Effects of Whole-body Electromyostimulation on knee pain and physical function in adults with symptomatic knee osteoarthritis: a randomized controlled trial

Abstract

In a randomized, controlled study, whole-body electromyostimulation (WB-EMS) was investigated as a promising alternative to conventional strength training for the treatment of knee osteoarthritis (OA). 72 overweight participants with symptomatic knee OA were randomly assigned to WB-EMS (n = 36) or a usual care control group (CG, n = 36). For seven months, the WB-EMS group received three times per fortnight a WB-EMS training, while the CG was prescribed 6x physiotherapeutic treatments. The primary outcome, change in the pain subscale of the Knee injury and Osteoarthritis Outcome Score (KOOS), significantly improved in favour of the WB-EMS group, with a mean increase of 16.7 points versus 7.0 points in the CG (absolute difference between groups 9.0 points, 95%CI 2.9 to 15.1, p = 0.004). Secondary outcomes, including the other KOOS subscales (symptoms, function in daily living, function in sports/recreational activities and quality of life), 7-day pain diary, isometric muscle strength and lower limb function (30s sit-to-stand test), were also in favour of WB-EMS. With few dropouts and no reported adverse events, WB-EMS had a participation rate of 88% ± 10%. Overall, WB-EMS was found to be effective in relieving knee pain symptoms and improving physical function in individuals with symptomatic knee OA compared to usual care treatment.

Full text

Effects of Whole-body Electromyostimulation on
knee pain and physical function in adults with
symptomatic knee osteoarthritis: a randomized
controlled trial
Stephanie Kast
University of Erlangen-Nuremberg
Wolfgang Kemmler
Universitätsklinikum Erlangen
Frank W. Römer
Universitätsklinikum Erlangen
Matthias Kohl
Furtwangen University
Adam G. Culvenor
La Trobe University
Ali Mobasheri
University of Oulu
Michael Uder
Universitätsklinikum Erlangen
Simon von Stengel
Universitätsklinikum Erlangen
Article
Keywords:
Posted Date: April 27th, 2024
DOI: https://doi.org/10.21203/rs.3.rs-4150052/v1
License:   This work is licensed under a Creative Commons Attribution 4.0 International License. 
Read Full License
Additional Declarations: No competing interests reported.

1
Effects of Whole-body Electromyostimulation on knee pain and 1
physical function in adults with symptomatic knee osteoarthritis: a 2
randomized controlled trial 3
Stephanie Kast1,2,*, Wolfgang Kemmler1,2, Frank W. Roemer2,4, Matthias Kohl3, Adam G. Culvenor5, Ali 4
Mobasheri6,7, Michael Uder2, Simon von Stengel2 5
1 Institute of Medical Physics, Friedrich-Alexander University Erlangen-Nürnberg, Germany 6
2 Institute of Radiology, University Hospital Erlangen, Erlangen, Germany 7
3 Department of Medical and Life Sciences, University of Furtwangen, Schwenningen, Germany 8
4 Department of Radiology, Boston University Chobanian & Avedisian School of Medicine, Boston, MA, 9
United States 10
5 La Trobe Sport and Exercise Medicine Research Centre, School of Allied Health, Human Services and Sport, 11
La Trobe University, Bundoora, VIC, Australia 12
6 Oulu University, Oulu, Finland 13
7 Department of Regenerative Medicine, State Research Institute Centre for Innovative Medicine, Vilnius, 14
Lithuania 15
* stephanie.kast@fau.de 16
17
Correspondence: 18
Stephanie Kast 19
Institute of Medical Physics, Friedrich-Alexander University Erlangen-Nürnberg 20
Henkestrasse 91, 91052 Erlangen 21
Tel: 09131-8525531 22
Fax: 09131-8522824 23
Email: stephanie.kast@fau.de 24
25
26
27

2
Abstract 28
29
In a randomized, controlled study, whole-body electromyostimulation (WB-EMS) was investigated as 30
a promising alternative to conventional strength training for the treatment of knee osteoarthritis 31
(OA). 72 overweight participants with symptomatic knee OA were randomly assigned to WB-EMS 32
(n=36) or a usual care control group (CG, n=36). For seven months, the WB-EMS group received three 33
times per fortnight a WB-EMS training, while the CG was prescribed 6x physiotherapeutic 34
treatments. The primary outcome, change in the pain subscale of the Knee injury and Osteoarthritis 35
Outcome Score (KOOS), significantly improved in favour of the WB-EMS group, with a mean increase 36
of 16.7 points versus 7.0 points in the CG (absolute difference between groups 9.0 points, 95%CI 2.9 37
to 15.1, p=0.004). Secondary outcomes, including the other KOOS subscales (symptoms, function in 38
daily living, function in sports/recreational activities and quality of life), 7-day pain diary, isometric 39
muscle strength and lower limb function (30s sit-to-stand test), were also in favour of WB-EMS. With 40
few dropouts and no reported adverse events, WB-EMS had a participation rate of 88% ± 10%. 41
Overall, WB-EMS was found to be effective in relieving knee pain symptoms and improving physical 42
function in individuals with symptomatic knee OA compared to usual care treatment. 43
44

3
Introduction 45
Knee osteoarthritis (OA) is a leading cause of global disability [1]. The individual burden and 46
socioeconomic impact of knee OA is profound and is expected to increase in the coming decades [2-47
4]. With no cure for OA currently, clinical guidelines emphasize treatments that relieve symptoms of 48
the disease and improve function, such as exercise, weight loss (for those overweight) and education 49
[5-7]. 50
Various exercise programs, such as resistance and endurance training, have a positive effect on pain 51
and function in knee OA [8]. In a recent systematic review, resistance training was effective in reducing 52
pain and/or improving function in daily living in 11 out of 12 studies (with a moderate to large effect 53
size) [9]. However, despite the high level of evidence regarding the benefits of physical activity and 54
exercise for knee OA, the majority of individuals with knee OA do not meet recommendations for 55
physical activity [10]. 56
In individuals with knee OA, a vicious cycle of pain, avoidance of physical activity, reduced muscle 57
strength and further functional limitations has been proposed [11]. As such, there can be barriers for 58
participation in resistance training to improve strength [12]. In contrast to conventional resistance 59
exercise, Whole-body Electromyostimulation (WB-EMS) is an approach characterized by intense 60
activation of muscles via an adjustable impulse delivered via surface electrodes with low voluntary 61
effort. This approach may be an attractive alternative for individuals with knee OA who may have an 62
inability to sufficiently voluntarily contract muscles to facilitate muscle strength gains and associated 63
symptomatic relief. In previous studies, WB-EMS has shown positive effects on muscle strength, 64
muscular morphology and fat mass in healthy, sarcopenic and/or functionally impaired participants 65
[13-19]. 66
The majority of existing EMS studies in individuals with knee OA concentrated on the effects of local 67
EMS. A systematic review by de Oliveira et al. [20] showed moderate evidence in favour of 68
neuromuscular electrical stimulation (NMES) alone or in combination with exercise for isometric 69
quadriceps strengthening. A recent meta-analysis by Carvalho et al. [21] reported insufficient evidence 70
on the effects of NMES combined with exercise compared to exercise alone on patient-reported 71
outcomes (e.g. pain). Due to the lack of comparability between studies (methodological differences, 72
e.g. study design, training protocol, type of stimulation), the evidence for NMES in individuals with 73
knee OA remains limited. 74
WB-EMS could have some advantages compared to local EMS. WB exercise increases overall physical 75
performance and may also exhibit positive systemic anti-inflammatory effects by activating large 76
muscle groups [22,23]. 77
The aim of this study was to compare the effects of a 7-months WB-EMS application to a usual care 78
control group (CG) in overweight individuals with symptomatic knee OA. Our primary hypothesis was 79
that WB-EMS will result in significantly greater reductions in knee pain compared to the usual care CG. 80
We further hypothesized that, compared to the CG, WB-EMS will result in significantly greater 81
improvements in self-reported function in daily living, recreational activities and quality of life, 82
quadriceps strength and physical function. 83

