Diet and Sex-Hormone Binding Globulin,
Dysmenorrhea, and Premenstrual Symptoms
NEAL D. BARNARD, MD, ANTHONY R. SCIALLI, MD, DONNA HURLOCK, MD, AND
PATRICIA BERTRON, RD
Objective: To test the hypothesis that a low-fat, vegetarian
diet reduces dysmenorrhea and premenstrual symptoms by
its effect on serum sex-hormone binding globulin concentration and estrogen activity.
Methods: In a crossover design, 33 women followed a
low-fat, vegetarian diet for two menstrual cycles. For two
additional cycles, they followed their customary diet while
taking a supplement placebo pill. Dietary intake, serum
sex-hormone binding globulin concentration, body weight,
pain duration and intensity, and premenstrual symptoms
were assessed during each study phase.
Results: Mean (6
6 standard deviation [SD]) serum sexhormone binding globulin concentration was higher during
the diet phase (46.7 6 23.6 nmol/L) than during the supplement phase (39.3 6 19.8 nmol/L, P < .001). Mean (6
6 SD)
body weight was lower during the diet (66.1 6 11.3 kg)
compared with the supplement phase (67.9 6 12.1 kg, P <
.001). Mean dysmenorrhea duration fell significantly from
baseline (3.9 6 1.7 days) to diet phase (2.7 6 1.9 days)
compared with change from baseline to supplement phase
(3.6 6 1.7 days, P < .01). Pain intensity fell significantly
during the diet phase, compared with baseline, for the worst,
second-worst, and third-worst days, and mean durations of
premenstrual concentration, behavioral change, and water
retention symptoms were reduced significantly, compared
with the supplement phase.
Conclusion: A low-fat vegetarian diet was associated with
increased serum sex-hormone binding globulin concentration and reductions in body weight, dysmenorrhea duration
and intensity, and premenstrual symptom duration. The
symptom effects might be mediated by dietary influences on
estrogen activity. (Obstet Gynecol 2000;95:245–50. © 2000
by The American College of Obstetricians and Gynecologists.)
Up to 10% of women in their teens and early twenties
suffer from severe, mostly primary, dysmenorrhea.1 Its
From the Physicians Committee for Responsible Medicine, Washington, DC, and the Department of Obstetrics and Gynecology, Georgetown
University School of Medicine, Washington, DC.
VOL. 95, NO. 2, FEBRUARY 2000
cause is believed to be related to uterine muscle contraction and ischemia induced by prostaglandins (PGs)
(particularly PGE2 and PGF2a) produced in endometrial
tissue under the influence of estrogens and progesterone.2 Cyclic hormonal changes also affect premenstrual
symptoms, which affect an estimated 20 – 40% of
women, with 2.5–5% reporting adverse effects on work
or social adjustment.3
Dietary factors alter serum sex-hormone concentrations and activity. Plant-based and vegetarian diets
increase serum concentration of sex-hormone binding
globulin, which binds and inactivates estrogens.4 – 6 In
addition, independent of the effect of diet, serum sexhormone binding globulin concentration is inversely
associated with body weight, which is typically lower
in vegetarians.7 Low-fat and vegetarian diets also reduce serum estrogen concentrations in premenopausal
and postmenopausal women.6,8 –12
Anecdotal reports indicate that a low-fat, vegetarian
diet might reduce menstrual pain in some individuals.13
The present study tested the hypothesis that a low-fat,
vegetarian diet reduces dysmenorrhea and premenstrual symptoms in women with moderate to severe
menstrual pain, by its effect on serum sex-hormone
binding globulin and estrogen activity.
Materials and Methods
The study was approved by the institutional review
board and the Department of Obstetrics and Gynecology of the Georgetown University. Volunteers were
recruited through newspaper advertisements and notices mailed to gynecologists in the Washington, DC,
area. All participants gave written informed consent
before enrollment. Participants were to be at least 18
years of age, with menstrual periods accompanied
consistently by moderate to severe abdominal pain
causing significant distress or impairment in social or
occupational functioning, with pain-free intervals of at
0029-7844/00/$20.00
PII S0029-7844(99)00525-6
245
least 2 weeks between periods. Premenstrual symptoms
were not a criterion for participation. No monetary
compensation was provided.
