THE JOURNAL OF ALTERNATIVE AND COMPLEMENTARY MEDICINE

Volume 11, Number 4, 2005, pp. 681-687

© Mary Ann Liebert, Inc.

A Randomized, Double-Blinded, Placebo-Controlled Pilot Study

to Investigate the Effectiveness of a Static Magnet

to Relieve Dysmenorrhea

NYJON K. ECCLES, MRCP, Ph.D.

ABSTRACT

Objectives: The aim of this study was to investigate the hypothesis that a specially designed, static magnet

of 2700 gauss, attached over the pelvic area, could relieve menstrual pain.

Design: This was a randomized, double-blind, placebo-controlled, postal questionnaire study.

Setting: The study was conducted in a primary care, single center.

Participants: Sixty-five (65) women (mean age 29.1 ± 1.52 years) were recruited from an advertisement in

a London newspaper. The entry criterion was regular dysmenorrhea. The exclusion criterion was known sec

ondary dysmenorrhea. Of the 65 women who were enrolled, 35 completed the study.

Interventions: A questionnaire-based assessment was completed by each subject and checked by telephone

before and after random allocation to use of either the static magnet device (2700 gauss) or an identical, weaker

magnetic placebo device (140 gauss). Assessment was made by telephone before and after a complete men

strual cycle. None of the participants was examined or seen face-to-face.

Main outcome measures: The main outcome measures were level of pain, using the McGill Pain and Vi

sual Analogue Scales, and ratings of associated symptoms such as irritability, restriction of usual activities, and

painkiller consumption.

Results: There was a significant reduction (p < 0.02) in pain in the magnet group compared to the placebo

group. Pain score differences (McGill pain score before - pain score after use of device) were —17 (—53, 13)

(median and interquartile ranges) in the magnet group and -5.0 (-29, 27) in the placebo group. The 95%

Mann-Whitney confidence intervals for the median difference between the magnet and placebo groups (mag

net — placebo) were —53.0 to 23.38. A reduction in irritability symptoms in the magnet group approached sta

tistical significance (p = 0.056).

Conclusions: Despite the small number of participants, the level of significance reached in the reduction of

pain merits reporting. This is a pilot study to a much larger study of the same device as an analgesic in women

with primary dysmenorrhea.

INTRODUCTION itated for 1-3 days each month and may result in absence

from school or work.3 Menstrual pain is the most common

On a worldwide basis, dysmenorrhea affects 40%-70% reason the women miss work. Dysmenorrhea is a leading

of women of reproductive age and their daily activi- cause of absenteeism for women younger than 30 years.4

ties.1 Some investigators have estimated prevalence rates to A recent critical review of randomized control trials of

be as high as 90%.2 In the United States approximately 40% static magnets for pain relief revealed that 73.3% of 21 stud-

of adult women have menstrual pain, and 10% are incapac- ies reported statistically significant analgesic effect5 across

The Chiron Clinic, London, England.

681

682 ECCLES

a broad range of different types of pain (neuropathic, in

flammatory, musculoskeletal, fibromyalgic, rheumatic, and

post surgical).

Four (4)-week application of 500-gauss unidirectional

static magnets to trigger points for 24 hours per day was

studied in 14 women with chronic pelvic pain of nonspeci-

fied duration. This was conducted as a 2-week double-blind

study with a 2-week single-blind extension. Pain was as

sessed by McGill pain inventory and the pain disability in

dex. The study demonstrated a 50% reduction in the level

of pain in 60% of subjects after 4 weeks compared to a 33%

reduction in the level of pain after 2 weeks.6 There are no

other double-blind studies that have looked specifically at

static magnets in dysmenorrhea.

The current study was prompted on the basis of increas

ing anecdotal evidence for the efficacy of static magnets in

relieving dysmenorrhea both from the author's clinical ex

perience and from a telephone survey of 193 women with

primary dysmenorrhea7 who had owned or purchased the

device from the manufacturer. The aim of this study was to

examine this apparent efficacy to relieve dysmenorrhea in a

controlled manner in volunteers who had regular monthly

dysmenorrhea.