4
Method 84
Study design 85
The EMSOAT (Whole-Body Electromyostimulation for the Treatment of knee OA) study is a parallel-86
group (1:1 allocation) superiority randomized controlled trial (RCT) conducted at the Institute of 87
Medical Physics (IMP), Friedrich-Alexander University of Erlangen-Nürnberg (FAU), and the 88
Department of Radiology, University Hospital Erlangen Germany. The RCT was approved by the FAU 89
ethics committee (Nr. 352_20 B) and all participants provided written informed consent prior to 90
enrolment. The project fully complies with the Helsinki Declaration [24] and was prospectively 91
registered at clinicaltrials.gov, NCT05672264, on 05/01/2023. 92
Participants 93
Participants were recruited between March and June 2022 in the metropolitan area of Erlangen-94
Nürnberg, Germany. As in previous studies, we recruited potential participants by reports and expert 95
interviews on knee OA and corresponding study calls in local newspapers and social media. The call 96
listed the key study eligibility criteria, contact person and an email address. Furthermore, we contacted 97
eight medical practices (practitioners with qualification in sports medicine and orthopaedists) via letter 98
and provided information flyers for their patients. 99
Inclusion criteria were (1) men or women 40-70 years of age, with (2) overweight (BMI>25 kg/m2), (3) 100
confirmed femorotibial OA equivalent to Kellgren-Lawrence grades (KL) 2 and 3 [25] (see explanation 101
below), (4) knee pain for at least 3 months, (5) pain in the last 30 days at least on 50% of the days and 102
(6) an average pain intensity > 2.5 [26] on a scale 0-10 (NRS). 103
Exclusion criteria were: (1) Any WB-EMS training or more than 60min of resistance exercise training 104
per week in the last year, (2) glucocorticoid or opioid medication, (3) trauma to the knee joint within 105
the last 3 months, (4) intra-articular knee injection in the last 3 months, (5) conditions and diseases 106
(and corresponding medication) with relevant impact on study outcomes (i.e. other rheumatic diseases 107
e.g. rheumatoid arthritis, fibromyalgia, serious cardiovascular diseases), (6) conditions or diseases that 108
are contraindications for WB-EMS (e.g. electric implants, epilepsy, cardiac pacemakers [27]) and (7) 109
absence ≥4 weeks during the intervention period. 110
As radiographs could not be obtained for study purposes only [28], potential participants were asked 111
to provide externally acquired anterior-posterior radiographs of their index (more painful) knee when 112
available. These were assessed by an experienced musculoskeletal radiologist (FWR) and those with 113
KL 2 or KL3 were included [25]. Participants without externally acquired radiographs or radiographs 114
older than 2 years were screened by MRI and those with full-thickness cartilage damage at both the 115
femur and tibia in at least one compartment (grades 3.2 or 3.3 in at least one central femoral and one 116
subregion of the anterior, central and posterior tibial subregions on the MOAKS (MRI Osteoarthritis 117
Knee Score) [29] scale) were excluded. Also, those with no or only focal cartilage damage (maximum 118
of 1.0 or 1.1. in the 10 femorotibial subregions of the MOAKS instrument) were excluded. Using these 119
MRI definitions, the likelihood of including KL 0 and 1 knees or knees with end stage structural OA (KL4) 120
was minimized [30]. 121
If both knees of a single participant were eligible, we defined the side that caused more pain as the 122
“index limb” (affected knee). 123

5
Intervention 124
WB-EMS application 125
WB-EMS was applied using a system with medical device approval (miha bodytec®, Type II, Gersthofen, 126
Germany) that enables simultaneous stimulation of up to 10 main muscle groups (thighs and upper 127
arms, hip/bottom, abdomen, chest, lower back, upper back, latissimus dorsi and two free options) with 128
an overall area of stimulation of about 2600 cm2. The system allows intensities to be chosen for each 129
region. We established a consistently supervised, video-guided WB-EMS program 3 times per fortnight 130
(e.g. every Monday or Tuesday and every second Thursday or Friday) for 6 months (from August 2022 131
to January 2023) plus one month of conditioning (July 2022; see below). All participants started the 132
intervention at the same time. We used an impulse protocol that was applied in research 133
[14,15,17,18,31-33] and most commercial settings in order to allow transferability of our approach. 134
Bipolar electric current with a frequency of 85Hz, an impulse-width of 350 µs and a rectangular impulse 135
pattern was used for 20 minutes in an interval approach with 6 sec of EMS stimulation and 4 sec of 136
rest. Participants completed two sets with 6-8 repetitions of seven exercises (e.g. light dynamic 137
squatting with knee angles ≥ 120° and arm curls) in a standing position (Figure 1). Of importance, we 138
designed low-intensity movements/exercises to keep the effect of the voluntary movements itself as 139
low as possible. 140
141
The intensity of the EMS was regulated based on the rate of perceived exertion (RPE) scale. We applied 142
a perceived exertion rate to generate and maintain a sufficient but tolerable intensity of the EMS 143
Figure 1. WB-EMS training session (Written informed consent was obtained
from the participants to publish this picture)

6
application. Before the 6 months of WB-EMS training, we implemented 4 weeks of conditioning with 144
lower impulse intensity and shorter sessions (July 2022). We started with 12minutes in the first session 145
and increased time by 2 minutes per session. After conditioning, participants were encouraged to 146
exercise at an EMS-induced RPE of “6-7” (i.e. “hard+ to very hard”) on the Borg CR10 Scale [34]. Impulse 147
intensity was individually adapted for each body region in close interaction with the participant. During 148
the session, instructors slightly increased (impulse) intensity every 2-3 min in close cooperation with 149
the participants to maintain the prescribed RPE (“6-7”) during the session. From mid-September 2022, 150
all participants used a second pair of circular electrodes for the thighs, to adequately stimulate the 151
thighs and maintain the intensity. All training sessions took place in the Institute of Medical Physics. 152
We applied a personal training setting with one licensed and experienced instructor responsible for 153
two participants. Instructors monitored compliance with the prescribed exercise intensity and 154
recorded attendance rate accurately. In case of non-participation, participants reported absence by 155
email or telephone. Possible adverse events were recorded on a weekly basis during the entire course 156
of the study. Further, the international guideline of safe and effective WB-EMS application was strictly 157
respected [35]. 158
Control intervention (referral to physiotherapy) 159
The participants received a prescription for 6 physiotherapy treatment sessions (20 min each) with the 160
recommendation have those within the first three months at a frequency of 1x/week. Physiotherapy 161
treatment was carried out individually in the sense of "usual care" in a diagnosis-orientated manner. 162
The specific content was at the decision of the treating physiotherapists containing techniques and 163
exercises for reducing pain and detonisation of muscle tissue, increasing mobility of the knee joint and 164
strengthening leg muscles. It was recommended that the therapy be carried out in one of three co-165
operating practices. However, participants were free to take the prescription to another practice of 166
their choice. All practices were informed about the study and the aims of the study in a letter 167
accompanying the prescription. 168
Education (both groups) 169
Both groups were invited to participate in a training program for self-management of OA [36]. Six units 170
(60min each) were offered over a period of 12 weeks. Before each of the 6 sessions, an invitation with 171
a brief information was sent via email to the participants of both groups. The 6 sessions were led by 172
different experts, each of them was blinded to the group allocation. The aim of the program was 173
education, information and counselling to improve quality of life and mobility. Self-management, 174
personal responsibility and coping strategies of the participants to cope with bio-psycho-social (stress) 175
factors was promoted and supported. Overall, we intended to reduce fear and avoidance behaviour. 176
Outcomes 177
Primary outcome 178
• Changes in the pain subscale of the Knee injury and Osteoarthritis Outcome Score (KOOS-Pain) 179
from baseline to 7-month follow-up (FU) 180
Secondary outcomes 181
• Changes in the other four subscales of the KOOS over 7 months covering (a) symptoms, (b) 182
function in daily living, (c) function in sports/ recreational activities and (d) quality of life. 183
• Changes in knee pain intensity over 7 months as determined by a 7-day knee pain protocol 184
applying the numerical rating scale (NRS) [37,38]. 185