Exclusionary criteria included menstrual cycles that
were irregular or consistently shorter than 25 days or
longer than 35 days; physical illness affecting eating
behavior or causing pain; any history of a hormonerelated illness, such as diabetes or thyroid disease;
history of mental illness or alcohol or drug abuse; use in
the previous 6 months of oral contraceptives or other
drugs known to affect hormonal function; or a structural abnormality that could account for pain. The
study was conducted in two successive replications
with 24 and 27 participants, respectively. After medical
history and physical examination, volunteers were assigned randomly, using a computer-generated random
number list, to two groups that received the same
treatments (a diet intervention phase and a placebo
supplement phase), but in opposite order in a crossover
design.
During the baseline period of one full menstrual
cycle, participants were asked to make no diet changes.
On day 2 of the next menstrual cycle (counting the first
day of menstruation as day 1), one group was asked to
begin an intervention diet for two full menstrual cycles,
after which they were asked to resume their customary
diets and take a daily supplement pill for two full
menstrual cycles. The remaining participants had the
conditions in the reverse order.
The intervention diet consisted of grains, vegetables,
legumes, and fruits, with no quantitative restrictions.
Animal products, added oils, fried foods, avocados,
olives, nuts, nut butters, and seeds were proscribed. The
diet was generally adequate in all nutrients except for
vitamin B-12, for which supplementation was recommended for any participants who chose to follow the
intervention diet after the study’s conclusion. The diet
provided approximately 10% of calories from fat.
The supplement consisted of 2 mg of vitamin B-12.
Participants were told that the supplement they received could consist of a vitamin or a placebo, and that
if it were a vitamin, its identity and dose would not be
revealed until the end of the study.
Participants attended weekly 1-hour meetings of
their assigned groups for the duration of the study.
Meetings included nutrition lectures, cooking demonstrations, informal discussions, and emotional support.
Spouses or partners were invited to specially scheduled
cooking demonstrations and lectures. Samples of groceries, such as brown rice, canned beans, and instant
soups, were provided occasionally during the intervention diet phase.
During the baseline cycle and the second cycle of
each intervention phase, the following data were col-
246 Barnard et al
Diet and Dysmenorrhea
lected: On day 6, blood samples for sex-hormone binding globulin and serum lipids were collected after an
overnight fast. Serum was separated and frozen so
laboratory determinations for all participants in each
replication could be run in the same batch at the end of
the study.
Serum sex-hormone binding globulin was measured
by DSL-6300 radioimmunoassay (Diagnostic Systems
Laboratories, Inc., Webster, TX.)14 Methods and results
for serum lipids are reported elsewhere. An exploratory
assessment of serum estrogen concentrations, including
estradiol (E2) (total, free, and albumin-bound), estrone
(E1), and E1 sulfate, was made using radioimmunoassay techniques in a subsample of 17 participants. No
clear pattern emerged in those determinations, and the
assays were discontinued.
Between days 6 and 14, each participant completed a
3-day dietary record, including 2 weekdays and 1
weekend day.15 Records were analyzed using Nutritionist V for Windows 95 (First DataBank Inc., Hearst
Corporation, San Bruno, CA). At the same time, participants also completed a food-supplement questionnaire
that addressed acceptability of the diet and any beneficial or adverse effects experienced.
Between days 14 and 21, body weight in street clothes
but without shoes was measured to the nearest 0.2 kg.
Premenstrual symptoms, including pain, altered concentration, behavioral change, autonomic reactions, water retention, negative affect, arousal, sleep disturbances, and food cravings, were assessed daily from
day 14 through the end of menstrual flow using the
Menstrual Symptom Diary, a modification of Moos’
Menstrual Distress Questionnaire.16,17
Menstrual pain was assessed on the same schedule as
the Menstrual Symptom Diary, using a short form of
the Brief Pain Inventory developed by the Pain Research Group of the World Health Organization Collaborating Centre for Symptom Evaluation in Cancer
Care.18 It records pain at its worst during the preceding
24 hours using a scale from 0 (no pain) to 10 (pain as
bad as you can imagine). Throughout the study, participants were asked to use pain medications only when
pain was experienced, rather than prophylactically.
At the conclusion of the supplement phase, each
participant was asked to describe on a four-point scale
the extent to which the supplement reduced menstrual
pain. Participants also were asked to guess if the
supplement they had received was a placebo or an
active compound. Participants who believed it to be an
active compound were asked if it seemed to reduce pain
or to cause side effects.
