p < 0.01) during the early/mid-luteal phase as well as during the premenstrual period (global mood: p=.45, p<0.05; depression: p =.55, p < 0.01). Higher levels of interpersonal stress during the early/mid-luteal phase were also associated with somatic complaints such as headache (p = .36, p < 0.05) and nausea (p = .39, p < 0.05). In the premenstrual phase, higher ratings of interpersonal stress were significantly related to headache (p=.54, p<0.01), back pain (p=.50, p<0.01), worse sleep quality (p =.44, p<0.05), and increased appetite (p =.37, p <0.05). Higher ratings for stress at work/study correlated significantly with increased appetite (p = .51, p < 0.01) and with decreased sexual interest (p =-.44, p < 0.05). DISCUSSION Our data do not support the commonly held assumption that affective fluctuations related to the menstrual cycle do exist for most normal women. These results confirm the findings of Abplanalp et al. (1979), who also had assessed cycle phases endocrinologicaUy. They also are in line with data from prospective studies in which subjects had been unaware that menstrual symptoms would be analysed (Parlee, 1982; Slade, 1984; AuBuchon & Calhoun, 1985). Therefore, it may be concluded that many previous reports on cycle-dependent negative affect reflect expectations and beliefs rather than changes which were truly experienced. This interpretation is supported by the study of Olasov and Jackson (1987) in which expectations concerning a

136

R.G. LAESSLEet al.

mood-cycle relationship were experimentally manipulated, yielding significant effects. The same phenomenon also had been demonstrated by Ruble (1977). Similar to the results of Abplanalp et al. (1979), mood was unrelated to levels of gonadal hormones, which also was the case in a study of PMS sufferers and normal controls (Rubinow et al., 1988), in whom the presence or absence of affective symptoms related to PMS was not related to differences in hormone levels. However, the existence of an association between endocrinological events during the cycle and emotional state cannot be definitely ruled out by our data. In particular, analyses of rates of change of estradiol or progesterone during premenstrual or menstrual phase in relation to psychological state are of interest but could not he performed in our study, because blood samples, which would have taken into account the pulsatile nature of progesterone secretion, were not collected frequently enough for this purpose. On the other hand, the affective state of the women in our sample was significantly correlated with aspects of their daily life, especially with interpersonal or work-related stress events. This had been observed by Wilcoxon et al. (1976) and Straul3 and Appelt (1983). It seems very plausible that stress has a great influence on affective state, but one should also take into account the possibility that alterations in affective state may contribute to the perception of stress. In contrast to the results on affective state, specific physical symptoms were significantly related to the early/mid-luteal, the premenstrual, and the menstrual phases, These symptoms, including in particular abdominal pain and breast tenderness, have been reported consistently in previous prospective studies (Janowsky et al., 1973; Beumont et al., 1975). Therefore, it seems reasonable to assume that such somatic complaints indeed vary systematically over the menstrual cycle, with an exaggeration in the premenstrual period. Pain symptoms during the premenstrual phase were more severe the lower the progesterone levels were. Because symptom levels were low during the follicular phase, when progesterone levels were also low, it is unlikely that this correlation reflects a relationship between symptoms and absolute hormone levels during the premenstrual period. Our sampling intervals were too wide to determine the rate of decline of progesterone concentration, but we might speculate that the physical symptoms are due to the premenstrually occurring decrease in progesterone levels. This interpretation is consistent with data of Halbreich et al. (1986) showing a high correlation (p = .74) between decrease of progesterone and physical discomfort in the premenstrual phase. Nevertheless, it has to he considered that stress factors also accounted for a significant amount of variance in physical symptoms and should not be neglected. Analysis of change in specific vegetatively regulated functions revealed a significant effect of cycle phase on appetite ratings, with peaks in the periovulatory and in the premenstrual phases. For the premenstrual phase this has been similarly demonstrated for normal women in the prospective studies of Abraham et al. (1981) and Both-Orthman et al. (1988). Although animal studies suggest coincidence between levels of gonadal hormones and food intake (Czaja, 1975), no convincing empirical evidence of such a relationship exists in humans. We did not find a significant correlation of appetite rating either with progesterone or with estradiol. Instead, during the premenstrual period, stress was related to increased appetite. Nevertheless, some caution is warranted when attempting to generalize the present results for the appetite ratings. It has to be considered that the subjects, although not dieting during the time the data were collected, had originally been recruited for the investigation of the effects of dieting on cycle function. This may have influenced self-selection of the subjects in a way that women who were more sensitive to their eating behavior were more likely to volunteer for the study. In accordance with the studies of Schreiner-Engel et al. (1981) and Bancroft et al. (1983), no increase in sexuality associated with ovulation was observed. Instead, sexual interest showed a tendency to be highest in the post-menstrual period. This is consistent with the majority of

