Intervention Review

You have free access to this content

Electromagnetic fields for treating osteoarthritis

  1. Shasha Li1,
  2. Bo Yu2,
  3. Dong Zhou3,
  4. Chengqi He1,*,
  5. Qi Zhuo4,
  6. Jennifer M Hulme5

Editorial Group: Cochrane Musculoskeletal Group

Published Online: 14 DEC 2013

Assessed as up-to-date: 3 OCT 2013

DOI: 10.1002/14651858.CD003523.pub2


How to Cite

Li S, Yu B, Zhou D, He C, Zhuo Q, Hulme JM. Electromagnetic fields for treating osteoarthritis. Cochrane Database of Systematic Reviews 2013, Issue 12. Art. No.: CD003523. DOI: 10.1002/14651858.CD003523.pub2.

Author Information

  1. 1

    West China Hospital, Sichuan University, Department of Rehabilitation Medicine, Chengdu, Sichuan Province, China

  2. 2

    Sichuan Provincial Hospital for Women and Children, Department of Paediatrics, Chengdu, China

  3. 3

    West China Hospital, Sichuan University, Department of Neurology, Chengdu, Sichuan, China

  4. 4

    Chinese PLA General Hospital, Department of Orthopaedic Surgery, Beijing, China

  5. 5

    McGill University School of Medicine, Class of 2012, Montreal, Quebec, Canada

*Chengqi He, Department of Rehabilitation Medicine, West China Hospital, Sichuan University, No.37 Guo-xue-xiang Street, Chengdu, Sichuan Province, 610041, China. chengqi.he623477@gmail.com.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 14 DEC 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Fary 2011

MethodsRandomised, double-blind, placebo-controlled trial
Sample size at entry: 70 patients (34 in the active and 36 in the placebo group)
Withdrawals: 3 participants (2 in the active and 1 in the placebo group)
Treatment duration: 7 hours daily, preferably overnight, for 26 weeks
Follow-up: 26 weeks


ParticipantsInclusion criteria: diagnosis of osteoarthritis of the knee was in accordance with the American College of Rheumatology modified clinical classification system. Plain radiographs available for 64 participants confirmed the diagnosis. Persistent and stable pain (defined as not getting worse or better overall despite short-term fluctuations) for a minimum of 3 months prior to study entry was confirmed in all participants by telephone interview.
Exclusion criteria: co-existing inflammatory arthropathies, contraindications to electrical stimulation, skin disorders in the
vicinity of the knee to be treated, total knee replacement scheduled during the study period, and/or insufficient English
to follow instructions and complete forms
Number of patients who finished this study: 67
Male/female: 17/17; 20/16
Mean age in years (range): 70.7 ± 8.9; 68.9 ± 11.4

Intervention group number: 34

Control/sham group number: 36


InterventionsActive group: a commercially available TENS stimulator (Metron Digi-10s) was modified by a biomedical engineer to deliver pulsed electrical stimulation current parameters as follows: pulsed, asymmetrically biphasic, exponentially decreasing waveform with a frequency of 100 Hz and pulse width of 4 ms
Placebo group: placebo devices
Current was delivered via 120 mm x 80 mm multiple-use conductive silicone electrodes inserted into larger calico pockets (175 mm x 100 mm) to increase the contact surface area and reduce current density

The placebo device was identical in appearance and method of use; however, the current flow was programmed to turn off after 3 minutes. Since this was a sub-sensory treatment, this change was not detectable by participants
Patients were advised to use the instrument for 7 hours daily, preferably overnight, for 26 weeks


OutcomesPrimary outcomes:
(1) Pain: change in pain score over 26 weeks measured on a 100 mm VAS
(2) Physical function (Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Likert format 3.1)
(3) Patient's global assessment of disease activity (on a 100 mm VAS)

Secondary outcomes:
(1) Quality of life (the 36-item Medical Outcomes Study Short-Form 36 version 2 (SF-36 v. 2) health survey)
(2) Joint stiffness (WOMAC 3.1)
(3) Physical activity (Human Activity Profile and Actigraph GT1M accelerometers worn for 7 consecutive days)
(4) An 11-point global perceived effect scale


NotesSupported by an Arthritis Australia and State & Territory Affiliate Grant and a Physiotherapy Research Foundation Research Seeding grant, and by a Curtin University School of Physiotherapy Early Career Researcher grant to Dr. Fary. Dr. Fary was the recipient of an Australian Government Postgraduate PhD scholarship and a Curtin University School of Physiotherapy Movement Through Life Top-Up scholarship


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "using computer-generated randomization in blocks of 6."