Sample size was based on the primary outcome of
pain duration and was estimated to provide 80% power
in a repeated measures analysis of variance with a 5
Obstetrics & Gynecology
Table 1. Baseline Demographic and Clinical Characteristics
Mean age (y)
Age range (y)
White, non-Hispanic
Black, non-Hispanic
White, Hispanic
Black, Hispanic
Asian
Smoker
Single
Married
Separated, divorced, widowed
Age at menarche (mean, y)
Pregnancies (median, range)
Live births (median, range)
Living children (median, range)
Completers
(n 5 33)
Noncompleters
(n 5 18)
36.1
22– 48
20
8
4
0
1
2
22
8
3
12.3
0 (0 –13)
0 (0 –2)
0 (0 –2)
32.8
18 –50
10
7
0
1
0
1
13
4
1
13.2
0 (0 –5)
0 (0 –5)
0 (0 –5)
5% to detect a medium to small effect. This calculation
called for 50 participants, but interim analysis revealed
significant effects with a lower number. Statistical analyses included parametric and nonparametric tests. For
most measures (dietary characteristics, body weight,
body mass index (BMI), sex-hormone binding globulin,
and duration of pain, flow, and premenstrual symptoms), a within-subjects analysis of variance was done
on the basis of two change scores, the difference between baseline and intervention diet phase values and
the difference between baseline and supplement phase
values. A significant effect indicated that the change
from baseline to the intervention diet phase was greater
or smaller than the change from baseline to the supplement phase. Because some participants had the intervention diet phase first and others had the supplement
phase first, we also tested for order effect.
For pain intensity, we conducted Wilcoxon matchedpairs, signed-rank tests to compare baseline with supplement phase values, baseline with intervention diet
phase values, and supplement phase with intervention
diet phase values for pain intensity for the worst,
second-worst, and third-worst days of pain. The
matched-pairs test is a within-subjects test in which an
individual’s score in one phase is compared with the
same individual’s score in another phase. A significant
effect indicated that over all participants, pain intensity
was different in one phase than the comparison phase.
Differences between the intervention diet and supplement phases in responses to the food-supplement questionnaire were assessed using McNemar test.
Results
Fifty-one volunteers met the criteria for participation.
Their demographic characteristics are listed in Table 1.
VOL. 95, NO. 2, FEBRUARY 2000
Table 2. Dietary Characteristics (Per Day)
Energy (MJ)
Protein (g)
Carbohydrates (g)
Total sugar (g)
Fat (g)
Saturated fat (g)
Cholesterol (mg)
Total fiber (g)
Caffeine (mg)
Baseline
Supplement
Vegetarian
7.85 6 1.86
67.2 6 20.1
293 6 78
108 6 42
51.8 6 24.9
15.5 6 12.0
121.2 6 130.0
26.7 6 9.4
95.1 6 109.0
7.33 6 2.02
59.8 6 17.7
284 6 93
101 6 45
43.2 6 23.9
13.8 6 10.2
130.8 6 93.4
23.1 6 11.5
81.2 6 104.2
6.42* 6 1.62
43.5†‡ 6 11.5
313 6 91
108 6 49
16.6†‡ 6 9.8
2.5†§ 6 2.4
4.7†‡ 6 7.8
31.3† 6 9.2
54.2 6 69.6
Data presented as mean 6 standard deviation.
n 5 33.
* P , .05 for difference between change from baseline to intervention
diet, compared with change from baseline to supplement phase.
†
P , .001 for difference between change from baseline to intervention diet, compared with change from baseline to supplement phase.
‡
Order effect, P , .05.
§
Order effect, P , .01.
Twelve volunteers failed to complete the study due to
menopause (1), pregnancy (1), repeatedly missing
group meetings (2), noncompliance with the intervention diet (2), extended overseas travel (1), death in the
family (1), major financial problems (1), spouse objections to the diet (1), scheduling difficulties (1), and
unknown reasons (1). In addition, six participants were
excluded from the data analysis for failure to return all
data forms. There were no significant demographic
differences between completers and noncompleters.
Nutrient intake is presented in Table 2. The reductions in energy, protein, fat, and cholesterol intake and
the increase in fiber intake from baseline to the intervention diet phase were significantly greater than the
corresponding changes from baseline to the supplement
phase.