PSYCHOSOMATICCHANGESANDTHE MENSTRUALCYCLE

137

previous studies (Sanders & Bancroft, 1982). A widely accepted explanation is that this pattern results from the abstinence that usually accompanies the days of menstruation, but it also has been suggested that there exists a biologically determined basic cyclical pattern which is only amplified by the effects of menstrual abstinence (Bancroft et al., 1983). In the present study no evidence was found to support this latter assumption, because there were no significant correlations between estradiol or progesterone levels and sexual interest. However, we did not measure testosterone, which also has been hypothesized to be linked to sexuality in women (Bancroft et al., 1983). Similar to the results of Straui3 and Appelt (1983), our data revealed a significant relationship between stress factors and sexual interest. Considering the rather inconsistent results of many previous studies which tried to link hormone levels to sexual appetite, a strong connection between endocrine factors and sexuality in normal young women seems unlikely; the psychosocial factors of everyday life appear to have a much more important role in determining sexual motivation. With regard to the pathogenesis of PMS (or Late Luteal Phase Dysphodc Disorder), it may be concluded from our data that the premenstrual affective changes in those patients are not exaggerations of normally occurring hormone related mood swings, but rather seem to be of a distinct origin. Premenstrual physical symptoms, however, are present in most normal women, but may be perceived as more severe by PMS patients due to their altered affective state.

REFERENCES Abplanalp JM, Donnelly AF, Robert MR (1979) Psychoendocrinology of the menstrual cycle: I. Enjoyment of daily activities and moods. Psychosom Med 41: 587-615. Abraham SF, Beumont PJV, Argall WJ, Haywood P (1981) Nutrient intake and menstrual cycle. Aust NZ J Med 11: 210-211. AuBuchon P, Calhoun K (1985) Menstrual cycle symptomatology: the role of social expectancy and experimental demand characteristics. Psychosom Med 47: 35-45. B~ickstr6m T, Baird DT, Bancroft J, Bixo M, Hammarb~ick S, Sanders D, Smith S, Zetterlund B (1983) Endocrinological aspects of cyclical mood changes during the menstrual cycle or the premenstrual syndrome. J Psychosom Obstet Gynecol 2: 8-20. Bancroft J, Sanders D, Davidson D, Wamer P (1983) Mood, sexuality, hormones, and the menstrual cycle. III. Sexuality and the role of androgens. Psychosorn Med 45: 509-516. Beumont PJV, Richards DH, Gelder MG (1975) A study of minor psychiatric and physical symptoms during the menstrual cycle. Br J Psychiatry 126:431- 434. Both-Orthman B, Rubinow DR, Hoban C, Malley J, Grover GN (1988) Menstrual cycle phase-related changes in appetite in patients with premenstrual syndrome and in control subjects. Am J Psychiatry 145: 628-631. Cella DF, Perry SW (1986) Reliability and concurrent validity of three visual-analogue mood scales. Psychol Rep 59: 827-833. Czaja JA (1975) Food rejection by female rhesus monkeys during the menstrual cycle and early pregnancy. PhysiolBehav 14: 579-587. Dennerstein L, Abraham SF (1982) Affective changes and the menstrual cycle. In: Beumont PJV, Burrows GD (Eds) Handbook of Psychiatry and Endocrinology. Elsevier, Amsterdam, pp 367-400. Frank RT (1931) The hormonal causes of premenstrual tension. Arch Neurol Psychiatry 26" 1053-1057. Halbreich U, Endicott J (1985) Methodological issues in studies of premenstrual changes. Psychoneuroendocrinology 10: 15-32. Halbreich U, Endicott J, Goldstein S, Nee J (1986) Premenstrual changes in gonadal hormones. Acta Psychiatr Scand 74: 576-586. Janowsky DS, Berens SC, Davis JM (1973) Correlations between mood, weight, and electrolytes during the menstrual cycle: a reninangio-tension-aldosterone hypothesis of premenstrual tension. Psychosom Med 35: 143-154. Metcalf M (1983) Incidence of ovulation from the menarche to the menopause: observations of 622 New Zealand women. N Z Med J 96: 645-648. O'Brien RG, Kaiser MK (1985) MANOVA method for analyzing repeated measures designs: an extensive primer. Psychol Bull 97: 316-324.