MATERIALS AND METHODS

Seventy (70) women responded to an advertisement in

the newspaper to take part in this double-blind controlled

study of menstrual pain. These respondents were screened

by telephone and excluded if they did not experience pain

every menstrual cycle or if they had been diagnosed with

secondary dysmenorrhea (endometriosis, fibroids, or pelvic

inflammatory disease). The remaining 65 women were sent

detailed information by mail about the study, a consent form

confirming their understanding of the nature of the study

and their agreement to take part, and a home McGill ques

tionnaire with instructions on how to fill this in. They were

told over the phone that they would receive a coded device

and that neither they nor the assessor would know whether

it was a live device or a placebo. Furthermore they were in

formed verbally and in the literature sent to them that this

was a study involving a metallic device and that they should

continue their usual treatment for menstrual pain if neces

sary. Magnetic devices (LadyCare,® Magnopulse, Bristol,

UK) were neodymium magnets of 2700 gauss (surface mea

surement made by the manufacturer at the body surface of

the magnet using a gauss meter) with patented directional

plate designed to allow high gauss rating in proportion to

small size (Fig. 1), whereas placebo devices were low-power

magnets of approximately 140 gauss. The two devices were

identical in appearance and designed to be self-secured to

the underwear by magnetic force between two parts of the

device applied on either side of the underwear material an

terior to the pubis. The magnetic force of the placebo de

vice was powerful enough to hold the device in place se

curely and without comparison with a live device so that

subjects would not know that this was a weaker device. None

of the subjects enrolled in the study knew each other.

Both placebo and live devices were applied in an identi

cal manner. After returning a completed consent form, each

woman received either a live or placebo device that had been

selected randomly by computer (randomization generated by

the supplier of the devices) from a batch of mixed live and

placebo devices, each bearing a 3-digit code. Sequences and

codes were concealed by the manufacturer until after the

study was completed and only revealed after the all mea

surements of pain and associated symptoms had been made.

Women were asked to apply the device 2 days before their

anticipated start of next menses and to leave the device in

place except during bathing until the end of menses. After

4-5 weeks the volunteers were contacted by telephone and

their detailed inventory, including a McGill pain question

naire, was confirmed with them over the telephone. This was

done to optimize data gathering and the researcher was

blinded to the type of device that the subject had used. Sub

jects were asked to rate their experience of pain throughout

their menstrual periods comparing the experience after wear

ing the device to their usual experience. They were then

asked to send in their home McGill questionnaires as part

of their assessment.

Statistical methods

Results were analyzed by an independent university-

based statistician. To compare pain score differences and all

other outcomes between the magnet and placebo groups, the

Mann-Whitney test (nonparametric test) was used, as the

distribution of the pain scores differences and other data in

both groups did not appear to be symmetric (that is, nor

mally distributed). For all hypothesis tests a 5% significance

level (p < 0.05) and two-tailed tests were used. Ninety-five

percent (95%) Mann-Whitney confidence intervals (CI) for

the median differences between the magnet and placebo

groups were determined. Where improvement versus no im

provement was compared the x2 test was used.

This analysis was on an intention-to-treat basis insofar as

data for the participants were analyzed within the group to

which the patients were allocated. However dropout and

missing data were not explored and this analysis was of those

individuals who completed the study.

This was unlikely to cause bias unless the reason for data

being missing was different in the two arms of the study.

RESULTS

The advertisement was placed in July 2002 and 35 women

had completed the study by April 2003. Because of poor fol

low-through with initially expressed interest the advertise-

STATIC MAGNET FOR DYSMENORRHEA 683

ment was placed in February 2003 to recruit more volunteers

for the study. Eighteen (18) women had received placebo de

vices and 17 had received the test magnetic device. Thirty

(30) women declined for the reasons shown in Table 1.

Both groups were well matched for age; mean age was

29.1 ± 1.52 years in the placebo group and 30.4 ± 1.37

years in the magnet group (p = 0.531).