Allocation concealment (selection bias)Low riskQuote: "Allocation, stratified by sex, age (60 years, 60–75 years, and 75 years), and baseline VAS pain scores (25–40 mm, 41–60 mm, and 61–100 mm), was performed independently by an administrator, not otherwise involved in the study"

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "This process ensured that all study investigators and participants remained blinded to allocation until analysis was complete."

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: "This process ensured that all study investigators and participants remained blinded to allocation until analysis was complete."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Unclear riskNo information provided

Incomplete outcome data (attrition bias)
All outcomes
Low risk4 weeks: 1/34 missing from active group (failed to return week 4 data)

16 weeks: 1/34 missing from active group (device uncomfortable to wear) but participant who completed week 4 data returned; 1/36 missing from placebo group (protocol too onerous)

26 weeks: 1/ 34 missing from active group (failed to attend final appointment)

Selective reporting (reporting bias)Unclear riskNo information provided

Garland 2007

MethodsRandomised, double-blind, placebo-controlled trial
Sample size at entry: 58 patients (39 in the active and 19 in the placebo group)
Withdrawals: 2 patients
Treatment duration: 6 to 14 hours per day during the 3-month treatment period
Follow-up: 3 months


ParticipantsInclusion criteria: age 18 years or greater, the intellectual ability to understand and sign an informed consent and complete the study questionnaire, and willingness to maintain stable doses of analgesics and NSAIDs for 1 month prior to study entry and during the 3-month double-blind period
Exclusion criteria: knee instability and/or valgus or varus deformities of > 20; pregnancy; infectious arthritis; implantable electronic devices; a diagnosis of gout; recurrent inflammatory episodes of pseudogout; malignancy (other than basal cell carcinoma) in the prior 3 years; inflammatory arthritis such as rheumatoid arthritis; psoriatic arthritis; Reiter's syndrome; haemochromatosis; inflammatory bowel disease, etc.
Number of patients who finished this study: 58
Male/female: 12/27; 8/11
Mean age in years (range): 64.3 ± 10.2; 69.9 ± 11.4

Intervention group number: 39

Control/sham group number: 19


InterventionsActive group: pulsed electrical stimulation (BioniCare Medical Technologies, Inc., Sparks, Maryland)
Placebo group: placebo devices
100 Hz, negative pulsed signal, 0 and 12 V, active treatment remained imperceptible and indistinguishable from placebo. The placebo devices shut off after the amplitude was reduced. All active and placebo devices contain a timer that recorded the cumulative hours when the device was in use
Patients were advised to use the instrument for 6 to 14 hours/day during the 3-month treatment period


OutcomesPrimary outcomes:
(1) Per cent change from baseline on a 0 to 100 VAS measuring patient global evaluation of arthritis symptoms in the treated knee
(2) Per cent change from baseline on a 0 to 100 VAS measuring pain and other symptoms in the treated knee
(3) Per cent change from baseline on the Western Ontario and McMaster Universities (WOMAC) pain (0 to 500), stiffness (0 to 200) and function (0 to 850) sub-scales as measured on a 100 mm VAS
Secondary outcomes: the percentage of patients that experienced 50% improvement in patient global evaluation, pain and symptoms in the treated knee, and in the 3 WOMAC outcome scale measures
The occurrence of rashes and other adverse device effects was solicited and recorded at each visit


NotesSupported by a grant from BioniCare Medical Technologies, Inc


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "Each placebo and active device was assigned a unique number."

Allocation concealment (selection bias)Low riskQuote: "A master random number chart was generated and maintained by an independent observer who had no interaction with the sponsors pendent observer who had no interaction with the sponsors"; "The number coincided with active or placebo units according to the dictates of the random number table."

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "A master random number chart was generated and maintained by an independent observer who had no interaction with the sponsors pendent observer who had no interaction with the sponsors."

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: "A master random number chart was generated and maintained by an independent observer who had no interaction with the sponsors pendent observer who had no interaction with the sponsors"; "The clinical investigators had no influence on the assignment of any specific patient to an active or placebo device. The blind has never been revealed to any of the authors except the statistician."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Low riskQuote: "A master random number chart was generated and maintained by an independent observer who had no interaction with the sponsors pendent observer who had no interaction with the sponsors"; "The clinical investigators had no influence on the assignment of any specific patient to an active or placebo device. The blind has never been revealed to any of the authors except the statistician."

