Journal of Rehabilitation Medicine 51-11 | Page 9

Pulsed electromagnetic field therapy for knee osteoarthritis Records identiffied through database searching (n=241) Additional records identiffied through other sources (n=0) Records after duplicates removed (n=163) Records excluded by title and abstract (n=149) Records screened (n=163) Full-text articles assessed for eligibility (n=14) Studies included in qualitative synthesis (n=8) Full-text articles excluded, with reasons: Not classical PEMF:1 Not knee OA separately:1 Data unextractable:4 (n=6) Studies included in quantitative synthesis (meta-analysis) (n=8) Fig. 1. Study flow diagram of search results and study selection. tent of the assessment consisted mainly of the following items: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. There were 3 evaluation options for each item: low risk, high risk, and unclear, based on the original research. A third reviewer participated in the assessment when differences could not be resolved through discussion. Statistical analysis Weighted mean differences (WMD) and 95% CIs were used to estimate the pooled effects for continuous outcomes, including VAS for pain and WOMAC scores. In addition, Cochrane’s Q 823 test and the I 2 test were used to evaluate heterogeneity among all studies. When the I 2 test value exceeded 50%, indicating high heterogeneity, a meta-analysis was performed using a random effects model, whereas a fixed effects model was chosen when the I 2 test value was 50% or less. In addition, sensitivity analysis was performed by eliminating each study in turn so as to estimate the effect of a single study on overall heterogeneity and pooled effects. Publication bias was not performed, due to the limited number of studies (n < 10) included in this analysis. p < 0.05 was defined as a significant difference, and all statistical analyses were performed by RevMan version 5.3. RESULTS Search results A total of 241 studies were retrieved through searching the electronic databases listed (Fig. 1). No additional records were identified from other sources. Seventy-eight duplicate studies were excluded initially, and a further 149 studies were exclu- ded after reading the title and abstract. This left 14 studies that needed to be further screened by reading the full text. As a result, 6 studies were excluded due to unextractable data, non-classical PEMF therapy, or knee OA not reported separately. A final total of 8 studies was included in this meta-analysis (15–22). Study characteristics The baseline characteristics, PEMF therapy parame- ters, and treatment regimen of the 8 included studies, with a total sample size of 421 patients, are shown in Table I. These studies were all randomized placebo- controlled trials involving patients of similar age, sex ratio, and body mass index, which were published between 1994 and 2016. There were some differences in the parameters and treatment regimen of PEMF therapy among all studies. In addition, 3 studies re- Table I. Baseline characteristics, PEMF parameters and treatment regimen of included studies PEMF therapy group Placebo group Disease BMI, duration, BMI, Age, years Female, kg/m 2 years Age, years Female, kg/m 2 n Mean (SD) n Mean (SD) Mean (SD) n Mean (SD) n Mean (SD) Disease duration Year Treatment Mean (SD) Frequency Intensity regimen Bagnato et al. 2016 (15) Wuschech et al. 2015 (16) Nelson et al. 2013 (17) Ay & Evcik 2009 (18) 30 68.6 (11.9) 21 27.7 (4.6) 12.4 (9.1) 30 66.9 (10.0) 22 27.1 (4.1) 11.9 (7.4) 44 63.4 (12.1) 15 N/A N/A 13 55.5 (10.8) 5 N/A N/A 15 55.5 (2.5) N/A 33.5 (1.9) N/A 19 58.4 (2.5) N/A 34.7 (1.7) N/A 30 58.9 (8.8) 21 N/A 3.6 (4.6) 25 57.7 (6.5) 19 N/A 3.5 (4.1) Thamsborg et al. 2005 (19) 42 60.4 (8.7) 20 27.0 (4.0) 7.5 (5.2) 41 59.6 (8.6) 25 27.5 (5.7) 7.9 (7.7) N/A N/A 17 67.0 (7.0) 8 N/A N/A N/A N/A 35 64.0 7 N/A N/A N/A 7.4 (6.7) 44 65.8 (11.7) 31 N/A 8.1 (8.0) Study Nicolakis et al. 15 69.0 (5.0) 11 2002 (20) Pipitone & 34 62.0 12 Scott 2001 (21) Trock et al. 42 69.2 (11.5) 29 1994 (22) 1 MHz 9.5 mT 12 h a day, daily, 4 weeks 4–12 Hz 105 mT 5 min twice a day, daily, 18 days 6.8 MHz 34±8 V/m 15 min twice a day, daily, 6 weeks 50 Hz 105 uT 30 min/session, 5 sessions/week, 3 weeks 50 Hz 10 mV/cm 2 h/session, 5 sessions/week, 6 weeks 1–3,000 Hz 40 uT 30 min twice a day, daily, 6 weeks 3–7.8 Hz 50 uT 10 min 3 times a day, daily, 6 weeks 5–12 Hz 1–2.5 mT 30 min/session, 3–5 sessions/ week, 18 sessions SD: standard deviation; BMI: body mass index; N/A: Not applicable. J Rehabil Med 51, 2019