The usual level of menstrual pain experienced was rated

similarly in both groups (placebo group 7.6, magnet group

7.2, p = 0.468). Figure 2 illustrates the rating of menstrual

pain in relation to the worst of other types of pain ever ex

perienced by the women. The majority of women in both

groups (70% of the placebo group, 53% of the magnet

group) described an onset of pain since the commencement

of menstruation. Mean duration of menstruation was 5.25

days in the placebo group and 5.41 days in the magnet group

(p = 0.518) and mean pain duration was 3.08 days and 3.56

days in the placebo and magnet groups, respectively. Men

strual flow, rated on an analog scale of 1-3 (1 = light and

3 = heavy), was rated as a mean of 2.1 in the placebo and

2.5 in the magnet group.

Days taken off work because of menstrual pain appeared

to be greater in the magnet group but did not reach statisti

cal significance (p = 0.384). A significant proportion of

women in both groups described restriction across a broad

range of activities because of menstrual pain (Fig. 2). The

commonly associated symptoms (Fig. 3) were experienced

to the same degree in both groups. Regarding consultations

with a professional for dysmenorrhea, 73% of all women

had consulted their own doctors about menstrual pain. There

was no difference in consultation rate between the magnet

and placebo groups (p = 0.188). Fourteen (14) of the 35

women (39%) had consulted a specialist (gynacologist)

about menstrual pain and again there was no difference in

consultation rate between the magnet and placebo groups

(p = 0.589). Seven (7) of the 35 women (21%) had con

sulted a complementary and alternative medicine practi

tioner about menstrual pain.

Seven of the 35 (21%) had taken prescribed medicine for

period pain (five of 17 of the magnet group and two of 18

of the placebo group, p = 0.153). Prescribed medicines in

cluded the oral contraceptive pill and the nonsteroidal anti-

Table 1. Reasons Given by Volunteers for Not

Completing the Study

FIG. 1. A. The LadyCare® (LC; Bristol, UK) magnetic device

used designed for attachment to the underwear by magnetic force.

LC is plastic coated and is comprised of two parts. The pear-shaped

piece is worn inside of the ladies' underwear, directly against the

pelvis. The LC contains a 20 X 5 mm iron neodymium-boron mag

net within the pear-shaped piece. The second part is a circular plas

tic case that contains a 20 X 1 mm iron neodymium-boron mag

net with a 20 X 2 mm, 400-grade stainless steel directional plate

adherent to its outside. This second part positioned on the outside

of the underwear (as shown in B) attaching securely with the force

of the magnetic field (manufacturer's patent). This directional plate

increases the magnetic power of the north pole (standard scientific

notation) of the magnetic field into the body to 2700 gauss (sur

face measurement made by manufacturer at the body surface of

the magnet (both parts) using a gauss meter. (LadyCare is a reg

istered Class 1 Medical Device. U.K. and U.S. patents are as fol

lows: U.K. Patent No. GB2377179; U.S. Patent No. 6-67659182.)

B. Schematic diagram shows how the device is attached to the

underwear.

inflammatory drugs, mefenamic acid, and naproxen. Only

six of the 35 women (18%), took painkillers preventively

before the anticipated onset of pain. There was no signifi

cant difference between the two groups in this respect (p =

0.642). Ten (10) of the 35 women (29%) had tried herbal or

natural medicines for menstrual pain (seven in the placebo

group and three in the magnet group, p — 0.182). Only nine

(26%) had obtained complete pain relief with various com

binations of therapies (over-the-counter preparations plus

painkillers, painkillers alone, a combination of painkillers

with hot water bottle, Chinese herbal medicine). Twenty-

nine (29) of the 35 (82%) said that they would be very much

interested in a natural alternative to treat menstrual pain.

Static magnets versus placebo

Pain score differences (McGill pain score before —

McGill pain score after device) were, in median with in

terquartile ranges, 17.0 (—53.0, 13.0) in the magnet group

684 ECCLES

[■Period pain BToothache QHeadache UStomachache

Excruciating Horrible

FIG. 2. Pain descriptions. Compared to other sources of pain,

menstrual pain was rated as one of the most severe types of pain

reported, with 85% of subjects describing the level of pain as hor

rible or excruciating.

