- Email: [email protected]
Dietary calcium and manganese effects on menstrual cycle symptoms
Dietary calcium and manganese effects on menstrual cycle symptoms James G. Penland, PhD, and Phyllis E. Johnson, PhD Grand Forks, North Dakota OBJECTIVE: This exploratory study was designed to determine whether dietary calcium and manganese affect menstrual symptoms in healthy women. STUDY DESIGN: Ten women with normal menstrual cycles completed the Menstrual Distress Questionnaire each cycle during a 169-day, live-in metabolic study of calcium and manganese nutrition. Women were assigned in a double-blind, Latin-square manner to each of four 39-day dietary periods: 587 or 1336 mg calcium per day with 1.0 or 5.6 mg maoganese per day. Responses were analyzed by repeated-measures analysis of variance. RESULTS: Increasing calcium intake reduced mood, concentration, and behavior symptoms generally (p :s 0.05), reduced pain during the menstrual phase of the cycle (p = 0.034), and reduced water retention during the premenstrual phase (p = 0.041). In spite of increasing calcium intake, lower dietary manganese increased mood and pain symptoms during the premenstrual phase (p :s 0.05). CONCLUSION: Dietary calcium and manganese may have a functional role in the manifestation of symptomatology typically associated with menstrual distress. (AM J OBSTET GYNECOL 1993;168:1417-23.)
Key words: Menstrual symptoms, calcium, manganese, nutrition
The normal menstrual cycle is characterized by numerous neuroendocrine and physiologic changes. I In addition to hormonal variations, body weight and basal metabolic rate increase during the luteal phase, and reliable although complex differences among cycle phases have been found in brain electrical activity and sensory function. 2 Less consistent findings have emerged from studies attempting to relate cycle phase to disease occurrence and psychopathologic conditions, and numerous behavioral studies have yielded conflicting data regarding phase-related differences in cognitive performance and mood states among the normal population. 3 However, clinical observation and controlled studies continue to find that psychologic and behavioral disturbances are among the most prevalent symptoms reported by women with menstrual disorders. Research during the past 20 years suggests that nutritional intake, metabolism, and status may play a From the Grand Forks Human Nutrition Research Center, Agricultural Research Service, United States Department of Agriculture. Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by the United States Department ofAgriculture and does not imply its approval to the exclusion of other products that may also be suitable. Received for publication August 16, 1992; revised October 13, 1992; accepted December 16, 1992. A Letter to the Editors by Rossignol and Bonnlander pertaining to this article appears on page 1640. Reprint requests: James G. Penland, PhD, Research Psychologist, USDA, ARS Grand Forks Human Nutrition Research Center, P.O. Box 9034, University Station, Grand Forks, ND 58202. 6/1/45012
role in regulating the normal menstrual cycle and in the cause and treatment of menstrual disorders. Variations in total energy intake and the consumption of protein, fat, and carbohydrates have been correlated with cycle phase:· 7 and recent data indicate that tryptophan metabolism differs between follicular and luteal phases. s Deficiencies in a-tocopherol (vitamin E), magnesium, and zinc have all been implicated in the cause of premenstrual tension,9 whereas pyridoxine (vitamin B6 ) and fatty acids such as T-linolenic acid (primrose oil) have been prescribed in the treatment of menstrual disorders. 10 Unfortunately, most studies designed to test the efficacy of various form of nutritional intervention have suffered from multiple methodologic problems. I 1 The research reported here examines the relationship between dietary intake of calcium and manganese among healthy women with normal cycles and numerous physiologic, psychologic, and behavioral symptoms often reported to occur in relation to menstruation and the menstrual cycle. As part of a larger study,12 this research was exploratory in nature with no specific or formal a priori hypotheses. However, as described in the Comment section, biochemical concentrations of calcium and manganese have been correlated with phase of the menstrual cycle; therefore both nutrients were regarded as strong candidates to affect menstrual symptomatology. Subjects and methods
Fourteen women were recruited through regional and national advertising to participate in a study of 1417
1418 Penland and Johnson
May 1993 Am J Obstet Gynecol
Table I. Subject characteristics* Characteristic Age (yr) Height (cm) Weight (kg) Body fat (%)t Length of menstrual cycle (days) Duration of menstrual phase (days)
Mean ± SE
27.2 165.2 63.0 31.9 27.0 5.0
± ± ± ± ±
1.6 2.5 3.8 1.9 1.3 ± 0.44
*All subjects were white.
t Assessed by underwater weighing.
calcium and manganese nutrition. All subjects were in good physical and mental health, as determined by extensive medical and psychologic screening. Medical screening consisted of a medical history and review, clinical examination including electrocardiogram, chest radiography, and a standard metabolic panel performed on fasting blood and urine samples. Psychologic screening consisted of a psychosocial history, clinical interview, the Eating Disorders Inventory, Eysenck Personality Inventory, Minnesota Multiphasic Personality Inventory, Psychiatric Diagnostic Interview, and cursory tests for normal visual and auditory function. In addition, dietary patterns and preferences were determined by dietary history and interview, and body composition was determined by anthropometry. All subjects reported normal menstrual patterns with no history or current symptoms of premenstrual tension or dysmenorrhea. Subjects did not use oral contraceptives for 30 days before or during their participation in this study. Data from four subjects were excluded from analysis: one woman withdrew during the first month of the study; one woman, who was cycling every 20 days, was excused from completing the symptomatology questionnaire; and two women who had cycles of abnormal length (>60 days). Relevant characteristics of the 10 participants whose data are reported here are shown in Table I. Mter receiving written and verbal explanation, subjects gave written consent to participate. This study was approved by the Institutional Review Board of the University of North Dakota and the Human Studies Review Committee of the United States Department of Agriculture and conformed to guidelines of the Helsinki Declaration. Diet. Throughout the study subjects were fed a diet consisting of conventional foods on a 3-day menu rotation as described by Johnson and Lykken. 12 The basal diet was designed to reflect intakes of nutrients that did not necessarily meet the Recommended Dietary Allowances" but that are typical of intakes by adult women in the United States and that, in our experience, were sufficient to prevent deficiency of those nutrients. To achieve adequate iron intake and to prevent iron de-
pletion from phlebotomy, a supplement containing 8.5 mg of iron as ferrous gluconate was added to the beverage at each of the three daily meals. Thus, when this was combined with the diet, which contained 6.54 mg of iron per 2000 kcaVday the total daily intake of iron was approximately 32 mg. Demineralized water was consumed ad libitum. During an initial 13-day equilibration period subjects were fed a diet containing 800 mg of calcium/day and 2.97 mg of manganese/day, by analysis. 12 Subjects were then assigned in a doubleblind, Latin-square manner to each of four 39-day dietary periods created by the factorial combination of 587 and 1336 mg of calciUm/day and 1.0 and 5.6 mg of manganese/day, constant across subjects whose range of energy intakes was 1600 to 2800 kcaVday. Calcium and manganese intakes were varied by supplementing the basal diet with calcium lactate and manganese sulfate, respectively. Energy in the diet was distributed as 16% protein, 48% carbohydrate, and 36% fat. Initial energy expenditures were determined separately for each subject and adjusted during the course of the study to maintain body weight (measured daily) within ± 2% of initial weight. Individually prescribed exercise was performed three times weekly to maintain body composition and physical work capacity. Procedure. Subjects resided 24 hours/day in the metabolic research ward at the Grand Forks Human Nutrition Research Center to permit strict control of dietary intakes, exercise, and data collection and to provide a common environment. Subjects were chaperoned on all outings to ensure compliance with study protocol. All urine, feces, and menses were collected for biochemical analyses. In addition, blood samples were obtained by venipuncture during the middle and at the end of each dietary period for analysis of hematologic variables, minerals, lipids, enzymes, and hormones. Results of analyses on biologic samples were reported by Johnson and Lykken. 12 Collection of all menses allowed precise determination of the length of each cycle and duration of menstrual flow. The Menstrual Distress Questionnaire was administered individually to each subject at the completion of the menstrual phase of each cycle. The Menstrual Distress Questionnaire is a retrospective self-report instrument that assesses the presence and severity of 47 symptoms occurring during the menstrual, premenstrual (late luteal), and intermenstrual (follicular and early luteal) phases of the most recent menstrual cycle. 14 Operationally, the premenstrual phase was defined as the 7 days before the onset of menstruation. The Menstrual Distress Questionnaire was developed to help characterize changes women experience during the different phases of the menstrual cycle and to facilitate diagnosis of menstrual disorders such as premenstrual tension and dysmenorrhea. Symptoms are
