Bond University

ePublications@bond
Humanities & Social Sciences papers

Faculty of Humanities and Social Sciences

1-1-1994

Aerobic exercise, mood states and menstrual cycle

symptoms
Julie A. Aganoff
University of Queensland

Gregory J. Boyle
Bond University, [email protected]

Follow this and additional works at: http://epublications.bond.edu.au/hss_pubs

Part of the Biological Psychology Commons
Recommended Citation
Julie A. Aganoff and Gregory J. Boyle. (1994) "Aerobic exercise, mood states and menstrual cycle
symptoms" ,, .
http://epublications.bond.edu.au/hss_pubs/37
This Journal Article is brought to you by the Faculty of Humanities and Social Sciences at ePublications@bond. It has been accepted for inclusion in
Humanities & Social Sciences papers by an authorized administrator of ePublications@bond. For more information, please contact Bond University's
Repository Coordinator.

Aerobic exercise, mood states and menstrual cycle symptoms

By Julie A. Aganoff and Gregory J. Boyle
Based on a paper presented at the 28th Annual Conference of the Australian
Psychological Society, Gold Coast, Qld., 29 September2 October 1993.
Address correspondence to: Dr G.J. Boyle, Associate Professor of Psychology, Bond
University, Gold Coast, Qld 4229, Australia

This study examined the effects of regular, moderate exercise on mood states and
menstrual cycle symptoms. A group of female regular exercisers (N = 97), and a
second group of female non-exercisers (N = 159), completed the Menstrual Distress
Questionnaire (MDQ) and the Differential Emotions Scale (DES-IV) premenstrually,
menstrually and intermenstrually. Multivariate analyses of covariance (MANCOVAs)
revealed significant effects for exercise on negative mood states and physical
symptoms, and significant effects on all measures across menstrual cycle phase. The
regular exercisers obtained significantly lower scores on impaired concentration,
negative affect, behaviour change and pain. No differences were found between
groups on positive affect and other physical symptoms.
Introduction
Evidence exists for behavioral and somatic changes across the menstrual cycle [ 1, 2].
Some studies have found significant changes in mood states [3-7], while others have
found no significant changes [8, 9]. The premenstrual phase* occurs over at least 4
days prior to onset of menstruation [8, 10-12]. Up to 97% of women experience some
physical symptoms and mood changes premenstrually. Some 50% experience minor
changes premenstrually, while 35% experience symptoms and mood changes that
disrupt work, social and family life. Approximately 5-10% experience severely
debilitating symptoms that cause major disruptions to their lives [14]. Mood changes
such as anxiety, depression, confusion, emotional lability, irritability, loss of
concentration, lethargy, and aggression/hostility, have been associated with the
menstrual cycle [10]. Physical symptoms reported mainly during the
premenstrual/paramenstrual phases [14], include skin disorders, oedema, pelvic pain,
breast tenderness, headaches, muscle pain, weight increase and vomiting [15]. Corney
and Stanton [6] reported that 63% of women experienced symptoms and mood
changes up to 3 days after the onset of menstruation while 5% reported debilitating
effects continuing until the end of menstruation. All women reported their
symptoms/mood changes as lasting from 2 to 8 days premenstrually. Treatments
include administration of antidepressants and tranquillisers, hormonal treatments such
as oral contraceptives, counselling and psychotherapy [14].
*The terms premenstrual syndrome (PMS) and premenstrual tension (PMT) have been used
interchangeably in the literature. There is no single definition of what is meant by premenstrual [8].
This paper uses the classification of Dalton [13] where PMS refers to the whole collection of physical
symptoms and psychological mood states while PMT refers to mood states only.