7
• Changes in maximum strength of the hip/leg extensors (“leg press”) over 7 months 186
• Changes in objective lower-limb function (30s sit-to-stand test) over 7 months 187
Exploratory outcomes 188
• Changes of total body-fat content and lean body mass over 7 months as determined by a direct 189
segmental multi-frequency bioelectrical impedance analysis (DSM-BIA) 190
• Changes in pain medication use as determined by 7-day knee pain protocol over 7 months 191
Outcome measures 192
Participants were requested to refrain from intense physical activity and exercise 48 hours before the 193
assessments. Baseline and FU assessments were consistently performed by the same research 194
assistant using the identically calibrated devices, in exactly the same setting and at about the same 195
time of the day (±90 min). 196
Knee pain diary and questionnaire 197
Knee pain and self-reported functional status was determined using the KOOS questionnaire [39,40] 198
which comprises five subscales (dimensions): pain, other symptoms, activities of daily living (ADL), 199
sports and recreation function (Sport/Rec) and knee-related quality of life (QoL). Each of these 200
dimensions is scored separately, using a Likert scale with five possible answers ranging from 0 (no 201
problems) to 4 (extreme problems). According to a formula, described in detail by Roos [39,40], scores 202
are transformed to a 0–100 scale, with zero representing extreme knee problems and 100 representing 203
no knee problems. 204
In addition to the KOOS subscale pain, the intensity of knee pain was monitored using a numerical 205
rating scale from 0 (no pain) to 10 (worst possible pain) [37,38] conducted over 7 days, before and 206
during the last week of the intervention. We provided standardized logs and requested the participants 207
to rate their highest daily knee pain intensity every evening. The average 7-day pain intensity at 208
baseline and FU was included in the analysis. Additionally, participants were asked to record pain 209
medication daily in their logs. Average numbers of days using analgesics during the 7-day periods of 210
monitoring were included in the analysis. 211
Lastly, we asked all participants in a baseline questionnaire for demographic parameters, diseases, 212
medication and confounding lifestyle factors (physical activity, exercise and nutrition). The follow-up 213
questionnaire specially addressed changes of this parameters in order to detect factors that may 214
confound our results. 215
Functional testing 216
Maximum isokinetic hip-/leg-extension strength was tested using a linear isokinetic leg press (CON-217
TREX LP, Physiomed, Laipersdorf, Germany). Maximum strength was measured unilateral on the index 218
limb (affected knee). Participants were sitting in a slightly supine (seatback 55°) position, fixed by hip 219
and chest straps. Using the standard velocity of 0.5 m/s, range of motion was within 30° to 90° knee 220
angle. After briefing and one familiarization trial with low effort, participants were requested to 221
conduct two sets of five repetitions each with maximum voluntary effort (“push as strongly as 222
possible”) separated by 60 s of rest. The highest force value of the two trials was included in the 223
analysis. The present protocol has been applied in prior studies (e.g. [15,16,41,42]). 224
In order to determine the physical function of the lower extremities (objective lower-limb function), 225
the 30-second sit-to-stand test (“Chair Rise Test”) was used, which is a recommended performance-226
based test in individuals with knee OA [43]. With arms folded across their chests, participants were 227

8
instructed to complete as many sit-to-stand movements as possible from a chair within 30s. Knees and 228
hips had to be extended in the standing position, while the buttocks had to touch the seat in the lower 229
position. Following a demonstration by the tester, a practice trial of one repetition was given to check 230
proper form, followed by the 30s test trial. We did not adjust the seat height for lower extremity 231
length. The same standard chair was used for all assessments [44,45]. 232
Anthropometry 233
Body mass and composition was determined through direct-segmental, multi-frequency Bio-234
Impedance-Analysis (DSM-BIA; InBody 770, Seoul, Korea). This device measures impedance of the 235
trunk, arms and legs separately using an eight-point tactile electrode system that applies six 236
frequencies between 1 and 1000 kHz. 237
Sample size calculation 238
The sample size analysis was based on the primary endpoint of KOOS-Pain. Since there is a lack of data 239
on the effect of WB-EMS in OA, the power analysis was based on the effects of conventional strength 240
training on pain in knee OA. In the meta-analysis by Goh et al. [46], a sub-analysis (89 studies; n = 7184) 241
on the effect of strength training compared to "usual care" showed an SMD of 0.73 (0.49 - 0.98). With 242
a power of 80% and an -level of 5%, a two-sided t-test results in a required number of cases of n = 243
31/group. Since the meta-analysis of Goh et al. included predominately passive control groups, while 244
our study implemented a usual care control group (6 physiotherapeutic sessions), we designed our 245
sample size analysis more conservatively by increasing the number of cases by 15% which is equivalent 246
to assuming an SMD of 0.67. Correspondingly, we aimed to include 36 subjects per group (WB-EMS: 247
n=36, CG: n=36). 248
Randomization and blinding 249
Using two strata for pain intensity (NRS, assessed as inclusion criteria), the 72 eligible participants were 250
allocated to the study groups based on drawing small opaque capsules placed in a bowl. In detail, 36 251
capsules of WB-EMS and 36 capsules of CG were put in the bowl, prepared by a researcher not involved 252
in the trial. Thus, neither participants nor researcher knew the allocation beforehand (allocation 253
concealment). After the randomization procedure, the principal investigator (SK) registered 254
participants and instructed them in detail about study specifications. 255
Our blinding strategy focused on research assistants who assessed the outcome parameters and were 256
kept unaware of the participants’ group status (WB-EMS or CG) and were not allowed to ask, either. 257
Statistical analysis 258
Intention to treat (ITT) analyses were applied. Multiple imputation (ITT) was performed using R 259
statistics software (R Development Core Team Vienna, Austria [47]) in combination with Amelia II [48]. 260
We used the full data set for multiple imputations, with imputation repeated 100 times. Over 261
imputation diagnostic plots (“observed versus imputed values”) were checked by Amelia II. For 262
pooling, the results R package mice [49] was used. Additionally, we applied per protocol (PP) analyses 263
for all participants with complete datasets (baseline and 7-months assessment), independent of their 264
compliance, for all the primary and secondary study outcomes. The results of PP and ITT analyses were 265
similar and identical with respect to significances. Assumptions, such as normal distribution, were 266
checked graphically (qq-plots, residual plots). The changes over time within groups were analysed by 267
paired t-tests. The group differences at follow-up (”effects“) were determined by ANCOVA, adjusting 268