Incomplete outcome data (attrition bias)
All outcomes
Low risk1/39 missing from active group (discontinued study participation before the second month follow-up visit); 1/19 missing from placebo group (discontinued study participation before the second month follow-up visit)

Selective reporting (reporting bias)Unclear riskNo information provided

Nelson 2013

MethodsDouble-blind, placebo-controlled, randomised pilot study
Sample size at entry: 34 patients (15 in the active and 19 in the sham group)
Withdrawals: 10 participants (3 in the active and 7 in the sham group)
Treatment duration: 15 minutes twice daily, for 42 days
Follow-up: 42 days


ParticipantsInclusion criteria: patient selection required that participants had knee pain for at least 3 months with an imaging study that confirmed articular cartilage loss, an initial VAS score ≥ 4, and at least 2 hours of daily standing activity in a physical occupation

Exclusion criteria: patients with rheumatoid arthritis, gout and pregnancy; cortisone injections, surgery or an effective visco supplementation series within the past 6 months; implanted electronic devices. Patients on disability or with third party claims were excluded

Male/female: 5/10, 5/14
Mean age in years (range): 55.5 ± 2.5, 58.4 ± 2.5

Intervention group number: 15

Control/sham group number: 19


InterventionsActive group: a pulsed electromagnetic field signal consisting of a 7 ms burst of 6.8 MHz sinusoidal waves repeating at 1 burst/s delivering a peak induced electrical field of 34 ± 8 V/m in the knee from the portable battery operated device (Palermo, Ivivi Health Sciences, LLC, San Francisco, CA)

Sham group: sham devices were activated with a switch, same as active devices, and both sham and active units had blinking indicator lights. The pulsed electromagnetic field signal from these devices did not produce heat or cause any other sensation in tissue

Once manually activated, treatment was automatically applied for 15 minutes. Device was used for 15 minutes twice daily for 42 days


OutcomesPrimary outcomes: pain: self report maximum daily VAS pain scores on an unmarked horizontal 10 cm line (0 is no pain and 10 is worst possible pain)


NotesPartially supported by the Department of Orthopaedic Surgery, Henry Ford Hospital, Detroit Michigan, and Ivivi Health Sciences, LLC, San Francisco, CA, who manufactured the pulsed electromagnetic field devices utilised in this study


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "Randomization was performed by the blinded assignment of devices according to their serial numbers."

Allocation concealment (selection bias)Low riskQuote: "Device randomization was performed by the manufacturer (Ivivi Health Sciences, LLC) and all devices with the randomization code were sent to the Epidemiology Dept at Henry Ford Hospital, from which assignment to patients was controlled. General unblinding occurred after all data were collected."

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "PEMF signal employed in this study is at least 1,000-fold below the ambient magnetic field and cannot be detected using standard Gauss meters. Therefore, only measurements with specialized laboratory equipment, not readily available to the patient or health care practitioner, could determine whether a device was active. General unblinding occurred after all data were collected."

Comment: probably done.

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: ""PEMF signal employed in this study is at least 1,000-fold below the ambient magnetic field and cannot be detected using standard Gauss meters. Therefore, only measurements with specialized laboratory equipment, not readily available to the patient or health care practitioner, could determine whether a device was active. General unblinding occurred after all data were collected."

Comment: probably done.

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Unclear riskNo information provided

Incomplete outcome data (attrition bias)
All outcomes
High riskDay 42: 3/15 missing from active group (lack of perceived benefit as the reason, confirmed by VAS scores); 7/19 missing from sham group (lack of perceived benefit as the reason, confirmed by VAS scores)

Selective reporting (reporting bias)Unclear riskNo information provided

Nicolakis 2002

MethodsRandomised, double-blind, controlled trial
Sample size at entry: 36 patients
Withdrawals: 4 patients
Treatment duration: 30 minutes per session twice a day for 6 weeks
Follow-up: 2 months


ParticipantsInclusion criteria: patients with symptomatic osteoarthritis of the knee. All patients had radiographic signs of osteoarthritis according to the classification of Kellgren, grade 2 or 3
Exclusion criteria: included clinical evidence of meniscal tears, polyarthritis, cardiovascular, pulmonary or neurological disorders, cardiac pacemakers or malignancies
Number of patients who finished this study: 36
All patients were on long-term intra-articular steroid injections every 2.6 ± 0.8 (verum) and 2.4 ± 1.1 (sham) weeks (mean ± SD) before inclusion into the study
Male/female: 11/4, 8/9
Mean age in years (range): 69 ± 5, 67 ± 7

Intervention group number: 15

Control/sham group number: 17


InterventionsTreatment group: PMF
Sham group: PMF devices had disconnected coils and a shunt resistor in the mat, so they did not produce a PMF. PMF was administered to the whole body using a mat 1.7 m x 0.65 m in size. The mat produced a pulsating electromagnetic field with a mean intensity of 40 μT (wave ranger professional, program 12, Mediscan GmbH, Bad Haller StraBe34, 4500 Kremsmünster, Austria). The frequency of the PMF ranged from 1 Hz to 3000 Hz
During the treatment the patients lay on the mat for 30 minutes per session twice a day for 6 weeks


Outcomes1) Primary outcomes: change between baseline and the end of treatment in the 240-point WOMAC Osteoarthritis Index: pain score (maximum score, 50), stiffness score (maximum score, 20) and physical function score (maximum score, 170)
2) Secondary outcomes included objective and subjective measures: gait test, isokinetic dynamometry strength tests


NotesThe study was supported by a grant from the "Institut zurwissenschaftlichen Evaluierung alternativer Heilmethoden" (Institute for Research in Complementary Medicine), Rokitan-skygasse 40/8, 1170 Vienna, Austria


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "The study coordinator assigned the devices on a random basis to patient according to a list created by randomization function (Matlab, Mathworks Incor-poration, Natick, Massachusetts)."

