Participants

The recruitment of participants for this study took place between 10/01/2022 and 10/03/2022. We recruited asymptomatic female college students who were physically active and free from injuries but had DKV (see below). The inclusion criteria were female gender, age between 18 and 30 years, and the presence of DKV subjectively evaluated by a physiotherapist, using the single leg squat test. Exclusion criteria were current injuries and past surgeries in the spine, hip, knee, and ankle joints, or pain of orthopedic origin. When the actual quantitative DKV measurements were preformed, all participants demonstrated greater than 8° of DKV in at least one of the two DKV tests. Considering recent literature suggesting 7–10° as thresholds for ACL injuries, we made no further exclusions [30]. Finally, 20 participants (age: 21.3 ± 2.51 years; height: 167.8 ± 5.76 cm; mass: 60.6 ± 6.17 kg; exercise training history: 15.4 ± 3.32 years) were involved in the study. Participants pursued ground-contact sports 6.3 ± 2.42 hours per week at the club level, but none of them competed at a national or international level. Participants received verbal and written explanations of the experimental procedures and were informed about the potential risks. They provided written consent under the principles outlined in the Declaration of Helsinki. The University Ethics Committee approved the protocol (approval number, 7961-PTE2019). No participants withdrew from the study.

Experimental procedure

The duration of the study was five weeks, including all measurements and training in the following order: pre-testing, 4-week-long training intervention, and post-testing (Fig 1). Such a short-term intervention is known to produce changes in lower limb kinematics during sidestep cutting, a task closely associated with ACL injury risk in female athletes [31]. Participants underwent a standardized warm-up of 5-min low intensity cycling on a cycle ergometer, followed by whole body stretching before any testing or training. Before the pre-test, a single leg squat test was conducted for each leg to select the less stable limb, which was subsequently tested and trained.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of the study.

CMJ = countermovement jump, MVC = maximal voluntary contraction.

https://doi.org/10.1371/journal.pone.0330140.g001

After the pre-test session, participants were assigned either to the eccentric or the concentric hip abduction exercise training experimental groups following a matched group design based on the pre-testing data: we applied the z-score method, which is a direct group allocation procedure. That is, we calculated the z-scores and ranked the participants based on the maximal voluntary eccentric torque, DKV values for the two types of jumps, and single leg CMJ propulsive mechanical impulse. As a result, we created two groups for the measured variables (between-group difference in eccentric torque: p = 0.88, single leg CMJ DKV: p = 0.59, single leg landing DKV: p = 0.66, and single leg CMJ propulsive mechanical impulse: p = 0.57) for the eccentric and the concentric training groups (Table 1). Thus, each participant in the eccentric group had her matched pair in the concentric group with nearly identical abilities. Within these pairs, we equalized the mechanical work achieved in the eccentric and the concentric hip abductions during the entire training period. The post-test was conducted with the same measurements used in the pre-test, 72 hours after the last training session. Participants were unaware about the identity of their matched partner in the other experimental group. Also, participants were informed about their individual improvements only after the completion of the study.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Participants’ group-specific descriptive data.

https://doi.org/10.1371/journal.pone.0330140.t001

Maximal voluntary eccentric hip abduction contraction test

We measured the maximum voluntary hip abduction eccentric and concentric torque using a computer-controlled isokinetic dynamometer with a sampling frequency of 1.0 kHz (Multicont II, Mediagnost, Budapest and Mechatronic Ltd, Szeged, Hungary). Participants were positioned on their side opposite of the tested leg on the dynamometer. The lever arm starting position was set at 23° frontal-plane hip angle, and a constant velocity of 10°/s was used for eccentric measurements. Participants resisted the lever arm with maximal force until it reached the final position of 0°. The lever arm starting position was set at 0° frontal-plane hip angle, and a constant velocity of 10°/s was used for concentric measurements. Participants resisted the lever arm with maximal force until it reached the final position of 23°. Prior to testing, participants executed two submaximal warm-up trials at ~60 and ~80% of the estimated maximum effort. Then, two maximal effort contractions were executed with two minutes of inter-trial rest. Peak torque values were recorded, and the higher value was used for further analyses. To ensure maximal effort, participants received verbal encouragement.

Single leg CMJ

Participants performed the single leg CMJ test on a force platform (Tenzi Ltd., Pilisvörösvár, Hungary). Standing on one leg, with arms akimbo and the knee of the free leg slightly flexed, participants performed three jumps with one minute of inter-trial rest. Participants were instructed to jump as high as possible with hands kept on the hips, but no instructions were given on jumping strategy. We recorded the vertical ground reaction force at a sampling rate of 420 Hz. The propulsive mechanical impulse was then calculated from the force-time curve and the impulse values normalized for body mass.

