The effects of fixation and contraction mode on prone isokinetic knee flexor tests

2.1 Participants

Fourteen healthy male participants (mean ± SD, age: 21.9 ± 2.8 years, height: 1.81 ± 0.06 m, mass: 76.6 ± 7.3 kg) signed an informed consent to voluntarily join the study. They were familiar with isokinetic strength tests and lower extremity resistance training. All participants were free of pain, injuries or other physical restrictions that might have prevented maximum force exertion. The local ethics commission confirmed that the requirements of the Declaration of Helsinki were met.

2.2 Procedures

All participants performed two sessions (familiarisation and testing) which followed the identical procedures. The sessions were separated by 72–96 h. After determining their mass, the participants individually warmed up (jogging, dynamic stretching) for ten minutes. This general warm-up was followed by the participant’s positioning on the dynamometer (IsoMed 2000, D&R Ferstl GmbH, Hemau, Germany). The participants were asked to lie prone, with shoes off to reduce acceleration errors, with a 0 ∘ flexed hip (full extension). By the aid of a laser pointer the mechanical axis of the dynamometer was aligned with the transepicondylar axis of the participant’s right knee. This was conducted in an isometrically-contracted muscular state at maximum knee extension. The anatomical landmarks were palpated and marked with retro-reflective spheres (Ø 8–14 mm) (Fig. 1, points 1–5). Additionally, spherical markers (points 6–8) were attached on the lever arm of the dynamometer, on the thigh and on the fibula head to enable the reconstruction of small gaps ( ⩽ 3 frames).

The individual settings of each participant were stored by the manufacturer’s computer software IsoMed analyze V.2.0 to guarantee an accurate repositioning. Afterwards, a rigid three-dimensional calibration frame (LxWxH: 40 × 40 × 40 cm) with twelve retro-reflective markers (Ø 10 mm) was mounted on the rotation axis of the isokinetic dynamometer to calibrate and to determine axis alignment. Another calibration frame (LxWxH: 180 × 60 × 60 cm) covered the space for the recording of hip and knee joint kinematics. A static calibration was conducted by four cameras (Basler A602fc-2, Basler AG, Ahrensburg, Germany) (Fig. 2). Four headlights (Bar Fly 200, Kino Flo ® Lightning Systems, Burbank, CA, USA) improved marker visibility. The movement planes were defined as shown in Fig. 2.

OPEN IN VIEWER

OPEN IN VIEWER

After removing both calibration frames, a double shin pad for unilateral knee extension was attached to the motor-driven axis, whereas its distal part was fastened by straps approximately 2–3 cm proximal to the lateral malleolus. A static gravity correction in horizontal lever arm position and at a relaxed muscular state was executed by the integrated software. Each participant had to perform concentric and eccentric knee flexion movements at three fixation methods: minimal (hand grips), moderate (grips, hip strap) and maximal (grips, thigh and hip strap, hip wedges, shoulder pads) (Fig. 1). To ensure the highest reproducibility the fixation was applied by the same examiner. The order of the fixation condition was randomly assigned. After each fixation was applied the participants performed 15 submaximal ( ∼ 50%) concentric and eccentric knee flexion movements at 60 ∘ /s. For both contraction modes, an isometric pre-activation threshold of 10 Nm was used [2, 9]. Within the following uni-directional tests, concentric (con) knee flexion movements were conducted discretely at 30 ∘ /s with maximum effort. After a rest of one minute the eccentric tests succeeded. The range of motion was set at 110 ∘ starting at full knee extension (0 ∘ ). A very quick acceleration and a hard deceleration ensured the largest possible isokinetic range of motion. Each set consisted of five repetitions (two practise and three with maximum effort) following a testing procedure with proven reliability [10]. The last three repetitions were selected for further analysis. During the tests, laboratory lighting was turned off.

OPEN IN VIEWER

Kinetic raw data were recorded at a sampling rate of 200 Hz (IsoMed analyze V.2.0, D&R Ferstl GmbH, Hemau, Germany). Kinematic data were captured with 100 Hz (Basler A602fc-2, Basler AG, Ahrensburg, Germany) and processed in Vicon Motus (V9.2, VICON TM, Oxford, UK). Kinetic and kinematic data were synchronized (TEMPLO 8.2.358, Contemplas, Kempten, Germany). Roll ( α ) and yaw ( β ) angles were defined as shifts – related to the dynamometer axis (y-axis) – of the transepicondylar axis in the transverse and frontal plane (Fig. 2). Positive values represented relative internal hip rotation (anterior motion of the lateral femoral epicondyle in relation to the medial one) and adduction (caudal motion of the lateral femoral epicondyle in relation to the medial one). Sagittal trajectory lines of the lateral femoral epicondyle were plotted for each condition (Fig. 3). In addition to this qualitative measure, their lengths were computed as quantitative measure of the combined rotational and translational motion of the knee axis (sagittal trajectory length = stl). The trajectory centre was calculated by the arithmetic means in antero-posterior (trajectory centre shift along x-axis = tcsx) and cranio-caudal direction (trajectory centre shift along z-axis = tcsz). The sagittal distance to the trajectory centre demonstrated the mean two-dimensional displacement of the transepicondylar track to its respective centre and thus the spatial expansion of the knee joint axis (sagittal distance to trajectory centre = sdtc).

