Journal of Rehabilitation Medicine 51-11 | Page 10

ported the outcome of pain using a VAS score, while
4 studies reported joint pain using a WOMAC score,
as well as providing the outcomes of joint stiffness and
physical function, and one other study used VAS score
and WOMAC score together to report the outcomes.
Quality assessment
L. Chen et al.
824
Randomization was unclear in one study, and allocation
concealment was unclear in 5 studies. With regards to
blindness, one study was unclear with regard to detec-
tion bias, while all the other studies showed low risk
with regard to performance bias and detection bias.
In addition, all studies showed low risk in terms of
attribution bias and reporting bias, but were unclear
for other bias. Overall quality assessment indicated
that all included studies had a low or moderate risk of
bias (Fig S1 1 and Fig. S2 1 ).
Meta-analysis
WOMAC total score. Four studies reported the WOMAC
total score of 218 patients (Fig. 2A). Meta-analysis
was performed using the fixed effects model due to the
lack of heterogeneity among studies (χ 2  = 1.14, df = 3,
I 2 =0%, p = 0.77). Pooled results showed that there was
no significant difference in WOMAC total score bet-
ween the PEMF treatment group and the placebo group
(WMD = −7.80, 95% CI −16.08 to 0.47, p = 0.06).
WOMAC pain score and VAS pain score. As shown in
Fig. 2B and 2C, there were 5 studies with data from
a total of 301 patients that were used to evaluate the
WOMAC pain score through a fixed effects model for
no significant heterogeneity (χ 2  = 4.33, df = 4, I 2  = 8%,
p = 0.36); no significant effect on joint pain was ob-
served in the PEMF treatment group (WMD = −1.06,
95% CI −2.30 to 0.17, p = 0.09). In addition, data from
5 studies, involving 233 patients, were used to analyse
VAS pain score by random effects mode. The pooled
results showed that the VAS pain score was not notably
reduced in the PEMF treatment group compared with
the placebo group (WMD = −0.88, 95% CI −2.06 to
0.31, p = 0.15) and also show significant heterogeneity
(χ 2  = 32.01, df=3, I 2 =91%, p < 0.00001).
WOMAC stiffness score. Analysis of the WOMAC
stiffness score was achieved through 5 studies involv-
ing 301 patients. There was no significant difference in
the pooled results between the PEMF treatment group
and the placebo group, which demonstrated that PEMF
therapy had no advantage in improving joint stiffness
(WMD = −0.50, 95% CI −1.09 to 0.09, p = 0.1, Fig.
http://www.medicaljournals.se/jrm/content/?doi=10.2340/16501977-2613
1
www.medicaljournals.se/jrm
2D). In addition, a fixed effects mode was used, which
indicated no heterogeneity (χ 2  = 2.17, df = 4, I 2  =  0%,
p = 0.7, Fig. 2D).
WOMAC physical function score. WOMAC physical
function score was reported by 5 studies involving 301
patients. Pooled results showed that the WOMAC func-
tion score in the PEMF treatment group was signifi-
cantly lower than in the placebo group (WMD = −5.28,
95% CI −9.45 to −1.11, p = 0.01, Fig. 2E). Since no
significant heterogeneity was found (χ 2  = 0.86, df = 4,
I 2  = 0%, p = 0.93, Fig. 2E), the fixed effects model was
used to perform this analysis.
Sensitivity analysis
Analysis of WOMAC total score, WOMAC pain
score, WOMAC stiffness score and WOMAC physical
function score showed no significant statistical hetero-
geneity; thus it was not necessary to perform sensitivity
analysis for these parameters. Nevertheless, there was
a significant heterogeneity in the analysis of VAS pain
score. As a result, significant heterogeneity remained
after sensitivity analysis was conducted by eliminating
each study individually, but this heterogeneity was not­
ably reduced if the study published by Ay & Evcik (18)
was omitted (χ 2  = 4.25, df = 2, I 2  = 53%, p = 0.12, Fig. 3).
Moreover, pooled results showed that there was also
a significant difference between the PEMF treatment
group and the placebo group if the study published by
Ay & Evcik (18) was omitted (WMD = −1.47, 95% CI
−2.14 to −0.80, p < 0.0001, Fig. 3).
DISCUSSION
Knee OA is currently one of the most common
chronic joint diseases and often causes joint pain, joint
dysfunction, and even disability. PEMF therapy is an
accepted physical therapy and is an effective method
for the treatment of various pathological conditions
and diseases, especially in trauma, orthopaedics and
rheumatology (23). Although PEMF therapy is not
recommended for treating OA by the American College
of Rheumatology due to the lack of clinical studies, it
has become popular among patients with knee OA in
recent years (18). In the past 2 decades, PEMF therapy
has been used more and more frequently for OA (24).
Moreover, current clinical trials have yielded different
results regarding the efficacy of PEMF therapy in the
treatment of knee OA, including some systematic re-
views (8–12). Consequently, it is necessary for us to
verify the efficacy of PEMF therapy in the management
of joint pain, joint stiffness, and physical function
for patients with knee OA. In this study, an effect
on improving physical function was observed, while