With regard to physical symptoms of dysmenorrhoea (painful periods), physical

exercise has been advocated as a therapeutic treatment [16-18]. However, surprisingly
little research has evaluated the effects of aerobic exercise on menstrual cycle
symptoms and mood states [19]. Metheny and Smith [18] measured positive and
negative affect and found that women who exercised regularly reported more positive
affect than non-exercisers. Gannon et al. [1] found that the length of time women had
been exercising correlated significantly with lower levels of menstrual symptoms.
Keye [20] reported lower levels of anxiety in women who exercised regularly
compared with non-exercisers, while Schwartz, et al. [21] found that women runners
reported a decrease in premenstrual symptoms.
Evidence suggests that aerobic exercise reduces negative affect [22-241 and except
for the Metheny and Smith [18] study, the evidence [l, 20, 21] suggests that women
who exercise regularly exhibit lower levels of negative affect and physical symptoms
across the menstrual cycle. This study tests the hypothesis that women who
participate in regular, aerobic exercise will report less negative affect and lower levels
of physical symptoms, throughout the menstrual cycle, than non-exercisers.
METHOD
Subjects and procedure
Two health and fitness clubs were contacted and 124 regular exercisers volunteered to
participate in the study (volunteers were requested at the end of aerobic exercise
sessions). A total of 97 out of 124 women who indicated that they were regular
exercisers (at least 5 hr per week) at the health and fitness clubs completed the
questionnaires. Exercising women ranged in age from 15 to 48 yr (M = 26.35 yr; SD
= 6.44 yr). Of the 27 women who did not return questionnaires, five became pregnant,
seven stopped exercising and 15 either could not be contacted or misplaced the
questionnaire. In addition, 159 healthy, but generally non-exercising women aged
between 16 and51 yr (M = 21.23 yr; SD = 6.88 yr) were recruited from various
sources, including undergraduate students. Questionnaires were handed out to
volunteers during a short briefing session. Participants were instructed to fill out the
DES-IV and MDQ scales on three separate occasions (menstrually, premenstrually
and intermenstrually). Aside from age, demographic information including
contraceptive pill use, marital status, number of children and amount of regular
weekly exercise was also collected.
Measures
The Menstrual Distress Questionnaire or MDQ [10, 11] is a forty-seven-item selfreport instrument, scored on a five-point Likert-type scale. The eight MDQ subscales
assess menstrual cycle symptoms such as cramps, headache and backache, mood
states such as depression and irritability, and behavior changes such as difficulty in
concentrating and decreased efficiency. Evidence of reliability and validity has been
provided by Boyle [25, 26]. The Differential Emotions Scale or DES-IV [27], is a
thirty-six-item self-report measure of twelve fundamental emotions. Boyle [4] has
provided evidence supporting the reliability and validity of the DES-IV. To enable a
clear and parsimonious examination of the effects of exercise on physical symptoms
and mood states across the menstrual cycle, individual MDQ and DES-IV subscales
were grouped according to higher-order factors identified by Boyle [27].

In investigating the factor structure of the MDQ, Boyle had reported two second-order
factors: a Psychological Factor-loading on the impaired concentration, behaviour
change, negative affect, and control subscales; and a Physical Symptoms Factorloading on the pain, water retention, autonomic reactions and arousal subscales.
In an investigation of the factor structure of the DES-IV instrument, Boyle [27]
reported three second order factors: Extraversion-loading on interest, joy, and surprise
subscales; Hostility-loading on sadness, anger, disgust, and contempt; and
Neuroticism-loading on hostility, fear, shame, shyness, and guilt subscales. These
higher-order factors are used in the present study to facilitate interpretability of
findings, by providing a more parsimonious account of the links between exercise,
menstrual-cycle phase and mood-state changes. As the higher-order factors load on an
increased number of items, they are necessarily more reliable than the primary MDQ
and DES-IV subscales.
RESULTS
Median test-retest reliability coefficients across the respective menstrual cycle phases
for the higher-order factors were 0.57 (premenstrual/menstrual), 0.53 (premenstrual/
intermenstrual), and 0.48 (menstrual/intermenstrual), and for the primary subscalesmedian coefficients were 0.55, 0.43, and 0.42 respectively (see Table I). Median
internal reliability coefficients for the higher-order factors were 0.83 (premenstrual),
0.83 (menstrual), and 0.78 (intermenstrual), respectively. As both instruments
measure state variables, the test-retest reliabilities are moderate only, as would be
expected if the MDQ and DES-IV subscales are truly sensitive to variability across
cycle phases. However, the physical symptom variables were less stable across the
menstrual cycle than were psychological and mood-state variables (median
coefficients being 0.8 1 (Psychological), 0.53 (Physical), 0.84 (Hostility), 0.78
(Extraversion), and 0.89 (Neuroticism).

