hypericum pms study

8/9/2019 Hypericum Pms study

1/20

The Efficacy ofHypericum perforatum(St Johns Wort) for the Treatment ofPremenstrual SyndromeA Randomized, Double-Blind, Placebo-Controlled Trial

Sarah Canning,1 Mitch Waterman,1 Nic Orsi,2 Julie Ayres,3 Nigel Simpson2 andLouise Dye1

1 Institute of Psychological Sciences, University of Leeds, Leeds, UK2 Perinatal Research Group, The YCR and Liz Dawn Pathology and Translational Sciences Centre,

Leeds Institute of Molecular Medicine, St Jamess University Hospital, Leeds, UK

3 Rosalind Bolton PMS Clinic, St Jamess University Hospital, Leeds, UK

Abstract Background:Premenstrual syndrome (PMS) is a common condition. Some ofthe most widely prescribed medications are selective serotonin reuptake in-

hibitors (SSRIs), based on the hypothesized role of serotonin in the produc-

tion of PMS symptoms. PMS sufferers, especially those experiencing mild to

moderate symptoms, are often reluctant to take this form of medication andinstead buy over-the-counter preparations to treat their symptoms, for which

the evidence base with regard to efficacy is limited. Hypericum perforatum

(St Johns wort) influences the serotonergic system. As such, this widely

available herbal remedy deserves attention as a PMS treatment.

Objective: To investigate the effectiveness of Hypericum perforatum on

symptoms of PMS.

Study design: This randomized, double-blind, placebo-controlled, crossover

study was conducted between November 2005 and June 2007.

Setting:Institute of Psychological Sciences, University of Leeds, Leeds, UK.

Participants: 36 women aged 1845 years with regular menstrual cycles(2535 days), who were prospectively diagnosed with mild PMS.

Intervention: Women who remained eligible after three screening cycles

(n= 36) underwent a two-cycle placebo run-in phase. They were then randomly

assigned to receive Hypericum perforatum tablets 900 mg/day (standardized

to 0.18% hypericin; 3.38% hyperforin) or identical placebo tablets for two

menstrual cycles. After a placebo-treated washout cycle, the women crossed

over to receive placebo or Hypericum perforatumfor two additional cycles.

Main outcome measures: Symptoms were rated daily throughout the trial

using the Daily Symptom Report. Secondary outcome measures were the

State Anxiety Inventory, Beck Depression Inventory, Aggression Question-naire and Barratt Impulsiveness Scale. Plasma hormone (follicle-stimulating

h [FSH] l t i i i h [LH] t di l t l ti

ORIGINAL RESEARCHARTICLE CNS Drugs 2010; 24 (3): 207-225

1172-7047/10/0003-0207/$49.95/0

2010 Adis Data Information BV. All rights reserved.


2/20

and testosterone) and cytokine (interleukin [IL]-1b, IL-6, IL-8, interferon

[IFN]-g and tumour necrosis factor [TNF]-a) levels were measured in the

follicular and luteal phases during Hypericum perforatum and placebo treatment.

Results: Hypericum perforatum was statistically superior to placebo in im-

proving physical and behavioural symptoms of PMS (p < 0.05). There were

no significant effects of Hypericum perforatum compared with placebo

treatment for mood- and pain-related PMS symptoms (p> 0.05). Plasma

hormone (FSH, LH, estradiol, progesterone, prolactin and testosterone) and

cytokine (IL-1b, IL-6, IL-8, IFNg and TNFa) levels, and weekly reports of

anxiety, depression, aggression and impulsivity, also did not differ sig-

nificantly during theHypericum perforatumand placebo cycles (p >0.05).

Conclusion: Daily treatment with Hypericum perforatum was more effectivethan placebo treatment for the most common physical and behavioural

symptoms associated with PMS. As proinflammatory cytokine levels did not

differ significantly between Hypericum perforatum and placebo treatment,

these beneficial effects are unlikely to be produced through this mechanism of

action alone. Further work is needed to determine whether pain- and mood-

related PMS symptoms benefit from longer treatment duration.

Trial registration number (International Standard Randomised Controlled

Trial Number Register) ISRCTN31487459

Background

The premenstrual syndrome (PMS) refers to awide range of cyclic and recurrent psychological,physical and behavioural symptoms that occur inthe 710 days prior to the onset of menstruation

and which are relieved shortly following the onsetof menstruation.[1-3] PMS severity falls along acontinuum.[4,5] Women with particularly severesymptoms may be diagnosed with premenstrual

dysphoric disorder (PMDD).[6] Epidemiologi-cal surveys have estimated that up to 75% ofwomen experience mild to moderate premenstrualsymptoms,[7,8] while approximately 38% of womenexperience symptoms at a severity sufficient to

meet a diagnosis for PMDD.[9,10] The term PMSis used in this article when discussion is relevant toall levels of premenstrual symptom severity, andthe term PMDD is used when discussion is re-

levant only to the severe end of the PMS spectrum.There is some dispute in the literature surround-

ing the nature of PMS.[8,11] From a diagnosticstandpoint, retrospective accounts and self-di i d t l t h ti t

reporting.[12-14] PMS diagnostic criteria should dif-ferentiate women who experience problematic pre-menstrual symptoms from women who do not.[13]

However, current diagnostic criteria for PMS donot consistently differentiate these two groups ofwomen.[13-16] The conceptualization of PMS isalso problematic and varies across disciplines.[15] For

example, the medical perspective pathologizeswomens distress and positions it as a reproductivedisorder, whereas the socio-cultural perspective re-

jects this standpoint.[14] These conceptualizationslead to quite different treatment approaches.

Various biomedical and psychological interven-tions are available for the treatment of PMS.While selective serotonin reuptake inhibitors(SSRIs) are the medical treatment of choice bymany practitioners for women who experiencesevere symptoms,[5,17,18] the use of over-the-counter(OTC) preparations is often advocated for womenwith mild to moderate symptoms.[5,19] Although

various dietary supplements and herbal remedies

have been proposed to alleviate PMS symptomsand are widely consumed, rigorous scientific

t di t t t th i ffi l ki [20 23] A

208 Canning et al.


3/20

variety of psychological interventions, which focuson changing the way in which women appraiseand cope with their premenstrual changes, arealso available.[8] Cognitive behavioural and nar-rative therapy,[24,25] coping skills training[26,27]

and relaxation training[28] have all been shown to

effectively reduce moderate to severe PMS. Thesedifferent approaches to PMS treatment reflectthe view that the syndrome could have a multi-factorial aetiology. A variety of biopsychosocialmodels of PMS have been proposed, which positthat symptoms arise from a complex interaction

between various biological, psychological andsocial factors.[29-33]

A range of biological factors have been im-plicated in the production of PMS symptoms.These include abnormal steroid hormone func-

tion,[34] disturbed prostaglandin metabolism,[35]

calcium dysregulation,[36,37] poor magnesiumstatus,[38] abnormal fatty acid metabolism,[39] andan increased sensitivity to[35] and elevated levelsof[40] prolactin. Increasing evidence suggests sero-tonergic dysfunction plays a role in the produc-

tion of PMS symptoms.[23,41,42]

Women with PMShave reduced serotonin levels during the lutealphase and in comparison with women without

PMS.[43-46] Moreover, PMS symptoms, includingaggression,[47,48] depression,[49] anxiety[50] andirritability,[51] have been linked to serotonergicdysfunction. Furthermore, SSRIs successfully re-

duce mood and somatic PMS symptoms.[17,18,23]

The serotonergic system interacts with the cyto-kine network.[52-58] Cytokine release varies overthe menstrual cycle.[59,60] Serum levels of inter-

leukin (IL)-1[61,62] and IL-6,[63] for example, in-crease during the luteal phase. These cytokineshave been associated with symptoms character-istic of PMS, including depression,[64,65] anxiety,[66]

fatigue,[67] headaches[68] and sleep disturbances.[69,70]

Endotoxin administration, which causes the pro-duction and secretion of tumour necrosis factor(TNF)-a, IL-1 and IL-6, has been shown to in-duce anxiety and depression.[71] It is therefore

possible that cytokines contribute to the produc-tion of PMS symptoms,[63] through interaction

with the serotonergic system, in a similar mannerto that suggested for the aetiology of depres-

i [52 72] SSRI li t ki l l [73 75]

and could therefore reduce PMS symptoms throughthis mechanism of action. However, no studieshave yet examined individual cytokine levels orcytokine interactions in women with PMS.

Hypericum perforatumis a herbal remedy thatinfluences the serotonergic system[76-78] and sup-

presses proinflammatory cytokine levels.[79,80]

Hypericum perforatumhas been demonstrated tobe an effective treatment for mild to moderatedepression,[81-84] presumably because of its actionon these systems.[76,79,80] As some similarities ex-ist between the symptoms of clinical depression

and mood-related PMS symptoms, such as feelingdown, hopeless, easily annoyed and fatigued,[85-87]

the hypothesis that the mood symptoms com-monly associated with PMS should respond toHypericum perforatumis plausible.

Three previous randomized, double-blind,placebo-controlled trials have assessed the effec-tiveness ofHypericum perforatumfor PMS.[86,88,89]

Two of these[88,89] found Hypericum perforatumto significantly benefit physical PMS symptoms.However, the omission of methodological details

from these articles meant that their quality couldnot be assessed. Hicks et al.[86] did not find

Hypericum perforatum (600 mg/day) to signifi-

cantly reduce PMS symptoms in comparisonwith placebo treatment. Combining symptomscores from both treatment cycles resulted in

Hypericum perforatum appearing more beneficial

than placebo treatment for each symptom sub-group assessed, but these combined effects did notreach statistical significance. Most research thathas assessed the efficacy ofHypericum perforatum

for depression has employed a higher dosage of900mg/day.[90-93] This is also the dosage recom-mended by most manufacturers ofHypericum per-

foratumto improve mood. Consequently, the aimof this study was to investigate whetherHypericum

perforatum900 mg/day was more beneficial thanplacebo treatment in relieving symptoms of PMS.