Allocation concealment (selection bias)Low riskQuote: "All devices were numbered; only the study coordinator was aware of the code."

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "The persons involved in patient administration, the evaluator, as well as the person instructing the patients in the use of the device and giving technical support if required, were blinded. At the end of the study the code was broken."

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: "The persons involved in patient administration, the evaluator, as well as the person instructing the patients in the use of the device and giving technical support if required, were blinded. At the end of the study the code was broken."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Low riskQuote: "The persons involved in patient administration, the evaluator, as well as the person instructing the patients in the use of the device and giving technical support if required, were blinded. At the end of the study the code was broken."

Incomplete outcome data (attrition bias)
All outcomes
Low risk3/18 missing from treatment group (1 did not complete treatment, 2 further patients had to be excluded from the statistical analysis as they did not use the device properly); 1/18 missing from the sham group (withdrew because of excessive pain)

Selective reporting (reporting bias)Unclear riskNo information provided

Pipitone 2001

MethodsRandomised, double-blind, placebo-controlled trial
Sample size at entry: 75 patients
Withdrawals: 16 patients
Treatment duration: 3 times a day for 30 minutes of 6 weeks duration
Follow-up: not clear


ParticipantsInclusion criteria: patients had radiographic signs and symptoms of osteoarthritis as judged by the criteria of the ACR
Exclusion criteria: pregnancy, use of pacemaker, insulin pump or of any implanted electrical device, etc.
Number of patients who finished this study: 69
Male/female: 22/12, 28/7
Mean age in years (range): 40 to 84, 48 to 84

Intervention group number: 39

Control/sham group number: 36

(6 patients failed to attend after the screening visit and were excluded from the analysis)


InterventionsActive group: pulsed electromagnetic fields (Snowden Healthcare Ltd., Nottingham, UK)
Sham group: sham devices
Pulsed electromagnetic field devices that generate pulses of magnetic energy via a soft iron core treated with 62 trace elements. Pulses are selected at their base frequencies (3 Hz, 7.8 Hz and 20 Hz). They have a rise time of 1 μs, a low magnetic output (< 0.5 gauss) and a range of activity of up to 30 cm around the unit. Medicur devices run on 9 V batteries and switch off automatically after a 10-minute period
Patients were instructed not to change their basic therapeutic regimen
Number of sessions: 30 minutes per session 3 times a day for 6 weeks


Outcomes1) Primary outcomes: reduction in overall pain assessed on a 4-point Likert scale ranging from none to severe
2) Secondary outcomes: Lequesne Index; WOMAC Osteoarthritis Index (Likert scale); UK 36-item short form of the Medical Outcomes Study (SF-36) and the EuroQol (Euro-Quality of Life, EQ-5D)


NotesThis study was supported by an educational grant from Snowden Healthcare


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "Patients fulfilling the study criteria were randomised on study entry to receive active pulsed electromagnetic fields treatment with the Medicur magnetic devices or placebo using a 12 * 12 randomisation table."

Allocation concealment (selection bias)Low riskQuote: "Neither the patients nor the medical assessor were aware of the treatment group". "The code numbers were not broken until all patients had complete the study."

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "Neither the patients nor the medical assessor were aware of the treatment group". "The code numbers were not broken until all patients had complete the study."

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: "Neither the patients nor the medical assessor were aware of the treatment group". "The code numbers were not broken until all patients had complete the study."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Low riskQuote: "Neither the patients nor the medical assessor were aware of the treatment group". "The code numbers were not broken until all patients had complete the study."

