Measurements were performed at one session either at the Neuromechanics Laboratory of the University of Potsdam (Germany) or at the Center for Rehabilitation (Vitalis, Brandenburg, Germany) by the same examiners using the same equipment.
Participants
A priori sample size analysis performed with G*Power (V3.1.9.4, Düsseldorf, Germany) revealed a minimum sample size of n = 17 per group for the main hypothesis (ART vs. CON) to detect a substantial effect size of Cohen’s d = 1.0 (α = 0.05; 1 – β = 0.8; unpaired t-test, two-tailed). Due to probably arising dropouts (e.g., pain during measurements) 20 patients were included.
20 male patients diagnosed with knee osteoarthritis (ART) were recruited from the University Hospital Brandenburg (Center of orthopedics and trauma surgery, Brandenburg a.d.H., Germany) and from a local orthopedic practice. Inclusion criteria were age between 55 and 80 years and knee OA diagnosed by orthopedists based on X-ray of grade ≥ 2 (Kellgren-Lawrence Score) on at least one side. Exclusion criterion was total endoprosthesis (TEP) of the knee. Seven patients reported of knee complaints on left, six on right and seven on both sides. Two patients had a knee TEP on one side (excl. from measurements). Four patients had an operation due to OA, four patients stated previous knee operations (cruciate ligament, meniscus, tibial plateau). The more affected side showed a moderate edema and crepitation in 2 and 8 patients, respectively. Pain intensity in daily life (not during measurements) was given by averagely 4.32 ± 1.99 on analogue scale (0 = no to 10 = worst). Current health condition was given by 2.42 ± 0.75 (0 = excellent to 4 = bad). Further general health problems were reported in 18 patients: hypertension (11), obesity (10), heart disease (6), diabetes (4), back pain (5), asthma/lung disease (2), cancer (2), depression (2).
19 male controls (CON) were recruited from an association for health promotion in Potsdam (Brandenburgischer Verein für Gesundheitsförderung e.V.) and the University of Potsdam. Inclusion criteria were age between 55 and 80 years without knee symptoms and any history of knee trauma or surgery. The knees were examined by a qualified therapist. Controls had no knee edema, crepitation or pain. Current health condition was given by 1.63 ± 0.68. General health problems were reported in 15 controls (hypertension (8), obesity (2), heart disease (1), back pain (6), asthma/lung disease (1).
Any kind of pain during measurements led to exclusion of the single trial (see below). Further information on participants is given in Table 1. Age did not differ significantly between ART vs. CON (p = 0.116, unpaired t-test, two-tailed). ART showed significantly higher body mass (p = 0.043) and body mass index (BMI) (p = 0.003) than CON. Consequently, torque values were normalized to body mass for statistical evaluation.
Table 1. Participants' information
|
ART
|
CON
|
ART1
|
ART2
|
CON1
|
CON2
|
n
|
20
|
19
|
age (years)
|
65.75 ± 9.15
|
61.79 ± 5.84
|
body mass (kg)
|
93.75 ± 19.13†
|
83.00 ± 9.92†
|
body height (cm)
|
175.15 ± 6.78
|
180.74 ± 6.14
|
BMI (kg/m²)
|
30.54 ± 5.94†
|
25.62 ± 2.99†
|
Dominant side
(left/right/both)
|
2 / 13 / 4
|
1 / 15 / 3
|
number left/right
|
8 / 10*
|
11 / 9
|
10 / 8*
|
9 / 9*
|
Anthropometric data, foot preference and number of measured side for patients with knee osteoarthritis (ART) and asymptomatic controls (CON).
ART1 = less affected/asymptomatic/1st measured side;
ART2 = more affected/2nd measured side;
CON1 = 1st measured side;
CON2 = 2nd measured side.
† Significant differences: p < 0.05
*partly one side had to be excluded (see data processing).
Clinical examinations
The patients were clinically examined prior to the measurements by an orthopedic surgeon regarding their knees. Furthermore, the neuromuscular function of hamstrings were determined in all participants on both sides by a manual muscle test (MMT) in the sense of a ‘break-test’.[21, 23–25, 27, 30] MMTs were performed by an experienced tester before and after the measurement series while the participant was sitting on the measurement chair. If the participant was able to maintain the isometric position during the whole force rise applied by the tester the MMT was rated as ‘stable’. If the muscle started to lengthen during force increase the test was rated as ‘unstable’.[23–25, 27]
Technical equipment and measurement principle
Figure 1 depicts the pneumatically driven measurement system to detect the AF of hamstring muscles. The measurement principle was previously described in detail for elbow extensors.[20] Such system can detect the AF parameters in a reliable way (ICC = 0.896 – 0.966) with acceptable random errors.[20] Table 2 lists the system components and technical specifications. The system consisted of a swing including two levers (left/right) and was connected to two bellows cylinders by a cross strut (Figure 1a, 1b and 1c). The bellows cylinders can be pneumatically actuated: one works in direction of knee flexion and the other one in direction of knee extension. The latter one was used for the present investigation. An interface with strain gauge on each lever recorded the force between device and lower leg. A motor-controlled throttle (Figure 1d) avoided an abrupt pressure increase at the beginning. Three acceleration sensors (ACCs) captured position changes of each lever and of the participant’s leg. The A/D converter buffered data with 1000 Hz recorded by the software NITM DIAdem 12.0 (National Instruments, Austin, TX, USA).