Penland and Johnson
Volume 168, Number 5 Am J Obstet Gynecol
grouped into eight scales on the basis of factor analytic studies with several independent groups, The Autonomic Reactions scale includes dizziness, faintness, cold sweats, nausea, vomiting, and hot flashes. Behavior Change includes poorer work or school performance, increased napping and time in bed, increased time at home, avoiding social activities, and decreased overall efficiency. Concentration includes insomnia, forgetfulness, confusion, poorer judgment, difficulty concentrating, distractibility, increased accidents, and poorer motor coordination. Negative Affect includes crying, loneliness, anxiety, restlessness, irritability, mood swings, depression, and tension. Pain includes muscle stiffness, headache, cramps, backache, fatigue, and general aches and pains. Water Retention includes weight gain, skin disorders, painful breasts, and swelling. The Arousal scale reflects the positive feelings that some women report during certain phases of their cycle and includes increased affection, orderliness, excitement, feelings of well-being, and bursts of energy or activity. Last, the Control scale reflects a general tendency to complain and consists of symptoms commonly found in postmenopausal rather than premenopausal women; these include feelings of suffocation, chest pains, ears ringing, heart pounding, numbness, tingling, blind spots, and fuzzy vision. Because the 39-day dietary periods sometimes spanned more than one menstrual period, some subjects completed the Menstrual Distress Questionnaire twice during a single dietary period. The Menstrual Distress Questionnaire was amended slightly to include a 100 mm visual analog scale used to assess perceived volume or rate of flow during each day of menstruation. Data analyses. Dietary effects on Menstrual Distress Questionnaire scale scores, length of cycle, duration of menstrual flow, and perceived rate of flow were analyzed by repeated-measures analysis of variance by means of the General Linear Models procedure of the Statistical Analysis System package. 15 Mean scores for the symptom scales were at the low end of the possible range (Table II) because subjects in this study were not experiencing the degree of distress found in women with premenstrual syndrome or dysmenorrhea. However, parametric analysis was deemed appropriate because the variability around the means approximated normal. Post hoc comparisons of cell means for significant interaction effects were made with the Bonferroni t test with the experiment-wise error rate adjusted to p :S 0.05. Results
In contrast to when they were fed the high calcium diets, women fed low calcium diets reported significantly increased negative affect and more undesirable behavioral changes during all three phases of the men-
1419
strual cycle (Tables II through IV). Low calcium intake resulted in reports of poorer concentration during the premenstrual phase (Table III) and marginally during the menstrual !p < 0.052) and intermenstrual !p < 0.063) phases. Low calcium intake also resulted in reports of greater pain during the menstrual phase (Table II) and increased water retention during the premenstrual phase (Table III). A calcium x manganese interaction was observed for negative affect and, marginally !p < 0.068), for pain during the premenstrual phase. In spite of high calcium intake, low manganese intake resulted in increased symptomatology similar to that reported with the low calcium intake (Table III). Responses to Menstrual Distress Questionnaire control items and to arousal and autonomic reactions scale items were not significantly related to dietary calcium or manganese during any phase of the cycle. Neither calcium nor manganese intake influenced the length of the menstrual cycle or the duration of the menstrual phase. In spite of significantly increased menstrual hemoglobin losses during low dietary intakes of manganese,IZ the visual analog scale on the Menstrual Distress Questionnaire showed no effect of dietary calcium or manganese intakes on perceived menstrual volume. Comment
In general, these findings are consistent with other studies showing a relationship between macro and micro nutrients and symptoms among women with premenstrual tension. For example, Wurtman et al. 16 found that increased carbohydrate intake during the premenstrual phase improved depression, tension, anger, confusion, sadness, fatigue, alertness, and calmness scores. There have been several reports of reduced depression, irritability, bloating, and overall symptom scores with vitamin B6 administration, either alone or in combination with other vitamins and minerals. 17 However, numerous other studies have shown mixed or no effects of B6. II Vitamin E was found to reduce breast tenderness and total symptom scores,18 although a recent study failed to document a vitamin E deficiency in women with premenstrual distress. 19 Recently, Facchinetti et al."o reported that dietary magnesium supplementation relieved negative affective symptoms. The current findings are unique in showing a nutritional effect on the relatively mild physiologic and psychologic symptoms experienced by women with normal menstrual cycles. In addition, the data support a possible role for the minerals calcium and manganese in the cause and management of menstrual disorders. Specifically, the data indicate the existence of a relationship between dietary intake of calcium and the occurrence of several classes of symptoms commonly associated with menstruation and the menstrual cycle.
1420 Penland and Johnson
May 1993 Am J Obstct Cynccol
Table II. Menstrual Distress Questionnaire scale scores in relation to daily dietary calcium and manganese intakes: menstrual phase Intake
Menstrual Distress Questionnaire scale
Calcium (mg/day)
Manganese (mg/day)
587 587 1336 1336
1.0 5.6 1.0 5.6
Calcium intake Manganese intake Calcium x Manganese intake Mean ± SE
No.
Arousal (5-30)*
Autonomic Reactions (4-24)
Behavior Change (5-30)
9 8.32 ± 0.70 6.56 ± 1.73 10.97 ± 0.84 7 7.95 ± 0.82 5.49 ± 2.04 9.89 ± 0.98 9 7.72 ± 0.70 4.77 ± 1.73 7.96 ± 0.84 7.31 ± 0.75 8 6.85 ± 0.90 8.72 ± 1.87 Summary of analysis of variance effects: probabilities < 0.05 0.003 NS NS NS NS NS NS NS NS 4.22 26.16 6.08
Concentration (8-48)
14.26 14.56 12.99 12.52
0.74 0.87 0.74 0.80
± ± ± ±
NS NS NS 4.82
Data are presented as means ± SEM. NS, Not significant. *Range of possible values for each scale of the Menstrual Distress Questionnaire.
Table III. Menstrual Distress Questionnaire scale scores in relation to daily dietary calcium and manganese intakes: premenstrual phase Menstrual Distress Questionnaire scale
Intake Calcium (mg/day)
Manganese (mg/day)
587 587 1336 1336
1.0 5.6 1.0 5.6
Calcium intake Manganese intake Calcium x manganese intake Mean ± SE
No.
Arousal (5-30)*
Autonomic Reactions (4-24)
Behavior Change (5-30)
6.14 ± 1.69 10.00 ± 0.78 7.44 ± 0.43 9 9.35 ± 0.91 7.46 ± 0.51 4.49 ± 1.98 7 7.83 ± 0.78 4.63 ± 1.69 7.75 ± 0.43 9 6.73 ± 0.84 7.28 ± 0.47 8.48 ± 1.82 8 Summary of analysis of variance effects: probabilities < 0.05 0.010 NS NS NS NS NS NS NS NS 3.51 1.63 24.81 5.25
Concentration (8-48)
13.33 13.76 12.57 11.63
± ± ± ±
0.64 0.75 0.64 0.69
0.047 NS NS 1.71
Data are presented as means ± SEM. NS, Not significant. *Range of possible values for each scale of the Menstrual Distress Questionnaire. t, tBonferroni paired comparisons of dietary means for calcium x manganese interactions with pooled within-subject error term. Means in the same column with different superscripts are statistically different (a. adjusted to sO.05).