Method

Ethical and Regulatory ApprovalEthical approval to carry out the study was

t d b th I tit t f P h l i l S i

Hypericum perforatum for the Treatment of PMS 209


4/20

(16/12/2004) and the Leeds (West) ResearchEthics Committee (23/6/2005). The Medicinesand Healthcare products Regulatory Agency(MHRA) also granted approval for the clinicaltrial (23/9/2005).

Protocol, Design and Objectives

This was a ten-cycle, randomized, double-blind,crossover, placebo-controlled study, designed toevaluate the efficacy ofHypericum perforatumforthe management of mild PMS. After three un-treated screening cycles (cycles 13) women who

met a prospective PMS diagnosis underwent asingle-blind placebo run-in phase of two menstrualcycles (cycles 4 and 5). Thereafter, women wererandomized to receive two cycles of double-blindtreatment with Hypericum perforatum extract or

placebo (cycles 6 and 7). Following a single-blind,placebo-treated washout cycle (cycle 8), womenwere crossed over to placebo or Hypericum per-

foratumfor two additional cycles (cycles 9 and 10).

Participants

Women were recruited for the trial betweenNovember 2005 and August 2006. Respondents

to advertisements and publicity in the local media,which requested volunteers from the Leedsarea experiencing premenstrual symptoms, wereasked to complete a preliminary screening ques-

tionnaire. Women aged between 18 and 45 yearswho had regular menstrual cycles (2535 days)and who were able to make frequent study visitswere eligible for the study. Women who were

photosensitive, using hormonal contraception ortreatment, pregnant or breast-feeding, takingprescribed (e.g. estrogens, progestogens, danazol,

gonadotropin-releasing hormone agonists) orOTC (e.g. calcium, evening primrose oil, vitaminB6, Vitex agnus castus) medication for PMS, re-porting menstrual cycle irregularity, or takingprescribed drugs that are known to interact with

Hypericum perforatumwere excluded.

Women who met these selection criteria wereinvited for their first study visit at the Rosalind

Bolton PMS clinic, Leeds General Infirmary.During this visit the inclusion and exclusion cri-t i h k d d d t

clinical interview. Those who were deemed to bein good physical and psychological health wereasked to provide written informed consent andwere entered onto the screening phase. Duringthis phase women were required to demonstrateat least a 30% increase between their follicular

and luteal Daily Symptom Report (DSR)[94] totalscale scores (30% increase criterion) in at leasttwo of the three cycles to be eligible to continue inthe study.[95] Follicular scores were obtained byaveraging the DSR total scale scores from cycledays 510, with day 1 being the first day of

menstruation. Luteal scores were obtained byaveraging the DSR total scale scores from the6 days prior to the onset of menstruation.[96]

To confirm the diagnosis, PMS symptomsshould be limited to the luteal phase.[1,2,6,97,98]

Prospective symptom reports do not always con-firm PMS.[5,99,100] Some women meet the criteriafor a psychiatric diagnosis, most commonly an-xiety and depression,[1,99-101] but report PMS dueto premenstrual magnification of their symp-toms.[5] Women were excluded from this study if

they showed elevated anxiety and depressionlevels in their follicular phase. Women who hadan average total score that exceeded the clinical

threshold for caseness for depression (12)[102,103]

on the Beck Depression Inventory (BDI)[104]

and/or an average total score that exceeded theupper 90% probability limit (1.645 standard de-

viation [SD] above the mean)[105] for the generaladult population (51)[106] on the Trait scale ofthe State-Trait Anxiety Inventory (STAIT)[106]

were excluded from entering the placebo run-in

phase. A letter was sent to the family practi-tioners of all participants who remained eligibleand willing to continue in the trial.

Intervention

TheHypericum perforatumextract Li160 (80%methanolic dry extract, drug-to-extract ratio36 : 1, containing 0.18% hypericin and 3.38%hyperforin) was supplied by Lichtwer PharmaAG (Berlin, Germany). The coated active

(Hypericum perforatum 450 mg) and control(placebo) tablets were identical, yellow and oval in

d li d i id ti l bli t k

210 Canning et al.


5/20

of 15 tablets. The blister packs were packaged inplain boxes containing 75 tablets, which wereclearly labelled with each participants treatmentcode and study cycle number (cycles 410). Everyparticipant was given seven boxes of tabletsthroughout the entire study, supplied at regular

intervals. Participants were instructed to beginthe appropriate box of tablets on the first day ofeach menstrual cycle and to take two tablets perday from this box throughout that cycle.

The probability of assignment to treatmentgroups was equal. The randomization code was

generated using a random permuted block meth-od.[107] Two permutations were used: (i)Hypericumperforatum followed by placebo; and (ii) placebofollowed by Hypericum perforatum. The schedulewas produced using a random number table.[107] A

block size of 10 was employed, with the first per-mutation being represented by numbers 04 andthe second by numbers 59. For every ten partici-pants randomized, five received Hypericum perfor-atumfollowed by placebo and five received placebofollowed by Hypericum perforatum. A pharmacist

from Leeds General Infirmary, otherwise notinvolved in the trial, performed the randomiza-tion and dispensed the medication. The alloca-

tion codes were kept from the principal investigatorwho had contact with the participants until afterthe last participant completed the trial, when thesecodes were revealed. The randomization code was

broken for one participant during the trial, as shebecame pregnant. However, as random num-bers were used to assign women to conditions, thisdid not result in the unblinding of the remaining

participants.

Measurements

The primary outcome measure was the DSR, achecklist of 17 premenstrual symptoms ratedfrom 0 (not present at all) to 4 (severe: symptomis overwhelming and/or unable to carry out dailyactivity). The DSR comprises four subscales

which describe mood (anxiety, irritability, de-pression, nervous tension, mood swing, feeling

out of control), behavioural (poor coordination,insomnia, confusion, headache, crying, fatigue),

i ( h b t t d ) d h

sical (food craving, swelling) symptoms.[94]

Women were asked to complete DSRs throughoutthe duration of the trial and to note down onthese whether they had their period.

Anxiety, depression, aggression and impulsi-vity are common premenstrual symptoms, which

cause great distress to women with PMS.[1,108-110]

Moreover, these are the symptoms that oftenpropel women to seek treatment.[15,111] Hence, sec-ondary outcome measures included the State scaleof the State-Trait Anxiety Inventory (STAIS),[106]

BDI,[104] Aggression Questionnaire (BPAQ)[112] and

Barratt Impulsiveness Scale version 11 (BIS-11)[113]

to determine whether Hypericum perforatumspecifically benefited these common and dis-tressing PMS symptoms. Women were asked tocomplete these secondary measures weekly through-

out the trial. All self-report measures were pro-vided to participants in diary booklets containingseven copies of the DSR and one copy of theBDI, STAIS, BPAQ and BIS-11, which were sentto participants biweekly throughout the trial.The diary booklets supplied during the screen-

ing cycles also contained one copy of theSTAIT[106] for diagnostic purposes (as describedearlier). The DSRs provided between cycles 4

and 10 included a box for women to tick to con-firm that they had taken their study medica-tion that day. Women were instructed to begina new diary booklet on the first day of each

menstrual cycle, and to complete the weeklymeasures from their previous booklet if applic-able. They were asked to return each diarybooklet immediately after its completion in one

of the freepost envelopes provided.In order to assess the effectiveness ofHypericum

perforatumfor PMS, each cycle was divided intofour phases: menses, follicular, rest and luteal. Thesephases were determined based on the onset of men-

struation. DSR total and subscale scores fromcycle days 14 were averaged and identifiedmenses, days 510 the follicular phase and days

-1 to -6 the luteal phase. The average of the re-

maining cycle days were referred to as rest (ofcycle).[96,114,115] The secondary measures asked

women to report how they had been feeling overthe previous week. Hence those completed on

l d 7 t k t t th



6/20

on day 14 the follicular phase and those on cycleday 1 of the next cycle the luteal phase. If twodiaries had been completed during the restphase, the one immediately before the lutealdiary was excluded, as this usually overlappedwith days in the luteal diary. In the rare case

where three diaries were completed (i.e. in a cycle

>35 days) the middle diary was used.Women were required to visit the principal

investigator at the Institute of Psychological Sci-ences, University of Leeds, during the follicularand luteal phases of study cycles 3, 5, 7, 8 and 10.

During these visits, blood samples were obtainedby venepuncture to assess hormone (follicle-stimulating hormone [FSH], luteinizing hormone[LH], estradiol, progesterone, prolactin and tes-tosterone) and cytokine (IL-1b, IL-6, IL-8, in-

terferon [IFN]-gand TNFa) levels. Samples werecollected in edetic acid (EDTA) tubes and im-mediately centrifuged for 10 minutes at 4200 rpm.Plasma was removed and aliquoted into eightEppendorfs, which were stored at -80C in theInstitute of Psychological Sciences, until they

were transferred to Leeds Institute of MolecularMedicine, St Jamess University Hospital, wherethe samples were thawed and analysed. Hormone

levels were detected in vitro from each samplesimultaneously using the automated biochiparray analyser, Evidence (Randox Laboratories,Crumlin, UK) using the method previously

described by FitzGerald and colleagues.[116]

Cytokine levels were determined by fluid-phasemultiplex immunoassay using custom kits ac-cording to manufacturers instructions (Linco

Research Inc., St Charles, MO, USA). Sampleswere run on a Luminex-100 cytometer (Lumi-nex Corporation, Austin, TX, USA), equipped withBio-Plex software (Bio-Rad, Hemel Hempstead,Hertfordshire, UK).[117] All samples from each

individual were analysed together to minimizeinterassay variability.