Table 2. Components and technical specifications of the pneumatic AF measuring system.
System components (company)
|
Specification
|
Basic construction: pivoted levers
|
range: extension/flexion: 83° – 97°
|
Compressor
(JUN-AIR International A/S, Nørresundby, Denmark)
|
Model 6, Serial-No. 702997; Condor MDR2 EN 60947-4-1; max. system pressure: 8 bar; calibrated to max. 3 bar
|
2 bellows cylinders
(Zitec Industrietechnik GmbH, Plattling, Germany)
|
SP-2 B04 R; 2-fold, Ø 165 mm, max force: 9 kN, stroke length: 1–110 mm (adjustable), rise time: 0.1–30 s continuously
|
Control unit
(Seifert Drucklufttechnik GmbH, Bernsbach, Germany)
|
Pressure reduction to max. 1 bar
|
Pressure sensor in control unit (Seifert Drucklufttechnik GmbH, Bernsbach, Germany)
|
Linear 1 V = 1.05 bar
|
2 strain gauges and amplifiers
(modified by Biovision, Wehrheim, Germany)
|
MLMZ_2000N_67 linearly 1 V = 191.72 N
MLMZ_2000N_36 linearly, 1 V = 194.72 N
|
3 acceleration sensors (ACC)
(modified by co. Biovision, Wehrheim, Germany)
|
Sensitivity 312 mV/g (range ± 2g) cosinusoidal, between 70–110° approx. linear, linearity: ± 0.2%
|
A/D converter
(National Instruments, modified by Biovision, Wehrheim, Germany)
|
14-bit, range: -5 to 5 V
|
Motor for throttle valve control
(RS Components GmbH, Frankfurt/Main, Germany)
|
Trident DC Geared Motor (12 V dc / 13.9 W, 0.75 Nm, 360 – 1020 rpm)
|
For AF measurements, the compressed air was led via control unit to one bellows cylinder which actuated the swing in direction of knee extension. The participant’s task was to prevent the movement of the lever by adapting isometrically to the external increasing load (detailed description see below). The maximal pressure of the system was adjusted so that each participant was forced into eccentric muscle action thereafter (safety stop at 97°). The velocity of air inflow into the bellows cylinder was standardized in dependance of the individual MVIC: under stable conditions 70% of the MVIC would be reached after 3 s. The actual velocity depended on the individual ability to adapt to the external force rise. This characterizes the special feature of AF assessment.
Setting and procedure
After clinical examination, a warm up followed (10 min ergometer bicycling; 1 W/kg; 75 rpm). Afterwards, the participant was positioned in the measuring system sitting upright with both feet hanging down (hip flexion ~90°). Knee joint was flexed in 93° (controlled by hydrogoniometer) and forearms were crossed in front of the chest (starting position of all trials). The rotation center of the participant’s knee was aligned to the pivot at the hinge of the swing. To prevent a knee movement out of the pivot during muscle tension, a cushioned fixture was placed on the thighs. The interface was adjusted in height so that it contacted most comfortably the lower leg from posterior between ankle and calf above the Achilles tendon. Lever length was measured from rotational axis to the middle of the interface to calculate torques. The ACC sensor was fixed on the tibial tuberosity with double-sided tape (Figure 1a and 1b).
A specific warm up in the measurement device followed directly before measurements of each side for familiarization with hamstring activation (2 x 10 reps, 1 min rest). For this the pressure system was closed in the most extended position. Thus, air was in the system but could not stream out. The participant performed concentric contractions with the hamstrings by pushing the lower leg against the interface. He was instructed to increase the force up to 50% of the self-estimated maximum. The air in the bellows cylinder was compressed and served as resistance thereby.