Table IV. Menstrual Distress Questionnaire scale scores in relation to daily dietary calcium and manganese intakes: intermenstrual phase Intake Calcium (mg/day)
Manganese (mg/day)
587 587 1136 1336
1.0 5.6 1.0 5.6
Calcium intake Manganese intake Calcium x manganese intake Mean ± SE
Menstrual Distress Questionnaire scale No.
9 7 9 8 Summary
Arousal (5-30)*
Autonomic Reactions (4-24)
7.24 ± 0.57 7.47 ± 0.67 7.37 ± 0.57 7.24 ± 0.61 of analysis of variance NS NS NS 2.82
Behavior Change (5-30)
8.79 ± 0.83 5.66 ± 0.96 8.83 ± 0.98 4.77 ± 1.13 7.31 ± 0.83 4.33 ± 0.96 6.09 ± 0.90 6.65 ± 1.04 effects: probabilities < 0.05 0.028 NS NS NS NS NS 6.01 8.02
Concentration (8-48)
12.47 13.55 11.43 10.75
± 0.90
± 1.06
± 0.90 ± 0.98
NS NS NS 7.11
Data are presented as means ± SEM. NS, Not significant. *Range of possible values for each scale of the Menstrual Distress Questionnaire.
The beneficial effects of high dietary calcium on behavior, mood, and concentration were present during all phases of the cycle, suggesting that calcium intake may influence psychologic processes in general (i.e., be in-
dependent of menstruation). Alternatively, these findings may simply confirm the absence of menstrual disorders among the women participating in this study. However, symptoms related to water retention and
Penland and Johnson
Volume WH. !';lImoer " Am J Obstet (;\,necol
Menstrual Distress Questionnaire scale Control (6-36)
7.37 7.30 7.16 7.01
:±: :±: :±: :±:
0.47 0.55 0.47 0.51
Negative Affect
Water Retention
(8-48)
(4-24)
0.78 0.91 0.78 0.84 Summary of analysis of variance effects: probabilities < 0.05 NS 0.018 0.034 NS
NS NS
1.93
19.93:±: 21.01:±: 18.34:±: 11.50:±:
1.99 2.34 1.99 2.15
NS NS
34.61
13.58:±: 13.59:±: 11.41:±: 10.89:±:
1.0 1.17 1.0 1.08
NS NS
8.66
8.37 :±: 9.34:±: 7.52 :±: 7.25:±:
NS NS
5.23
Menstrual Distress Questionnaire scale Control (6-36)
7.18:±: 7.04 :±: 6.99:±: 6.52 :±:
0.44 0.52 0.44 0.48
Negative Affect
Water Retention
(8-48)
(4-24)
0.67 0.79 0.67 0.72 Summary of analysis of variance effects: probabilities < 0.05 NS 0.008 NS 0.041
NS NS
27.67
18.91 :±: 21.46 :±: 17.60:±: 11.46:±:
1.78t 11.I4:±: 0.91 2.09t 12.39:±: 1.07 1.78t.+ 11.48:±: 0.91 1.92t 8.96 :±: 0.98
NS
0.035 7.24
8.98:±: 8.15:±: 7.71:±: 6.26:±:
NS NS
3.91
NS NS
Menstrual Distress Questionnaire scale Control (6-36)
7.02 6.97 6.81 6.56
:±: 0.45 ± 0.53 ± O.4S ± 0.49
Negative Affect (8-48)
16.31 19.37 14.56 8.73
:±: 2.28 ± 2.67 ± 2.28 :±: 2.46
NS NS NS
0.020
1.76
45.11
NS NS
Pain
Water Retention
(6-36)
(4-24)
11.29 10.61 10.10 8.02
:±: 1.21 ± 1.42 ± 1.21 :±: 1.31
NS NS NS
12.78
6.41 :±: 0.58
6.S7 ± 0.68 6.38:±: 0.58 5.37 :±: 0.62
2.88
NS NS NS
However, symptoms related to water retention and pain, which might be thought of as more physiologic in nature, showed beneficial effects of high dietary calcium that were phase-specific; these findings more clearly suggest a role for calcium in regulating menstrual symptoms.
1421
Reports of increased water retention with low calcium intake were found only for the premenstrual phase of the cycle, although a similar tendency was also observed for the menstrual phase (p < 0.094). Unfortunately, water retention was not directly assessed in the current study. Analysis of daily body weight showed only slight variation with phase of cycle. A marked dietary effect was evident, however, with the lowest weights associated with the high calcium-high manganese diet. Because water retention and weight gain are characteristic of the premenstrual phase, reports of increased water retention during this phase when women were fed low calcium diets might simply reflect an exacerbation of this normal process. Reports of greater pain associated with low calcium were found only during the menstrual phase, although a marginal calcium x manganese interaction was also observed for the premenstrual phase. Post hoc analysis of subject requests for pain medication during the course of the study (ratio of number of days pain medication was requested to number of days in each phase of the cycle) showed more requests during the menstrual than premenstrual or intermenstrual phases, except when the low calcium-low manganese diet was used. Requests for pain medication for the premenstrual and intermenstrual phases were approximately equal to those for the menstrual phase when the women were given this diet. Mild pain during menstruation corresponds to the vasomotor activity and endometrial changes that lead to menstrual effluvium and is a normal occurrence.' Reduced calcium in the diet during this phase may aggravate this common symptom by its effects on either of these processes (see below). The clinical observations regarding body weight and requests for pain medication suggest that low manganese intakes increase the effect of low calcium intakes on both water retention and pain and thus affect both premenstrual and menstrual phases of the cycle. Examination of mean scores on the Menstrual Distress Questionnaire Negative Affect scale for each dietary combination (Tables II through N) reveals that dietary intake of manganese also is related to symptoms reflecting mood, but only when calcium intake is high. Therefore low intakes of either manganese or calcium may increase the occurrence of negative mood states. Because this finding is statistically significant only for the premenstrual phase, it can be argued that the relationship between manganese and mood states is also mediated by the menstrual cycle. As with calcium, dietary manganese thus may be a factor in the changes in mood commonly reported before the onset of menstruation. It should be noted that the amount of calcium in the diets designated as low in this study represent typical intakes among women in this age group, whereas the amount of manganese in the low-manganese diets is roughly 50% of typical intakes; the high calcium and manganese intakes used here only slightly exceed those recommended.'"