Each week women were asked to provide in-formation regarding any adverse events they ex-

perienced and/or concomitant medications taken,on the relevant page in their diary booklet. Ad-

verse events were also assessed at each study visitthrough a semi-structured interview. Participantsbl d l d d t h i it

Sample Size

When this study was designed, only one pilotstudy testing the efficacy of Hypericum perfor-

atumfor PMS had been published.[118] This studyused a pre-post test design and did not employ a

placebo group or crossover design. Therefore, inorder to approximate the effect size necessary forthe purpose of sample size estimation in the cur-rent study, response to placebo had to be esti-mated. This was done by calculating the halfway

point between the total DSR score at baseline andthe first treatment cycle. The effect size was

estimated using the formula d= (treatment -placebo)/s within, where s within=S withinO(degress of freedom within/number of partici-pants).[119] Sample size was calculated using thetables provided by Machin and Campbell[120]

using the equation d=|m0-m1|/population SD.Hence it was estimated that 22 women would berequired to provide the study with 80%power todetect a treatment difference corresponding to aneffect size of 0.63 (using a two-sided test with

a= 0.05). A similar randomized, controlled trialthat assessed the efficacy of soy isoflavones forPMS screened 41 women which, following with-drawals, left 23 women completing the study.[114]

On this basis, it was estimated that at least 41women should enter the screening phase to en-

sure the required sample size of 22.

Statistical Analysis

All data analysis was carried out according to a

pre-established analysis plan. The primary analysisused to assess the effectiveness of Hypericum

perforatumfor PMS compared the DSR data col-

lected during Hypericum perforatum and placebotreatment, during four cycle phases (menses, folli-cular, rest and luteal) of two treatment cycles. DSRtotal scale scores were analysed using a univari-ate repeated measures analysis of variance (RMANOVA) model, while subscale scores were ana-

lysed using a doubly multivariate ANOVA model.Prespecified secondary analyses were conducted on

the BPAQ, BIS-11, STAIS and BDI scores to ex-amine whether a treatment effect operated on thePMS t i i l i it i t

212 Canning et al.


7/20

and depression. The STAIS and BDI scales eachcomprise one total scale score and therefore wereanalysed in separate univariate analyses. TheBPAQ comprises four subscales and the BIS-11three. Hence BPAQ and BIS-11 scores were ana-lysed in separate multivariate analyses. The nor-

mality of the distribution of the variables studiedwas assessed by consulting the Z scores for skew-ness and the Kolmogorov-Smirnov test for eachdependent variable (DSR, BPAQ, BIS-11, STAISand BDI total and subscale scores) at each time-point. Transformations were not necessary for the

STAIS, BDI, BPAQ and BIS-11 data. However,the majority of dependent variables constitutingthe DSR were positively skewed. Hence, the DSRtotal and subscale scores were transformedthrough a square root transformation. Analysis of

the DSR data was based on the transformedscores. However, the results interpreted and thevalues presented in tables and figures are based onuntransformed data.

Results

Three hundred and seventy-one women ex-pressed an interest in the study and completed the

preliminary screening questionnaire. Of these,91 women began screening and 36 were random-ized, of whom 34 completed the protocol. Figure 1shows reasons for nonrandomization, premature

termination and exclusion. Two women wereconsidered placebo responders and excludedfrom the analyses, as they no longer met the 30%increase criterion during each of the two placebo

run-in cycles. This resulted in a similar percentageof exclusions (6%) as has previously been re-ported.[87,121] All women who completed the trial

were of Northern European origin, with theexception of one British-born Asian and oneEastern European. Baseline demographic char-acteristics were examined using independentsamples t-tests and Pearsons chi-square tests,according to the order in which participants

received active and placebo treatment. No sig-nificant differences or associations with treat-

ment order were found between the treatmentgroups for any characteristic, including age, body

i d b f ith PMS lli

parity, employment, smoking and relationshipstatus (table I). Baseline symptom severity (table II)was also examined according to the order inwhich participants received active and placebotreatment, through performing a 4 (phase) 2(treatment order) doubly multivariate ANOVA

on the mean DSR subscale scores reported dur-ing the three screening cycles. A nonsignifi-cant main effect of treatment order (multivariateF [4, 27] = 1.10; p= 0.38) and a nonsignificantphase by treatment order interaction (multivariateF [12, 19]= 1.33; p= 0.28) were produced, indica-

ting that randomization to treatment order didnot result in women with different profiles atscreening receiving different treatment orders.

Total Daily Symptom Report Scores

The mean standard error of the mean (SE)DSR total scale scores reported during the fourphases of the two cycles ofHypericum perforatumand placebo treatment are shown in figure 2. A 2(treatment) 2 (cycle) 4 (phase) 2 (treatmentorder) RM ANOVA was conducted on these

scores. This revealed a nonsignificant main effectof treatment order (F [1, 30]= 0.96; p= 0.34).

There was a significant main effect of treatment(F [1, 30] = 4.82; p= 0.04; partial Z2= 0.14), suchthat women reported significantly fewer symp-toms duringHypericum perforatum(mean= 5.80)

than placebo (mean= 6.58) treatment. The treat-ment by cycle (F [1, 30]= 0.73; p= 0.40) andtreatment by phase (F [3, 90]= 1.16; p=0.33) inter-actions were nonsignificant.

Daily Symptom Report Subgroup Scores

The mean SE DSR subscale scores reported

during the four phases of the two cycles ofHypericum perforatumand placebo treatment areshown in figures 3ad. A 2 (treatment) 2 (cy-cle) 4 (phase) 2 (treatment order) doublymultivariate ANOVA was performed on thesescores. This revealed a nonsignificant main effect

of treatment order (multivariate F [4, 27] = 0.33;p= 0.86). There was a nonsignificant main effect

of treatment for the linear combination of theDSR subscales (multivariate F [4, 27] = 1.99;

0 12) d i ifi t t t t b l



8/20

(multivariate F [4, 27] = 1.27; p= 0.31) and treat-ment by phase (multivariate F [12, 19]= 0.66;p= 0.77) interactions. Given the limited previous

research that has assessed the efficacy ofHyper-icum perforatumfor PMS, it is reasonable to hy-pothesize that this herbal remedy may affectspecific components of PMS, masked by exami-

nation of only global scores. Hence, it was con-sidered appropriate to explore the univariate tests

of the main effect of treatment and of the treat-ment by cycle and treatment by phase inter-

ti f h b l (t bl III)

Significant univariate main effects of treatmentwere produced for the behavioural (p= 0.032)and physical (p = 0.013) DSR subscales, such that

women reported significantly fewer of thesesymptoms while receivingHypericum perforatumthan placebo treatment (figures 3b and 3d). Thetreatment by cycle and treatment by phase inter-

actions were nonsignificant for each of thesesubscales (p> 0.05). Nonsignificant univariate

main effects of treatment were produced for themood and pain DSR subscales (p>0.05). Thet t t b l d t t t b h i t

Responded to advertisementsand completed the online

questionnaire (n =371)

Three screening cycles(n =91)

Two placebo run-in cycles(n =38)

Women randomized (n =36)

Two cycles ofHypericum

perforatum(n =19)

Two cycles ofHypericum

perforatum(n =17)

Two cycles of

placebo(n =17)

Two cycles ofplacebo

(n =18)

Included inprimary analysis

(n =17)

Included inprimary analysis

(n =15)

Removed:defined as a

placebo responder(n =1)

Removed:defined as a

placebo responder(n =1)

Discontinuationdue to planning apregnancy (n =1)

Discontinuation (n =2) One due to indigestion One moved away

Withdrew (n =43)Criteria no longer met (n = 10): Follicular STAIS 51 (n =3) Follicular BDI 12 (n =2) Follicular STAIS 51 and follicular BDI 12 (n =3) 30% increase criterion not

met (n =2)

Removed fromtrial due to

pregnancy (n =1)

Failed initial screeningor unwilling to

continue (n =280)

One washoutcycle (n =18)

One washoutcycle (n =17)

Completed trial(n =18)

Completed trial(n =16)

Fig. 1. Participant flow through the trial.BDI=

Beck Depression Inventory;STAIS=

State-Trait Anxiety Inventory, State scale.

214 Canning et al.


9/20

actions were also nonsignificant for these sub-scales (p> 0.05).

Secondary Treatment Analyses

The mean (SD) STAIS, BDI, BPAQ and BIS-11 total and subscale scores reported during thefour phases of the two cycles of Hypericum per-

foratum and placebo treatment are shown intable IV. STAIS and BDI scores were examinedthrough 2 (treatment) 2 (cycle) 4 (phase) 2(treatment order) RM ANOVAs. BPAQ andBIS-11 scores were examined through 2 (treat-ment) 2 (cycle) 4 (phase) 2 (treatment order)

doubly multivariate ANOVAs. A nonsignificantmain effect of treatment, and nonsignificanttreatment by cycle and treatment by phase inter-actions, were produced from each of these ana-

lyses (all p > 0.05) [table V].

Biochemical Measures

The hormone (estradiol, progesterone, testo-sterone, LH, FSH and prolactin) and cytokine(IL-1b, IL-6, IL-8, IFNgand TNFa) levels mea-sured during the follicular and luteal phasesof Hypericum perforatum and placebo treatmentare shown in table VI. Paired samples t-tests

revealed that there were no significant differencesbetween the follicular (t [26]= 0.21; p = 0.84)

nor luteal (t [17] = 1.57; p= 0.14) cycle days onwhich blood was sampled during Hypericum

f d l b t t t Th f th

effect of treatment on these biochemical measureswas examined during the follicular and lutealphases through conducting 2 (treatment) 2(phase) RM ANOVAs for each hormone and

cytokine under analysis. A nonsignificant maineffect of treatment and a nonsignificant treat-ment by phase interaction was produced foreach hormone and cytokine that was analysed

(p> 0.05).

Adverse Events

Thirty-five adverse events were reported dur-ing the treatment cycles. Of these, 15 were re-

ported duringHypericum perforatumtherapy and20 during placebo. The difference in the fre-quency of each type of adverse event reportedduring the Hypericum perforatum and placebo

treatment cycles was nonsignificant (table VII).