Mean body weight was 2.7 kg lower during the
intervention diet phase than at baseline, a significant
difference compared with the change from baseline to
supplement phase (Table 3). Mean BMI fell from baseline to intervention diet phase. An order effect was
Table 3. Clinical Measures
Weight (kg, n 5 33)
BMI (kg/m2, n 5 33)
SHBG (nmol/L) (n 5 32)
Flow duration (d, n 5 32)
Pain duration (d, n 5 31)
Baseline
Supplement
Intervention
68.8 6 11.9
25.5 6 5.1
41.9 6 20.8
4.9 6 1.4
3.9 6 1.7
67.9 6 12.1
25.1 6 5.1
39.3 6 19.8
5.1 6 1.4
3.6 6 1.7
66.1 6 11.3*†
24.5 6 4.9*†
46.7 6 23.6*
4.8 6 1.8
2.7 6 1.9‡
BMI 5 body mass index; SHBG 5 sex-hormone binding globulin.
Data presented as mean 6 standard deviation.
* P , .001 for difference between change from baseline to intervention diet phase, compared with change from baseline to supplement
phase.
†
Order effect, P , .05.
‡
P , .01 for difference between change from baseline to intervention
diet phase, compared with change from baseline to supplement phase.
Barnard et al
Diet and Dysmenorrhea
247
evident, such that weight lost by those having the
intervention diet phase first was not regained fully
during the subsequent supplement phase.
Mean serum sex-hormone binding globulin concentration was 19% higher during diet phase, compared
with supplement phase (Table 3). The change from
baseline to diet phase was significantly greater than
from baseline to supplement phase. There were no
significant associations between serum sex-hormone
binding globulin concentration and body weight or
intake of any single nutrient.
Menstrual flow duration did not differ significantly
by treatment condition. However, pain duration (defined as number of days during which participants
reported scores of 1 or greater on the Brief Pain Inventory for pain at its worst during the preceding 24 hours,
from 3 days before the onset of menses to 5 days after)
dropped significantly more from baseline to diet phase
compared with the change from baseline to supplement
phase (Table 3).
Pain intensity, defined as the Brief Pain Inventory
value for pain at its worst during the preceding 24
hours on each of the 3 days of maximal severity during
each study phase, shifted toward lower values during
the diet phase. Median values for the 3 days of maximal
severity were 7, 5, and 3 at baseline; 6, 4, and 2 during
the supplement phase; and 6, 3, and 0 during the
intervention phase. Using Wilcoxon matched-pairs,
signed-rank test, we found that differences between
baseline and intervention phases were significant for
each of the 3 days (P 5 .05 for the worst day, P , .01 for
the second- and third-worst days). There were no
significant differences between baseline and supplement phase values. Differences between diet and supplement phases were significant for the second-worst
day (P 5 .05). There were no significant associations
between serum sex-hormone binding globulin concentration and pain duration or intensity.
When asked the extent to which the supplement had
reduced their pain, 28 participants reported that it had
minimal or no effect, whereas five described its effect as
moderate. When asked to guess, 25 participants believed the supplement to be a placebo. Eight believed it
to be an active compound, of whom six thought it
reduced their pain, whereas two believed it caused side
effects.
The mean duration of premenstrual symptoms, defined as the number of days during which any symptom was rated as present during the 5 days preceding
the onset of menses, regardless of symptom intensity,
was significantly reduced during the diet phase for
three symptom clusters (comparing the change from
baseline to diet phase with the change from baseline to
the supplement phase). For water retention, mean
248 Barnard et al
Diet and Dysmenorrhea
symptom durations in the baseline, supplement, and
intervention phases were 2.9, 2.5, and 1.3 days, respectively (P , .01). For behavioral change symptoms, mean
durations were 1.7, 1.7, and 1.1 days, respectively (P ,
.05), and for concentration symptoms, mean durations
were 1.9, 1.5, and 0.7 days, respectively (P , .01).
On the food-supplement questionnaire, 23 participants noted weight loss during the intervention diet
phase, compared with three during the supplement
phase (McNemar Q 5 18.18, P , .001), and 17 reported
improved energy during diet phase, compared with
five during supplement phase (McNemar Q 5 10.28,
P , .001.) Related to those perceived benefits, several
participants who had the vegetarian phase first in
sequence refused to return fully to their customary diets
during the supplement phase, despite the fact that the
crossover design required it.
Discussion
In women with moderate to severe dysmenorrhea, a
low-fat, vegetarian diet was associated with a significant increase in mean serum sex-hormone binding
globulin concentration and reductions in mean body
weight and BMI as well as significant reductions in
menstrual pain duration, pain intensity, and duration of
premenstrual symptoms related to concentration, behavioral change, and water retention.