1422 Penland and Johnson Am
Dietary manganese and iron are known to interact so that iron intakes are inversely related to manganese absorption and use. Therefore it is possible that relatively high daily intakes of approximately 32 mg of iron, used in this study to prevent iron depletion from phlebotomy, may have inhibited manganese absorption or use. Because absorption of a nutrient is typically less with higher compared with lower intakes of that nutrient, this effect would have been most pronounced during manganese supplementation. Further, manganese sulfate, the form of manganese used in this study, is poorly absorbed in comparison with manganese chloride. Thus the amount of available manganese supplementation and the difference between dietary treatments may have been less than originally planned. The apparent absence in this study of an effect of calcium on manganese absorption 12 could also be caused by the relatively poor solubility of the manganese salt. If iron intake and the form of manganese supplement were factors in the current study, then their effect most likely was to reduce statistical power and minimize the effects of manganese and thus make more noteworthy the few significant effects of manganese that were observed on menstrual symptoms. The potential for dietary intakes of calcium and manganese to influence menstrual cycle symptoms can be found in numerous studies showing the disruptive effects of low intakes of these nutrients on reproduction and the estrus cycle in animals and showing a relationship between biochemical concentrations of these nutrients and phase of the estrus or menstrual cycle in both animals and humans. In animals, calcium concentrations in the blood vary over the course of the estrus cycle in turkeys, sheep, and cows, and low dietary manganese results in reproductive problems and absent or irregular estrus cycles in rats, guinea pigs, rabbits, chicks, pigs, goats, sheep, and cows. Further, manganese uptake by the ovaries, fallopian tubes, uterine horn, and vagina of ewes varies during the estrus cycle, which suggests that manganese is needed for normal cyclic function. In humans, Puskulian 21 reported decr~a~ed ~ali~ary calcium at midcycle compared with the menstrual phase in women with normal cycles, apparently independent of calcium intake. Although failing to find a change in total calcium during the menstrual cycle, Pitkin et al!2 reported significant phase-related changes in ionic calcium and the calcium-regulating hormones, serum calcitonin and plasma parathyroid hormone, in normally menstruating women. Most recently, Lehtovirta et a1. 23 reported that serum calcium concentrations were significantly elevated during menstruation in women with normal cycles and those with anovulatory cycles. The latter group also showed elevated calcium concentrations during the premenstrual phase. Further, both groups showed a negative corre-
May 1993 Gynecol
J Obstet
lation between calcium and progesterone concentrations during the premenstrual phase. In a series of studies with healthy menstruating women, Hagenfeldt et a1. 24 found elevated manganese concentrations in the endometrium during the early proliferative phase of the cycle and higher plasma manganese during the proliferative phase than during the secretory phase. It has been proposed that increased food consumption, particularly carbohydrates, during the premenstrual phase of the cycle is an attempt to improve mood states by increasing brain tryptophan and serotonin. 16 Perhaps higher intakes of calcium and manganese in interaction with changes in progesterone during this phase would yield a similar result, with additional positive consequences during this and other phases of the cycle. Interestingly, peak blood glucose during intravenous glucose tolerance tests was highest when subjects were fed the high calcium-low manganese diet'2; when contrasted with the high calcium-high manganese diet, this diet resulted in increased reports of negative affect. If women experiencing normal menstruation and normal menstrual cycles show a relationship between calcium and manganese intakes and the symptoms often used to characterize clinical syndromes such as premenstrual distress and dysmenorrhea, would that relationship be more clearly observed in women suffering menstrual-related disorders? For example, would calcium intakes significantly in excess of the recommended 800 to 1200 mg/day be of any therapeutic value in the treatment of clinical syndromes related to menstruation? A recent study by Thys-Jacobs et a1."5 directly addresses the latter question. In a double-blind, crossover study of 33 women (14 black and 12 Hispanic) diagnosed as having premenstrual syndrome, these investigators found that daily supplements of 1000 mg of calcium, as calcium carbonate, resulted in reduced reports of negative affect and water retention during the premenstrual phase and of pain during both the premenstrual and menstrual phases of the cycle. These findings are remarkably consistent with the phase-related effects of calcium intake found in the current study. Thys-Jacobs et a1. 25 speculated that their findings may reflect the actions of calcium on neurotransmitters, smooth muscle responsiveness, or hormone secretion. Clearly, these same mechanisms also may be responsible for the current findings in women with normal mesntruation and cycles. We are currently in the process of replicating this study on a more extensive cross section of women and using a prospective measure of symptomatology with the hope of being able to further delineate the effects of calcium and manganese on behavior, concentration, and other symptoms relevant to menstruation. Considering the potential for retrospective reports to include variability resulting from inaccurate recall of symptom
Volume 168. Number 5 Am J Obstet Gynecol
occurrence and severity, daily reporting of symptoms is recommended for future studies. At present, the data reported here underscore the importance of menstruating women consuming a diet that fully conforms to current National Research Council guidelines" for both calcium and manganese. We thank Mary Lou Eberhardt for coordinating data collection, LoAnne Mullen and the dietary staff, Betty Vetter and the nursing staff, Bill Beckwith and the psychology staff, LuAnn Johnson and the data processing staff, and, most important, the women who participated as subjects in this study. REFERENCES 1. Kase NG. The endocrinology of the ovarian cycle. In: Kase NG, Weingold AB, Gershenson DM, eds. Principles and practice of clinical gynecology. 2nd ed. New York: Churchill Livingstone, 1990:121-49. 2. Asso D. Physiology and psychology of the normal menstrual cycle. In: Brush MG, Goudsmit EM, eds. Functional disorders of the menstrual cycle. New York: John Wiley, 1988: 15-36. 3. Gannon LR. Covariates of the menstrual cycle in normal populations. In: Menstrual disorders and menopause. New York: Praeger, 1985:36-66. 4. Lissner L, Stevens J, Levitsky DA, Rasmussen KM, Strupp BJ. Variations in energy intake during the menstru.al cycle: implications for food-intake research. Am J Clm Nutr 1988;48:956-62. 5. Tarasuk V, Beaton GH. Menstrual-cycle patterns in energy and macronutrient intake. AmJ Clin Nutr 1991;53:442-7. 6. Bowen DJ, Grunberg NE. Variations in food preference and consumption across the menstrual cycle. Physiol Behav 1990;47:287-91. 7. Gallant MP, Bowering J, Short SH, Turkki PR. Pyridoxine and magnesium status of women with premenstrual syndrome. Nutr Res 1987;7:243-52. 8. Hrboticky N, Leiter LA, Anderson GH. Menstrual cycle effects on the metabolism of tryptophan loads. Am J Clin Nutr 1989;50:46-52. 9. Chuong CJ, Kinch RAH. Premenstrual syndrome. In: Kase NG, Weingold AB, Gershenson DM, eds. Principles and practice of clinical gynecology. 2nd ed. New York: Churchill Livingstone, 1990:669-84. 10. Brush MG. Vitamins, essential fatty acids and minerals in relation to the aetiology and management of premenstrual
Penland and Johnson
1423
syndrome. In: Brush MG, Goudsmit EM, eds. Functional disorders of the menstrual cycle. New York: John Wiley, 1988:69-85. 11. Kleijnen J, Ter Riet G, Knipschild P. Vitamin B6 in the treatment of premenstrual syndrome-a review. Br J Obstet Gynaecol 1990;97:847-52. 12. Johnson PE, Lykken Gl. Manganese and calcium absorption and balance in young women fed diets with varying amounts of manganese and calcium. J Trace Elements Exp Med 1991;4:19-35. 13. National Research Council. Recommended dietary allowances. 10th ed. Washington, DC: National Academy Press, 1989. 14. Moos RH. Typology of menstrual cycle symptoms. AM J OBSTET GYNECOL 1969;103:390-402. 15. SAS Institute. SAS user's guide: statistics, version 5 edition. Cary, North Carolina: SAS Institute, 1985. 16. Wurtman JJ, Brzezinski A, Wurtman ~, Laferrere B. Effect of nutrient intake on premenstrual depression. AM J OBSTET GYNECOL 1989; 161: 1228-34. 17. Abraham G, Rumley R. The role of nutrition in managing the premenstrual tension syndrome. J Reprod Med 1987; 32:405-22. 18. London RS, Murphy L, Kitlowski KE, Reynolds MA. Efficacy of a-tocopherol in the treatment of premenstrual syndrome. J Reprod Med 1987;32:400-4. 19. Chuong CJ, Dawson EB, Smith ER. Vitamin E levels in premenstrual syndrome. AM J OBSTET GYNECOL 1990;163: 1591-5. 20. Facchinetti F, Borella P, Sances G, Fioroni L, Nappi RE, Genazzani AR. Oral magnesium successfully relieves premenstrual mood changes. Obstet Gynecol 1991 ;78: 17781. 21. Puskulian L. Salivary electrolyte changes during the normal menstrual cycle. J Dent Res 1972;5 I (suppl 5):1212-6. 22. Pitkin RM, Reynolds WA, Williams GA, Hargis GK. Calcium-regulating hormones during the menstrual cycle. J Clin Endocrinol Metab 1978;47:626-32. 23. Lehtovirta P, Apter D, Stenman UH. Serum CA 125 levels during the menstrual cycle. Br J Obstet Gynaecol 1990; 97:930-3. 24. Hagenfeldt K, Plantin LO, Diczfalusy E. Trace elements in the human endometrium. 2. Zinc, copper and manganese levels in the endometrium, cervical mucus and plasma. Acta Endocrinol 1973;72: 115-26. 25. Thys-Jacobs S, Ceccarelli S, Bierman A, Weisman H, Cohen M, Alvir J. Calcium supplementation in premenstrual syndrome: a randomized crossover trial. J Gen Intern Med 1989;4: 183-9.