Given the mixed design (2 between groups x 3 occasions), preliminary ANOVAs

checked differences between groups on several independent variables including age,
contraceptive pill use, and menstrual distress between the exercise and non-exercise
groups, as these variables have been associated with menstrual cycle symptoms [I,
281. The groups differed significantly only on age [two-tailed t(254) = - 5.99, p <
O.OOl]. Consequently, effects of age-the exercise group was older (M = 26.35 yr)
than the non-exercise group (M = 21.23 yr)-was controlled statistically. Separate
Exercise (exercise, non-exercise) x Menstrual Cycle Phase (menstrual, premenstrual,
intermenstrual) MANCOVAs were conducted on each of the higher-order factors with
age as the covariate. Age was a significant covariate for two of the higher-order
DESIV factors: Hostility, F(4,250) = 6.03, p < 0.001; and Extraversion, F(3,251) =
7.67, p < 0.001. The multiple dependent measures in each analysis were scores on the
subscales associated with the particular higher-order factor. As five separate
MANCOVAs were performed on the same data, a Bonferroni correction was applied,
and the alpha level was reduced from 0.05 to a more conservative 0.01, since
heterogeneity of variance was indicated by Boxs M tests. The mean MDQ and DESIV higher order factor scores are shown in Figs l-5, while those for the primary
subscales are presented in Table II. In every instance, irrespective of menstrual cycle
phase, the exercise group obtained lower mean scores than did the non-exercise
group, suggesting that membership of a health and fitness club is associated with
reduced menstrual cycle symptomatology. Mean scores accord with expectations that
negative moods and symptoms would be elevated premenstrually and menstrually.
Conversely, mean scores for the positive states are elevated intermenstrually.

A MANCOVA examined the MDQ Psychological factor with four subscales

concentration, behavior change, negative affect and control. The multivariate analysis,
using Pillais criterion [29], showed a highly significant between groups effect for
exercise, F(4, 250) = 6.58, p < 0.001; and repeated-measures effect for menstrual
cycle phase, F(8, 1012) = 21.69, p < 0.001. Across exercise conditions, three of the
MDQ subscales (impaired concentration, behaviour change, and negative affect) were

highly significant at the p < 0.001 level. Each of these three subscales was also highly
significant at the p < 0.001 level across menstrual cycle phases. The MDQ control
subscale, which includes items that measure infrequently reported symptoms
indicative of a respondents tendency to complain [l0], was not significant at the
adjusted significance level, suggesting that this measure was not affected by exercise.
A second MANCOVA was performed on the Physical factor of the MDQ, with the
pain, autonomic reactions, water retention and arousal subscales as dependent
measures. Highly significant effects for exercise, F(4, 250) = 5.2, p < 0.001; and for
menstrual cycle phase, F(8, 1012) = 52.23, p < 0.001, were obtained. Among the
psychological subscales, only the MDQ pain subscale was significant across exercise
conditions, F(1, 253) = 17.67, p -C 0.01. Across menstrual cycle phases, each of the
four subscales was highly significant at the p < 0.001 level.
The third MANCOVA examined the Neuroticism factor of the DES-IV with hostility,
fear, shame, shyness and guilt subscales as dependent measures. Highly significant
effects for exercise, F(5, 248) = 6.97, p < 0.001; and for menstrual cycle phase, F( 10,
1006) = 3.19, p < 0.01, were obtained. For the exercise condition, all subscales were
significant at the p < 0.01 level. Across menstrual cycle phases, hostility, fear, and
shyness were significant at the p < 0.01 level.

The fourth MANCOVA was performed on the DES-IV Extraversion factor with
interest, joy and surprise subscales as dependent measures. The age covariate was
significant, F(3, 251) = 7.67, p < 0.01. After controlling for the effect of age, the
multivariate analysis showed no significant effect for exercise, but a significant effect
for menstrual cycle phase, F(6, 1014) = 8.09, p < 0.01. Univariate tests on menstrual
cycle phase showed that interest and joy were significant at thep < 0.01 level. Results
indicate that positive mood states (interest, joy, surprise) covary with age and appear
unaffected by exercise, although interest and joy show significant changes over the
menstrual cycle.