Incomplete outcome data (attrition bias)
All outcomes
High risk10/39 missing from active group (5 attended only for screening and were excluded from the analysis, 5 withdrew before completing treatment); 6/36 missing from sham group (1 attended only for screening and was excluded from the analysis, 5 withdrew before completing treatment)

Selective reporting (reporting bias)Unclear riskNo information provided

Thamsborg 2005

MethodsRandomised, double-blind, placebo-controlled trial
Sample size at entry: 90 patients (pulsed electromagnetic field 45, placebo 45)
Withdrawals: 7 patients
Treatment duration: 2 hours daily treatment 5 days per week for 6 weeks
Follow-up: 6 weeks


ParticipantsInclusion criteria: patients older than 45 years with painful knee osteoarthritis of the femorotibial compartment fulfilling the combined clinical and radiological criteria of the American College of Rheumatology were included
Exclusion criteria: inflammatory joint disease, acromegaly, Charcot's arthropathy, haemochromatosis, Wilson's disease, ochronosis, terminal illnesses/malignancies, pregnancy or lack of contraception use in women of childbearing age, and use of pacemaker or any implanted electrical device
Number of patients who finished this study: 83
Male/female: 23/20, 16/25
Mean age in years (range): 60.4 ± 8.7, 59.6 ± 8.6

Intervention group number: 45

Control/sham group number: 45


InterventionsActive group: pulsed electromagnetic fields (Biofields Aps, Copenhagen, Denmark)
Sham group: sham devices
A pulse generator yields G50V in 50 Hz pulses changing voltage in 3 ms intervals. The current in the coils that create the pulsed electromagnetic fields causes the coils to become slightly warmer than the surroundings after 30 minutes (28 to 35). For the group of patients not receiving active treatment, direct current (DC) was applied to the coils yielding a permanent magnetic field that does not evoke changing electrical potentials in tissue
Treatment duration: 2 hours daily treatment 5 days per week for 6 weeks


Outcomes1) Primary outcomes: WOMAC osteoarthritis index: a questionnaire addressing the severity of joint pain (5 questions), stiffness (2 questions) and limitation of physical function (17 questions)
2) Secondary outcome measures: WOMAC sub-score of joint pain (0 to 25); WOMAC sub-scores of stiffness (0 to 10); activities of daily living (ADL) (0 to 85) and scintigraphic results


NotesEconomic support from IMK Almene Fond and from Københavns Amts Erhvervskontor


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskQuote: "This was a 1:1 randomized, controlled, double-blind add-on study." No further information provided

Allocation concealment (selection bias)Low riskQuote: "This was a 1:1 randomized, controlled, double-blind add-on study." "Blinding was maintained until the final database was cleaned and locked."

Comment: probably done

Blinding (performance bias and detection bias)
Blinding of patients?
Unclear riskQuote: "This was a 1:1 randomized, controlled, double-blind add-on study." No further information provided

Blinding (performance bias and detection bias)
Blinding of physicians?
Unclear riskQuote: "This was a 1:1 randomized, controlled, double-blind add-on study." No further information provided

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Unclear riskQuote: "This was a 1:1 randomized, controlled, double-blind add-on study." No further information provided

Incomplete outcome data (attrition bias)
All outcomes
Low risk3/45 missing from active group; 4/45 missing from sham group

Selective reporting (reporting bias)Unclear riskNo information provided

Trock 1993

MethodsRandomised, placebo-controlled trial
Sample size at entry: 27 patients (pulsed electromagnetic field 15, placebo 12)
Withdrawals: 7 patients
Treatment duration: 18 half-hour periods of exposure over about 1 month
Follow-up: 2 months


ParticipantsInclusion criteria: all patients met the criteria for the diagnosis of osteoarthritis published by Altman. Patients with osteoarthritis, primarily of the knee with symptoms for greater than 1 year, who had not started any new treatment within 1 month of study
Exclusion criteria: pregnancy or lack of contraception use in women of childbearing age, and use of pacemaker or the presence of any serious unstable medical illness
Number of patients who finished this study: 25
Male/female: unclear
Age: at least 18 years

Intervention group number: 15

Control/sham group number: 12


InterventionsActive group: pulsed electromagnetic fields (MFG)
Sham group: sham devices
Magnetic therapy system with extremely low-frequency (ELF) pulsed waves consisting of a magnetic field generator with an electronic interface to freely moving air coil. Separate systems for peripheral joints and the axis skeleton. Placebo therapy applied by not energising the air coil
Treatment duration: 18 half-hour periods of exposure over about 1 month


OutcomesSeverity of pain, difficulty performing activities of daily living (ADL), pain performing ADL, worst discomfort in previous week, pain on joint motion by MD exam, joint tenderness by MD exam, assessment of improvement by MDs


NotesFunded by Bio-Magnetic Therapy Systems, Inc.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "Patients were randomized to receive active pulsed electromagnetic fields or placebo using a table of 1000 random digits."

Allocation concealment (selection bias)Low riskQuote: "A code master record sheet was kept by an office administrator."

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "Both the patients and the examining physician remained blinded as to whether active pulsed electromagnetic fields or placebo was given."

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: "Both the patients and the examining physician remained blinded as to whether active pulsed electromagnetic fields or placebo was given."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Low riskQuote: "A code master record sheet was kept by an office administrator."