Digestive and respiratory symptoms were the

Table I. Background characteristics of clinical trial completers

Characteristic Total (n= 32) Active first (n = 17) Placebo first (n = 15)

Age in years [mean (SD)] 35.3 (5.9) 33.6 (6.5) 37.2 (4.6)

No. of PMS years [mean (SD)] 12.7 (7.5) 13.4 (7.1) 11.8 (8.1)

Body mass index [mean (SD)] 25.1 (4.3) 25.7 (5.0) 24.5 (3.5)

Nulliparous (no.) 11 8 3

Employed (no.) 26 12 14

Smoker (no.) 2 1 1

Relationship status (no.)

single 5 4 1

partner 5 3 2

married/cohabiting 17 9 8

divorced/separated 5 1 4PMS= premenstrual syndrome.

Table II. Baseline symptom severity [mean (SD)], across the men-

strual cycle (Daily Symptom Report [DSR][70] total and subscale

scores) of clinical trial completers (n =32)

Measure Phase

menstrual follicular rest luteal

Total DSR score 11.4 (7.1) 3.4 (2.5) 4.3 (2.6) 12.6 (7.2)

Subscale scores

mood 4.5 (3.6) 1.6 (1.4) 2.0 (1.4) 5.4 (3.7)

behavioural 3.5 (2.6) 1.4 (1.0) 1.4 (0.9) 3.4 (2.3)pain 2.1 (1.1) 0.2 (0.2) 0.4 (0.3) 1.9 (1.3)

physical 1.3 (0.8) 0.2 (0.2) 0.5 (0.3) 1.8 (1.0)



10/20

most common adverse events, but these were al-most equally distributed between the Hypericum

perforatum and placebo treatment cycles. One

serious adverse event occurred duringHypericumperforatum treatment (chest pains) but this was

believed to be unrelated toHypericum perforatum,since a cause other than the administration ofHypericum perforatumwas evident.

Discussion

Principal Findings

The primary objective of this ten-cycle, double-blind, randomized, controlled trial was to de-

termine whetherHypericum perforatumwas morebeneficial than placebo treatment in relievingPMS symptoms in women experiencing mild

PMS. Hypericum perforatum was shown to ben-efit physical (food craving, swelling) and beha-vioural (poor coordination, insomnia, confusion,headaches, crying, fatigue) PMS symptoms incomparison with placebo treatment duringboth menstrual cycles in which it was taken.

These beneficial effects occurred with minimaladverse effects, which did not statistically differ

from those experienced during placebo treat-ment. Mood-related (anxiety, irritability, de-

i t i d i f li

out of control) and pain-related (aches, cramps,breast tenderness) PMS symptoms were not sig-nificantly relieved by Hypericum perforatum incomparison with placebo during two cycles oftreatment. However, pain-related symptoms ap-peared to improve with Hypericum perforatumin

comparison with placebo towards the end of thetreatment period (figure 3c). It is therefore pos-sible thatHypericum perforatummay benefit painsymptoms after longer treatment duration thanin the current study. Although the separationbetween the mood-related symptoms that were

reported during Hypericum perforatum and pla-cebo treatment was less than that for the othersymptom subgroups, a similar pattern emerged,with women reporting fewer of these symptomsduring Hypericum perforatum than placebo

treatment towards the end of the treatment peri-od (figure 3a). Future studies should determinewhether longer treatment duration could increasethe potential effect suggested. The secondaryanalyses revealed thatHypericum perforatumdidnot benefit the PMS symptoms of anxiety, de-

pression, aggression or impulsivity. However, asthese are all mood-related PMS symptoms,

longer treatment duration could be necessary forsuch benefits to become apparent.

Aetiology

Biopsychosocial models of PMS posit thatsymptoms arise from a complex interaction be-tween various biological, psychological andsocial factors.[29-33] This proposed multifactorial

aetiology is supported by the finding thatHypericum perforatumsignificantly reduced physi-cal and behavioural but not pain- or mood-related

PMS symptoms. Research has demonstratedHypericum perforatum to effectively reduce mildto moderate depression.[81-84] However, in thisstudyHypericum perforatumdid not significantlyimprove premenstrual depression or other mood-related PMS symptoms. This lack of effect of

Hypericum perforatumon mood symptoms raisesthe possibility that PMS-related mood symptoms

have a different aetiology from those seen in clini-cal depression. The physiological basis of depres-

i t i ll t bli h d [49 122 124] hil

2

Cycle 1 Cycle 2

M F R L M F R L

MeantotalDSR

score

3

4

5

6

7

8

9

10

1112

Hypericum perforatum

Placebo

Fig. 2. Mean total Daily Symptom Report (DSR) score is the sum ofseverity ratings for each of the 17 DSR symptoms within each phasedivided by the number of days in that phase.F = follicular;L = luteal;M =menses;R = rest.

216 Canning et al.


11/20

mood-related PMS symptoms may be psycho-social in origin. If this is the case, then this would

explain why Hypericum perforatum did not sig-

nificantly reduce PMS mood symptoms but iseffective in other types of depression.

Various biological factors have been proposedto contribute to PMS symptom production.[34-40]

Increasing evidence suggests that serotonin plays

an important role.[23,41,42]

SSRIs have been shownto effectively reduce psychological and physical

Table III. Multivariate and univariate statistics for the main effect of treatment, and for the treatment by cycle and treatment by phase

interactions for the treatment analysis of the Daily Symptom Report subscales

Measure Treatment Treatment by cycle Treatment by phase

statistic partialZ2 statistic partialZ2 statistic partialZ2

Multivariate F (4, 27)= 1.99 0.23 F (4, 27)= 1.27 0.16 F (12, 19) = 0.66 0.29

Univariate

mood F (1, 30)= 2.33 0.07 F (1, 30) = 1.17 0.04 F (3, 90) = 1.18 0.04

behavioural F (1, 30)= 5.07* 0.15 F (1, 30) = 0.75 0.02 F (3, 90) = 0.66 0.02

pain F (1, 30)= 0.20 0.01 F (1, 30) = 1.12 0.04 F (3, 90) = 1.16 0.04

physical F (1, 30)= 7.00* 0.19 F (1, 30) = 0.15 0.01 F (3, 90) = 0.90 0.03

* p


12/20

PMS symptoms within a few days.[5,18,23,101,125]

Hypericum perforatum increases brain serotonin

levels[76] and yields an upregulation of serotonin

5-HT1A and 5-HT2A receptors[78]

in the frontalcortex.[77] Hence, the rapid beneficial effects ofHypericum perforatum on the physical and be-havioural PMS symptoms observed here mayhave arisen from an increase in serotonin pro-

duction, biomarkers of which should be assessedin future evaluations ofHypericum perforatum.

Increased production of proinflammatory cyto-kines may also be involved in PMS symptomproduction. Cytokines, which have been associ-ated with symptoms characteristic of PMS,[64-70]

have been shown to increase during the lutealphase.[61-63] Hypericum perforatum suppressesproinflammatory cytokine production.[79,80] Intheory, it is plausible to suggest that Hypericum

perforatum may benefit PMS symptoms by re-

ducing proinflammatory cytokine levels duringthe luteal phase. The beneficial effects ofHyper-icum perforatum found in this study did notappear to operate through this mechanism ofaction, as Hypericum perforatum did not affectluteal cytokine levels in comparison with placebo

treatment. However, any demonstration of thismechanism must rely on peripheral measures,

hi h d t il fl t t l h

and cytokine action. Moreover, although thisresearch suggests that individual cytokines do not

play a major role in the beneficial effects of

Hypericum perforatum on physical and behav-ioural PMS symptoms, further analysis of cyto-kine networks might reveal relevant interactions.

The beneficial properties ofHypericum perfor-atum may arise from a combination of mecha-

nisms, each singularly too weak to result in atherapeutic effect.[90,126,127] Hypericum perfor-atumcould benefit some PMS symptoms throughproducing an effect on the interaction betweenthe cytokine and serotonergic systems, in a simi-lar manner to that suggested for the aetiology of

depression.[52,72] Therefore, the sole assessmentof the influence of Hypericum perforatum onproinflammatory cytokine levels may not be en-lightening. Effects may only become apparentwhen these levels are assessed in conjunction with

effects on other systems, such as the serotonergicsystem. To elucidate the mechanism of actionthrough which Hypericum perforatum producesits beneficial effects, further studies assessing theefficacy ofHypericum perforatum for PMS areneeded, while assessing its impact upon sero-

tonergic transmission in combination with pro-inflammatory cytokine production. Furthermore,f t t di ld tili lit ti

Table IV. Secondary outcome symptom scores across four phases of Hypericum perforatumand placebo treatment [mean (SD)]

Measure Hypericum perforatum Placebo

M F R L M F R L

STAIS 40.6 (12.0) 36.7 (10.6) 38.3 (10.4) 45.2 (11.8) 41.4 (12.1) 36.6 (9.7) 40.0 (9.7) 45.6 (12.4)

BDI 6.8 (7.1) 5.3 (5.5) 6.0 (5.5) 8.9 (5.5) 8.4 (7.7) 4.7 (7.7) 7.6 (6.4) 10.5 (8.6)

BPAQ

total 57.0 (21.2) 51.9 (17.2) 53.1 (16.1) 61.8 (19.3) 59.0 (19.8) 50.9 (19.8) 54.8 (15.2) 63.4 (21.5)

verbal 10.7 (4.9) 9.0 (3.4) 9.4 (3.2) 12.2 (4.7) 10.9 (4.9) 9.1 (4.0) 9.9 (3.8) 12.2 (5.0)

physical 15.0 (5.2) 14.4 (5.1) 14.3 (4.6) 15.2 (5.5) 15.1 (5.2) 13.4 (4.4) 14.7 (5.0) 15.9 (6.0)

anger 16.6 (6.7) 15.0 (5.5) 15.5 (4.9) 19.0 (6.2) 17.2 (6.2) 14.7 (5.3) 15.9 (4.3) 19.3 (6.4)

hostility 14.8 (6.3) 13.6 (5.4) 14.0 (5.9) 15.3 (6.3) 15.8 (6.8) 13.6 (5.5) 14.4 (5.9) 16.0 (7.6)