There are several ways that diet might affect PG
synthesis. Diet influences sex hormone concentration
and activity. Populations on plant-based or vegetarian
diets typically have increased serum sex-hormone binding globulin concentrations.4 – 6 Such diets also often
reduce body weight, which is correlated inversely with
serum sex-hormone binding globulin concentrations.7
Low-fat diets also reduce serum estrogen concentrations in premenopausal and postmenopausal women.10 –12 Estrogen conjugates are excreted in bile and are
subject to enterohepatic circulation, which can be interrupted by dietary fiber, encouraging fecal estrogen
elimination.6 Elevated serum sex-hormone binding
globulin or reduced serum estrogen concentrations
might reduce estrogenic stimulation of the endometrium, limiting proliferation of tissues that produce
PGs.
In the current study, mean serum sex-hormone binding globulin concentration during the diet phase was
19% higher than that during the supplement phase.
Sex-hormone binding globulin concentration was not
associated with intake of nutrients or body weight, nor
was it associated with pain duration or intensity during
any phase. We found no significant changes in serum
estrogen concentrations in a subsample of participants.
Pulsatile release of estrogens might have prevented us
Obstetrics & Gynecology
from detecting changes in estrogen concentrations.
Also, we measured serum hormones earlier than in
previous studies, after only about 5 weeks on the diet.
Vegetables and legumes, particularly soy products,
are rich in phytoestrogens, which compete with other
estrogens for receptor binding, potentially affecting the
pituitary-ovarian axis.6 The addition of 60 g of soy
protein to the diets of premenopausal nonvegetarian
women was associated with a lengthening of the follicular phase by a mean of 2.5 days and lower midcycle
peaks of luteinizing hormone and FSH.19
Limited evidence suggests that ovulatory disturbances might be less frequent among vegetarians. In a
6-month study, 15% of cycles among 22 nonvegetarians
were anovulatory, compared with fewer than 5% of
those of 23 vegetarians. The distribution of ovulatory,
anovulatory, and short-luteal-phase cycles differed significantly between groups (x2 5 9.64, P , .01).9
Vegetables, fruits, and legumes are typically low in
total fat, but their content of omega-3 fatty acids,
relative to other fats, is often high. Omega-3 fatty acids
are precursors of 3-series PGs, which have antiinflammatory actions. In contrast, diets rich in animal
fats and cooking oils are proportionately richer in
omega-6 fatty acids, which promote the formation of
PGE2 and PGF2a.20,21 In a group of Danish women, a
higher intake of omega-3 fatty acids or a higher ratio of
omega-3/omega-6 fatty acids was associated with reduced menstrual pain.22
Duration of premenstrual symptoms related to concentration, behavioral change, and water retention was
significantly reduced during the diet phase. In a previous study of 30 healthy women, a reduction in dietary
fat from 40% to 20% of energy was associated with
reduced premenopausal water retention symptoms.23
Interpretation of our data is limited by several factors.
A placebo pill is clearly not as engaging as a diet
change. We did not believe that an alternative diet, such
as a gluten-free diet, could be presented credibly as a
potential treatment for menstrual symptoms. Our participants appeared to believe that an active substance
was being tested during the supplement phase, at least
on some participants, and eight believed they received
an active compound.
Some individuals with dysmenorrhea might have
endometriosis or other conditions that are not detectable on routine examination. We excluded volunteers
who had been diagnosed with such conditions, but we
did not conduct laparoscopic examinations or biopsies.
Our control over analgesics was limited to our request that they be used only during pain, rather than
prophylactically. We believed that proscription of analgesics was not ethically permissible with an experimental treatment and placebo. For future studies it would
VOL. 95, NO. 2, FEBRUARY 2000
help to provide a standardized medication and dose
that would facilitate comparisons of medication use
across different study phases. In future research, it
would be useful to measure changes in the quantity of
endometrial flow and its prostaglandin concentration in
the context of a low-fat, vegetarian diet and to explore
effects of that regimen on biochemical indices in other
hormone-related conditions, such as endometriosis,
leiomyomas, or hormone-related cancers.
References
1. Task Force on Taxonomy of the International Association for the
Study of Pain. Diseases of the bladder, uterus, ovaries, and adnexa.
In: Merskey H, Bogduk N, eds. Classification of chronic pain, 2nd
ed. Seattle, Washington: IASP Press, 1994:164 – 6.
2. Benedetto C. Eicosanoids in primary dysmenorrhea, endometriosis
and menstrual migraine. Gynecol Endocrinol 1989;3:71–94.
3. American College of Obstetricians and Gynecologists. ACOG
committee opinion: Premenstrual syndrome. Int J Gynaecol Obstet
1995;50:80 – 4.
