Population: patients with OA of one or multiple joints who underwent PEMF therapy alone or in combination with other therapeutic modalities.
Intervention: studies reporting on the influence of PEMF alone or in combination with other modalities.
Outcome: studies reporting on the influence of PEMF or any outcome associated with OA.
Study designs including systematic reviews and meta-analyses of RCTs.
Design other than a systematic review (narrative reviews).
Unavailability of data to be extracted, in this case the corresponding author has been contacted. If no information was available from the corresponding author, the study was excluded.
Systematic reviews of observational studies.
Systematic reviews of non-clinical studies or animal model studies.
Full text articles in a language other than English or German.
Data extraction and critical appraisal
Author and year of publication
Participants; % female; Mean age
Duration of intervention and follow up
Magnetic field parameters
Quality assessment outcome (tool)
Outcome and outcome measures
Yang et al. (2020) 
Cochrane Central Register of Controlled Trials, PubMed, CINAHL, Embase, PEDro
Overall 16 studies included. Studies in English reporting on studies in adults with OA who received PEMF as primary treatment and reported on pain, stiffness, physical function and QOL
Adults with OA with a total population N = 1078 (554 treatment vs 524 controls). % female: NR.
Mean age: 59.5 years
PEMF alone, PEMF with routine PT
Sham PEMF, routine PT with sham PEMF, no treatment and medicine (analgesia when needed)
Duration of intervention (range): 10–36 days; Follow up (range): 10–84 days
Frequency 0.1 Hz‑6.8 MHz; intensity varying from 10–80 Gauss; 7.8 × 10–8 to 1.5 × 10–7; 2–30mT, 40–105 mcT
Knee (n = 14); ankle (n = 1); hand (n = 2); cervical spine (n = 2)
Only 2 studies low risk of bias, other studies 14 unclear, 2 high risk (Cochrane, GRADE)
Pain (n = 15): VAS or WOMAC pain subscale; Stiffness (n = 7): WOMAC stiffness subscale; Physical function (n = 8): Lequesne index or WOMAC function subscale; QOL (n = 3): EuroQoL or SF-36 scale
PEMF was beneficial in pain reduction regardless of treatment duration, field intensity or frequency. Significant improvements were also seen in stiffness and physical function, albeit for the latter only for duration between 4 and 6 weeks. No differences were seen for QOL. Using MCID results suggest effects on pain reduction are clinically relevant
Viganò et al. (2020) 
MEDLINE, Embase, Web of Science, Cochrane Database
Overall 13 studies included. Studies in English reporting on patients with knee OA treated with PEMF that reported pain (VAS) and disability/activity (WOMAC scale)
Adults with knee OA with a total population N = 914 (472 treatment vs 442 controls). % female: NR
Mean age: NR
PEMF alone, PEMF with PT, SW, TENS and PEMF with ESWT and HP
9 placebo with inactive PEMF device, 4 different combination of other therapeutic modalities (PT, TENS, SW)
Duration of intervention (range): 14–42 days; with some studies reporting 1–30 sessions range between 30 min to 1 h duration. Follow up (range): 0–24 weeks
Frequency between 1–3000 Hz; intensity 3.4 mcT–105 mT, 0.5–30 Gauss, 34 V/m–100 V/cm
3 studies low risk, 5 studies unclear, 5 studies high risk (Cochrane)
Pain (n = 13): VAS, disability associated with knee OA (n = 6): WOMAC index
PEMF provided significant pain reduction and WOMAC disability score in knee OA patients in placebo controlled trials. Subgroup analyses showed no difference between PEMF and other theraputic modalities. PEMF is a safe therapeutic option but improvements are similar to other therapeutic modalities (PT, TENS, hyperthermia and ultrasound)
Chen et al. (2019) 
PubMed, Embase, Web of Science, Cochrane Library
Overall 8 studies included. Studies in English comparing patients with knee OA treated with PEMF, that reported pain, stiffness, and physical function (WOMAC total, WOMAC stiffness, WOMAC physical function, VAS score)
Adults with OA with a total population N = 421 (252 IG vs. 224 CG; there were inconsistency in the results table, authors not available for comment). % female: NR.
Mean age: Treatment group range: 55.5–69.2 years; placebo group range: 55.5–67.0 years
PEMF vs placebo
Duration of intervention (range): 18–45 days; from 5 min 2x/day up to 12 h/day; 1 study reported 18 sessions. Follow-up not reported
Frequency: 1 Hz—6.8 MHz; intensity from 40 mcT to 105mT; one study reporting 34 ± 8 V/m and one 10 mV/cm
All studies low or moderate risk of bias (Cochrane)
Pain: VAS (n = 3), Pain + function: WOMAC (n = 4); VAS + WOMAC (n = 1)
PEMF had a statistically significant positive effect on physical function as measured by the WOMAC physical function score, compared to placebo. There was no improvement in stiffness or pain. PEMF may be a useful and cost-effective addition to non-interventional treatment of knee OA
Wu et al. (2018) 
PubMed, Embase, Web of Science, Cochrane Library
Overall 12 studies included: 10 knee OA, 2 cervical OA, 1 hand OA
Knee N = 634; Cervical N = 115; Hand N = 50; % female: Knee: 8–87.9% (PEMF), 11.5–88.2% (placebo); Cervical: 28.6–64.7% (PEMF), 30.8–66.7% (placebo).