Single leg drop landing

For the single leg landing test, participants stepped off from a 46 cm-high plyometric box (Capital Sports Plyo Crew Box) and landed on the force plate with the stepping leg. Participants were instructed to step off the box without jumping up and after landing on one leg, hold the landing position for 5 s. Participants performed three trials with 2 min of inter-trial rest.

Knee joint kinematics

A 3D motion tracking system (Noraxon, Scottdale, USA, sampling frequency: 100 Hz) was used to analyze the kinematics of single leg CMJs and single leg drop landings. Sensors comprising accelerometer-gyroscope-magnetometer complex were affixed with Velcro straps on the shank and the thigh. We followed the manufacturer’s recommendations in sensor positioning and calibration. The sensors provided data to compute DKV from frontal plane orientation angle and maximum knee flexion in the sagittal plane during single leg CMJs and landings.

Hip abduction exercise training on an isokinetic dynamometer

Participants trained 3 times/week for four weeks on the same dynamometer used for testing (Multicont II, Mediagnost, Budapest and Mechatronic Ltd, Szeged, Hungary). To control for mechanical load, we used a mechanical work-matched design [32–35]. In a session, both the eccentric and the concentric group performed 4 sets of 10 maximal effort repetitions of eccentric and concentric hip abduction contractions with 5 min of inter-set rest. The range of motion and the contraction velocity were identical to those used in the testing protocol described above. Because in a concentric-eccentric subject pair the eccentric subject was able to do more mechanical work, concentric subjects performed an additional set (set 5 with individual repetition adjustments) to equate mechanical work in the two groups. This means, that within a subject pair, the eccentric subject did the given training session (4 sets of 10 reps) first, and her total mechanical work was calculated immediately, using the dynamometer’s data output for every contraction. Afterwards, her concentric pair performed the session also with 4 sets of 10 reps, for which, we also calculated the total mechanical work. Using then the total work difference between a concentric and an eccentric subject in a session, we estimated how many more repetitions the concentric subject needed to perform in set 5. This procedure was administered in every training session, and the total load in terms of mechanical work for the four-week training, therefore, was identical in the two groups (see Results).

Data processing and statistical analyses

We present the data as mean ± standard deviation (SD). We checked each variable with the Shapiro-Wilk test for normality. When a variable revealed not normal distribution, we log-transformed the data for analyses, but we report the non-transformed data in Tables and Figures. The key analysis was a group (eccentric, concentric) by time (pre, post) analysis of variance (ANOVA) with repeated measures on Time. A significant a group by time interaction was followed with a Tukey’s post-hoc contrast to determine the means that differed at p < 0.05. We characterize these statistical effects with partial eta squared (ηp², F values) and pair-wise time-point and between-group differences with Cohen’s d effect sizes.

Cutoffs for ηp² are ≥ 0.01 (small), ≥ 0.06 (medium), and ≥0.14 (large) [36]. Cutoffs for d are small = 0.20; moderate = 0.50; large = 0.80 [36].

Statistical power was calculated post hoc (G*Power 3.1.9.7), and the tests revealed very high statistical power values: DKV during single leg CMJ = 1.0, DKV during single leg landing = 0.99, eccentric hip abduction torque = 1.0, concentric hip abduction torque = 1.0.

The level of significance was set at p < 0.05.

Effects of hip abductor strength training on hip abduction torque

Eccentric and concentric hip abduction training induced favorable but similar adaptations in eccentric hip abduction torque (20% and 28%, or 1.7% and 2.3% per session, respectively), as well as in concentric hip abduction torque (33% and 42%, or 2.7% and 3.5% per session, respectively). These changes fail to support Hypothesis 1, as there was a lack of training-specificity with respect to contraction mode. Improved hip abduction torque following hip abduction resistance training has been observed previously, with improvements varying between 0.8 to 41.0% from pre- to post-intervention, or 0.04 to 2.06% improvements per session. [18,22,37]. While the torque increases we observed are within this range, we note that previous studies used an isometric hip abduction test contraction.

To our knowledge, this is the first study reporting adaptation in eccentric hip abduction torque after pure eccentric and concentric hip abduction strength training. The lack of group by time interaction for eccentric hip abduction torque can perhaps be explained with our dynamometric exercise protocol: the range of motion during training was set to a narrow 23°, which perhaps prevented contraction-specific adaptations in torque generation to emerge. It is also possible that training with either contraction mode was unusual so that performing these unaccustomed contractions produced similar neural adaptations, underlying the torque changes.