and 5.0 (-29.0, 27.0) in the placebo group (Fig. 4). This

difference was statistically significant (p < 0.02) and rep

resents an average of 53% reduction in pain in the magnet

group compared to an average of 15% reduction in pain in

the placebo group. Seventy percent (70%) of the subjects

in the magnet group had at least a 50% reduction in pain,

47% of whom had a >75% reduction in pain (Fig. 5). Fifty

percent (50%) of the placebo group had a reduction in pain

but this was <25% in magnitude. The median difference

in total number of painkillers consumed during the course

of menses (painkillers before — painkillers after using the

magnet) was 0.0 ± —8.0, 32 (interquartile range) for the

placebo group, n = 8, and —5.0 ± -14.0, 0 for the mag

net group (n = 5). This reduction in painkiller consump

tion in the magnet group did not reach significance (p <

0.071). The numbers were small here because of missing

or unreliable recording of these data by volunteers. It would

seem on enquiry that this was primarily caused by the vol

unteers being unclear as to the instructions given on this

matter.

Ten (10; 57%) women in the magnet group reported a re

duced tendency to be awakened by pain compared to five

(28%) in the placebo group. This apparent difference ap

proached statistical significance (p = 0.093). One woman in

the magnet group reported more sleep disturbance.

Associated symptoms

In all, 71.5% of the magnet group reported a decrease in

irritability compared to 32.5% of the placebo group. Figure

3 illustrates the high proportion of women that reported this

as an associated symptom (91% of the magnet and 89% of

the placebo group). This difference did not reach statistical

significance (p = 0.056). There was no significant differ

ence between the magnet and placebo groups in reported

breast tenderness (p = 0.800), water retention and bloating

(p = 0.180), nausea or vomiting (p = 0.228), or loss of ap

petite, (although the small numbers of patients experiencing

the latter precluded meaningful statistical analysis). Five (5)

of 13 women in the magnet group (38%) noticed a reduc

tion in facial spots compared to two of 10 (20%) in the

placebo group (p = 0.064).

Side-effects

Six (6) women in the placebo group reported unusual

symptoms after wearing the device compared to two women

in the magnet group. In the placebo group these symptoms

were listed as more painful menstrual periods (two women),

heavier periods (one), nausea (one), rumbling stomach (one),

and diarrhea (one). In the magnet group these were listed as

nausea (one) and "fuzzy-headedness" (one).

Role of the funding source

The study sponsor Magnopulse provided the live and

placebo devices in coded form. The magnetic devices are

commercially sold under the brand name of LadyCare

(Fig. 1). The manufacturer had no role in the design of the

| m Whole Group B Placebo §1 LadyCare

100-

90

80

70

60

50-

40-

30

20

Irritability PMS Aching/ Bloating/ Spots Nausea/ Loss of

Symptoms tender water vomiting appetite

breasts retention

Symptoms

FIG. 3. Percentage of patients with various associated symptoms.

The commonly associated symptoms were experienced to the same

degree in both groups. With regard to premenstrual syndrome

(PMS) symptoms, subjects were asked if they experienced these,

but no definition of PMS was given to the patients. LadyCare®

(Magnopulse, Bristol, UK).

STATIC MAGNET FOR DYSMENORRHEA 685

Placebo LadyCare

FIG. 4. Comparison of the analgesic effect of static magnet

against placebo for dysmenorrhea. Pain score differences (McGill

pain score before — McGill pain score after device) were —17

(-53, 13) (median and interquartile ranges) in the magnet group

and —5.0 (—29, 27) in the placebo group. This represents an av

erage of 53% reduction in pain in the magnet group compared to

an average of 15% reduction in pain in the placebo group. This

difference was statistically significant, p < 0.02. The 95% Mann-

Whitney confidence interval (CI) for the median difference be

tween the magnet and placebo groups (magnet — placebo) were

-53.0 to 23.38. LadyCare® (Magnopulse, Bristol, UK).

study, nor in collection, analysis or interpretation, of the

data. The manufacturer also had no role in the write-up of

this report or in the decision to submit this study for pub

lication.