9
for baseline data using the group as covariate. Categorical variables were addressed using the Chi-269
Square test. Differences in use of pain medication (yes vs no) were determined by a two-way Analysis 270
of Deviance (logistic regression) using the likelihood-ratio-test. All tests were 2-tailed and significance 271
accepted at p <0.05. According to the suggestion of Li et al. [50], we did not adjust secondary outcomes 272
for multiplicity. Standardized Mean Difference (SMD) according to Cohen (Cohen’s d) [51] was also 273
calculated to indicate the size of the effect for primary and secondary outcome variables. SMDs ≥0.2, 274
0.5 and 0.8 represent small, medium and large effect sizes. 275
Results 276
A total of 440 women and men responded by email or telephone. After sending detailed study 277
information via email, potential participants were further assessed for eligibility by phone calls. Of the 278
remaining 113 participants, 12 were unwilling to be randomly assigned to the groups, 6 were unwilling 279
to attend MRI and 23 declined to participate for other reasons. Finally, 72 participants could be 280
included in the study. Participant flow through the study is displayed in Figure 2. 281
282
Table 1 lists the baseline data for the two groups. Of the 72 subjects randomized, 4 subjects were lost 283
to FU for reasons unrelated to the study (CG: n=1; WB-EMS: n=3) (Fig. 2). Two participants of the WB-284
EMS group quit the intervention. One of these persons quit the trial after 11 weeks of training because 285
Figure 2. Study flow diagram (according to CONSORT, Consolidated Standards of Reporting Trial)

10
of orthopaedic problems unrelated to the exercise program. The second person quit after 5 months of 286
training because of personal reasons. 287
Please add Table 1 about here.
288
On average, participants attended 88% ± 10% of WB-EMS sessions (3 times per fortnight) over the 289
period of 7 months (including condition). In most cases, the reason given for the absence was illness, 290
whereby three participants had longer periods (4-8 weeks) of inactivity due to viral infections. No 291
adverse or unintended effects or injuries were observed during the WB-EMS sessions, and no 292
participant reported any WB-EMS-related discomfort during or after WB-EMS application. More than 293
90% of the participants in the CG have redeemed the prescription with the 6 physiotherapy 294
treatments. The participation rate regarding the self-management program was around 50%. Both 295
groups participated equally. 296
Table 2 displays the results of primary and secondary outcomes. KOOS-Pain scores improved 297
significantly more (18.2% difference) in the WB-EMS group compared with the CG (mean difference 298
(MD) 9.0 points, 95% CI 2.9 to 15.1, p=0.004). In Detail, the score improved by 12.5% in CG (p=0.003) 299
and by 30.7% in the WB-EMS (p<0.001). Thus, we confirmed our primary hypothesis that 7 months of 300
WB-EMS application positively changes knee OA pain as assessed by KOOS-Pain subscale more than 301
control. 302
Please add Table 2 about here.
303
All secondary outcomes (other KOOS subscales, NRS, sit-to-stand test, muscle strength) also improved 304
significantly more in the WB-EMS group compared to the control group at 7-month FU (Table 2). More 305
in detail, in KOOS-Symptoms score there was a net benefit in favour of the WB-EMS group of 14,7% 306
(MD 8.6 points, 95% CI 2.8 to 14.4). The result for KOOS-ADL score was similar: WB-EMS improved the 307
score by 16.2% compared to CG (MD 10.8 points, 95% CI 5.3 to 16.3). The fourth and fifth KOOS 308
dimensions Sport/REC and QoL also changed more favourably in the WB-EMS. The Sport/REC score 309
was 49.2% (MD 11.5 points, 95% CI 3.3 to 19.6) and the QoL score was 33.9% (MD 9.5 points, 95% CI 310
3.1 to 16.0) higher in the WB-EMS than in the CG. 311
In parallel, the average knee pain intensity (NRS), which was recorded via 7-day diary, decreased 312
significantly in WB-EMS by 25.3% compared to the CG (MD -1.04, 95% CI -1.75 to -0.33). The number 313
of “sit-to-stands” in 30s (Chair Rise) developed in favour of the WB-EMS compared to the CG (MD 3.9 314
reps, 95% CI 2.0 to 5.8). In line with the changes in sit-to-stand test, there was a significant between-315
group difference for change in maximum isokinetic hip/leg extensor strength (MD 79.0 N, 95% CI 6.9 316
to 151.2) favouring WB-EMS group. 317
Table 3 displays the results of the exploratory outcomes. In contrast to the results described above, 318
the WB-EMS program did not lead to a significant change or between-group differences in body 319
weight. With respect to body composition, lean body mass remained stable in WB-EMS, whereas it 320
significantly decreased (p=0.02) in the CG. The difference between the groups was non-significant 321
(p=0.09). CG significantly gained fat mass (Tab. 3), whereas the increase in fat mas in WB-EMS group 322
was not significant. Again, the between group difference were not significant (Tab. 3). 323
Please add Table 3 about here.
324
No significant between-group differences with respect to physical activity (p=0.106), exercise or diet 325
were reported. The weekly intake of analgesics, assessed via 7-day protocol, tendentially increased in 326
the CG (BL: 0.81±2.47; FU: 1.36±2.85) and decreased in the WB-EMS (BL: 0.64±1.33; FU: 0.32±1.36). 327
The intergroup difference was borderline non-significant (p=0.059). Of note, the number of subjects 328

11
who took oral analgesics, as determined via the 7-day protocol, was 8 in CG and 9 in WB-EMS at 329
baseline. At FU 10 participants in CG and 2 in WB-EMS used oral analgesics. After 7 month of 330
intervention a significant reduction of number of participants taking analgesics in the WB-EMS 331
compared to CG was observed (p= 0.033). 332
Discussion 333
In the present study, we examined whether a 7-month WB-EMS training program improves knee pain 334
and function in individuals with symptomatic knee OA. In summary, our findings demonstrated that 335
WB-EMS was highly effective in alleviating pain (KOOS) as primary outcome and improving the other 336
four KOOS scores. Along with the enhancement of the KOOS scores, WB-EMS was more effective in 337
improving pain intensity (NRS), objective lower-limb function (30s sit-to-stand) and maximum strength 338
of hip-/leg extensors compared to a usual care approach. 339
To our knowledge, only one other study investigated the effect of WB-EMS in individuals with knee OA 340
[22]. However, the pilot study of Park et al. included individuals with early knee OA (KL 1-2) and pain 341
was not an inclusion criterion. Accordingly, the baseline KOOS-Pain score in their study was on average 342
18 points higher compared to our study. The study of Park [22] also pursued a fundamentally different 343
approach: they examined the effectiveness of isometric strength exercise superimposed by WB-EMS 344
compared to the exercises alone and a passive control. Worth mentioning, the isometric exercises 345
alone showed an effect on maximum knee extension strength and the KOOS scores symptoms, ADL, 346
Sports/Rec and QoL compared to passive control. However, the WB-EMS application led to additional 347
effects. The KOOS scores for pain, symptoms and ADL were significant higher in the combined WB-348
EMS group compared to exercise alone [22]. 349
We pursued a low-threshold approach in which the muscles are activated predominantly via EMS while 350
performing light and less strenuous movements. This method might be attractive especially for the 351
large target group of people who are not willing or able (e.g. because of pain) to perform intensive and 352
strenuous strength training exercises. Following our philosophy of low barriers, the training frequency 353
was 3 sessions per fortnight, compared to 3 sessions per week in Park’s study. 354
All other studies that have investigated the effect of EMS – mostly the term neuromuscular electrical 355
stimulation (NMES) is used in literature – in knee OA have only used a local stimulation. The results of 356
two recent meta-analysis on the effect of local EMS in individuals with knee OA indicate an increase in 357
quadriceps muscle strength [52], but no significant reduction in pain [21,52]. 358
It has to be noted that WB-EMS is not comparable with local EMS. The difference is not just that WB-359
EMS stimulates all major muscle groups at the same time. By using cuff electrodes, agonists and 360
antagonists (e.g. quadriceps and hamstrings) are activated simultaneously over a large area. In most 361
of the local EMS studies, the quadriceps muscle was stimulated in isolation with adhesive electrodes. 362
This approach appears suboptimal, considering the importance of the hamstring muscles and 363
intermuscular and proprioceptive coordination for the stability of the knee joint [53]. Strengthening 364
the hamstring muscles in addition to strengthening the quadriceps muscles has even been shown to 365
be beneficial for pain symptoms, mobility and function in knee OA [54]. In our study, we combined 366
WB-EMS with dynamic functional movements because it leads to more pronounced effects on muscle 367
mass and function than static, passive WB-EMS [55]. In the majority of studies on local EMS, the 368
muscles are stimulated statically without movement or passively without movement and without 369
voluntary activation of the muscles. 370
We focussed on overweight participants, because overweight/obesity is a strong risk factor for the 371
development and progression of knee OA [3,56,57]. Study results suggest that not only the higher 372