Mean age: Knee: 57.7–68.6 (placebo), 52.1–69.2 (PEMF); Cervical: 43.2–61.2 (PEMF), 42.1–67.4 (placebo)
Knee: 7x PEMF, 1x PEMF + hot pack + TENS, 1x PEMS + standard; Cervical: 1x PEMF, 1x PEMS + regimen; Hand: PEMF + AROM + resistive exercise
Knee: 3–6 weeks, 1 × 20 sessions. Cervical: 3–5 weeks. Hand: 10 days. Dailty time between 10 min × 3 times a day and a minimum of 12 h
Knee, cervical spine, hand
1 study low quality (Cochrane)
Pain (unclear which scale, change from baseline), physical function (WOMAC, SF-36 social function score, SF-36 global score, or physician global assessment score), adverse events
PEMF proved beneficial for pain reduction and function improvement for knee and hand OA, but not for cervical OA. Overall, a treatment duration of less than 30 mins may be more effective in reducing pain and improving function
We et al. (2013) 
MEDLINE, Scopus, Cochrane Central Register of Controlled Trials
Overall 14 studies included with knee OA patients; all were placebo-controlled RCTs
N = 930 (482 vs. 448 placebo); % female: 9.8–100%. Mean age: 60.0–73.0 years
Various protocols between 2–6 weeks in duration (one study with 20 session no duration in days indicated); time reported between 6–30 min duration, between 1–8 times a day
Pulse frequency between 1–3000 Hz, intensity: 0.034–69 Gauss, Pule length between 10 mcs and 6 ms
5 trials low quality (score < 6), 9 high quality (PEDRo, Jadad scale)
Pain and function
No evidence that PEMF was more effective in treating knee pain but was more effective than placebo in improving knee function after 8 weeks following treatment initiation. High quality studies, do however, also suggest effects on pain improvement
Hug, Roosli (2011) 
PubMed, Embase, ISI Web of Knowledge, Cochrane Library
4 PEMF RCTs; 3 on knee OA, 1 trial on cervical OA. All studies were double-blind placebo-controlled RCTs
N = 255 for OA (131 IG vs. 124 CG); n = 223 for knee OA (114 IG vs 109 CG) and n = 32 for cervical OA (17 IG vs 15 CG); % female: Knee OA: 49–80%, cervical OA: 66%. Mean age: Knee: 25.2–68, Cervical spine: 42.5 years
Whole-body PEMF, whole-body PEMF + intraarticular steroid injections
3–6 weeks (16–30 min/day), 1x only a single session
Intensity from 3.4–105 mcT, frequency from 0.1–3000 Hz (not individually reported)
Knee, cervical spine
Knee: VAS-Score, Lequesne index, KSS, WOMAC OA index and ist subscores, sensory and pain thresholds. Cervical: pain, NPDS, ROM, cervical muscle spasm
Evidence for an effect of whole-body PEMF is insufficient and can thus not be recommended. No short-term side-effects, long-term effects were not examined
Vavken et al. (2009) 
PubMed, EMBASE, Cochrane Controlled Trials Register
9 RCTs, comparison of PEMF with placebo
Adults with knee OA (N = 483); n = 239 (IG), n = 244 (CG); % female: 35–91% (IG), 20–72% (CG). Mean age: 58.1–72.7 (IG), 58.3–73.3 (CG) years
PEMF alone (pulsed short wave and classical PEMF)
Duration (range): 2–8 weeks; from 15 min 3x/week up to 2 h/day; 1 study reported 18 sessions. Follow up: 4–12 weeks
1 Hz—27 MHz, 3.4 mcT—2.5 mT
Jadad 5/5 (Jadad)
Pain (n = 9): VAS, stiffness (n = 4), activities of daily living (n = 5), clinical scores (n = 4, WOMAC Index, Arthritis Impact Measurement Scale AIMS), end-points abstracted for a time-point as close as possible to 6 weeks of follow-up
No significant effects were reported for pain or stiffness, however there was evidence to support improvements in clinical scores in patients with knee OA. PEMF could be considered as adjuvant therapy modality for knee OA
Bjordal et al. (2007) 
Medline, Embase, Cochrane Controlled
Trials Register for RCTs, CINAHL, Database of
Abstracts of Reviews of Effectiveness, International
Network of Agencies for Health Technology Assessment database, PEDro, National Guideline Clearinghouse, PRODIGY Guidance, NICE, hand search
7 PEMF RCTs, placebo-controlled, knee OA verified by clinical examination according to ACR criteria and/or X‑ray with pain > 3 months
N = 487; Adults with clinical and radiological confirmation of knee OA; n = 255 (IG), n = 232 (CG); % female: NR. Mean age: 64.2 years
PEMF (short wave therapy (n = 1) and other PEMF (n = 6))
Short wave therapy: duration NR, 3x/week over 2 weeks; PEMF: 0.5–2 h/day over 6 weeks or 0.5 h 3–5x/week total 18 sessions; one study reported only the performance of 8 sessions without further information
Short wave therapy: 400 Hz, treatment dose 20 k; PEMF: 1–3000 Hz, intensities: 3 × 10–7Gauss, < 0.5–15 Gauss, 10 mV, 40 mT
Pain reduction (VAS pain scale, WOMAC Pain Scale) during the first 4 weeks after initiation of treatment and follow up at 1–12 weeks after end of treatment; global health status 1–12 weeks after end of treatment