The measurement series started with the less affected/asymptomatic followed by the more affected side for OA patients. For controls the order of measured sides was randomized. The measurements were guided by two raters (device/software control; instruction/positioning of participant). The following measurement series were conducted subsequently for each side:
- 3 MVIC trials: For MVIC-tests the swing was fixed and the system was passive. The participant should push as strong as possible against the interface (not explosively), reach the maximum within 3 s and maintain this for 1 – 2 s. (Resting periods: 60 s).
- 5 AF trials: The pneumatic system was active and provided an increasing load against the participant’s leg in direction of knee extension. The starting position of 93° was adjusted so that the lever was held with 10% of the MVIC. The participant’s task was to maintain this isometric position for as long as possible despite the rising external load of the system. It increased so that the participant merged into eccentric motion after the maximal holding capacity (breaking point; AFisomax) was exceeded. The participant should try to decelerate the movement during this eccentric phase, whereby the force increased further until the maximal AF (AFmax) was reached. One trial was terminated as the security stop was reached (resting periods: 120 s).
- 2 MVIC trials: Those were performed again for comparison with MVICpre.
Data processing and statistical analyses
Data were processed using NITM DIAdem 2017. Raw signals were filtered: low-pass Butterworth, filter order 10, cut-off frequency 3 Hz (force) or 1 Hz (ACC). Force and ACC signals were converted from volts to Nm and angles, respectively. Thereby, torque (Nm) was calculated by τ = r * F, whereby r refers to the lever and F to force (conversion from V to N see Table 2). Torque values were normalized to body mass. The following parameters were extracted for further considerations:
- MVIC (Nm/kg): Peak value of each MVIC-test. The maximal value of the three MVIC values before (MVICpre) and of the two MVIC values after (MVICpost) AF trials, respectively, was used for statistical evaluation.
- AFmax (Nm/kg): Peak value of each AF trial.
- AFisomax (Nm/kg): For this, no peak value exists. The angle signals were used to determine the force at the breaking point. This refers to the highest force value under isometric condition. The algorithm was previously described in detail.[20]
- Duration of pre-force adjustment in AF trials: time period (s) of the contraction phase before pressure increase was extracted from start of contraction to start of pressure increase. Exceptional long-lasting pre-phases might have influenced the AF.
Furthermore, the ratios AFisomax to AFmax (%), AFisomax to MVICpre (%) and AFmax to MVICpre (%) were calculated. Arithmetic means (M) and standard deviations (SD) were calculated for AF parameters and 95%-confidence intervals (CI) for all parameters. Additionally, the difference (strength deficit) between CON and ART was calculated for all parameters by (according to Alnahdi et al.[4]). This difference was regarded analogues between MVIC and AFmax as well as between MVIC and AFisomax within groups.
Statistical comparisons were executed in SPSS Statistics 29 (IBM, Armonk, New York, USA). Normality was checked by Shapiro-Wilk-test. Only the ratio AFisomax to AFmax showed violation of normality for ART2 and CON2, hence, Mann-Whitey-U-test was used for group comparison. T-tests (unpaired/paired) were chosen for all other comparisons. If homogeneity of variance was not fulfilled Welch correction was used (tW).
The main comparison was the more affected side of ART (ART2, 2nd measured side) vs. 2nd measured side of CON (CON2) (unpaired, one-tailed). This was chosen since 1st and 2nd measured side might show differences. Moreover, the less affected/asymptomatic side of ART (ART1) was compared to 1st measured side of CON (CON1) (unpaired, one-tailed). Paired tests (two-tailed) were executed to compare ART1 vs. ART2 and CON1 vs. CON2 as well as MVICpre vs. MVICpost within each group. Significance level was α = 0.05. Effect size Cohen’s d was interpreted as small (0.2), moderate (0.5), large (0.80) or very large (1.3).[31, 32]
Exclusion of trials for evaluation
For ART group, one side of two patients was excluded completely because of knee TEP. Hence, ART1 consisted of 18 and ART2 of 20 patients. For controls, one side of two participants was excluded completely because of existing knee complaints (before measurements) or permanent difficulties in understanding and executing the AF task (e.g., pushed against the lever or similar, which resulted in invalid trials). Hence, 18 participants were included for CON1 and CON2. Regarding single trials, all MVIC tests were included. Due to unfamiliarity with AF measurements, the first trial was excluded in general. Furthermore, five single trials of four ART patients had to be excluded due to pain during measurements (knee, hamstrings, lumbar spine). Moreover, single invalid AF trials were excluded, e.g., if participants pushed against the lever (holding task was not executed appropriately; visible in force signals in 34 of 300 trials) or due to problems during measurements (technical, execution of AF; 12 trials). In total, 71 of 84 AF trials were included for ART2, 64 of 72 for ART1, 59 of 72 for CON1 and 55 of 72 for CON2.