Key words: Menstrual symptoms, calcium, manganese, nutrition
The normal menstrual cycle is characterized by numerous neuroendocrine and physiologic changes. I In addition to hormonal variations, body weight and basal metabolic rate increase during the luteal phase, and reliable although complex differences among cycle phases have been found in brain electrical activity and sensory function. 2 Less consistent findings have emerged from studies attempting to relate cycle phase to disease occurrence and psychopathologic conditions, and numerous behavioral studies have yielded conflicting data regarding phase-related differences in cognitive performance and mood states among the normal population. 3 However, clinical observation and controlled studies continue to find that psychologic and behavioral disturbances are among the most prevalent symptoms reported by women with menstrual disorders. Research during the past 20 years suggests that nutritional intake, metabolism, and status may play a From the Grand Forks Human Nutrition Research Center, Agricultural Research Service, United States Department of Agriculture. Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by the United States Department ofAgriculture and does not imply its approval to the exclusion of other products that may also be suitable. Received for publication August 16, 1992; revised October 13, 1992; accepted December 16, 1992. A Letter to the Editors by Rossignol and Bonnlander pertaining to this article appears on page 1640. Reprint requests: James G. Penland, PhD, Research Psychologist, USDA, ARS Grand Forks Human Nutrition Research Center, P.O. Box 9034, University Station, Grand Forks, ND 58202. 6/1/45012
role in regulating the normal menstrual cycle and in the cause and treatment of menstrual disorders. Variations in total energy intake and the consumption of protein, fat, and carbohydrates have been correlated with cycle phase:· 7 and recent data indicate that tryptophan metabolism differs between follicular and luteal phases. s Deficiencies in a-tocopherol (vitamin E), magnesium, and zinc have all been implicated in the cause of premenstrual tension,9 whereas pyridoxine (vitamin B6 ) and fatty acids such as T-linolenic acid (primrose oil) have been prescribed in the treatment of menstrual disorders. 10 Unfortunately, most studies designed to test the efficacy of various form of nutritional intervention have suffered from multiple methodologic problems. I 1 The research reported here examines the relationship between dietary intake of calcium and manganese among healthy women with normal cycles and numerous physiologic, psychologic, and behavioral symptoms often reported to occur in relation to menstruation and the menstrual cycle. As part of a larger study,12 this research was exploratory in nature with no specific or formal a priori hypotheses. However, as described in the Comment section, biochemical concentrations of calcium and manganese have been correlated with phase of the menstrual cycle; therefore both nutrients were regarded as strong candidates to affect menstrual symptomatology. Subjects and methods
Fourteen women were recruited through regional and national advertising to participate in a study of 1417
1418 Penland and Johnson
May 1993 Am J Obstet Gynecol
Table I. Subject characteristics* Characteristic Age (yr) Height (cm) Weight (kg) Body fat (%)t Length of menstrual cycle (days) Duration of menstrual phase (days)
Mean ± SE
27.2 165.2 63.0 31.9 27.0 5.0
± ± ± ± ±
1.6 2.5 3.8 1.9 1.3 ± 0.44
*All subjects were white.
t Assessed by underwater weighing.
calcium and manganese nutrition. All subjects were in good physical and mental health, as determined by extensive medical and psychologic screening. Medical screening consisted of a medical history and review, clinical examination including electrocardiogram, chest radiography, and a standard metabolic panel performed on fasting blood and urine samples. Psychologic screening consisted of a psychosocial history, clinical interview, the Eating Disorders Inventory, Eysenck Personality Inventory, Minnesota Multiphasic Personality Inventory, Psychiatric Diagnostic Interview, and cursory tests for normal visual and auditory function. In addition, dietary patterns and preferences were determined by dietary history and interview, and body composition was determined by anthropometry. All subjects reported normal menstrual patterns with no history or current symptoms of premenstrual tension or dysmenorrhea. Subjects did not use oral contraceptives for 30 days before or during their participation in this study. Data from four subjects were excluded from analysis: one woman withdrew during the first month of the study; one woman, who was cycling every 20 days, was excused from completing the symptomatology questionnaire; and two women who had cycles of abnormal length (>60 days). Relevant characteristics of the 10 participants whose data are reported here are shown in Table I. Mter receiving written and verbal explanation, subjects gave written consent to participate. This study was approved by the Institutional Review Board of the University of North Dakota and the Human Studies Review Committee of the United States Department of Agriculture and conformed to guidelines of the Helsinki Declaration. Diet. Throughout the study subjects were fed a diet consisting of conventional foods on a 3-day menu rotation as described by Johnson and Lykken. 12 The basal diet was designed to reflect intakes of nutrients that did not necessarily meet the Recommended Dietary Allowances" but that are typical of intakes by adult women in the United States and that, in our experience, were sufficient to prevent deficiency of those nutrients. To achieve adequate iron intake and to prevent iron de-
pletion from phlebotomy, a supplement containing 8.5 mg of iron as ferrous gluconate was added to the beverage at each of the three daily meals. Thus, when this was combined with the diet, which contained 6.54 mg of iron per 2000 kcaVday the total daily intake of iron was approximately 32 mg. Demineralized water was consumed ad libitum. During an initial 13-day equilibration period subjects were fed a diet containing 800 mg of calcium/day and 2.97 mg of manganese/day, by analysis. 12 Subjects were then assigned in a doubleblind, Latin-square manner to each of four 39-day dietary periods created by the factorial combination of 587 and 1336 mg of calciUm/day and 1.0 and 5.6 mg of manganese/day, constant across subjects whose range of energy intakes was 1600 to 2800 kcaVday. Calcium and manganese intakes were varied by supplementing the basal diet with calcium lactate and manganese sulfate, respectively. Energy in the diet was distributed as 16% protein, 48% carbohydrate, and 36% fat. Initial energy expenditures were determined separately for each subject and adjusted during the course of the study to maintain body weight (measured daily) within ± 2% of initial weight. Individually prescribed exercise was performed three times weekly to maintain body composition and physical work capacity. Procedure. Subjects resided 24 hours/day in the metabolic research ward at the Grand Forks Human Nutrition Research Center to permit strict control of dietary intakes, exercise, and data collection and to provide a common environment. Subjects were chaperoned on all outings to ensure compliance with study protocol. All urine, feces, and menses were collected for biochemical analyses. In addition, blood samples were obtained by venipuncture during the middle and at the end of each dietary period for analysis of hematologic variables, minerals, lipids, enzymes, and hormones. Results of analyses on biologic samples were reported by Johnson and Lykken. 12 Collection of all menses allowed precise determination of the length of each cycle and duration of menstrual flow. The Menstrual Distress Questionnaire was administered individually to each subject at the completion of the menstrual phase of each cycle. The Menstrual Distress Questionnaire is a retrospective self-report instrument that assesses the presence and severity of 47 symptoms occurring during the menstrual, premenstrual (late luteal), and intermenstrual (follicular and early luteal) phases of the most recent menstrual cycle. 14 Operationally, the premenstrual phase was defined as the 7 days before the onset of menstruation. The Menstrual Distress Questionnaire was developed to help characterize changes women experience during the different phases of the menstrual cycle and to facilitate diagnosis of menstrual disorders such as premenstrual tension and dysmenorrhea. Symptoms are