Kinetic raw data were stored as ASCII files. By the aid of a custom software written in C++, a recursive 5 th order Butterworth low-pass filter with a cut-off frequency of 6 Hz was applied and the isokinetic range of motion was identified ( ± 1% deviation of the pre-selected angular velocity) [10]. Afterwards, the relevant test parameters from the obtained moment-angle curves were computed: highest gravity-corrected peak moments (PM), angles of peak moment (APM) and contractional work (CW).

Kolmogorov-Smirnov ( p ⩽ 0.05) and Levene’s tests ( p ⩽ 0.10) verified normal distribution and variance homogeneity. A two-way ANOVA with repeated measures (3 × 2) examined significant effects of fixation (Fix min , Fix mod and Fix max ) and contraction mode (con and ecc) on kinetic (PM, CW) and kinematic parameters (APM, sagittal trajectory length, shift of trajectory centre in antero-posterior and cranio-caudal direction, sagittal distance to trajectory centre, roll and yaw angle). By the aid of Bonferroni post hoc tests ( p ⩽ 0.05) actual p -values were calculated.

The effect size of the ANOVA was assessed by the partial eta-squared ( η ⁢ p 2 ) representing the meaningfulness of the results. With respect to Cohen’s rule of thumb [11], a small, medium and large effect size was reached at 0.02, 0.13 and 0.26, respectively. The effect size of single comparisons was rated by the Cohen’s d ( $>$ 0.8 high, 0.8–0.5 medium, 0.5–0.2 low, 0.2 negligible) [11]. All statistical tests were executed with SPSS (V22.0, SPSS Inc., Chicago, IL, USA).

3.1 Kinetic results

Moderate fixation led to significant reductions of peak moments (PM) ( - 6 to - 10%, 0.36 ⩾ d ⩾ 0.52) throughout all contraction conditions compared to the minimal and maximal fixation (Table 1). Concerning contractional work (CW), moderate and maximal fixation showed lower values than minimal fixation ( - 7 to - 11%, 0.37 ⩾ d ⩾ 0.56). Both kinetic parameters demonstrated large effects of contraction mode as values obtained from eccentric exercise were on average 28% higher than concentric ones ( 0.001, 0.883 ⩾ η ⁢ p 2 ⩾ 0.910).

3.2 Kinematic results

Compared to minimal and moderate fixation, isokinetically determined angles of concentric and eccentric peak moments (APM iso ) were significantly reduced at maximal fixation ( - 2 ∘ to - 6 ∘ , 0.75 ⩾ d ⩾ 1.31) (Table 1). With regard to kinematic APMs (APM kin ), this divergence was still apparent for concentric tests ( - 8 ∘ to - 9 ∘ , 1.43 ⩾ d ⩾ 2.15), but disappeared for eccentric exercise ( p ⩾ 0.05), while contraction mode showed no influence. Hip flexion at peak moment (HF PM ) was significantly higher at maximal fixation ( - 3 ∘ to - 5 ∘ , 1.28 ⩾ d ⩾ 2.69).

Concerning the kinematic range of motion (ROM KF ) and the mean angular velocity of the knee joint ( ω KF ), lowest values appeared under maximal fixation for concentric tests ( - 3 ∘ to - 4 ∘ and - 1 ∘ /s, 0.67 ⩾ d ⩾ 1.16) and under minimal fixation for eccentric exercise ( - 3 ∘ to - 5 ∘ and - 1 ∘ /s, 0.65 ⩾ d ⩾ 1.14). Generally, eccentric ROM KF and ω KF were lower compared to concentric tests ( - 4 ∘ to - 13 ∘ and - 3 ∘ /s, 0.001, 0.848 ⩾ η ⁢ p 2 ⩾ 0.882). Hip motion (ROM HF ) was significantly reduced when maximal fixation was applied ( - 1 ∘ to - 7 ∘ , 0.90 ⩾ d ⩾ 1.82), while this effect was more pronounced during concentric exercise.