The fifth MANCOVA examined the DES-IV Hostility factor with the anger, disgust,
contempt and sadness subscales as dependent measures. The age covariate was
significant, F(4, 250) = 6.03, p < 0.01, After controlling for age, the multivariate
analyses showed a highly significant effect for exercise, F(4, 250) = 6.16, p < 0.01;
and menstrual cycle phase, F(8, 1012) = 8.85, p < 0.01. Across exercise conditions, all
subscales were significant at the p < 0.01 level. The menstrual cycle phase effect
showed that anger, disgust, and sadness were significant at the p < 0.01 level.

Hence, the multivariate analyses revealed significant differences between groups and
across menstrual cycle phases. All univariate analyses showed significant differences
between the exercise conditions. Only the MDQ contempt subscale did not change
significantly over menstrual cycle phase. Mean scores on the three higher order
factors that measure negative affect (Hostility and Neuroticism; Psychological factor)
and the Physical Symptoms factor were significantly lower in the exercise group. The
positive affect (Extraversion) did not differentiate between the groups. All higherorder factors revealed significant menstrual cycle phase effects, however.

DISCUSSION
Results suggest that women who undertake regular, moderate, aerobic exercise show
significantly lower levels of negative mood states, (anger, contempt, disgust, sadness,
hostility, fear, shame, shyness, and guilt), as measured by the DES-IV, than nonexercisers. No significant differences between groups were found for positive affect.
The exercise group showed significantly lower levels of negative affect, impaired
concentration, behaviour change, and pain, as measured by the MDQ, than the nonexercise group. Autonomic reactions, water retention, and arousal did not differ
significantly between groups. Age was included as a covariate as a decrease in
menstrual and premenstrual symptoms has been reported with increasing age [28].
Gannon et al. [l] also reported a reduction in pain, negative affect, impaired
concentration, and behaviour change with increasing age. However, no significant
effect for age was found for these subscales in the present study.
The results support the literature suggesting that moderate regular exercise attenuates
negative mood states [30-36]. This agrees with previous findings that moderate,
aerobic exercise may decrease negative mood states [24, 37, 38]. The present results
suggest links between exercise and improvements in negative mood states across the
menstrual cycle in accord with Gannon et al. [l], and Keye [20]. In the present study,
many more significant differences in MDQ subscale scores were found across
exercise groups and menstrual cycle phases than previously has been reported [1, 18].
Furthermore, the present finding that regularly exercising women reported
significantly less pain than non-exercisers contradicts the findings of Metheny and
Smith [18] who found higher levels of dysmenorrhoea associated with regular
exercise (cf. ref. [16]). There are several methodological problems associated with
their study, as it was correlational, and retrospective. Numerous studies have found
that retrospective reporting of menstrual cycle symptoms results in an exaggeration of
responses compared with prospective reporting [9, 39, 40].
Although it appears that there is some amelioration of negative mood states with
moderate, regular exercise, there is no clear physiological explanation for these
findings. Maximum fluctuation in hormones occurs during the
premenstrual/menstrual phases [13] and premenstrual symptoms are related to
hormone levels [5, 42], which, in turn, may be dependent on endorphin functioning
[14]. Accordingly, Fremont and Craighead [32] suggested that exercise may produce
increased levels of endorphins which influence mood. Raised endorphin levels have
been associated with significant reductions in depression [43]. However, Farrell et al.
[44] found that endorphins activated by exercise had no effect on mood alterations.
Psychological explanations for mood improvement after exercise have been the
subject of several reviews [45-47]. Cognitive-behavioral theory proposes that
cognitive distortions and repeated intrusive thoughts are major components of
depression [48]. Exercise may act as a distraction from intrusive thoughts, and allow
positive thoughts to surface, decreasing depression in the short-term [49]. Another
possible explanation is that exercise improves body image and self-efficacy which
impact on self-concept and self-esteem [45, 47]. The increased social contact offered
by exercise. groups may be particularly relevant. Kirkcaldy and Shephard [47]
suggested that women desire and derive greater benefit from social interaction than
men.