Penland and Johnson
Volume 168, Number 5 Am J Obstet Gynecol
grouped into eight scales on the basis of factor analytic studies with several independent groups, The Autonomic Reactions scale includes dizziness, faintness, cold sweats, nausea, vomiting, and hot flashes. Behavior Change includes poorer work or school performance, increased napping and time in bed, increased time at home, avoiding social activities, and decreased overall efficiency. Concentration includes insomnia, forgetfulness, confusion, poorer judgment, difficulty concentrating, distractibility, increased accidents, and poorer motor coordination. Negative Affect includes crying, loneliness, anxiety, restlessness, irritability, mood swings, depression, and tension. Pain includes muscle stiffness, headache, cramps, backache, fatigue, and general aches and pains. Water Retention includes weight gain, skin disorders, painful breasts, and swelling. The Arousal scale reflects the positive feelings that some women report during certain phases of their cycle and includes increased affection, orderliness, excitement, feelings of well-being, and bursts of energy or activity. Last, the Control scale reflects a general tendency to complain and consists of symptoms commonly found in postmenopausal rather than premenopausal women; these include feelings of suffocation, chest pains, ears ringing, heart pounding, numbness, tingling, blind spots, and fuzzy vision. Because the 39-day dietary periods sometimes spanned more than one menstrual period, some subjects completed the Menstrual Distress Questionnaire twice during a single dietary period. The Menstrual Distress Questionnaire was amended slightly to include a 100 mm visual analog scale used to assess perceived volume or rate of flow during each day of menstruation. Data analyses. Dietary effects on Menstrual Distress Questionnaire scale scores, length of cycle, duration of menstrual flow, and perceived rate of flow were analyzed by repeated-measures analysis of variance by means of the General Linear Models procedure of the Statistical Analysis System package. 15 Mean scores for the symptom scales were at the low end of the possible range (Table II) because subjects in this study were not experiencing the degree of distress found in women with premenstrual syndrome or dysmenorrhea. However, parametric analysis was deemed appropriate because the variability around the means approximated normal. Post hoc comparisons of cell means for significant interaction effects were made with the Bonferroni t test with the experiment-wise error rate adjusted to p :S 0.05. Results
In contrast to when they were fed the high calcium diets, women fed low calcium diets reported significantly increased negative affect and more undesirable behavioral changes during all three phases of the men-
1419
strual cycle (Tables II through IV). Low calcium intake resulted in reports of poorer concentration during the premenstrual phase (Table III) and marginally during the menstrual !p < 0.052) and intermenstrual !p < 0.063) phases. Low calcium intake also resulted in reports of greater pain during the menstrual phase (Table II) and increased water retention during the premenstrual phase (Table III). A calcium x manganese interaction was observed for negative affect and, marginally !p < 0.068), for pain during the premenstrual phase. In spite of high calcium intake, low manganese intake resulted in increased symptomatology similar to that reported with the low calcium intake (Table III). Responses to Menstrual Distress Questionnaire control items and to arousal and autonomic reactions scale items were not significantly related to dietary calcium or manganese during any phase of the cycle. Neither calcium nor manganese intake influenced the length of the menstrual cycle or the duration of the menstrual phase. In spite of significantly increased menstrual hemoglobin losses during low dietary intakes of manganese,IZ the visual analog scale on the Menstrual Distress Questionnaire showed no effect of dietary calcium or manganese intakes on perceived menstrual volume. Comment
In general, these findings are consistent with other studies showing a relationship between macro and micro nutrients and symptoms among women with premenstrual tension. For example, Wurtman et al. 16 found that increased carbohydrate intake during the premenstrual phase improved depression, tension, anger, confusion, sadness, fatigue, alertness, and calmness scores. There have been several reports of reduced depression, irritability, bloating, and overall symptom scores with vitamin B6 administration, either alone or in combination with other vitamins and minerals. 17 However, numerous other studies have shown mixed or no effects of B6. II Vitamin E was found to reduce breast tenderness and total symptom scores,18 although a recent study failed to document a vitamin E deficiency in women with premenstrual distress. 19 Recently, Facchinetti et al."o reported that dietary magnesium supplementation relieved negative affective symptoms. The current findings are unique in showing a nutritional effect on the relatively mild physiologic and psychologic symptoms experienced by women with normal menstrual cycles. In addition, the data support a possible role for the minerals calcium and manganese in the cause and management of menstrual disorders. Specifically, the data indicate the existence of a relationship between dietary intake of calcium and the occurrence of several classes of symptoms commonly associated with menstruation and the menstrual cycle.
1420 Penland and Johnson
May 1993 Am J Obstct Cynccol
Table II. Menstrual Distress Questionnaire scale scores in relation to daily dietary calcium and manganese intakes: menstrual phase Intake
Menstrual Distress Questionnaire scale
Calcium (mg/day)
Manganese (mg/day)
587 587 1336 1336
1.0 5.6 1.0 5.6
Calcium intake Manganese intake Calcium x Manganese intake Mean ± SE
No.
Arousal (5-30)*
Autonomic Reactions (4-24)
Behavior Change (5-30)
9 8.32 ± 0.70 6.56 ± 1.73 10.97 ± 0.84 7 7.95 ± 0.82 5.49 ± 2.04 9.89 ± 0.98 9 7.72 ± 0.70 4.77 ± 1.73 7.96 ± 0.84 7.31 ± 0.75 8 6.85 ± 0.90 8.72 ± 1.87 Summary of analysis of variance effects: probabilities < 0.05 0.003 NS NS NS NS NS NS NS NS 4.22 26.16 6.08
Concentration (8-48)
14.26 14.56 12.99 12.52
0.74 0.87 0.74 0.80
± ± ± ±
NS NS NS 4.82
Data are presented as means ± SEM. NS, Not significant. *Range of possible values for each scale of the Menstrual Distress Questionnaire.
Table III. Menstrual Distress Questionnaire scale scores in relation to daily dietary calcium and manganese intakes: premenstrual phase Menstrual Distress Questionnaire scale
Intake Calcium (mg/day)
Manganese (mg/day)
587 587 1336 1336
1.0 5.6 1.0 5.6
Calcium intake Manganese intake Calcium x manganese intake Mean ± SE
No.
Arousal (5-30)*
Autonomic Reactions (4-24)
Behavior Change (5-30)
6.14 ± 1.69 10.00 ± 0.78 7.44 ± 0.43 9 9.35 ± 0.91 7.46 ± 0.51 4.49 ± 1.98 7 7.83 ± 0.78 4.63 ± 1.69 7.75 ± 0.43 9 6.73 ± 0.84 7.28 ± 0.47 8.48 ± 1.82 8 Summary of analysis of variance effects: probabilities < 0.05 0.010 NS NS NS NS NS NS NS NS 3.51 1.63 24.81 5.25
Concentration (8-48)
13.33 13.76 12.57 11.63
± ± ± ±
0.64 0.75 0.64 0.69
0.047 NS NS 1.71
Data are presented as means ± SEM. NS, Not significant. *Range of possible values for each scale of the Menstrual Distress Questionnaire. t, tBonferroni paired comparisons of dietary means for calcium x manganese interactions with pooled within-subject error term. Means in the same column with different superscripts are statistically different (a. adjusted to sO.05).
Table IV. Menstrual Distress Questionnaire scale scores in relation to daily dietary calcium and manganese intakes: intermenstrual phase Intake Calcium (mg/day)
Manganese (mg/day)
587 587 1136 1336
1.0 5.6 1.0 5.6
Calcium intake Manganese intake Calcium x manganese intake Mean ± SE
Menstrual Distress Questionnaire scale No.