Incomplete outcome data (attrition bias)
All outcomes
High risk5/15 missing from active group; 2/12 missing from sham group

Selective reporting (reporting bias)Unclear riskNo information provided

Trock 1994

MethodsRandomised, double-blind, multicentre controlled trial
Sample size at entry: 86 patients with osteoarthritis of the knee and 81 patients with osteoarthritis of the cervical spine
Withdrawals: 11 patients withdrew from the double-blind trial of knee osteoarthritis and 11 patients from the trial of osteoarthritis of the cervical spine before completing treatment
Treatment duration: treatments were given for 30-minute periods, 3 to 5 times a week for 18 treatments
Follow-up: 1 month


ParticipantsInclusion criteria: patients with osteoarthritis of the knee who met the criteria published by Altman. 11 patients with cervical spine pain were admitted to the study if radiographs showed evidence of disk space narrowing with osteocyte formation and/or subchondral sclerosis in 1 or more locations; osteophyte formation and subchondral sclerosis of facet joints were also accepted as evidence of osteoarthritis, with local symptoms such as pain and stiffness of at least 1 year duration. Patients were instructed not to change their basic therapeutic regimen, including drugs and physical therapy, during the period of treatment and observation
Exclusion criteria: patients who had changed their therapeutic regimen within 1 month before evaluation were excluded

Number of patients who finished this study: 86 knee osteoarthritis and 81 cervical osteoarthritis
Male/female: unclear
Age: at least 35 years

Intervention group number: 86

Control/sham group number: 86


InterventionsActive group: pulsed electromagnetic fields (M-T System)
Sham group: sham devices
Magnetic therapy system with extremely low-frequency (ELF) pulsed waves consisting of 3 integrated components: a magnetic field generator, an electronic interface and a segmented single toroid coil with annular windings that produced pulsed DC elliptical magnetic fields
Wave form: quasi-rectangular with abruptly rising and deteriorating wave form: pulse burst duty cycle of up to 0.8. The number of pulsed bursts is determined by the frequency: the maximum was 20
Frequency: the following energy characteristics were applied stepwise to the area of the joint being treated: 5 Hz 10 to 15 gauss for 10 minutes; 10 Hz 15 to 25 gauss for 10 minutes; and 12 Hz 15 to 25 gauss for 10 minutes
Intensity and voltage control: coil current of < 2 A with 120 V
Placebo therapy applied by not energising the air coil
Treatment duration: 30 minutes
Number of sessions: 3 to 5 times a week for 18 treatments


Outcomes1) The physician recorded the degree of pain on motion (none, slight, moderate or severe) and tenderness using the Ritchie Scale as described for patients with osteoarthritis by Doyle. For both of these observations the score for 'none' was assigned as 0, 'slight' as 1, 'moderate' as 2 and 'severe' as 3
2) At each of the evaluations after the baseline, the patient was asked to quantify any improvement by marking a 10 cm VAS
3) At the end of treatment and the 1-month follow-up evaluation the observing physician was asked to record a global assessment of overall improvement using a 4-point scale (0 = none, 1 = slight, 2 = good and 3 = excellent)


NotesSupported by Bio-Magnetic Therapy Systems, Inc.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskQuote: "One research associate kept the list of 1000 random numbers, divided into pairs, the higher of which was to receive treatment and the lower placebo."

Allocation concealment (selection bias)Low riskQuote: "Separate lists of random numbers were kept for the patients."

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "Neither patient in the trial nor any patient in the center, nor any other staff could tell which patients were receiving active treatment."

Blinding (performance bias and detection bias)
Blinding of physicians?
High riskQuote: "Only the therapy technician at the treatment center was informed, by means of a code."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Low riskQuote: "Neither patient in the trial nor any patient in the center, nor any other staff could tell which patients were receiving active treatment." "Only the therapy technician at the treatment center was informed, by means of a code."

Incomplete outcome data (attrition bias)
All outcomes
Low riskDouble-blind trial of knee osteoarthritis: 5/42 missing from active group; 5/44 missing from sham group
Trial of osteoarthritis of the cervical spine: 4/42 missing from active group; 7/39 missing from sham group

Selective reporting (reporting bias)Unclear riskNo information provided

Zizic 1995

MethodsRandomised, multicentre,double-blind, placebo-controlled trial
Sample size at entry: 78 patients (41 in the active and 37 in the placebo group)
Withdrawals: 7 patients
Treatment duration: 6 to 10 hours per day during the 4-week treatment period
Follow-up: 6 months


ParticipantsPatients older than 20 years with painful knee osteoarthritis of the femorotibial compartment fulfilling the combined clinical
and radiological criteria
Number of patients who finished this study: 71
Male/female: 20/18, 18/15
Age: over 20 years