BIS-11

total 60.7 (10.0) 58.2 (9.1) 60.2 (8.6) 63.5 (7.7) 62.1 (10.4) 58.1 (8.6) 60.3 (9.7) 64.8 (9.7)attentional 13.9 (3.7) 13.1 (2.9) 13.5 (2.7) 14.7 (3.5) 14.5 (3.9) 12.7 (2.8) 13.4 (3.5) 14.9 (3.5)

motor 19.6 (2.3) 19.2 (2.6) 19.8 (2.7) 19.7 (2.2) 19.9 (2.9) 19.6 (2.4) 19.7 (2.9) 20.5 (3.4)

non-planning 27.2 (6.1) 25.9 (5.6) 27.0 (5.6) 29.1 (4.7) 27.7 (6.2) 25.8 (5.6) 27.2 (5.5) 29.5 (5.4)

BDI=Beck Depression Inventory[68]; BIS-11=Barratt Impulsiveness Scale version 11[72]; BPAQ =Aggression Questionnaire[71]; F = follicular;

L = luteal;M =menses;R = rest;STAIS=State-Trait Anxiety Inventory, State scale.[106]

218 Canning et al.


13/20

to assess womens experiences duringHypericumperforatum and placebo treatment, to furtherunderstanding of how psychological and socialfactors (e.g. life circumstances, relationships andcoping mechanisms) could contribute to PMSsymptoms and their response to treatment. These

measures could also provide insight into the lackof effect of Hypericum perforatum on mood-related PMS symptoms if these remained un-changed after a longer treatment duration.[8,14,32]

Previous Research

Few trials have assessed the effectiveness ofHypericum perforatum for PMS. When this pro-tocol was written, only one small pilot study hadbeen published.[118] Although these researchers

found Hypericum perforatum to significantly re-duce PMS symptoms, their study did not employa placebo group or crossover design. Therefore,their treatment effects cannot be separated from

those arising from the placebo response. Since thecommencement of this study, three randomized,

double-blind, placebo-controlled trials employ-ing a parallel design have been published.[86,88,89]

Hicks et al.[86] found that women treated withHypericum perforatumreported fewer symptomsthan those receiving placebo on all symptomsubgroups studied. However, these differencesdid not reach statistical significance. As a dosage

of 600mg/day was used, it is possible that ahigher dosage of 900 mg/day is needed for thebeneficial effects ofHypericum perforatumto be-

come apparent. The findings from the currentstudy appear compatible to those of Pakgoharet al.,[88,89] who also foundHypericum perforatumto reduce physical PMS symptoms.

Limitations

The diagnosis of PMS requires that symptoms

are confined to the luteal phase.[1,97] Manywomen who present with PMS are found not to havethe syndrome, but rather a premenstrual exacer-bation of a psychiatric disorder, most commonly

anxiety and/or depression.

[1,5,99,101]

Hence,women reporting clinical levels of anxiety and/ordepression during the follicular phase were ex-cluded from this trial. However, it is possible thatsome participants in this study did not have PMS,

but a psychiatric condition that was not directlyassessed, such as bulimia nervosa, substance abuseor seasonal affective disorder.[1,99,128,129] Althoughthis is unlikely, since a physician-led clinical

interview indicated that all participants werein good psychological health before they were

entered into the trial, future studies could for-mally verify this information through the ad-ministration of a structured clinical interviewsuch as the Structured Clinical Interview forDSM-IV.[130]

Trial participants were provided with diarybooklets containing seven DSRs on a weekly

basis. In order to minimize the possibility of retro-spective reporting and women inferring pat-terns in their symptom reporting, each DSR was

Table V. Multivariate and univariate statistics for the main effect of treatment and for the treatment by cycle and treatment by phaseinteractions for the secondary treatment analyses

Measure Treatment Treatment by cycle Treatment by phase

statistic partialZ2 statistic partialZ2 statistic partialZ2

BPAQ

total (UV) F (1, 30) = 0.45 0.01 F (1, 30) = 0.07


14/20

presented on a separate page. Furthermore,women were explicitly informed that they were tocomplete one DSR each evening, date it, and re-turn each booklet immediately after its comple-

tion. However, as each booklet contained sevenDSRs, it is possible that some women may not

have completed their daily ratings in a fully pro-spective manner. Moreover, some ratings mayhave been biased by women looking at previousDSR completions from that week. Although

some researchers ask women to return their rat-ings daily,[131] it was thought that applying thismethod of data collection to a study of this lengthwould have led to greater study burden, a higherattrition rate, and a larger amount of missing

data. Future studies could ask women to com-plete their ratings on the Internet at the end ofeach day, with no access to previous reports.

Implications for Clinicians

These results support the use ofHypericum per-foratum (900 mg/day) for women with mild PMSwho are experiencing symptoms predominantly

physical and behavioural in nature. In line withprevious studies that have assessed the effective-

ness of OTC preparations for PMS,[86,114,121,132-136]

this research utilized PMS diagnostic criteria thatid tifi d PMS ff i i t t

a mild severity. For example, the researchers whodeveloped the DSR used the 30% increase criterionto diagnose PMS in their treatment trial of thecarbohydrate-rich beverage Escape.[133] The women

identified as PMS sufferers in this research reportedpremenstrual symptoms at an almost identical

Table VI. Serum hormone and cytokine levels [mean (SD)] during the follicular and luteal phases of Hypericum perforatumand placebo

treatment (n = 12)

Hypericum perforatum Placebo

follicular luteal follicular luteal

Hormones

Estradiol (pmol/L) 231.8 (179.5) 443.6 (443.7) 311.0 (559.4) 213.6 (113.5)

Progesterone (nmol/L) 5.3 (5.7) 44.3 (50.9) 3.7 (5.2) 18.5 (7.6)

Testosterone (nmol/L) 2.2 (3.4) 3.4 (5.6) 1.1 (2.6) 0.7 (1.1)

LH (mIU/mL) 4.3 (2.5) 4.1 (5.1) 5.0 (3.4) 3.1 (2.3)

FSH (mIU/mL) 5.4 (2.5) 3.3 (3.1) 5.8 (2.2) 3.0 (1.5)

Prolactin (mIU/L) 296.9 (176.3) 321.0 (206.7) 329.7 (190.1) 373.2 (208.9)

Cytokines

IL-1b(pg/mL) 54.9 (80.8) 43.9 (59.5) 52.0 (67.9) 53.1 (70.7)

IL-6 (pg/mL) 393.4 (860.7) 368.4 (784.8) 332.6 (732.4) 373.5 (807.9)

IL-8 (pg/mL) 708.2 (1818.6) 638.8 (1637.8) 610.9 (1606.5) 609.0 (1642.3)

IFNg(pg/mL) 53.0 (46.2) 43.7 (50.9) 51.8 (39.5) 39.2 (31.4)

TNFa(pg/mL) 116.4 (199.6) 106.6 (191.9) 107.6 (181.8) 96.3 (167.7)

FSH= follicle-stimulating hormone;IFN = interferon;IL = interleukin;LH = luteinizing hormone;TNF = tumour necrosis factor.

Table VII. Adverse events reported as number per group during

Hypericum perforatumand placebo treatment

Adverse event Placebo

(n= 20)

Hypericum

perforatum

(n= 15)

Digestive (stomach cramps,

abdominal pain, nausea, diarrhoea,

dizziness)

3 4

Vasomotor (hot flushes, increased

sweating)

2 2

Respiratory (cold, sinus ache, sore

throat, swollen glands, viral infection,

laryngitis)

4 3

Other

headache/migraine 5 1

skin rash 2 0

spots 1 1

leg/back ache 2 0

vaginal discharge 0 1

menstrual flooding 1 1forgetfulness 0 1

chest pains 0 1

220 Canning et al.


15/20

severity (mean premenstrual total DSR score14.70, SD 8.36) to that previously reported byBritish women with mild PMS (mean 13.40, SD9.70),[114] and at a much lower severity than thatpreviously reported by women experiencing se-vere PMS (mean 23.00, SD 12.5),[137] and PMDD

(mean 30.17, SD 9.67) in US studies.[138] Thisapproach was considered appropriate in this re-search, as it is usually women with mild PMSsymptoms who buy,[133] and who are advised bytheir clinicians to take,[5,19] OTC preparations.PMS sufferers experiencing symptoms at a

greater severity are usually prescribed conven-tional treatment.[5,17,18] Many women with mild tomoderate premenstrual symptoms who believethey experience PMS do not meet strict PMS di-agnostic criteria (e.g. 75% increase criterion;

modified 30% increase criterion).[13,14] However,the 30% increase criterion has been shown toidentify the majority of women who perceivethemselves to experience PMS.[13] These womenwith mild PMS are most likely to buy OTC pre-parations to relieve symptoms they attribute to

their premenstrual phase. Stricter diagnostic cri-teria could have excluded these women, forwhom an evaluation of the efficacy ofHypericum

perforatum to relieve premenstrual symptoms ismost pertinent. The 30% increase criterion usedas the PMS diagnostic method in this researchdoes render the findings applicable only to PMS

sufferers experiencing PMS symptoms at a mildseverity. Future research is needed if researcherswish to generalize these findings to PMS suf-ferers experiencing PMS symptoms at a greater

severity.