Data overall scarce, however some evidence that PEMF offered small improvement in pain outcomes after 4 weeks of therapy, with conflicting results based on different time points. Therefore, it is difficult to provide conclusions on the duration of pain reduction. On patient in one study reported increased pain during treatment and withdrew
McCarthy et al. (2006) 
MEDLINE, AMED, EMBASE, HealthSTAR, CINAHL, PEDro, SPORTDiscus, Cochrane Controlled Trials Register (CCTR)
5 RCTs comparing PEMF with placebo; studies using non-validated outcome measures were excluded
Adults with clinical and radiological confirmation of knee OA; N = 276; n = 138 (IG), n = 138 (CG); % female: NR. Mean age: NR
PEMF alone; low frequency PEMF (n = 2), “pulsed short wave” high frequency PEMF (n = 3)
Duration (range): 2–6 weeks; treatment duration: 3–5 h/week (low frequency); high frequency PEMF: NR; follow up period: NR
Low frequency PEMF: 3–50 Hz, high frequency PEMF: NR; intensity: NR
Pain (n = 5, VAS Pain Scale, WOMAC Pain Scale), functional disability (n = 4, WOMAC Physical Function Scale, AIMS)
Low level of evidence that PEMF provides a substantial contribution to the management of knee OA, with no effects on pain improvement. Some trends point to low frequency and high duration of treatment as effective in improvement of the WOMAC function score
Li et al. (2013) 
The Cochrane Central Register of Controlled Trials (CENTRAL), PreMEDLINE, MEDLINE, CINAHL, PEDro, handsearch
9 RCTs, placebo-controlled, treatment duration ≥4 weeks; no language restrictions
Adults with clinical and/or radiological diagnosis of OA (ACR criteria), trials with previous surgical treatment of OA excluded; N = 636, n = 327 (IG), n = 309(CG); % female: NR. Mean age: NR
Electromagnetic field interventions (PEMF (n = 6) and pulsed electrical stimulation (n = 3))
PEMF: 0.3–1.5 h/day or 0.5 h 3–5x/week for 4–6 weeks; Pulsed electrical stimulation: 6–14 h/day for 4–26 weeks
PEMF: 1 Hz—6.8 MHz, intensities: 40 mcT—105 mT, 34 ± 8 V/m, 10 mV/cm pulsed electrical stimulation: 100 Hz
Knee (n = 7), OA in general (n = 1), knee, cervical spine (n = 1)
Inadequate reporting of study design and conduct (n = 9), high risk of bias for incomplete outcome data (n = 3). Overall risk of bias was low for the other domains across the 9 studies
Pain (n = 6, VAS Pain Scale), physical function (n = 3, WOMAC Physical Function Scale), health-related quality of life (n = 2, SF-36), radiographic joint structure changes (n = 1, bone scintigraphic examination), number of patients experiencing adverse events (n = 4), number of patients who withdrew because of adverse events (n = 1), number of patients experiencing any serious adverse event
Some evidence to support a moderate benefit for OA patients in terms of pain reduction, no conclusive evidence to support improvements in physical functioning or to general health and well-being
Conclusion and implications for future research
Further improve the quality of future studies, for example by aiming for a more meticulous study design and by ensuring proper blinding and randomization in larger and better defined samples, in order to further improve the quality and level of evidence for the use of PEMF in patients with OA.
Conduct future trials with homogeneous outcome assessment (to enable future meta-analysis).
Achieve an international consensus on the uniform reporting of the magnetic flux density of the applied electromagnetic fields, such as microtesla/millitesla or Gauss, in order to be able to better compare study protocols.
Standardize additional therapeutic modalities, such as physiotherapy, hyperthermia, TENS, or ultrasound if these modalities are used in conjunction with PEMF to enable meaningful comparisons between groups.
Provide sufficient information on the treatment protocol (e.g. frequency, intensity, waveforms, treatment duration) and on therapy adherence.
Evaluate the optimal type, frequency, intensity and duration of PEMF interventions in order to develop standardized protocols. It can make sense to homogenize interventions according to the particular physical parameters of the applied electromagnetic fields as well as according to the duration of treatment and treatment indication.
Evaluate the effect of PEMF on osteoarthritic conditions other than the knee, for example in patients with coxarthrosis
Continue to evaluate the safety of PEMF interventions (especially when high-intensity protocols are used over a long period of time)
Evaluate a shorter duration of the electromagnetic fields in RCTs, as there is limited evidence that they affect cellular changes. Similarly, evaluate protocols using high-intensity magnetic fields in the millitesla range that allow sufficient penetration of body tissues as they are likely to produce a stronger physiological response.