Intervention group number: 41

Control/sham group number: 37


InterventionsActive group: pulsed electrical stimulation
Control group: sham devices
The electrical impulse was generated by a portable, battery-operated device that delivers a low-frequency (100 Hz), low-amplitude, voltage sourced (mean = 6.2 peak volts), monophasic, spiked signal to the knee via skin surface electrodes. The voltage of the placebo device was adjusted to the subthreshold level using the same procedure as the active device. Patients were advised to use the instrument for 6 to 10 hours/day during the 4-week treatment period


Outcomes1) Primary efficacy outcomes: physician's global evaluation, patient's evaluation of pain, patient's evaluation of function of the treated knee
2) Secondary efficacy outcomes: patient estimate of the duration of morning stiffness in the affected knee, measured in minutes; range of motion of the knee, determined by goniometric measurement of flexion and extension; tenderness; swelling; knee circumference; 50-foot walking time


NotesThis study was supported by Murray Electronics


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskQuote: "Patients were randomized to receive an active device or an identical appearing placebo device."

Comment: no further information provided

Allocation concealment (selection bias)Unclear riskQuote: "Patients were randomized to receive an active device or an identical appearing placebo device."

Comment: no further information provided

Blinding (performance bias and detection bias)
Blinding of patients?
Low riskQuote: "...in a multicenter, prospective, randomized, placebo controlled, double blind study"

Comment: probably done

Blinding (performance bias and detection bias)
Blinding of physicians?
Low riskQuote: "A nurse coordinator instructed patients in the proper use of the device."

Blinding (performance bias and detection bias)
Blinding of outcome assessors?
Unclear riskNo further information provided

Incomplete outcome data (attrition bias)
All outcomes
Low risk3/41 missing from active group; 4/37 missing from sham group

Selective reporting (reporting bias)Unclear riskNo information provided

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Alcidi 2007The duration of treatment was only 5 days (once daily). Efficacy requires at least 4 weeks' treatment duration (suggestion of several physiotherapists)

Ay 2009All patients participated in the therapy programmes for 3 weeks (15 sessions). Pulsed electromagnetic field therapy for 30 minutes per day and combined hot-pack and transcutaneous electrical nerve stimulation (TENS)

Battisti 2004Treatments were given for 30 minutes for 15 days. The treatment duration did not match the inclusion criteria of this review

Danao-Camara 2001Primary report, not a RCT

Fischer 2005Randomisation was not mentioned. Li SS sent an email to Dr. Pelka, the corresponding author of the paper, on 10 December 2007, with 2 questions as follows: 1. Was the randomisation of the allocation procedure adequate? 2. Was the allocation concealment properly done? The author did not reply, so, we considered this a cohort study rather than a RCT

Fischer 2006The second version of Fischer 2005

Hinman 2002The aim of this article was to assess the effect of static magnetic fields for chronic knee pain but not osteoarthritis

Jack 2006Prospective, cohort study and not a RCT

Jacobson 2001Treatment duration: 48 minutes per treatment session 8 times in 2 weeks. This time frame may be too short to assess safety and efficacy based on biological plausibility

Kulcu 2009A therapy session lasted for 35 minutes and 15 sessions were performed during 3 weeks (5 sessions/week)

Liu 2004Randomisation was mentioned but not described in detail. Li SS telephone interviewed the original author, Wencai Liu, on 3 December 2007. The author could not remember how the randomisation was designed. Therefore we considered the study not to be a RCT

Ozgüçlü 2010Both the pulsed electromagnetic fields and sham PEMF treatments being evaluated were 55 minutes/session, 5 sessions per week for 2 weeks. The treatment duration did not match the inclusion criteria of this review

Pavlović 2012The treatment period was only 10 days. The treatment duration did not match the inclusion criteria of this review

Rigato 2002The study included patients with cervical spondylosis and shoulder periarthritis without separately reported results and we could not extract data on cervical osteoarthritis

Sutbeyaz 2006Treatment duration: twice a day for 30 minutes of 3 weeks duration. The duration did not match the inclusion criteria of this review

Tomruk 2007Both the pulsed electromagnetic field group (n = 16) and the control group (n = 16) participated in therapy, 30 minutes per session twice a day for 3 weeks. The treatment duration did not match the inclusion criteria of this review

 
Comparison 1. Electromagnetic fields versus placebo for osteoarthritis

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pain6434Mean Difference (IV, Random, 95% CI)15.10 [9.08, 21.13]

 2 Physical function3197Mean Difference (IV, Random, 95% CI)4.55 [-2.23, 11.32]

 3 Quality of life2139Std. Mean Difference (IV, Random, 95% CI)0.09 [-0.36, 0.54]

 4 Number of patients experiencing any adverse event4288Risk Ratio (M-H, Fixed, 95% CI)1.17 [0.72, 1.92]

 5 Number of patients who withdrew because of adverse events1Risk Ratio (M-H, Fixed, 95% CI)Totals not selected

 
Summary of findings for the main comparison. Electromagnetic field treatment compared to placebo for the treatment of osteoarthritis