Conclusions

Daily Hypericum perforatum treatment at adosage of 900mg/day significantly improvedphysical and behavioural premenstrual symp-toms in women experiencing mild PMS. Thesebeneficial effects became apparent during the first

menstrual cycle in which Hypericum perforatumwas taken, and occurred with minimal adverse

effects. The plasma hormone and cytokine anal-yses suggested that biochemical changes withinth t b bl t i l d i th

beneficial effects, although their involvement viainteraction with the serotonergic system is apossibly interesting avenue for future investiga-tion. Although Hypericum perforatum did notprove statistically superior to placebo for mood-and pain-related PMS symptoms, pain-related

symptoms appeared to abate by the luteal phaseof the second treatment cycle. Future research isneeded to assess the effectiveness of Hypericum

perforatum over longer treatment durations inorder to be able to recommend its use for suf-ferers of PMS who find these symptoms proble-

matic. Moreover, further studies incorporatingpsychosocial measures are required to determinewhether mood-related PMS symptoms wouldbenefit from longer treatment duration. Theconvincing results forHypericum perforatumthat

were found in this trial deserve independentconfirmation by other studies.

Acknowledgements

Thanks are due to Lichtwer Pharma AG (Berlin, Germany)for the donation of the supplements and the Rosalind BoltonBequest for funding the PhD studentship of the correspondingauthor. The Rosalind Bolton Bequest played no role in theproduction or submission of this manuscript. Thanks are alsodue to Biss Hartley for assisting Dr Julie Ayres with clinicalinterviews and to Uma Ekbote and Sarah Field for technicalassistance.

No other sources of funding were used to assist in thepreparation of this study. The authors have no conflicts ofinterest that are directly relevant to the content of this study.

Authors contributions:SC, LD, MW and NS designed thestudy and interpreted the data. SC collected the data. SC andLD analysed the data. SC wrote the paper. LD, MW, NS and

NO made suggestions for revisions. NO supervised the hor-mone and cytokine assays. JA undertook clinical interviewswith each participant. JA and NS were the physicians re-sponsible for the study.

References1. Freeman EW. Premenstrual syndrome and premenstrual

dysphoric disorder: definitions and diagnosis. Psycho-neuroendocrinology 2003 Aug; 28 Suppl. 3: 25-37

2. Halbreich U, Backstrom T, Eriksson E, et al. Clinicaldiagnostic criteria for premenstrual syndrome and guide-lines for their quantification for research studies. Gynecol

Endocrinol 2007 Mar; 23 (3): 123-303. Reed SC, Levin FR, Evans SM. Changes in mood, cogni-

tive performance and appetite in the late luteal and follic-



16/20

without PMDD (premenstrual dysphoric disorder). HormBehav 2008 Jun; 54 (1): 185-93

4. Borenstein JE, Dean BB, Yonkers KA, et al. Using thedaily record of severity of problems as a screening in-strument for premenstrual syndrome. Obstet Gynecol2007 May; 109 (5): 1068-75

5. Johnson SR. Premenstrual syndrome, premenstrual dys-phoric disorder, and beyond: a clinical primer for practi-tioners. Obstet Gynecol 2004 Oct; 104 (4): 845-59

6. Diagnostic and Statistical Manual of Mental Disorders.4th ed. Washington, DC: American Psychiatric Associa-tion; Department of Health and Human Services, 1994

7. Steiner M, Wilkins A. Diagnosis and assessment of pre-menstrual dysphoria. Psychiatr Ann 1996 Sep; 26 (9):571-5

8. Ussher JM. Processes of appraisal and coping in the de-velopment and maintenance of premenstrual dysphoricdisorder. J Community Appl Soc Psychol 2002 Sep-Oct;12 (5): 309-22

9. Steiner M, Born L. Advances in the diagnosis and treat-ment of premenstrual dysphoric disorder. CNS Drugs2000 Apr; 13 (4): 287-304

10. Steiner M, Pearlstein T. Premenstrual dysphoria and theserotonin system: pathophysiology and treatment. J ClinPsychiatry 2000; 61 Suppl. 12: 17-21

11. Ussher JM. The ongoing silencing of women in families: ananalysis and rethinking of premenstrual syndrome andtherapy. J Fam Ther 2003 Nov; 25 (4): 388-405

12. Ainscough CE. Premenstrual emotional changes: a pro-

spective study of symptomatology in normal women.J Psychosom Res 1990; 34 (1): 35-45

13. Gallant SJ, Popiel DA, Hoffman DM, et al. Using dailyratings to confirm premenstrual syndrome/late lutealphase dysphoric disorder: part II. What makes a realdifference? Psychosom Med 1992 Mar-Apr; 54 (2): 167-81

14. Ussher JM. The role of premenstrual dysphoric disorder inthe subjectification of women. J Med Humanit 2003 Jun;24 (1-2): 131-46

15. Bancroft J, Williamson L, Warner P, et al. Perimenstrualcomplaints in women complaining of PMS, menorrhagia,and dysmenorrhea: toward a dismantling of the pre-menstrual syndrome. Psychosom Med 1993 Mar-Apr; 55(2): 133-45

16. Morse CA, Dennerstein L, Varnavides K, et al. Menstrualcycle symptoms: comparison of a non-clinical samplewith a patient group. J Affect Disord 1988 Jan-Feb; 14 (1):41-50

17. Dimmock PW, Wyatt KM, Jones PW, et al. Efficacy ofselective serotonin-reuptake inhibitors in premenstrualsyndrome: a systematic review. Lancet 2000 Sep; 356(9236): 1131-6

18. Wyatt KM, Dimmock PW, OBrien PM. Selective sero-tonin reuptake inhibitors for premenstrual syndrome.Cochrane Database Syst Rev 2002; (4): CD001396

19. Connolly M. Premenstrual syndrome: an update on defi-nitions, diagnosis and management. Adv Psychiatr Treat

2001; 7 (6): 469-7720. Bendich A. The potential for dietary supplements to reduce

premenstrual syndrome (PMS) symptoms. J Am Coll

21. Canning S, Waterman M, Dye L. Dietary supplements

and herbal remedies for premenstrual syndrome (PMS):

a systematic research review of the evidence for their ef-ficacy. J Reprod Infant Psychol 2006 Nov; 24 (4): 363-78

22. Domoney CL, Vashisht A, Studd JW. Premenstrual syn-

drome and the use of alternative therapies. Ann N Y AcadSci 2003 Nov; 997: 330-40

23. Eriksson E, Endicott J, Andersch B. New perspectives on

the treatment of premenstrual syndrome and premenstrual

dysphoric disorder. Arch Womens Ment Health 2002;

4 (4): 111-9

24. Blake F, Salkovskis P, Gath D, et al. Cognitive therapy for

premenstrual syndrome: a controlled trial. J Psychosom

Res 1998 Oct; 45 (4): 307-18

25. Hunter M, Ussher J, Cariss M, et al. A randomised com-

parison of psychological (cognitive behaviour therapy),medical (fluoxetine) and combined treatment for women

with premenstrual dysphoric disorder. J Psychosom Ob-

stet Gynaecol 2002 Sep; 23: 193-9

26. Kirkby RJ. Changes in premenstrual symptoms and irra-

tional thinking following cognitive-behavioral coping

skills training. J Consult Clin Psychol 1994 Oct; 62 (5):

1026-32

27. Morse CA, Dennerstein L, Farrell E, et al. A comparison of

hormone therapy, coping skills training, and relaxation

for the relief of premenstrual syndrome. J Behav Med

1991 Oct; 14 (5): 469-89

28. Goodale IL, Domar AD, Benson H. Alleviation of pre-

menstrual syndrome symptoms with the relaxation re-

sponse. Obstet Gynecol 1990 Apr; 74 (4): 649-55

29. Anson O. Exploring the bio-psycho-social approach to

premenstrual experiences. Soc Sci Med 1999 Jul; 49 (1):

67-80

30. Keye Jr WR, Trunnell EP. A biopsychosocial model of

premenstrual syndrome. Int J Fertil 1986 Sep-Oct; 31 (4):

259-62

31. Rubinow DR, Schmidt PJ. Models for the development

and expression of symptoms in premenstrual syndrome.

Psychiatr Clin North Am 1989 Mar; 12 (1): 53-68

32. Ussher JM, Hunter M, Cariss M. A women-centred

psychological intervention for premenstrual symptoms,

drawing on cognitive-behavioural and narrative therapy.

Clin Psychol Psychother 2002; 9 (5): 319-3133. Walker AF. Theory and methodology in premenstrual

syndrome research. Soc Sci Med 1995 Sep; 41 (6): 793-800

34. Ba ckstro m T, Carstensen H. Estrogen and progesterone in

plasma in relation to premenstrual tension. J SteroidBiochem 1974 May; 5 (3): 257-60

35. Horrobin DF. The role of essential fatty acids and prosta-

glandins in the premenstrual syndrome. J Reprod Med

1983 Jul; 28 (7): 465-8

36. Shamberger RJ. Calcium, magnesium, and other elements

in the red blood cells and hair of normals and patients

with premenstrual syndrome. Biol Trace Elem Res 2003

Aug; 94 (2): 123-9

37. Thys-Jacobs S, Alvir MJ. Calcium-regulating hormonesacross the menstrual cycle: evidence of a secondary hyper-

parathyroidism in women with PMS. J Clin Endocrinol

222 Canning et al.


17/20

38. Abraham GE, Lubran MM. Serum and red cell magnesiumlevels in patients with premenstrual tension. Am J Clin

Nutr 1981 Nov; 34 (11): 2364-639. Brush MG, Watson SJ, Horrobin DF, et al. Abnormal

essential fatty acid levels in plasma of women with pre-menstrual syndrome. Am J Obstet Gynecol 1984 Oct; 150(4): 363-6

40. Carroll BJ, Steiner M. The psychobiology of premenstrualdysphoria: the role of prolactin. Psychoneuroendocrinol-ogy 1978 Apr; 3 (2): 171-80

41. Halbreich U. Premenstrual dysphoric disorders: a diversi-fied cluster of vulnerability traits to depression. ActaPsychiatr Scand 1997 Mar; 95 (3): 169-76

42. Rapkin AJ. The role of serotonin in premenstrual syn-drome. Clin Obstet Gynecol 1992 Sep; 35 (3): 629-36

43. Ashby CR, Carr LA, Cook CL, et al. Alteration of plateletserotonergic mechanisms and monoamine oxidase activ-ity on premenstrual syndrome. Biol Psychiatry 1988 Jun;24 (2): 225-33