12
mechanical stress is associated with obesity, but in particular the visceral fat with its pro-inflammatory 373
effect plays a role in the development and progression of OA [58]. In this context, it should be 374
mentioned that our WB-EMS program did not result in any significant intergroup differences in weight, 375
muscle mass and fat mass, even though an increase in fat mass and a decrease in LBM was recorded 376
within the CG. From this perspective, the effects of our WB-EMS training program on body composition 377
are rather small. Our WB-EMS approach was time-efficient and required only 30 minutes of training 378
per week. The low training volume was probably not sufficient to induce major body composition 379
changes. However, study results suggest that muscle activity is associated with the secretion of anti-380
inflammatory substances, which could be one mechanism of pain reduction [23,59]. There is some 381
evidence of positive effects of WB-EMS application on inflammatory biomarkers in elderly women with 382
early knee OA [22]. 383
The pain-relieving effect of WB-EMS could take place via different pathways. Another pathway could 384
be an improvement in knee joint stability and mechanics through an increase in muscle strength as we 385
observed in the study. Finally, the EMS current, which is a TENS current, may have contributed to the 386
effect [60]. 387
Our project has various strengths. Great emphasis was placed on the safety aspect. This refers to an 388
individual dosage and a slow progressive increase in intensity to ensure safety and adaptation of the 389
muscles. To achieve that, we conducted 1 month of conditioning with an initial lower intensity (i.e. 390
current intensity) and a shorter application duration to prepare the participants well for the WB-EMS 391
training. The aim of this method was to avoid high levels of creatine kinase (CK) after initial applications 392
[61]. Moreover, we wanted to ensure that the training sessions set over threshold stimuli for the whole 393
period of 6 months. After the initial phase, an RPE target of “6-7” on the Borg CR10 was used. Lastly, 394
the training was carried out by qualified trainers with a supervision ratio of 1:2 (trainer:participant) to 395
ensure a high level of safety through optimal assistance and monitoring. 396
We observed a high attendance rate (88%). Further it indicated that our exercise protocol was not only 397
effective but obviously attractive, even in this cohort with a low affinity to conventional resistance 398
training. The high attractiveness was confirmed by the low drop-out rate, as there were only 3 399
dropouts in the WB-EMS group (all were unrelated to the program). No participant showed intolerance 400
to electrical stimulation and no EMS related side effects were reported. 401
Apart from its effectiveness and safety, high importance was attached to generalizability and 402
transferability. We included a representative cohort of individuals with knee OA and we applied a WB-403
EMS protocol used in the majority of commercial settings. This ensures a good transferability of the 404
results and enables the findings to be applied more broadly using existing structures of commercial 405
providers. 406
In order to rule out the possibility of the use of pain medication distorting the results, we recorded the 407
medications as part of the pain diary. It was notable that the number of participants taking pain 408
medication significantly decreased in the WB-EMS group and the amount of medication taken 409
decreased tendentially, which excludes the possibility that the medication distorted the study results. 410
Some limitations of our trial should be noted. One limitation is that it was not blinded at participant 411
level. To be blinded, the CG would have had to receive the identical intervention as the training group, 412
with the difference that the WB-EMS devices would have provided electrical stimuli only below 413
motorical threshold. However, since low-threshold electrical stimuli, applied as transcutaneous 414
electrical nerve stimulation (TENS), showed pain-relieving effects in individuals with knee OA [60], we 415
did not use a blinded study design with low-intensity TENS, but pragmatically implemented a usual 416
care CG. In this context, it should be mentioned once again that the exercises performed during WB-417
EMS were designed in such a way that they should not lead to muscular adaptations. However, it 418
cannot be ruled out that the dynamic movements without electrical stimulation also had a pain-419
relieving effect. Our design does not allow us to separate the possible effects of WB-EMS and the 420

13
movements. Another limitation is that OA was not uniformly defined radiologically as an inclusion 421
criterion using the Kellgren-Lawrence score. Since, for reasons of time and economy, no application 422
was made to the Federal Office for Radiation Protection for the production of X-ray images, we 423
examined existing X-ray images and, if not available or too old, MRI images were taken. However, with 424
this procedure, the likelihood of including KL 0 and 1 knees or knees with end stage structural OA (KL4) 425
was minimized [30]. 426
According to various international guidelines [6,7,62], targeted physical training is a critical component 427
of the treatment of knee OA. In summary, we could show that 3 times per fortnight of WB-EMS 428
positively effects knee pain and function in individuals with knee OA. The effects in our study were at 429
least as pronounced as those in studies in which conventional strength training was used [46]. Due to 430
its time efficiency, low weight-bearing joint load and low subjective effort, WB-EMS has the potential 431
to reach the large target group of individuals with knee OA who are not receptive to physical training. 432
However, WB-EMS is an exclusive and more expensive form of training compared to conventional 433
training, which in turn restricts the target group. 434
Data availability 435
Data relative to this work will be available upon reasonable request to the corresponding author. 436
437