Electromagnetic field treatment compared to placebo for the treatment of osteoarthritis

Patient or population: patients with osteoarthritis
Settings: out-patients recruited from healthcare facilities in Australia, Denmark, UK and the US
Intervention: electromagnetic field treatment
Comparison: placebo

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

PlaceboElectromagnetic field treatment

Pain
100 mm VAS

Scale from: 0 to 100

(Higher scores mean worse pain)
Follow-up: mean 6 weeks
The mean change in pain in the control groups was 10.7The mean change in pain in the intervention groups was
15.10lower
(9.08 to 21.13 lower)
434
(6 studies)
⊕⊕⊕⊝
moderate1
MD 15.10 (95% CI 9.08 to 21.13)

Absolute risk difference: 15% (95% CI 9.08% to 21.13%)

Relative per cent change: 21.03% (95% CI 12.65% to 29.43%)

NNT: 2 (95% CI 1 to 6)

Physical function

WOMAC function

Scale from: 0 to 100

(Higher scores mean more severe limitation)
Follow-up: mean 3 months
The mean change in physical function in the control groups was
1.7
The mean change in physical function in the intervention groups was
4.55lower
(2.23 lower to 11.32 higher)
197
(3 studies)
⊕⊕⊝⊝
low2
MD 4.55 (95% CI -2.23 to 11.32)

Absolute risk difference: 4.55% (95% CI -2.23% to 11.32%)

Relative per cent change: 268% (95% CI -131% to 666%)

NNT: not statistically significant

Quality of life

SF-36 item

Scale from: 0 to 100

(Lower scores mean worse quality)

Follow-up: mean 16 weeks
The mean change in quality of life in the control groups was
2.4
The mean change in quality of life in the intervention groups was
0.09 lower
(0.36 lower to 0.54 higher)
145
(2 studies)
⊕⊕⊕⊝
moderate3
SMD 0.09 (95% CI -0.36 to 0.54)

Absolute risk difference: 1% (95% CI -2.92% to 4.37%)

Relative per cent change: 30.38% (95% CI -121.5% to 182.25%)

NNT: not statistically significant

Radiographic progression

Bone scintigraphic examinations

Follow-up: mean 2.5 months
See commentSee commentNot estimable78
(1 study)
See commentNo related data were available

Number of patients experiencing any adverse event

Follow-up: mean 1 month
167 per 1000195 per 1000
(120 to 320)
RR 1.17
(0.72 to 1.92)
288
(4 studies)
⊕⊕⊕⊝
moderate4
Absolute risk difference: 3% (95% CI -6% to 12%)

Relative per cent change:

17% (95% CI -28% to 92%)

NNT: not statistically significant

Number of patients who withdrew because of adverse events

Follow-up: mean 6 months
27 per 100024 per 1000

(2 to 376)
RR 0.90

(0.06 to 13.92)
78
(1 study)
⊕⊕⊝⊝
low5
Only 1 study: 1 participant withdrew from each group because of adverse skin reactions unrelated to the therapy

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: confidence interval; MD: mean difference; NNT: number needed to treat; RR: risk ratio; VAS: visual analogue scale;WOMAC: Western Ontario and McMaster Universities osteoarthritis index

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1Downgraded for moderate heterogeneity (I2 = 55%); unclear risk for random sequence generation (Zizic 1995), allocation concealment (Zizic 1995), blinding of outcome assessors (Fary 2011; Nelson 2013; Zizic 1995), selective reporting (all six studies) and high risk for incomplete outcome data (Zizic 1995).
2Downgraded for considerable heterogeneity (I2 = 84%); Zizic 1995: unclear risk for random sequence generation, allocation concealment, blinding of outcome assessors, selective reporting and high risk for incomplete outcome data. Fary 2011: unclear risk for blinding of outcome assessors and selective reporting. Garland 2007: unclear risk for selective reporting.
3Fary 2011: unclear risk for blinding of outcome assessors and selective reporting. Pipitone 2001: high risk for incomplete outcome data.
4Unclear risk for random sequence generation (Thamsborg 2005; Zizic 1995), allocation concealment (Zizic 1995), blinding of outcome assessors (Thamsborg 2005; Zizic 1995), selective reporting (all four studies) and high risk for incomplete outcome data (Garland 2007; Thamsborg 2005; Zizic 1995).
5Only Zizic 1995 reported this outcome. Downgraded for imprecision (wide confidence interval and few events); unclear risk for random sequence generation, allocation concealment, blinding of outcome assessors and selective reporting and high risk for incomplete outcome data.