138

R.G. LAESSLEet al.

Olasov B, Jackson J (1987) Effects of expectancies on womens' reports of moods during the menstrual cycle. Psychosom Med 49: 65-78. Osofsky I-IJ, Blumenthal SJ (1985) Premenstrual Syndrome: Current Findings and Future Directions. American Psychiatric Press, Washington DC. Parlee M (1982) Changes in moods and activation levels during the menstrual cycle in experimentally naive subjects. Psychol Women 7:119-131. Pirke KM, Schweiger U, Laessle RG, Dickhaut B, Schweiger M, Waechtler M (1986) Dieting influences the menstrum cycle: vegetarian versus nonvegetarian diet. Fertil Steri146: 1083-1088. Rubinow DR, Hobart MC, Grover GN, Galloway DS, Roy-Byme P, Anderson R, Merriam GR (1988) Changes in plasma hormones across the menstrual cycle in patients with menstrually related mood disorder and in control subjects. Am J Obstet Gyneco1158: 5-11. Ruble DN (1977) Premenstrual symptoms: a reinterpretation. Science 197: 291-292. Ruble DN, Brooks-Dunn J (1979) Menstrual symptoms: a social cognition analysis. J Behav Med 2:171-194. Sanders D, Bancroft J (1982) Hormones and the sexuality of women - - the menstrual cycle. Clin Endocrinol Metab 11: 639-660. Sanders D, Warner P, Biickstr/Sm T, Bancroft J (1983) Mood, sexuality, hormones and the menstrual cycle. I. changes in mood and physical state: description of subjects and method. Psychosom Med 45: 487-500. Sanders SA, Reinisch JM (1985) Behavioral effects on humans of progesterone-related compounds during development and in the adult. In: Ganten D, Pfaff D (Eds) Current Topics in Neuroendocrinology, Vol 5. Raven Press, New York, pp 176-205. Schreiner-Engel P, Schiavi RC, Smith H, White D (1981) Sexual arousability and the menstrual cycle. Psychosom Med 43: 199-214. Schweiger U, Laessle RG, Pfister H, Hoehl C, Schwingenschloegel M, Schweiger M, Pirke KM (1987) Dietinduced menstrual irregularities: effects of age and weight loss. Fertil Steri148: 746-751. Slade P (1984) Premenstrual emotional changes in normal women: fact or fiction? Psychosom Res 28: 1-7. Straul] B, Appelt H (1983) Psychological concomitants of the menstrual cycle. J Psychosom Obstet Gynecol 4: 215-220. Wilcoxon LA, Schrader SL, Sherif CW (1976) Daily self-reports on activities, life events, moods, and somatic changes during the menstrual cycle. Psychosom Med 38: 399-417.