4. Armstrong BK, Brown JB, Clarke HT, Crooke DK, Hahnel R,
Masarei JR, et al. Diet and reproductive hormones: A study of
vegetarian and novegetarian postmenopausal women. J Natl Cancer Inst 1981;67:761–7.
5. Belanger A, Locong A, Noel C, Cusan L, Dupont A, Prevost J, et al.
Influence of diet on plasma steroid and sex plasma binding
globulin levels in adult men. J Steroid Biochem 1989;32:829 –33.
6. Adlercreutz H. Western diet and Western diseases: Some hormonal and biochemical mechanisms and associations. Scand J Clin
Lab Invest 1990;50:3–23.
7. Moore JW, Clark GM, Bulbrook RD, Hayward JL, Murai JT,
Hammond GL, et al. Serum concentrations of total and nonprotein-bound oestradiol in patients with breast cancer and in
normal controls. Int J Cancer 1982;29:17–21.
8. Barbosa JC, Shultz TD, Filley SJ, Nieman DC. The relationship
among adiposity, diet, and hormone concentrations in vegetarian
and nonvegetarian postmenopausal women. Am J Clin Nutr
1990;51:798 – 803.
9. Barr SI, Janelle KC, Prior JC. Vegetarian vs nonvegetarian diets,
dietary restraint, and subclinical ovulatory disturbances: Prospective 6-mo study. Am J Clin Nutr 1994;60:887–94.
10. Rose DP, Boyar AP, Cohen C, Strong LE. Effect of a low-fat diet on
hormone levels in women with cystic breast disease. I. Serum
steroids and gonadotropins. J Natl Cancer Inst 1987;78:623– 6.
11. Woods MN, Gorbach SL, Longcope C, Goldin BR, Dwyer JT,
Morrill-LaBrode A. Low-fat, high-fiber diet and serum estrone
sulfate in premenopausal women. Am J Clin Nutr 1989;49:1179 –
83.
12. Prentice R, Thompson D, Clifford C, Gorbach S, Goldin B, Byar D.
Dietary fat reduction and plasma estradiol concentration in
healthy postmenopausal women. J Natl Cancer Inst 1990;82:129 –
34.
13. Barnard ND. Food for life. New York: Harmony, 1993.
14. Cunningham SK, McKenna TJ. Evaluation of an immunoassay for
plasma sex-hormone binding globulin: Comparison with steroidbinding assay under physiological and pathological conditions.
Ann Clin Biochem 1988;25:360 – 6.
15. Thompson FE, Byers T. Dietary assessment resource manual. J
Nutr 1994;124 Suppl:2245S–317S.
16. Moos RH. Typology of menstrual cycle symptoms. Am J Obstet
Gynecol 1969;103:390 – 402.
Barnard et al
Diet and Dysmenorrhea
249
17. Penland JG, Johnson PE. Dietary calcium and manganese effects on
menstrual cycle symptoms. Am J Obstet Gynecol 1993;168:1417–23.
18. Cleeland CS, Ryan KM. Pain assessment: Global use of the Brief
Pain Inventory. Ann Acad Med 1994;23:129 –38.
19. Cassidy A, Bingham S, Setchel KDR. Biological effects of a diet of
soy protein rich in isoflavones on the menstrual cycle of premenopausal women. Am J Clin Nutr 1994;60:333– 40.
20. Lundström V, Green K. Endogenous levels of prostaglandin F2a
and its main metabolites in plasma and endometrium of normal
and dysmenorrheic women. Am J Obstet Gynecol 1978;130:640 –7.
21. Hansen SR, Deutch B. Dietary habits and menstrual distress.
Näringsforskning 1990;34:140 – 6.
22. Deutch B. Menstrual pain in Danish women correlated with low
n-3 polyunsaturated fatty acid intake. Eur J Clin Nutr 1995;49:508 –
16.
23. Jones DY. Influence of dietary fat on self-reported menstrual
symptoms. Physiol Behav 1987;40:483–7.
250 Barnard et al
Diet and Dysmenorrhea
Address reprint requests to:
Neal D. Barnard, MD
5100 Wisconsin Avenue, Suite 404
Washington, DC 20016
E-mail: nbarnard@pcrm.org
Received June 4, 1999.
Received in revised form July 23, 1999.
Accepted August 12, 1999.
Copyright © 2000 by The American College of Obstetricians and
Gynecologists. Published by Elsevier Science Inc.
Obstetrics & Gynecology