14
1. Duong, V., Oo, W. M., Ding, C., Culvenor, A. G. & Hunter, D. J. Evaluation and Treatment of 438
Knee Pain: A Review. Jama 330, 1568-1580, doi:10.1001/jama.2023.19675 (2023). 439
2. Hunter, D. J., Schofield, D. & Callander, E. The individual and socioeconomic impact of 440
osteoarthritis. Nature Reviews Rheumatology 10, 437-441 (2014). 441
3. Allen, K. D., Thoma, L. M. & Golightly, Y. M. Epidemiology of osteoarthritis. Osteoarthritis 442
Cartilage 30, 184-195, doi:10.1016/j.joca.2021.04.020 (2022). 443
4. Long, H. et al. Prevalence Trends of Site-Specific Osteoarthritis From 1990 to 2019: Findings 444
From the Global Burden of Disease Study 2019. Arthritis Rheumatol 74, 1172-1183, 445
doi:10.1002/art.42089 (2022). 446
5. Hochberg, M. C. et al. American College of Rheumatology 2012 recommendations for the use 447
of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee. 448
Arthritis Care Res (Hoboken) 64, 465-474, doi:10.1002/acr.21596 (2012). 449
6. McAlindon, T. E. et al. OARSI guidelines for the non-surgical management of knee 450
osteoarthritis. Osteoarthritis Cartilage 22, 363-388, doi:10.1016/j.joca.2014.01.003 (2014). 451
7. Brophy, R. H. & Fillingham, Y. A. AAOS Clinical Practice Guideline Summary: Management of 452
Osteoarthritis of the Knee (Nonarthroplasty), Third Edition. J Am Acad Orthop Surg 30, e721-453
e729, doi:10.5435/JAAOS-D-21-01233 (2022). 454
8. Fransen, M. et al. Exercise for osteoarthritis of the knee: a Cochrane systematic review. Br J 455
Sports Med 49, 1554-1557, doi:10.1136/bjsports-2015-095424 (2015). 456
9. Turner, M. N. et al. The Role of Resistance Training Dosing on Pain and Physical Function in 457
Individuals With Knee Osteoarthritis: A Systematic Review. Sports Health 12, 200-206, 458
doi:10.1177/1941738119887183 (2020). 459
10. Wallis, J. A., Webster, K. E., Levinger, P. & Taylor, N. F. What proportion of people with hip and 460
knee osteoarthritis meet physical activity guidelines? A systematic review and meta-analysis. 461
Osteoarthritis Cartilage 21, 1648-1659, doi:10.1016/j.joca.2013.08.003 (2013). 462
11. Pisters, M., Veenhof, C., Van Dijk, G., Dekker, J. & Group, C. S. Avoidance of activity and 463
limitations in activities in patients with osteoarthritis of the hip or knee: a 5 year follow-up 464
study on the mediating role of reduced muscle strength. Osteoarthritis and Cartilage 22, 171-465
177 (2014). 466
12. Dieppe, P. A. & Lohmander, L. S. Pathogenesis and management of pain in osteoarthritis. The 467
Lancet 365, 965-973 (2005). 468
13. Kemmler, W., Bebenek, M., Engelke, K. & von Stengel, S. Impact of whole-body 469
electromyostimulation on body composition in elderly women at risk for sarcopenia: the 470
Training and ElectroStimulation Trial (TEST-III). Age 36, 395-406, doi:10.1007/s11357-013-471
9575-2 (2014). 472
14. Kemmler, W., Grimm, A., Bebenek, M., Kohl, M. & von Stengel, S. Effects of Combined Whole-473
Body Electromyostimulation and Protein Supplementation on Local and Overall Muscle/Fat 474
Distribution in Older Men with Sarcopenic Obesity: The Randomized Controlled Franconia 475
Sarcopenic Obesity (FranSO) Study. Calcified tissue international 103, 266-277, 476
doi:10.1007/s00223-018-0424-2 (2018). 477
15. Kemmler, W. et al. Effects of Whole-Body Electromyostimulation versus High-Intensity 478
Resistance Exercise on Body Composition and Strength: A Randomized Controlled Study. 479
Evidence-based complementary and alternative medicine : eCAM 2016, 9236809, 480
doi:10.1155/2016/9236809 (2016). 481
16. Kemmler, W. et al. Whole-body electromyostimulation to fight sarcopenic obesity in 482
community-dwelling older women at risk. Resultsof the randomized controlled FORMOsA-483

15
sarcopenic obesity study. Osteoporosis international : a journal established as result of 484
cooperation between the European Foundation for Osteoporosis and the National Osteoporosis 485
Foundation of the USA 27, 3261-3270, doi:10.1007/s00198-016-3662-z (2016). 486
17. von Stengel, S., Bebenek, M., Engelke, K. & Kemmler, W. Whole-Body Electromyostimulation 487
to Fight Osteopenia in Elderly Females: The Randomized Controlled Training and 488
Electrostimulation Trial (TEST-III). Journal of osteoporosis 2015, 643520, 489
doi:10.1155/2015/643520 (2015). 490
18. Kemmler, W. et al. Whole-body electromyostimulation and protein supplementation favorably 491
affect sarcopenic obesity in community-dwelling older men at risk: the randomized controlled 492
FranSO study. Clin Interv Aging 12, 1503-1513, doi:10.2147/CIA.S137987 (2017). 493
19. Kemmler, W. et al. Whole-body EMS to fight sarcopenic obesity–a review with emphasis on 494
body fat. Deutsche Zeitschrift fur Sportmedizin 68, 170-177 (2017). 495
20. de Oliveira Melo, M., Aragao, F. A. & Vaz, M. A. Neuromuscular electrical stimulation for 496
muscle strengthening in elderly with knee osteoarthritis - a systematic review. Complement 497
Ther Clin Pract 19, 27-31, doi:10.1016/j.ctcp.2012.09.002 (2013). 498
21. Carvalho, M. T. X., Guesser Pinheiro, V. H. & Alberton, C. L. Effectiveness of neuromuscular 499
electrical stimulation training combined with exercise on patient-reported outcomes 500
measures in people with knee osteoarthritis: A systematic review and meta-analysis. 501
Physiother Res Int, e2062, doi:10.1002/pri.2062 (2023). 502
22. Park, S., Min, S., Park, S.-H., Yoo, J. & Jee, Y.-S. Influence of isometric exercise combined with 503
electromyostimulation on inflammatory cytokine levels, muscle strength, and knee joint 504
function in elderly women with early knee osteoarthritis. Frontiers in Physiology 12 (2021). 505
23. Runhaar, J. et al. Inflammatory cytokines mediate the effects of diet and exercise on pain and 506
function in knee osteoarthritis independent of BMI. Osteoarthritis Cartilage 27, 1118-1123, 507
doi:10.1016/j.joca.2019.04.009 (2019). 508
24. World_Medical_Association. World Medical Association Declaration of Helsinki: ethical 509
principles for medical research involving human subjects. Jama 310, 2191-2194, 510
doi:10.1001/jama.2013.281053 (2013). 511
25. Kellgren, J. H. & Lawrence, J. S. Radiological assessment of rheumatoid arthritis. Ann Rheum 512
Dis 16, 485-493, doi:10.1136/ard.16.4.485 (1957). 513
26. Bennell, K. et al. What type of exercise is most effective for people with knee osteoarthritis 514
and co-morbid obesity?: The TARGET randomized controlled trial. Osteoarthritis and cartilage 515
28, 755-765 (2020). 516
27. Kemmler, W. et al. Recommended Contraindications for the Use of Non-Medical WB-517
Electromyostimulation. Dtsch Z Sportmed 70, 278-281 (2019). 518
28. Bundesamt für Strahlenschutz. BfS-Merkblatt - Genehmigungs- und Anzeigebedürftigkeit von 519
Strahlenanwendungen am Menschen in der medizinischen Forschung, 520
<https://www.bfs.de/SharedDocs/Downloads/BfS/DE/fachinfo/ion/forschung-521
einreichungsbeduerftigkeit.pdf?__blob=publicationFile&v=10> (2022). 522
29. Hunter, D. J. et al. Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS 523
(MRI Osteoarthritis Knee Score). Osteoarthritis and cartilage 19, 990-1002 (2011). 524
30. Roemer, F. W. et al. Heterogeneity of cartilage damage in Kellgren and Lawrence grade 2 and 525
3 knees: the MOST study. Osteoarthritis Cartilage 30, 714-723, doi:10.1016/j.joca.2022.02.614 526
(2022). 527
31. Micke, F. et al. Similar Pain Intensity Reductions and Trunk Strength Improvements following 528
Whole-Body Electromyostimulation vs. Whole-Body Vibration vs. Conventional Back-529