DISCUSSION

Primary dysmenorrhea is by far the most common gy

necologic problem in menstruating women.8 A recent

prospective study of college students, based on diaries kept

for 1 year, found that 72% of monitored menstrual periods

were painful, most commonly during the first day of

menses. Sixty percent (60%) of the women studied reported

at least one episode of severe pain.9 Dysmenorrhea is a

common cause of absenteeism and reduced quality of life

in women.2'10 In an earlier study,11 dysmenorrhea ac

counted for 600 million lost work hours and $2 billion in

lost productivity on an annual basis. The problem of ab

senteeism from school or work is also underappreciated.8

Dysmenorrhea is often underdiagnosed and undertreated.8

Nonsteroidal anti-inflammatory medications are the main

stay of treatment, with the addition of oral contraceptive

pills when necessary. Approximately 10% of affected

women do not respond to these measures.8 A relative lack

of physician awareness of the very high rates of prevalence

and the substantial morbidity of dysmenorrhea often leads

to inadequate treatment of this problem.8 With the wide

spread availability of over-the-counter nonsteroidal anti-in

flammatory drugs, it is often assumed that women are treat

ing their symptoms to an adequate extent. Unfortunately,

this is not always the case. The significant impact of pri

mary dysmenorrhea on women was confirmed in this study

(Fig. 2), with significant proportions of women in both

groups describing restriction across a broad range of daily

activities. A high percentage of women in the study (82%)

expressed interest in the concept of a natural alternative for

menstrual pain relief.

Only 26% of women in this study had obtained complete

pain relief with various combinations of prescribed and non-

prescribed therapies. This finding, together with the data on

absenteeism, would support the conclusions from other stud

ies (cited at the beginning of this discussion) that many

women may not be obtaining adequate treatment for their

dysmenorrhea and makes the findings of the current study

all the more interesting and relevant.

pPlacebo BLadyCarel

Increase in

pain

<25% 25-49% 50-75%

Level of pain reduction

>75%

FIG. 5. Degree of pain relief achieved. In all, 70% of subjects

in the magnet group had at least a 50% reduction in pain, 47% of

whom had a greater than 75% reduction in pain. A total of 50%

in the placebo group had a reduction in pain, but this was less than

25% in magnitude. LadyCare® (Magnopulse, Bristol, UK).

6S6 ECCLES

The relative lack of compliance (50% of the 70 women

who had expressed an initial interest in taking part in the

study) was one of the reasons that data collecting ceased af

ter 1 month. Despite this, the results obtained were believed

to be of sufficient importance to warrant publication. A

larger study over several menstrual cycles is proposed.

Both the magnet and placebo groups were well matched

demographically for age, pain severity and duration, degree of

associated symptoms, and doctor or specialist consultations.

Both devices were identical in appearance and were held in

place by their magnetic force on the underwear. It was not pos

sible therefore for the women to determine whether they had

?Ate&peutic or a placebo device. One of the difficulties raised

in other double-blind studies using magnets lies in masking

the obvious interaction of the magnet with metallic objects.12

It was therefore decided to use a static magnet of relatively

low power (140 gauss) as a control compared to the active de

vice (2700 gauss). It was anticipated that this would provide

effective masking without effective therapeutic power. How

ever, it is still possible that some of the reported analgesic ef

fect in the placebo group was in fact a magnetic effect even

at this low magnetic power.

Consistent with the finding of significant analgesia with

the magnet was the trend, although not statistically signifi

cant, for a reduction in usual painkiller consumption in the

magnet group (p = 0.071). Some women had not responded

or had not completed the question on painkillers used on the

home questionnaire, which reduced the statistical power of

the analysis on painkiller consumption. The tendency for a

reduction in irritability (p = 0.056) is perhaps likely to be

secondary to the reduction in pain as opposed to a specific

anxiolytic effect of the magnetic field.

A limition of the study is the small size of the study and

also that the effects were only determined over one menstrual

cycle. In a recent survey7 of 193 women with primary dys-

menorrhea (randomly selected from the manufacturer's cus

tomer database) (average duration, 11.6 years), all of whom

were using the same magnetic device as in this study. This

survey found a highly statistically significant reduction in pain

(p < 0.0001), consumption in painkillers (p < 0.0001), and a

54% reduction in days taken off work (p < 0.0001). Long-

term persistence of the analgesic effect of these magnets to re

lieve menstrual pain was suggested by 90% of the women us

ing them for > 1 year and still having pain relief. Furthermore,

these long-term users noted no side-effects.