Figure 3 shows the postero-cranial displacement of the lateral femoral epicondyle in relation to the dynamometer axis (coordinate origin). Initial axis misalignment resulted in an average antero-cranial shift (2.0 and 0.1 cm) during maximal contractions. Maximal fixation led to a significantly increased anterior trajectory centre shift (tcsx) (1.04 ⩾ d ⩾ 1.43) (Table 1) compared to the minimum and moderate fixation. The cranial and caudal displacements (tcsz) of the trajectory centre were greatest for concentric movements at minimum fixation (0.8 ± 1.5 ∘ ) and for eccentric movements at maximum fixation ( - 0.9 ± 0.9 ∘ ). Best cranio-caudal axis alignment and spatial expansion (sdtc) was achieved at moderate fixation for both contraction modes (Table 1). Sagittal trajectory length (stl) was significantly increased when minimal fixation was applied ( + 12% to + 34%, 0.76 ⩾ d ⩾ 1.75).

OPEN IN VIEWER

Figure 4 illustrates the roll (top) and yaw angle (bottom) histories of the transepicondylar axis in relation to the dynamometer axis (zero line). During concentric and eccentric knee flexion movements, there was an average misalignment of 0.1 ∘ and 6.1 ∘ towards external hip rotation and hip abduction. Roll angles at both contraction modes as well as yaw angles of concentric movements were unaffected by participant’s fixation method ( $p>$ 0.05). Eccentric exercise induced the largest yaw angles (hip abduction) at maximal fixation compared to minimal and moderate fixation ( + 1 ∘ to + 3 ∘ , 0.38 ⩾ d ⩾ 0.63).

4.1 Peak moments and contractional work would increase with higher hip flexion

Peak moments and contractional work were highest at minimal fixation when maximal hip flexion angles demonstrated mean values of 15 ∘ –17 ∘ (Table 1). Greater hip flexion at maximal fixation (23 ∘ –24 ∘ ) resulted in similar concentric and eccentric PMs ( - 2% to 2%), but significantly lower CW ( - 8% to - 11%) at both contraction modes. The latter parameter showed similar values between maximal and moderate fixation although hip flexion was considerably different ( - 13 ∘ to - 14 ∘ ). This may be attributable to the fact that all three fixation methods were performed across the identical device’s range of motion leading to significant differences in minimal and maximal knee flexion angles (Table 1). Consequently, higher hip flexion increased the range of motion with high maximal knee flexion angles ( + 7 ∘ to + 10 ∘ at Fix max ) which might have led to divergent internal lever arms at the hip and knee joint and/or to an unfavourably large actin-myosin overlap [12]. Previous research differentiated between greater ranges of hip flexion angles (e.g., 0 ∘ , 30 ∘ , 45 ∘ , 60 ∘ and 90 ∘ ) and did not verify the actual hip angle by a concomitant kinematic analysis [5, 6, 7, 8]. Guex et al. [8] examined a 27% increase of eccentric hamstring peak moments from 0 ∘ (85 Nm) to 30 ∘ (103 Nm) of hip flexion for a seated test setting at 60 ∘ /s. The present prone tests were performed with hip flexion angles which lay in between these values (6 ± 5 ∘ to 21 ± 7 ∘ ) and resulted in much higher peak moments (155–171 Nm). Contractional work was not reported in any of these studies [5, 6, 7, 8]. The rationale for this lack of information is uncertain.

Due to the above reasons, the first hypothesis can be partially confirmed for peak moments, but not for contractional work.

Maximal and moderate fixation induced the shortest sagittal trajectory length (Table 1). However, maximal fixation caused a mean antero-cranial shift of the trajectory centre (2.8 cm and - 0.7 cm) which significantly exceeded the axis misalignment of the minimal and moderate fixation method for both concentric and eccentric movements (Fig. 3). Best axis alignment and spatial expansion (sdtc) was achieved by moderate fixation since hip motion was restricted by the hip strap (Table 1). While there is no physiological explanation why maximal fixation resulted in a systematically greater misalignment, we speculate that it has to do with a poorer initial axis alignment. However, this issue requires further exploration.

Roll and yaw angles of concentric movements were unaffected by fixation method (Table 1). Eccentric exercise showed lowest values at minimal fixation (Fig. 4). These findings support the assumption that maximal fixation altered the movement patterns of the knee joint axis and its surrounding muscles [6, 13]. The present results of prone knee flexion movements can be well compared with previous research about supine concentric and eccentric knee extensions [2]. The average misalignment of 2.0 and 0.1 cm in antero-caudal direction (0.8 and 2.4 cm), mean roll and yaw angle tilts of - 0.1 ∘ and - 6.1 ∘ ( - 2.3 ∘ and - 2.3 ∘ ) as well as different angle characteristics (Fig. 4) revealed considerable muscle-specific differences. These divergences might also be promoted by the body position (prone vs. supine [14, 15]. The anterior displacement of the transepicondylar axis could not be avoided as the dynamometer axis was already at its lowest configurable position.