9 7 9 8 Summary
Arousal (5-30)*
Autonomic Reactions (4-24)
7.24 ± 0.57 7.47 ± 0.67 7.37 ± 0.57 7.24 ± 0.61 of analysis of variance NS NS NS 2.82
Behavior Change (5-30)
8.79 ± 0.83 5.66 ± 0.96 8.83 ± 0.98 4.77 ± 1.13 7.31 ± 0.83 4.33 ± 0.96 6.09 ± 0.90 6.65 ± 1.04 effects: probabilities < 0.05 0.028 NS NS NS NS NS 6.01 8.02
Concentration (8-48)
12.47 13.55 11.43 10.75
± 0.90
± 1.06
± 0.90 ± 0.98
NS NS NS 7.11
Data are presented as means ± SEM. NS, Not significant. *Range of possible values for each scale of the Menstrual Distress Questionnaire.
The beneficial effects of high dietary calcium on behavior, mood, and concentration were present during all phases of the cycle, suggesting that calcium intake may influence psychologic processes in general (i.e., be in-
dependent of menstruation). Alternatively, these findings may simply confirm the absence of menstrual disorders among the women participating in this study. However, symptoms related to water retention and
Penland and Johnson
Volume WH. !';lImoer " Am J Obstet (;\,necol
Menstrual Distress Questionnaire scale Control (6-36)
7.37 7.30 7.16 7.01
:±: :±: :±: :±:
0.47 0.55 0.47 0.51
Negative Affect
Water Retention
(8-48)
(4-24)
0.78 0.91 0.78 0.84 Summary of analysis of variance effects: probabilities < 0.05 NS 0.018 0.034 NS
NS NS
1.93
19.93:±: 21.01:±: 18.34:±: 11.50:±:
1.99 2.34 1.99 2.15
NS NS
34.61
13.58:±: 13.59:±: 11.41:±: 10.89:±:
1.0 1.17 1.0 1.08
NS NS
8.66
8.37 :±: 9.34:±: 7.52 :±: 7.25:±:
NS NS
5.23
Menstrual Distress Questionnaire scale Control (6-36)
7.18:±: 7.04 :±: 6.99:±: 6.52 :±:
0.44 0.52 0.44 0.48
Negative Affect
Water Retention
(8-48)
(4-24)
0.67 0.79 0.67 0.72 Summary of analysis of variance effects: probabilities < 0.05 NS 0.008 NS 0.041
NS NS
27.67
18.91 :±: 21.46 :±: 17.60:±: 11.46:±:
1.78t 11.I4:±: 0.91 2.09t 12.39:±: 1.07 1.78t.+ 11.48:±: 0.91 1.92t 8.96 :±: 0.98
NS
0.035 7.24
8.98:±: 8.15:±: 7.71:±: 6.26:±:
NS NS
3.91
NS NS
Menstrual Distress Questionnaire scale Control (6-36)
7.02 6.97 6.81 6.56
:±: 0.45 ± 0.53 ± O.4S ± 0.49
Negative Affect (8-48)
16.31 19.37 14.56 8.73
:±: 2.28 ± 2.67 ± 2.28 :±: 2.46
NS NS NS
0.020
1.76
45.11
NS NS
Pain
Water Retention
(6-36)
(4-24)
11.29 10.61 10.10 8.02
:±: 1.21 ± 1.42 ± 1.21 :±: 1.31
NS NS NS
12.78
6.41 :±: 0.58
6.S7 ± 0.68 6.38:±: 0.58 5.37 :±: 0.62
2.88
NS NS NS
However, symptoms related to water retention and pain, which might be thought of as more physiologic in nature, showed beneficial effects of high dietary calcium that were phase-specific; these findings more clearly suggest a role for calcium in regulating menstrual symptoms.
1421
Reports of increased water retention with low calcium intake were found only for the premenstrual phase of the cycle, although a similar tendency was also observed for the menstrual phase (p < 0.094). Unfortunately, water retention was not directly assessed in the current study. Analysis of daily body weight showed only slight variation with phase of cycle. A marked dietary effect was evident, however, with the lowest weights associated with the high calcium-high manganese diet. Because water retention and weight gain are characteristic of the premenstrual phase, reports of increased water retention during this phase when women were fed low calcium diets might simply reflect an exacerbation of this normal process. Reports of greater pain associated with low calcium were found only during the menstrual phase, although a marginal calcium x manganese interaction was also observed for the premenstrual phase. Post hoc analysis of subject requests for pain medication during the course of the study (ratio of number of days pain medication was requested to number of days in each phase of the cycle) showed more requests during the menstrual than premenstrual or intermenstrual phases, except when the low calcium-low manganese diet was used. Requests for pain medication for the premenstrual and intermenstrual phases were approximately equal to those for the menstrual phase when the women were given this diet. Mild pain during menstruation corresponds to the vasomotor activity and endometrial changes that lead to menstrual effluvium and is a normal occurrence.' Reduced calcium in the diet during this phase may aggravate this common symptom by its effects on either of these processes (see below). The clinical observations regarding body weight and requests for pain medication suggest that low manganese intakes increase the effect of low calcium intakes on both water retention and pain and thus affect both premenstrual and menstrual phases of the cycle. Examination of mean scores on the Menstrual Distress Questionnaire Negative Affect scale for each dietary combination (Tables II through N) reveals that dietary intake of manganese also is related to symptoms reflecting mood, but only when calcium intake is high. Therefore low intakes of either manganese or calcium may increase the occurrence of negative mood states. Because this finding is statistically significant only for the premenstrual phase, it can be argued that the relationship between manganese and mood states is also mediated by the menstrual cycle. As with calcium, dietary manganese thus may be a factor in the changes in mood commonly reported before the onset of menstruation. It should be noted that the amount of calcium in the diets designated as low in this study represent typical intakes among women in this age group, whereas the amount of manganese in the low-manganese diets is roughly 50% of typical intakes; the high calcium and manganese intakes used here only slightly exceed those recommended.'"