One caveat however, is that it cannot be assumed that women who exercise regularly
at health and fitness clubs come from the same population as non-exercising women.
It must be acknowledged that there are likely to be numerous differences between
these groups in terms of health histories, attitudes and interests that may account for
differences in menstrual cycle symptomatology. However, the nonexercising women
were drawn mainly from healthy undergraduate students. Consequently, the finding of
significant differences in menstrual cycle moods and symptoms between groups is all
the more impressive, and suggests that there is a beneficial impact of physical
exercise on menstrual cycle symptomatology and moods.
REFERENCES
1. L Gannon, T Luchetta, L Pardie and K Rhodes , Perimenstrual symptoms:
Relationships with chronic stress and selected lifestyle variables. J Behav Med 12
(1989), pp. 149159.
2. C Wood, L Larsen and R Williams , Menstrual characteristics of 2,343 women
attending the Shepherd Foundation. Aust NZ J Obstet Gynaecol 19 (1979), pp. 107
110.
3. GK Bains and PS Slade , Attributional patterns, moods, and the menstrual cycle.
Psychosom Med 50 (1988), pp. 469476.
4. GJ Boyle , The paramenstruum and negative moods in normal young women.
Person Indiv Diff 6 (1985), pp. 649652.
5. A Collins, P Eneroth and B Landgren , Psychoneuroendocrine stress responses and
mood as related to the menstrual cycle. Psychosom Med 47 (1985), pp. 513527.
6. RH Corney and R Stanton , A survey of 658 women who reported symptoms of
premenstrual syndrome. J Psychosom Res 35 (1991), pp. 471482.
7. S Goldstein, U Halbreich, J Endicott and E Hill , Premenstrual hostility, impulsivity
and impaired social functioning. J Psychosom Obstet Gynaecol 5 (1986), pp. 3338.
8. CE Ainscough , Premenstrual emotional changes: A prospective study of
symptomatology in normal women. J Psychosom Res 34 (1990), pp. 3545.
9. MB Parlee , Changes in moods and activation levels during the menstrual cycle in
experimentally naive subjects. Psycholoy Women Quart 7 (1982), pp. 119131.
10. RH Moos , Menstrual Distress Questionnaire Manual. , Department of
Psychiatry,Stanford University, Palo Alto (1977).
11. RH Moos , Perimenstrual Symptoms: A Manual and Overview of Research with
the Menstrual Distress Questionnaire. , Department of Psychiatry, Stanford
University, Palo Alto (1985).

12. D Llewellyn-Jones , Everywoman: A Gynaecological Guide for Life. (Third Edn.

ed.),, Faber & Faber, London (1982).
13. K Dalton , The Premenstrual Syndrome and Progesterone Therapy. (Second Edn.
ed.),, Heinemann, London (1984).
14. PM O'Brien , Premenstrual Syndrome. , Blackwell Scientific Publications, Oxford
(1987).
15. AW Chen and E Filsinger , Mood across the menstrual cycle and number of
menstrual symptoms reported: A cross-cultural study. Can J Psychiatry 32 (1987), pp.
429432.
16. E Fernandez and WL Brown , Temporal changes in pain and emotional state
during the course of dysmenorrhea among high school students. In: Ninth National
Behavioural Medicine Conference: Adolescence and Health (1991).
17. E Fernandez and DC Turk , Sensory and affective components of pain: Separation
and synthesis. Psychol Bull 112 (1992), pp. 205217.
18. WP Metheny and RP Smith , The relationship among exercise, stress, and primary
dysmenorrhea. J Behav Med 12 (1989), pp. 569586.
19. L Gannon , The potential role of exercise in the alleviation of menstrual disorders
and menopausal symptoms: A theoretical synthesis of recent research. Women and
Health 14 (1988), pp. 105217.
20. WR Keye , Medical treatment of premenstrual syndrome. Can J Psychiat 30
(1985), pp. 483487.
21. B Schwartz, DC Cumming, E Riordan, M Selye, SS Yen and RW Rebar ,
Exercise-associated amenorrhoea: A distinct entity?. Am J Obstet Gynecol 141
(1981), pp. 662670.
22. S Biddle and N Mutrie , Psychology of Physical Activity and Exercise: A Healthrelated Perspective. , Springer, New York (1991).
23. DG McDonald and JA Hodgdon , The Psychological Effects of Aerobic Fitness
Training: Research and Theory. , Springer, New York (1991).
24. SJ Petruzzello, DM Landers, BD Hatfield, KA Kubitz and W Salazar , A metaanalysis on the anxiety-reducing effects of acute and chronic exercise. Sports Med 11
(1991), pp. 143182.
25. GJ Boyle , Interset relationships between the Eight State Questionnaire and the
Menstrual Distress Questionnaire. Person Indiv Diff 12 (1991), pp. 703711.
26. GJ Boyle , Factor structure of the Menstrual Distress Questionnaire (MDQ):
Exploratory and LISREL analyses. Person Indiv Diff 12 (1991), pp. 565574.