Thus, the second hypothesis has to be rejected for all three movement planes. Best axis alignment was achieved by moderate fixation which prevented hip motion by a strap.

4.3 Eccentric knee flexion movements would result in a smaller range of motion and a reduced mean angular velocity compared to concentric exercise

It is well-established that under the same angular velocity, the peak moments produced during eccentric efforts are higher than their concentric counterparts [5, 6, 7, 8]. The present results demonstrated an expected difference of 26%–31% (Table 1). This enhanced muscular strength might have led to greater material deformation of the lounger’s edge (directly adjacent to the knee joint axis, see Figs 1 and 2) especially near full knee extension inducing an anterior shift of the knee joint axis (Fig. 2). Therefore, minimal knee flexion angles ( φ KF min ) are significantly greater ( 0.001; η ⁢ p 2 = 0.942) when comparing eccentric and concentric knee flexor tests (Table 1). This large effect resulted in a reduced mean angular velocity as participants tried to resist lever arm’s movement (26.0 ∘ /s vs. 29.4 ∘ /s; 0.001; η ⁢ p 2 = 0.882). Nevertheless, these deviations (2%–13%) from the pre-selected angular velocity of 30 ∘ /s are much smaller in contrast to concentric knee extensor and eccentric knee flexor tests at 150 ∘ /s revealing a loss of 18%–19%, respectively [3].

Thus, the third hypothesis can be confirmed underlining the importance of controlling joint kinematics during isokinetic tests.

4.4 Synopsis

Table 2 enumerates the advantages of minimal, moderate and maximal participant’s fixation as well as the unaffected parameters. As previously mentioned, some of the detected differences – especially concerning sagittal axis alignment at maximal fixation – might be induced by inappropriate initial alignment which likely befalls each examiner to a certain extent. The average deviations – an antero-cranial shift of 2.0 and 0.1 cm as well as mean roll and yaw angle tilts of 0.1 ∘ of external hip rotation and 6.1 ∘ of hip abduction – appeared to be tolerable considering the physiologically inherent nonconformity of the human knee joint with regard to a perfect hinge joint.

Depending on their purpose, prone isokinetic knee flexor tests or training sessions should be conducted with minimal (maximal force production) or moderate (axis alignment) fixation. Maximal fixation showed no essential advantages over the other fixation methods which does not justify the additional time requirement for appropriate positioning and fixation of the shoulder pads, the hip wedges as well as the hip and thigh strap (Fig. 1). Apart from the fixation method used, the examiner must always check if the patient/athlete executes unwanted evasive movements which might compromise measurement accuracy. Therefore, good instructions as well as a separate and properly conducted familiarisation session is crucial for appropriate standardisation of testing procedures and reliable data acquisition [10]. If muscular strength of the hands, arm and trunk muscles is not sufficient to ensure an adequate internal fixation by hand grips (e.g. among females), additional external fixation will be mandatory [13].

4.5 Limitations

The present study incorporated three major constraints. First, the material deformation-induced anterior displacement of the lateral femoral epicondyle could have been reduced if the plinth’s padding would have been more rigid and/or a tough covering of ∼ 2–3 cm height would have been attached onto the plinth. Second, the three fixation methods were performed across the device’s identical range of motion so that an identical knee joint’s range of motion was not achieved (Table 1). Particularly for the maximal fixation, an adjustment in the range near full knee extension would have been appropriate. Third, the calculation of resultant moments [1] was not possible because the position of both calibration frames to each other was not determined. Therefore, the effects of the kinematic measuring inaccuracies on the kinetic results was not examined.

4.6 Perspectives

It is well-established that hamstring peak moments are considerably increased ( + 34% to + 58%) when the rarely used prone position is compared to the standard seated position [14, 15]. Future studies should examine if there is a potential relation between these two testing procedures. Due to the body position induced higher muscle activation, the similarity to the commonly conducted prone leg curl exercise and its higher sensitivity in detecting strength imbalances, we suggest making the prone position to the standard one to assess maximal hamstring strength capacities by isokinetic devices [4, 14, 15]. The illustrated method should be applied as internal validation of every commonly used isokinetic test and of each examiner in order to check the respective measuring accuracy and to identify the best fitted test modalities. In order to reduce acceleration induced oscillations of the retro-reflective markers, a low angular velocity was used. It is useful to expand the present research to higher angular velocities. Following the kinematic analyses of knee extensor and knee flexor tests [2, 3], it is necessary to establish standardised procedures of initial axes alignment and reasonable fixation techniques concerning isokinetic test settings of more complex and thus more difficult joints (e.g., hip and shoulder). Furthermore, it is interesting to know if participant’s kinematic profiles (e.g., range of motion, axis alignment and spatial expansion) of a certain isokinetic exercise are reliable or rather random when intra- and inter-session-comparisons are executed.