16
Strengthening Training in Chronic Non-specific Low Back Pain Patients: A 3-armed randomized 530
controlled trial. Front Physiol 13, 664991, doi:10.3389/fphys.2021.664991 (2021). 531
32. Weissenfels, A. et al. Effects of whole-body electromyostimulation on chronic nonspecific low 532
back pain in adults: a randomized controlled study. Journal of pain research 11, 1949-1957, 533
doi:10.2147/JPR.S164904 (2018). 534
33. Zink-Rückel, C., Kohl, M., von Stengel, S. & Kemmler , W. Once weekly whole-body 535
electromyostimulation increase strength, stability and body composition in amateur golfers. A 536
randomized controlled study. Int. J. Environ. Res. Public Health accepted for publication 537
(2021). 538
34. Borg, E. & Kaijser, L. A comparison between three rating scales for perceived exertion and two 539
different work tests. Scand J Med Sci Sports 16, 57-69, doi:10.1111/j.1600-0838.2005.00448.x 540
(2006). 541
35. Kemmler, W. et al. Position statement and updated international guideline for safe and 542
effective whole-body electromyostimulation training-the need for common sense in WB-EMS 543
application. Front Physiol 14, 1174103, doi:10.3389/fphys.2023.1174103 (2023). 544
36. Nelson, A. E., Allen, K. D., Golightly, Y. M., Goode, A. P. & Jordan, J. M. in Seminars in arthritis 545
and rheumatism. 701-712 (Elsevier). 546
37. Alghadir, A. H., Anwer, S., Iqbal, A. & Iqbal, Z. A. Test-retest reliability, validity, and minimum 547
detectable change of visual analog, numerical rating, and verbal rating scales for measurement 548
of osteoarthritic knee pain. Journal of pain research 11, 851-856, doi:10.2147/JPR.S158847 549
(2018). 550
38. Hjermstad, M. J. et al. Studies comparing Numerical Rating Scales, Verbal Rating Scales, and 551
Visual Analogue Scales for assessment of pain intensity in adults: a systematic literature 552
review. J Pain Symptom Manage 41, 1073-1093, doi:10.1016/j.jpainsymman.2010.08.016 553
(2011). 554
39. Roos, E. M. & Lohmander, L. S. The Knee injury and Osteoarthritis Outcome Score (KOOS): from 555
joint injury to osteoarthritis. Health and quality of life outcomes 1, 1-8 (2003). 556
40. Roos, E. M., Roos, H. P., Lohmander, L. S., Ekdahl, C. & Beynnon, B. D. Knee Injury and 557
Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome measure. 558
J Orthop Sports Phys Ther 28, 88-96, doi:10.2519/jospt.1998.28.2.88 (1998). 559
41. Kemmler, W. et al. Effects of High Intensity Resistance Training on Fitness and Fatness in Older 560
Men with Osteosarcopenia. Front Physiol 11, 1014, doi:10.3389/fphys.2020.01014 (2020). 561
42. Hettchen, M. et al. Changes in Menopausal Risk Factors in Early Postmenopausal Osteopenic 562
Women After 13 Months of High-Intensity Exercise: The Randomized Controlled ACTLIFE-RCT. 563
Clin Interv Aging 16, 83-96, doi:10.2147/CIA.S283177 (2021). 564
43. Dobson, F. et al. OARSI recommended performance-based tests to assess physical function in 565
people diagnosed with hip or knee osteoarthritis. Osteoarthritis Cartilage 21, 1042-1052, 566
doi:10.1016/j.joca.2013.05.002 (2013). 567
44. Gill, S. et al. Thirty second chair stand test: Test-retest reliability, agreement and minimum 568
detectable change in people with early-stage knee osteoarthritis. Physiother Res Int 27, e1957, 569
doi:10.1002/pri.1957 (2022). 570
45. Jones, C. J., Rikli, R. E. & Beam, W. C. A 30-s chair-stand test as a measure of lower body 571
strength in community-residing older adults. Res Q Exerc Sport 70, 113-119, 572
doi:10.1080/02701367.1999.10608028 (1999). 573

17
46. Goh, S.-L. et al. Relative efficacy of different exercises for pain, function, performance and 574
quality of life in knee and hip osteoarthritis: systematic review and network meta-analysis. 575
Sports Medicine 49, 743-761 (2019). 576
47. R: A Language and Environment for Statistical Computing. v. 4.3.0 (R Foundation for Statistical 577
Computing, Vienna, Austria., 2023). 578
48. Honaker, J., King, G. & Blackwell, M. Amelia II: A program for missing data JSS 45, 1-47 (2011). 579
49. Van Buuren, S. & Groothuis-Oudshoorn, K. mice: Multivariate imputation by chained equations 580
in R. Journal of statistical software 45, 1-67 (2011). 581
50. Li, G. et al. An introduction to multiplicity issues in clinical trials: the what, why, when and how. 582
Int J Epidemiol 46, 746-755, doi:10.1093/ije/dyw320 (2017). 583
51. Cohen, J. Statistical power analysis for the behavioral sciences. 2nd edn, (Lawrence Earlbaum 584
Associate, 1988). 585
52. Bispo, V. A. et al. The effects of neuromuscular electrical stimulation on strength, pain, and 586
function in individuals with knee osteoarthritis: a systematic review with meta-analysis. 587
Fisioterapia e Pesquisa 28, 416-426, doi:10.1590/1809-2950/20028528042021 (2021). 588
53. Aslan, Ö., Batur, E. B. & Meray, J. The Importance of Functional Hamstring/Quadriceps Ratios 589
in Knee Osteoarthritis. Journal of Sport Rehabilitation 1, 1-5 (2019). 590
54. Hafez, A. R. et al. Treatment of knee osteoarthritis in relation to hamstring and quadriceps 591
strength. J Phys Ther Sci 25, 1401-1405, doi:10.1589/jpts.25.1401 (2013). 592
55. Kemmler, W., Teschler, M. & von Stengel, S. Effekt von Ganzkörper-Elektromyostimulation - 593
‚‚A series of studies". Osteologie 24, 20-29 (2015). 594
56. Reijman, M. et al. Body mass index associated with onset and progression of osteoarthritis of 595
the knee but not of the hip: the Rotterdam Study. Annals of the rheumatic diseases 66, 158-596
162 (2007). 597
57. Silverwood, V. et al. Current evidence on risk factors for knee osteoarthritis in older adults: a 598
systematic review and meta-analysis. Osteoarthritis Cartilage 23, 507-515, 599
doi:10.1016/j.joca.2014.11.019 (2015). 600
58. Chang, J. et al. Systemic and local adipose tissue in knee osteoarthritis. Osteoarthritis and 601
cartilage 26, 864-871 (2018). 602
59. Petersen, A. M. & Pedersen, B. K. The anti-inflammatory effect of exercise. J Appl Physiol (1985) 603
98, 1154-1162, doi:10.1152/japplphysiol.00164.2004 (2005). 604
60. Wu, Y., Zhu, F., Chen, W. & Zhang, M. Effects of transcutaneous electrical nerve stimulation 605
(TENS) in people with knee osteoarthritis: A systematic review and meta-analysis. Clin Rehabil 606
36, 472-485, doi:10.1177/02692155211065636 (2022). 607
61. Teschler, M. et al. Very high creatine kinase CK levels after WB_EMS. Are there implications 608
for health. Int J Clin Exp Med 9, 22841-22850 (2016). 609
62. Sabha, M. & Hochberg, M. C. Non-surgical management of hip and knee osteoarthritis; 610
comparison of ACR/AF and OARSI 2019 and VA/DoD 2020 guidelines. Osteoarthr Cartil Open 611
4, 100232, doi:10.1016/j.ocarto.2021.100232 (2022). 612
613