44. Rapkin AJ, Edelmuth E, Chang LC, et al. Whole-bloodserotonin in premenstrual syndrome. Obstet Gynecol1987 Oct; 70 (4): 533-7

45. Rasgon N, McGuire M, Tanavoli S, et al. Neuroendocrineresponse to an intravenous L-tryptophan challenge inwomen with premenstrual syndrome. Fertil Steril 2000Jan; 73 (1): 144-9

46. Taylor DL, Mathew RJ, Ho BT, et al. Serotonin levels andplatelet uptake during premenstrual tension. Neuro-psychobiology 1984; 12 (1): 16-8

47. Cleare AJ, Bond AJ. The effects of tryptophan depletionand enhancement on subjective and behavioural aggres-sion in normal subjects. Psychopharmacology 1995 Mar;118 (1): 72-81

48. Ho HP, Olsson M, Westberg L, et al. The serotoninreuptake inhibitor fluoxetine reduces sex steroid-relatedaggression in female rats: an animal model of premenstrualirritability? Neuropsychopharmacology 2001 May; 24 (5):502-10

49. Hirschfeld RM. History and evolution of the monoaminehypothesis of depression. J Clin Psychiatry 2000; 61Suppl. 6: 4-6

50. Kahn RS, van Praag HM, Wetzler S, et al. Serotonin and

anxiety revisited. Biol Psychiatry 1988 Jan; 23 (2): 189-20851. Russo S, Kema IP, Haagsma EB, et al. Irritability rather

than depression during interferon treatment is linked toincreased tryptophan catabolism. Psychosom Med 2005Sep-Oct; 67 (5): 773-7

52. Bonaccorso S, Marino V, Puzella A, et al. Increased de-pressive ratings in patients with hepatitis C receivinginterferon-alpha-based immunotherapy are related tointerferon-alpha-induced changes in the serotonergicsystem. J Clin Psychopharmacol 2002 Feb; 22 (1): 86-90

53. Gemma C, Imeri L, De Simoni MG, et al. Interleukin-1induces changes in sleep, brain temperature, and sero-tonergic metabolism. Am J Physiol 1997 Feb; 272 (2 Pt 2):R601-6

54. Gemma C, Imeri L, Opp MR. Serotonergic activation sti-mulates the pituitary-adrenal axis and alters interleukin-1(IL-1) mRNA expression in rat brain. Psychoneuro-

55. Linthorst AC, Flachskamm C, Holsboer F, et al. Localadministration of recombinant human interleukin-1 beta

in the rat hippocampus increases serotonergic neuro-transmission, hypothalamic-pituitary-adrenocortical axisactivity, and body temperature. Endocrinology 1994 Aug;135 (2): 520-32

56. OBrien SM, Scott LV, Dinan TG. Cytokines: abnormal-ities in major depression and implications for pharmaco-logical treatment. Hum Psychopharmacol 2004 Aug; 19(6): 397-403

57. Pousset F, Fournier J, Legoux P, et al. Effect of serotoninon cytokine mRNA expression in rat hippocampalastrocytes. Brain Res Mol Brain Res 1996 May; 38 (1): 54-62

58. Silverman DH, Imam K, Karnovsky ML.Muramyl peptide/serotonin receptors in brain-derived preparations. PeptRes 1989 Sep-Oct; 2 (5): 338-44

59. Gorai I, Taguchi Y, Chaki O, et al. Serum soluble inter-leukin-6 receptor and biochemical markers of bone meta-bolism show significant variations during the menstrualcycle. J Clin Endocrinol Metab 1998 Feb; 83 (2): 326-32

60. Vrieze A, Postma DS, Kerstjens HA. Perimenstrual asth-ma: a syndrome without known cause or cure. J AllergyClin Immunol 2003 Aug; 112 (2): 271-82

61. Cannon JG, Dinarello CA. Increased plasma interleukin-1activity in women after ovulation. Science 1985 Mar; 227(4691): 1247-9

62. Polan ML, Kuo A, Loukides J, et al. Cultured human lutealperipheral monocytes secrete increased levels of inter-leukin-1. J Clin Endocrinol Metab 1990 Feb; 70 (2): 480-4

63. Konecna L, Yan MS, Miller LE, et al. Modulation of IL-6production during the menstrual cycle in vivo and in vitro.Brain Behav Immun 2000 Mar; 14 (1): 49-61

64. Maes M, Meltzer H, Bosmans E, et al. Increased plasmaconcentrations of interleukin-6, soluble interleukin-6 re-ceptor, soluble interleukin-2 receptor and transferrin re-ceptor in major depression. J Affect Disord 1995 Aug; 34(4): 301-9

65. Schlatter J, Ortuno F, Cervera-Enguix S. Differences ininterleukins patterns between dysthymia and major de-pression. Eur Psychiatry 2001 Aug; 16 (5): 317-9

66. Connor TJ, Song C, Leonard BE, et al. An assessment ofthe effects of central interleukin-1beta, -2, -6, and tumornecrosis factor-alpha administration on some behaviour-

al, neurochemical, endocrine and immune parameters inthe rat. Neuroscience 1998 Jun; 84 (3): 923-33

67. Kronfol Z, Remick DG. Cytokines and the brain: im-plications for clinical psychiatry. Am J Psychiatry 2000May; 157 (5): 683-94

68. Martelletti P, Granata M, Giacovazzo M. Serum inter-leukin-1 beta is increased in cluster headache. Cephalalgia1993 Oct; 13 (5): 343-5

69. Kapas L, Krueger JM. Tumor necrosis factor-beta inducessleep, fever, and anorexia. Am J Physiol Regul IntegrComp Physiol 1992 Sep; 263 (3 Pt 2): 703-7

70. Vgontzas AN, Bixler EO, Lin HM, et al. IL-6 and its cir-cadian secretion in humans. Neuroimmunomodulation

2005 May; 12 (3): 131-4071. Reichenberg A, Yirmiya R, Schuld A, et al. Cytokine-

associated emotional and cognitive disturbances in humans.



18/20

72. Menkes DB, MacDonald JA. Interferons, serotonin andneurotoxicity. Psychol Med 2000 Mar; 30 (2): 259-68

73. Maes M. The immunoregulatory effects of antidepressants.Hum Psychopharmacol 2001 Jan; 16 (1): 95-103

74. Sluzewska A, Rybakowski JK, Laciak M, et al. Inter-leukin-6 serum levels in depressed patients before andafter treatment with fluoxetine. Ann N Y Acad Sci 1995Jul; 762: 474-6

75. Van Gool AR, Fekkes D, Kruit WH, et al. Serum aminoacids, biopterin and neopterin during long-term im-munotherapy with interferon-alpha in high-risk melano-ma patients. Psychiatry Res 2003 Jul; 119 (1-2): 125-32

76. Calapai G, Crupi A, Firenzuoli F, et al. Serotonin,norepinephrine and dopamine involvement in the anti-depressant action of hypericum perforatum. Pharma-copsychiatry 2001 Mar; 34 (2): 45-9

77. Muller WE, Singer A, Wonnemann M, et al. Hyperforinrepresents the neurotransmitter reuptake inhibiting con-stituent of hypericum extract. Pharmacopsychiatry 1998Jun; 31 Suppl. 1: 16-21

78. Teufel-Mayer R, Gleitz J. Effects of long-term adminis-tration of hypericum extracts on the affinity and densityof the central serotonergic 5-HT1 A and 5-HT2 A re-ceptors. Pharmacopsychiatry 1997 Sep; 30 Suppl. 2: 113-6

79. Fiebich BL, Hollig A, Lieb K. Inhibition of substanceP-induced cytokine synthesis by St. Johns wort extracts.Pharmacopsychiatry 2001 Jul; 34 Suppl. 1: S26-8

80. Thiele B, Brink I, Ploch M. Modulation of cytokine ex-pression by hypericum extract. J Geriatr Psychiatry

Neurol 1994 Oct; 7 Suppl. 1: S60-281. Clement K, Covertson CR, Johnson MJ, et al. St Johns

wort and the treatment of mild to moderate depression: asystematic review. Holist Nurs Pract 2006 Jul-Aug; 20 (4):197-203

82. Gaster B, Holroyd J. St Johns wort for depression: a sys-tematic review. Arch Intern Med 2000 Jan; 160 (2): 152-8

83. Kim HL, Streltzer J, Goebert D. St Johns wort fordepression: a meta-analysis of well-defined clinical trials.J Nerv Ment Dis 1999 Sep; 187 (9): 532-8

84. Williams Jr JW, Mulrow CD, Chiquette E, et al. A sys-tematic review of newer pharmacotherapies for depres-sion in adults: evidence report summary. Ann Intern Med

2000 May; 132 (9): 743-5685. Agha-Hosseini M, Kashani L, Aleyaseen A, et al. Crocus

sativus L. (saffron) in the treatment of premenstrualsyndrome: a double-blind, randomised and placebo-controlled trial. BJOG 2008 Mar; 115 (4): 515-9

86. Hicks SM, Walker AF, Gallagher J, et al. The significanceof nonsignificance in randomized controlled studies: adiscussion inspired by a double-blinded study on St.Johns wort (Hypericum perforatum L.) for premenstrualsymptoms. J Altern Complement Med 2004 Dec; 10 (6):925-32

87. Steiner M, Steinberg S, Stewart D, et al. Fluoxetine in thetreatment of premenstrual dysphoria. N Engl J Med 1995Jun; 332 (23): 1529-34

88. Pakgohar M, Mehran A, Salehi MS, et al. Comparison ofhypericum perforatum and placebo in treatment of phy-sical symptoms of premenstrual syndrome [in Persian].