Effects of hip abductor strength training on DKV

Internal rotation of the tibia and high values of DKV seem to play a significant role in the mechanism of ACL tears [38]. One effective way to decrease DKV during single-leg movements is by activating the hip abductor muscles to abduct the femur [3]. Consequently, having enough tension in the hip abductor muscles during single leg countermovement jumps and single leg landings can serve as a preventive measure against large DKV, decreasing the risk of an ACL tear. Hip abduction in resistance exercises can effectively reduce DKV when performing movements with a single leg. However, it should be noted that the DKV in those studies was measured using a single-leg squat test, which may not accurately capture the fast movements that can result in ACL tear [17,18]. In our study, maximal effort eccentric and concentric training each reduced DKV during single leg CMJ so that the reductions were greater after eccentric (−45%, d = 0.94) than concentric (−30%, d = 0.82). Such differential effects were not present in DKV when measured during single leg landing after training (−58%, d = 0.88, Table 2). These data partially support hypothesis 2.

One strategy to minimize the risk of severe knee injuries during single-leg movements is to consciously prepare and regulate the extent of DKV. Previous studies found either small [17] or no effects [18] of hip abduction exercise on DKV, or when favorable effects occurred, a low-velocity squat was used as a test [17,18,20], though making direct comparisons are difficult as we used different population in the current study. To our knowledge, the present study is the first that used eccentric hip abductor strength training to reduce DKV. Because there was a lack of specificity in the training adaptations (discussed previously) while eccentric strength training still induced greater reductions in DKV during single leg CMJ, we interpret these data to mean that participants learned to activate the hip abductors eccentrically. In other words, regardless of the training type (eccentric, concentric), strength gains could be more effectively utilized during an eccentric task. Why such ability was present only in single leg CMJ but not in single leg landing, remains to be further examined. The inconsistency between single leg CMJ and single leg landing adaptations might be related to differences between these two tasks in various neurokinematical mechanisms. First, we previously found that participants activated their gluteus medius muscle in a single leg CMJ 60% more than in a landing task, though our data is limited to males [39]. Difference in knee flexion strategy is also evident, showing that both males and females use considerably smaller knee flexions during landing compared with CMJ [40,41]. This latter suggests that DKV valgus develops faster in landing, and perhaps eccentric hip abductor training with faster contractions would reveal greater reductions in DKV. The neurokinematical differences between CMJ and landing altogether might be the reason that the two training modalities in the current study did not differentially affect the landing task.

Effects of hip abductor strength training on single leg CMJ performance

We found that improvements in single leg CMJ mechanical impulse did not differ after eccentric (9%) and concentric (6%) strength training. There is limited information on the effects of hip abductor strength training on vertical jump performance. Our preliminary data strengthens the report that 6 weeks of hip abduction training increased vertical jump performance by 11% in female volleyball players [42]. Giustino et al. reviewed in detail the potential mechanisms of how hip abduction training could improve vertical jump performance. Briefly, a standard inverse dynamics analysis revealed identical hip (31%), knee (34%), and ankle (35%) joint power contribution to leg CMJ propulsion, suggesting that any dysfunction in these joints would negatively affect vertical jump performance [43]. Furthermore, tibial-femoral misalignment can induce atypical loading patterns leading to reduced jump performance in volleyball players [44]. The above mechanisms together with the fact that hip abductor muscles control DKV [13–15] generally explain jump performance improvements after hip abductor strengthening.

In the present study, we found no superior effects of eccentric vs. concentric hip abductor strength training on single leg CMJ mechanical impulse, rejecting Hypothesis 3. Optimal femur alignment can enhance CMJ performance, and despite that our eccentric vs. concentric strength training reduced DKV more (see discussion below), the better femur position still seems insufficient to favor single leg CMJ performance. Given that the role of anterior and posterior leg muscle strength in single leg CMJ performance is beyond doubt [45], it can be still inferred that hip abductor strength training also has a notable impact on single leg CMJ performance.

Effect of hip abductor strength training on knee flexion during single leg CMJ

Because women tend to land with small knee flexion to avoid excessive DKV, we proposed that hip abductor strengthening would change knee flexion strategy during single leg CMJ. This was, however, not the case because neither eccentric nor concentric hip abduction training did modify knee flexion during single leg CMJs. It is also possible that our participants had weak knee extensor mechanism to generate impulse for single leg CMJs, preventing jumps executed with a greater knee flexion, producing higher single leg CMJ performance. Our participants used only 79–84° knee flexion, even though untrained female college students were reported to used 90° knee flexions during single leg CMJ [46]. It thus seems that changes in knee flexion did not contribute to the improved single leg CMJ performance.