18
Acknowledgements 614
We would like to thank the Else Kröner-Fresenius-Stiftung for providing funding for the present study 615
(2021_EKSE.22). Adam Culvenor is supported by a National Health and Medicine Council (NHMRC) of 616
Australia Investigator Grant (GNT2008523). 617
The present work was performed in (partial) fulfillment of the requirements for obtaining the degree 618
“Dr. hum. biol.” for the first author Stephanie Kast. 619
Author contributions 620
All authors made substantial contributions to the study conception and design, data acquisition, or 621
data analysis and interpretation; drafting the article or revising it critically for important intellectual 622
content; providing final approval of the manuscript for submission. The specific contributions of the 623
authors were as follows: Study conception and design: S.K., S.v.S. and W.K.; Data analysis and 624
interpretation: M.K., S.K., S.v.S. and W.K.; Drafting the article: S.K. and S.v.S.; Critically editing and 625
revising the article: S.K., S.v.S., W.K., F.R., A.G., A.M., M.K. and M.U. All authors reviewed and approved 626
the final version of the manuscript. 627
Additional information 628
Competing interests: The author(s) declare no competing interests. 629
630

19
Figure and table legend 631
632
Figure 1. WB-EMS training session (Written informed consent was obtained from the participants to 633
publish this picture) 634
Figure 2. Study flow diagram (according to CONSORT, Consolidated Standards of Reporting Trial) 635
Table 1. Baseline characteristics of the study participants 636
Table 2. Baseline data and changes of primary and secondary outcomes in the WB-EMS and CG. 637
Table 3. Baseline data and changes of exploratory outcomes in the WB-EMS and CG. 638
639

20
Table 1. Baseline characteristics of the study participants 640
Variable
CG (n=36)
WB-EMS (n=36)
Age (years)
57.9  7.0
58.3  7.2
Gender (women/men) [n]
24 / 12
22 / 14
Body mass index (BMI) [kg/m2]
29.3  3.6
31.1  4.6
Body height [cm]
174.3  9.0
173.2  9.9
Body mass [kg]
89.5  15.1
93.2  15.1
Lean body mass (LBM) [kg]
58.1  11.8
60.2  12.5
Total body fat [%]
35.0  7.7
35.2  9.2
Physical activity [Score] 1
3.70 ± 1.11
3.58 ± 1.28
No exercise [n] 2
12 (33%)
13 (36%)
Knee pain intensity [NRS] 3
4.07 ± 1.61
4.05 ± 1.45
All values are expressed as mean value ± standard deviation. 641
CG, control group; NRS, numeric rating scale (0-10); WB-EMS, whole-body electromyostimulation group. 642
1 self-rated physical activity (“very low” (1) to “very high” (7), assessed by questionnaire 643
2 assessed by questionnaire 644
3 average knee pain intensity, assessed by 7-day protocol 645
646

21
Table 2. Baseline data and changes of primary and secondary outcomes in the WB-EMS and CG. 647
648
All values are expressed as mean value (MV) ± standard deviation (SD). 649
CG, control group; CI, confidence interval; FU, 7-months follow-up; KOOS, Knee injury and Osteoarthritis Outcome 650
Score (0-100, 0=extreme problems, 100=no problems); NRS, numeric rating scale (0-10, 0=no pain, 10=worst 651
possible pain); SMD, standardized mean difference; WB-EMS, whole-body electromyostimulation group. 652
1 d≥ 0.2 small effect; d ≥ 0.5: moderate effect; d ≥ 0.8: high effect 653
2 measured unilateral (knee of interest) 654
* p<0.05; ** p<0.01; *** p<0.001; ns non-significant (changes within groups) 655
656
CG (n=36)
MV ± SD
WB-EMS (n=36)
MV ± SD
Difference
MV (95% CI)
SMD
d1
p-value
KOOS Pain
Baseline
56.1 ± 12.9
54.4 ± 12.4
FU
63.1 ± 15.1
71.1 ± 13.9
Changes
7.0 ± 13.6**
16.7 ± 13.9***
9.0 (2.9 to 15.1)
0.65
.004
KOOS Symptoms
Baseline
57.5 ± 15.4
57.7 ± 14.5
FU
61.7 ± 15.3
70.3 ± 13.4
Changes
4.1 ± 13.8 ns
12.6 ± 14.1***
8.6 (2.8 to 14.4)
0.62
.004
KOOS ADL
Baseline
64.6 ± 13.6.
65.1 ± 13.9
FU
68.0 ± 13.2
79.1 ± 12.6
Changes
3.4 ± 13.7 ns
14.0 ± 13.9***
10.8 (5.3 to 16.3)
0.78
<.001
KOOS Sports/REC
Baseline
33.1 ± 21.1
28.8 ± 20.8
FU
41.4 ± 22.5
50.2 ± 19.2
Changes
8.3 ± 18.7*
21.4 ± 19.1***
11.5 (3.3 to 19.6)
0.61
.007
KOOS QoL
Baseline
33.3 ± 16.5
31.4 ± 13.2
FU
39.1 ± 18.5
47.4 ± 13.6
Changes
5.7 ± 14.3*
16.0 ± 14.7***
9.5 (3.1 to 16.0)
0.66
.004
Knee pain intensity (NRS)
Baseline
4.07 ± 1.60
4.05 ± 1.45
FU
3.31 ± 1.87
2.26 ± 1.29
Changes
-0.76 ± 1.73*
-1.78 ± 1.75***
-1.04 (-1.75 to -0.33)
0.60
.005
Maximum isokinetic Hip/Leg Extensor Strength [N]2
Baseline
749.2 ± 224.8
798.5 ± 230.5
FU
778.5 ± 235.6
903.4 ± 278.9
Changes
29.3 ± 151.3 ns
104.9 ± 152.6***
79.0 (6.9 to 151.2)
0.52
.03
Sit-to-stand test (Chair Rise) [n]
Baseline
17.7 ± 6.6
18.7 ± 5.9
FU
18.2 ± 7.53
23.0 ± 5.74
Changes
0.53 ± 4.06 ns
4.30 ± 4.07***
3.9 (2.0 to 5.8)
0.96
<.001

22
Table 3. Baseline data and changes of exploratory outcomes in the WB-EMS and CG. 657
658
All values are expressed as mean value (MV) ± standard deviation (SD). 659
CG, control group; CI, confidence interval; FU, 7-months follow-up; SMD, standardized mean difference; WB-EMS, 660
whole-body electromyostimulation group. 661
1 d≥ 0.2 small effect; d ≥ 0.5: moderate effect; d ≥ 0.8: high effect 662
* p<0.05; ** p<0.01; *** p<0.001; ns non-significant (changes within groups) 663
CG (n=36)
MV ± SD
WB-EMS (n=36)
MV ± SD
Difference
MV (95% CI)
SMD
d1
p-value
Body fat content [%]
Baseline
35.0 ± 7.7
35.2 ± 9.2
FU
36.2 ± 8.1
35.6 ± 9.1
Changes
1.21 ± 1.95***
0.42 ± 2.02 ns
-0.79 (-1.73 to 0.15)
0.40
.098
Lean body mass [kg]
Baseline
58.1 ± 11.8
60.2 ± 12.5
FU
57.4 ± 11.7
60.1 ± 11.8
Changes
-0.62 ± 1.58*
-0.08 ± 1.62 ns
0.62 (-0.10 to 1.35)
0.39
.09
Pain medication [weekly dose]
Baseline
0.81±2.47
0.64±1.33
FU
1.36 ± 2.85
0.32 ± 1.40
Changes
0.56 ± 2.38 ns
-0.31 ± 2.43 ns
-0.98 (-1.97 to 0.04)
0.41
.059