89. Pakgohar M, Ahmadi M, Surmaghi MH, et al. Effect ofHypericum perforatum L. for treatment of premenstrual

syndrome [in Persian]. J Med Plants 2005 Sep; 4 (15):33-42

90. Bennett Jr DA, Phun L, Polk JF, et al. Neuropharmacol-ogy of St. Johns wort (Hypericum). Ann Pharmacother1998 Nov; 32 (11): 1201-8

91. Miller AL. St. Johns wort (Hypericum perforatum): clinicaleffects on depression and other conditions. Altern MedRev 1998 Feb; 3 (1): 18-26

92. Montgomery S, Hubener W, Gregoleit H. Efficacy andtolerability of St Johns wort extract compared withplacebo in patients with a mild to moderate depressivedisorder. Phytomedicine 2000; 7 Suppl. 2: 107

93. Shelton RC, Keller MB, Gelenberg A, et al. Effectivenessof St Johns wort in major depression: a randomizedcontrolled trial. JAMA 2001 Apr; 285 (15): 1978-86

94. Freeman EW, DeRubeis RJ, Rickels K. Reliability andvalidity of a daily diary for premenstrual syndrome.Psychiatry Res 1996 Nov; 65 (2): 97-106

95. Hamilton JA, Parry BL, Alagna S, et al. Premenstrualmood changes: a guide to evaluation and treatment.Psychiatr Ann 1984 Jun; 14 (6): 426-35

96. NIMH Workshop on Pre-menstrual Syndrome. Co-spon-sored by the Centre for Studies of Affective Disorders andthe Psychobiological Processes and Behavioral MedicineSection. Rockville (MD): 1983 Apr: 14-15

97. Frye GM, Silverman SD. Is it premenstrual syndrome?Keys to focused diagnosis, therapies for multiple symp-

toms. Postgrad Med 2000 May; 107 (5): 151-998. Lampe L. Antidepressants: not just for depression. Aust

Prescr 2005 Aug; 28 (4): 91-3

99. Pearlstein TB. Hormones and depression: what are thefacts about premenstrual syndrome, menopause, andhormone replacement therapy? Am J Obstet Gynecol1995 Aug; 173 (2): 646-53

100. Wyatt KM, Dimmock PW, Frischer M, et al. Prescribingpatterns in premenstrual syndrome. BMC WomensHealth 2002 Jun; 2 (4): 1-8

101. Landen M, Eriksson E. How does premenstrual dysphoricdisorder relate to depression and anxiety disorders? De-press Anxiety 2003; 17 (3): 122-9

102. Quah-Smith JI, Tang WM, Russell J. Laser acupuncturefor mild to moderate depression in a primary care setting:a randomised controlled trial. Acupunct Med 2005 Sep;23 (3): 103-11

103. Nielson A, Williams T. Depression in ambulatory medicalpatients, prevalence by self-report questionnaire andrecognition by non-psychiatric physicians. Arch GenPsychiatry 1980; 37: 999-1004

104. Beck AT, Ward CH, Mendelson M, et al. An inventory formeasuring depression. Arch Gen Psychiatry 1961 Jun;4: 561-71

105. Sheenan-Dare RA, Henderson MJ, Cotterill JA. Anxietyand depression in patients with chronic urticaria andgeneralised pruritus. Br J Dermatol 1990; 123 (6): 769-74

106. Spielberger CD. Manual for the State-Trait Anxiety In-

ventory: STAI (Form Y). Palo Alto (CA): ConsultingPsychologists Press, 1983

107. Pocock SJ. Clinical trials: a practical approach. Chichester:

224 Canning et al.


19/20

108. Chisholm G, Jung SO, Cumming CE, et al. Premenstrual

anxiety and depression: comparison of objective psycho-

logical tests with retrospective questionnaire. Acta Psy-chiatr Scand 1990 Jan; 81 (1): 52-7

109. Halbreich U. The etiology, biology, and evolving patho-logy of premenstrual syndromes. Psychoneuroendocrino-

logy 2003 Aug; 28 Suppl. 3: 55-99

110. Warner P, Bancroft J. Factors related to self-reporting ofthe pre-menstrual syndrome. Br J Psychiatry 1990 Aug;

157: 249-60

111. Hartlage SA, Arduino KE. Toward the content validity ofpremenstrual dysphoric disorder: do anger and irritability

more than depressed mood represent treatment-seekers

experience? Psychol Rep 2002 Feb; 90 (1): 189-202

112. Buss AH, Perry M. The Aggression Questionnaire. J Pers

Soc Psychol 1992 Sep; 63 (3): 452-9113. Patton JH, Stanford MS, Barratt ES. Factor structure of

the Barratt impulsiveness scale. J Clin Psychol 1995 Nov;51 (6): 768-74

114. Bryant M, Cassidy A, Hill C, et al. Effect of consumption

of soy isoflavones on behavioural, somatic and affective

symptoms in women with premenstrual syndrome. Br JNutr 2005 May; 93 (5): 731-9

115. Freeman EW, Rickels K, Sondheimer SJ, et al. Differential

response to antidepressants in women with premenstrual

syndrome/premenstrual dysphoric disorder: a randomi-zed controlled trial. Arch Gen Psychiatry 1999 Oct; 56 (10):

932-9

116. FitzGerald SP, Lamont JV, McConnell RI, et al. Develop-

ment of a high-throughput automated analyzer using

biochip array technology. Clin Chem 2005 Jul; 51 (7):

1165-76

117. Hildesheim A, Ryan RL, Rinehart E, et al. Simultaneous

measurement of several cytokines using small volumes ofbiospecimens. Cancer Epidemiol Biomarkers Prev 2002

Nov; 11 (11): 1477-84

118. Stevinson C, Ernst E. A pilot study of Hypericum perfor-

atum for the treatment of premenstrual syndrome. BJOG2000 Jul; 107 (7): 870-6

119. Rosenthal R, Rosnow RL. Essentials of behavioral re-

search: methods and data analysis. 2nd ed. Singapore:

McGraw-Hill, 1999

120. Machin D, Campbell MJ. Statistical tables for the design ofclinical trials. Oxford: Blackwell Scientific Publications,

1987

121. Pearlstein TB, Stone AB, Lund SA, et al. Comparison offluoxetine, bupropion, and placebo in the treatment of

premenstrual dysphoric disorder. J Clin Psychopharma-

col 1997 Aug; 17 (4): 261-6

122. Carlson NR. Foundations of Physiological Psychology.

Boston (MA): Allyn and Bacon, 1988

123. Dantzer R, Wollman E, Vitkovic L, et al. Cytokines anddepression: fortuitous or causative association? Mol Psy-

chiatry 1999 Jul; 4 (4): 328-32

124. Smith RS. The macrophage theory of depression. MedHypotheses 1991 Aug; 35 (4): 298-306

125. Yonkers KA, OBrien PMS, Eriksson E. Premenstrualsyndrome. Lancet 2008 Apr; 371 (9619): 1200-10

126. Nangia M, Syed W, Doraiswamy PM. Efficacy and safetyof St. Johns wort for the treatment of major depression.Public Health Nutr 2000; 3 (4A): 487-94

127. Raffa RB. Screen of receptor and uptake-site activity of hy-pericin component of St. Johns wort reveals sigma receptorbinding. Life Sci 1998 Mar; 62 (16): PL265-70

128. Lester NA, Keel PA, Lipson SF. Symptom fluctuation inbulimia nervosa: relation to menstrual-cycle phase andcortisol levels. Psychol Med 2003; 33 (1): 51-60

129. Price WA, Torem MS, DiMarzio LR. Premenstrual ex-acerbation of bulimia. Psychosomatics 1987; 28: 378-9

130. First MB, Spitzer RL, Gibbon M, et al. Structured ClinicalInterview for DSM-IV Axis I DisordersPatient Edition(SCID-I/P, version 2.0). New York: Biometrics ResearchDepartment, New York State Psychiatric Institute, 1995

131. Evans SM, Foltin RW, Fischman MW. Food cravingsand the acute effects of alprazolam on food intake inwomen with premenstrual dysphoric disorder. Appetite1999 Jun; 32 (3): 331-49

132. Facchinetti F, Borella P, Sances G, et al. Oral magnesiumsuccessfully relieves premenstrual mood changes. ObstetGynecol 1991 Aug; 78 (2): 177-81

133. Freeman EW, Stout AL, Endicott J, et al. Treatment ofpremenstrual syndrome with a carbohydrate-rich bev-erage. Int J Gynecol Obstet 2002; 77 (3): 253-4

134. De Souza MC, Walker AF, Robinson PA, et al. A syner-gistic effect of a daily supplementation for 1 month of200 mg magnesium plus 50 mg vitamin B6 for the relief ofanxiety-related premenstrual symptoms: a randomized,double-blind, crossover study. J Womens Health GendBased Med 2000 Mar; 9 (2): 131-9

135. Sayegh R, Schiff I, Wurtman J, et al. The effect of acarbohydrate-rich beverage on mood, appetite, and cog-nitive function in women with premenstrual syndrome.Obstet Gynecol 1995 Oct; 86 (4 Pt 1): 520-8

136. Walker AF, De Souza MC, Vickers MF, et al. Magnesiumsupplementation alleviates premenstrual symptoms offluid retention. J Womens Health 1998 Nov; 7 (9): 1157-65

137. Freeman EW, Rickels K, Sondheimer SJ, et al. A double-blind trial of oral progesterone, alprazolam, and placeboin treatment of severe premenstrual syndrome. JAMA1995 Jul; 274 (1): 51-7

138. Freeman EW, Sondheimer SJ, Sammel MD, et al. A pre-liminary study of luteal phase versus symptom-onsetdosing with escitalopram for premenstrual dysphoricdisorder. J Clin Psychiatry 2005 Jun; 66 (6): 769-73

Correspondence: Dr Sarah Canning, Institute of Psycholo-gical Sciences, University of Leeds, Leeds, LS2 9JT, UK.E-mail: [email protected]



20/20

Copyright of CNS Drugs is the property of ADIS International Limited and its content may not be copied or

emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.

However, users may print, download, or email articles for individual use.

hypericum pms study

Documents