1422 Penland and Johnson Am
Dietary manganese and iron are known to interact so that iron intakes are inversely related to manganese absorption and use. Therefore it is possible that relatively high daily intakes of approximately 32 mg of iron, used in this study to prevent iron depletion from phlebotomy, may have inhibited manganese absorption or use. Because absorption of a nutrient is typically less with higher compared with lower intakes of that nutrient, this effect would have been most pronounced during manganese supplementation. Further, manganese sulfate, the form of manganese used in this study, is poorly absorbed in comparison with manganese chloride. Thus the amount of available manganese supplementation and the difference between dietary treatments may have been less than originally planned. The apparent absence in this study of an effect of calcium on manganese absorption 12 could also be caused by the relatively poor solubility of the manganese salt. If iron intake and the form of manganese supplement were factors in the current study, then their effect most likely was to reduce statistical power and minimize the effects of manganese and thus make more noteworthy the few significant effects of manganese that were observed on menstrual symptoms. The potential for dietary intakes of calcium and manganese to influence menstrual cycle symptoms can be found in numerous studies showing the disruptive effects of low intakes of these nutrients on reproduction and the estrus cycle in animals and showing a relationship between biochemical concentrations of these nutrients and phase of the estrus or menstrual cycle in both animals and humans. In animals, calcium concentrations in the blood vary over the course of the estrus cycle in turkeys, sheep, and cows, and low dietary manganese results in reproductive problems and absent or irregular estrus cycles in rats, guinea pigs, rabbits, chicks, pigs, goats, sheep, and cows. Further, manganese uptake by the ovaries, fallopian tubes, uterine horn, and vagina of ewes varies during the estrus cycle, which suggests that manganese is needed for normal cyclic function. In humans, Puskulian 21 reported decr~a~ed ~ali~ary calcium at midcycle compared with the menstrual phase in women with normal cycles, apparently independent of calcium intake. Although failing to find a change in total calcium during the menstrual cycle, Pitkin et al!2 reported significant phase-related changes in ionic calcium and the calcium-regulating hormones, serum calcitonin and plasma parathyroid hormone, in normally menstruating women. Most recently, Lehtovirta et a1. 23 reported that serum calcium concentrations were significantly elevated during menstruation in women with normal cycles and those with anovulatory cycles. The latter group also showed elevated calcium concentrations during the premenstrual phase. Further, both groups showed a negative corre-
May 1993 Gynecol
J Obstet
lation between calcium and progesterone concentrations during the premenstrual phase. In a series of studies with healthy menstruating women, Hagenfeldt et a1. 24 found elevated manganese concentrations in the endometrium during the early proliferative phase of the cycle and higher plasma manganese during the proliferative phase than during the secretory phase. It has been proposed that increased food consumption, particularly carbohydrates, during the premenstrual phase of the cycle is an attempt to improve mood states by increasing brain tryptophan and serotonin. 16 Perhaps higher intakes of calcium and manganese in interaction with changes in progesterone during this phase would yield a similar result, with additional positive consequences during this and other phases of the cycle. Interestingly, peak blood glucose during intravenous glucose tolerance tests was highest when subjects were fed the high calcium-low manganese diet'2; when contrasted with the high calcium-high manganese diet, this diet resulted in increased reports of negative affect. If women experiencing normal menstruation and normal menstrual cycles show a relationship between calcium and manganese intakes and the symptoms often used to characterize clinical syndromes such as premenstrual distress and dysmenorrhea, would that relationship be more clearly observed in women suffering menstrual-related disorders? For example, would calcium intakes significantly in excess of the recommended 800 to 1200 mg/day be of any therapeutic value in the treatment of clinical syndromes related to menstruation? A recent study by Thys-Jacobs et a1."5 directly addresses the latter question. In a double-blind, crossover study of 33 women (14 black and 12 Hispanic) diagnosed as having premenstrual syndrome, these investigators found that daily supplements of 1000 mg of calcium, as calcium carbonate, resulted in reduced reports of negative affect and water retention during the premenstrual phase and of pain during both the premenstrual and menstrual phases of the cycle. These findings are remarkably consistent with the phase-related effects of calcium intake found in the current study. Thys-Jacobs et a1. 25 speculated that their findings may reflect the actions of calcium on neurotransmitters, smooth muscle responsiveness, or hormone secretion. Clearly, these same mechanisms also may be responsible for the current findings in women with normal mesntruation and cycles. We are currently in the process of replicating this study on a more extensive cross section of women and using a prospective measure of symptomatology with the hope of being able to further delineate the effects of calcium and manganese on behavior, concentration, and other symptoms relevant to menstruation. Considering the potential for retrospective reports to include variability resulting from inaccurate recall of symptom
Volume 168. Number 5 Am J Obstet Gynecol
occurrence and severity, daily reporting of symptoms is recommended for future studies. At present, the data reported here underscore the importance of menstruating women consuming a diet that fully conforms to current National Research Council guidelines" for both calcium and manganese. We thank Mary Lou Eberhardt for coordinating data collection, LoAnne Mullen and the dietary staff, Betty Vetter and the nursing staff, Bill Beckwith and the psychology staff, LuAnn Johnson and the data processing staff, and, most important, the women who participated as subjects in this study. REFERENCES 1. Kase NG. The endocrinology of the ovarian cycle. In: Kase NG, Weingold AB, Gershenson DM, eds. Principles and practice of clinical gynecology. 2nd ed. New York: Churchill Livingstone, 1990:121-49. 2. Asso D. Physiology and psychology of the normal menstrual cycle. In: Brush MG, Goudsmit EM, eds. Functional disorders of the menstrual cycle. New York: John Wiley, 1988: 15-36. 3. Gannon LR. Covariates of the menstrual cycle in normal populations. In: Menstrual disorders and menopause. New York: Praeger, 1985:36-66. 4. Lissner L, Stevens J, Levitsky DA, Rasmussen KM, Strupp BJ. Variations in energy intake during the menstru.al cycle: implications for food-intake research. Am J Clm Nutr 1988;48:956-62. 5. Tarasuk V, Beaton GH. Menstrual-cycle patterns in energy and macronutrient intake. AmJ Clin Nutr 1991;53:442-7. 6. Bowen DJ, Grunberg NE. Variations in food preference and consumption across the menstrual cycle. Physiol Behav 1990;47:287-91. 7. Gallant MP, Bowering J, Short SH, Turkki PR. Pyridoxine and magnesium status of women with premenstrual syndrome. Nutr Res 1987;7:243-52. 8. Hrboticky N, Leiter LA, Anderson GH. Menstrual cycle effects on the metabolism of tryptophan loads. Am J Clin Nutr 1989;50:46-52. 9. Chuong CJ, Kinch RAH. Premenstrual syndrome. In: Kase NG, Weingold AB, Gershenson DM, eds. Principles and practice of clinical gynecology. 2nd ed. New York: Churchill Livingstone, 1990:669-84. 10. Brush MG. Vitamins, essential fatty acids and minerals in relation to the aetiology and management of premenstrual
Penland and Johnson
1423
syndrome. In: Brush MG, Goudsmit EM, eds. Functional disorders of the menstrual cycle. New York: John Wiley, 1988:69-85. 11. Kleijnen J, Ter Riet G, Knipschild P. Vitamin B6 in the treatment of premenstrual syndrome-a review. Br J Obstet Gynaecol 1990;97:847-52. 12. Johnson PE, Lykken Gl. Manganese and calcium absorption and balance in young women fed diets with varying amounts of manganese and calcium. J Trace Elements Exp Med 1991;4:19-35. 13. National Research Council. Recommended dietary allowances. 10th ed. Washington, DC: National Academy Press, 1989. 14. Moos RH. Typology of menstrual cycle symptoms. AM J OBSTET GYNECOL 1969;103:390-402. 15. SAS Institute. SAS user's guide: statistics, version 5 edition. Cary, North Carolina: SAS Institute, 1985. 16. Wurtman JJ, Brzezinski A, Wurtman ~, Laferrere B. Effect of nutrient intake on premenstrual depression. AM J OBSTET GYNECOL 1989; 161: 1228-34. 17. Abraham G, Rumley R. The role of nutrition in managing the premenstrual tension syndrome. J Reprod Med 1987; 32:405-22. 18. London RS, Murphy L, Kitlowski KE, Reynolds MA. Efficacy of a-tocopherol in the treatment of premenstrual syndrome. J Reprod Med 1987;32:400-4. 19. Chuong CJ, Dawson EB, Smith ER. Vitamin E levels in premenstrual syndrome. AM J OBSTET GYNECOL 1990;163: 1591-5. 20. Facchinetti F, Borella P, Sances G, Fioroni L, Nappi RE, Genazzani AR. Oral magnesium successfully relieves premenstrual mood changes. Obstet Gynecol 1991 ;78: 17781. 21. Puskulian L. Salivary electrolyte changes during the normal menstrual cycle. J Dent Res 1972;5 I (suppl 5):1212-6. 22. Pitkin RM, Reynolds WA, Williams GA, Hargis GK. Calcium-regulating hormones during the menstrual cycle. J Clin Endocrinol Metab 1978;47:626-32. 23. Lehtovirta P, Apter D, Stenman UH. Serum CA 125 levels during the menstrual cycle. Br J Obstet Gynaecol 1990; 97:930-3. 24. Hagenfeldt K, Plantin LO, Diczfalusy E. Trace elements in the human endometrium. 2. Zinc, copper and manganese levels in the endometrium, cervical mucus and plasma. Acta Endocrinol 1973;72: 115-26. 25. Thys-Jacobs S, Ceccarelli S, Bierman A, Weisman H, Cohen M, Alvir J. Calcium supplementation in premenstrual syndrome: a randomized crossover trial. J Gen Intern Med 1989;4: 183-9.