This double blinded trial also provides preliminary pilot

evidence for the analgesic properties of static magnetic fields

in the treatment of dysmenorrhea. This requires large-scale

verification, perhaps with crossover, but the potential im

pact on what would seem to be for most inadequately treated

dysmenorrhea is exciting. These findings also add to the

growing weight of evidence from well-controlled trials that

static magnets can elicit analgesia. A very recent publica

tion in the British Medical Journal demonstrated that mag

netic bracelets relieved the pain of osteoarthritis of the hip

and knee,13 and a recent critical review of randomized con

trol trials of static magnets for pain relief revealed that

73.3% of 21 studies reported a statistically significant anal

gesic effect5 across a broad range of different types of pain.

ACKNOWLEDGMENTS

The author thanks Magnopulse for funding of this re

search. The author also thanks Ms. Philippa Martin for help

in construction of the graphs and to Dr. Lia Tzala, statisti

cian, for statistical analysis of the data.

CONFLICT OF INTEREST

Part of the funding included remuneration for the author's

time involved in conducting and writing up the project. The

author was approached by the company on the basis of his

interest and research in the field of magnotherapy. This had

become known to them by way of a comment on magnets

made in the National Press.14

REFERENCES

1. Zhang WY, Li Wan Po A. Efficacy of minor analgesics in pri

mary dysmenorrhoea: A systematic review. Br J Obstet Gy-

naecol 1998; 105:780-789.

2. Jamieson DJ, Steege JF. The prevalence of dysmenorrhea, dys-

pareunia, pelvic pain, and irritable bowel syndrome in primary

care practices. Obstet Gynecol 1996;87:55-58.

3. BUPA's Health Information Team. BUPA website. Online doc

ument at: http://hcd2.bupa.co.uk/fact_sheets/Mosby_factsheets/

Dysmenorrhoea.html. www.bupa.co.uk/ (Accessed June 2003.)

4. Clark AD. Dysmenorrhoea. Online document at: http://www.

emedicine.com/emerg/topicl56.htm Accessed December 2004.

5. Eccles NK. A critical review of randomized controlled trials

of static magnets for pain relief. J Altern Complement Med

2005; 11:495-509.

6. Brown CS, Parker N, Ling F, Wan J. Effect of magnets on

chronic pelvic pain. Obstet Gynecol 2000;95:S29.

7. Eccles NK, Price DR. A Survey to Determine the Long-term

Effects of LadyCare Static Magnets on Dysmenorrhoea (Pe

riod Pain). Online document at: http://www.magnopulse.com/

ladycare/scientific.htm Accessed September 2003.

8. Coco AS. Primary dysmenorrhea. Am Fam Phys 1999;60:

489-496.

9. Harlow SD, Park M. A longitudinal study of risk factors for

the occurrence, duration and severity of menstrual cramps in

a cohort of college women. Br J Obstet Gynaecol 1996; 103:

1134-1142.

10. Sundell G, Milson I, Andersch B. Factors influencing the

prevalence and severity of dysmenorrhea in young women. Br

J Obstet Gynecol 1990;97:588-594.

11. Andersch B, Milsom I. An epidemiologic study of young

women with dysmenorrhea. Am J Obset Gynecol 1982; 144:

655-660.

STATIC MAGNET FOR DYSMENORRHEA 687

12. Hong C, Lin JC, Bender LF, et al. Magnetic necklace: Its ther

apeutic effectiveness on neck and shoulder pain. Arch Phys

Med Rehab 1982;63:462-466.

13. Harlow T, Greaves C, White A, et al. Randomised controlled

trial of magnetic bracelets for relieving pain of osteoarthritis

of the hip and knee. BMJ 2004;329:1450-1454.

14. Halle, M. Magnets cured my period pains. Daily Mail

(femail.co.uk, 14th March 2002).

Address reprint requests to:

Nyjon K. Eccles, MRCP, Ph.D.

The Chiron Clinic

121 Harley Street

London WIG 6AX

England

E-mail: drnyjon@hotmail.com