27. GJ Boyle , Factor structure of the Differential Emotions Scale and the Eight State
Questionnaire revisited. Irish J Psychology 10 (1989), pp. 5666.
28. N Woods, A Most and G Dery , Prevalence of perimenstrual symptoms. Am J
Public Health 72 (1982), pp. 12571264.
29. BG Tabachnick and LS Fidell , Using Multivariate Statistics. (Second Edn. ed.),,
Harper Collins, New York (1989).
30. JA Blumenthal, S Williams, TL Needels and AG Wallace , Psychological changes
accompany aerobic exercise in healthy middle-aged adults. Psychosom Med 44
(1982), pp. 529536.
31. SH Boutcher and DM Landers , The effects of vigorous exercise on anxiety, heart
rate, and alpha activity of runners and nonrunners. Psychophysiology 25 (1988), pp.
696702.
32. J Fremont and LW Craighead , Aerobic exercise and cognitive therapy in the
treatment of dysphoric moods. Cognitive Ther Res 11 (1987), pp. 241251.
33. J Otto , The effects of physical exercise on psychophysiological reactions under
stress. Cognition and Emotion 4 (1990), pp. 341357.
34. DL Roth , Acute emotional and psychophysiological effects of aerobic exercise.
Psychophysiology 26 (1989), pp. 593602.
35. A Steptoe, S Edwards, J Moses and A Mathews , The effects of exercise training
on mood and perceived coping ability in anxious adults from the general population.
Psychosom Res 33 (1989), pp. 537547.
36. R Weinberg, A Jackson and K Kolodny , The relationship of massage and
exercise to mood enhancement. The Sport Psychologist 2 (1988), pp. 202211.
37. J Moses, A Steptoe, A Mathews and S Edwards , The effects of exercise training
on mental well-being in the normal population: A controlled trial. J Psychosom Res
33 (1989), pp. 4761.
38. R Norris, D Carroll and R Cochrane , The effects of aerobic and anaerobic
training on fitness, blood pressure, and psychological stress and well-being. J
Psychosom Res 34 (1990), pp. 367375.
39. GJ Boyle and AF Grant , Prospective versus retrospective assessment of menstrual
cycle symptoms and moods: Role of attitudes and beliefs. J Psychopathology Behav
Assess 14 (1992), pp. 307321.
40. J Endicott and V Halbreich , Retrospective report of premenstrual depressive
changes: Factors affecting confirmation by daily ratings. Psychopharmacol Bull 48
(1982), pp. 109112.

41. CM Logue and RH Moos , Positive premenstrual changes: Toward a new

perspective on the menstrual cycle. J Psychosom Res 32 (1988), pp. 3140.
42. U Halbreich, J Endicott, S Goldstein and J Nee , Premenstrual changes and
gonadal hormones. Acta Psychiat Scand 74 (1986), pp. 576586.
43. RH Gerner, GH Catlin, DA Gorelick, KK Hui and CH Li , Beta-endorphin:
Intravenous transfusion causes behavioral change in psychiatric patients. Arch Gen
Psychiat 37 (1980), pp. 635640.
44. PA Farrell, AB Gustafson, WP Morgan and CB Pert , Enkephalins,
catecholamines, and psychological mood alterations: Effects of prolonged exercise.
Med Sci Sports Exercise 19 (1987), pp. 347353.
45. de DM Coverley-Veale , Exercise and mental health. Acta Psychiat Scand 76
(1987), pp. 113120.
46. J Gleser and H Mendelberg , Exercise and sport in mental health: A review of the
literature. Israel J Psychiatry Related Sci 27 (1990), pp. 99112.
47. BD Kirkcaldy and RJ Shephard , Therapeutic implications of exercise. Int J Sport
Psychol 21 (1990), pp. 165184.
48. AT Beck , Cognitive models of depression. J Cognit Psychother 1 (1987), pp. 5
38.
49. JW Pennebaker and J Lightner , Competition of internal and external information
in an exercise setting. J Person Soc Psychol 39 (1980), pp. 165174.