Tendon injuries account for a substantial proportion of chronic injuries in sports,1 especially in explosive jump sports such as volleyball and basketball, which are characterized by high demands on speed and power of the leg extensors.2 In volleyball, a prevalence of patellar tendinopathy of 45% among elite male volleyball players has been reported.2 3 During ballistic movements of the knee joint, the patellar tendon is mechanically loaded according to a characteristic stress–strain curve,1 in which physiological forces usually cause less than 6% strain.4 The repeated activity of jumping in volleyball is often eccentric by nature, and given that the eccentric force production during the movement can potentially be three times that of concentric force production, is believed to be a primary cause of these cumulative microtraumas.5

A studyby Lian et al3 showed that for patients with “jumper’s knee” a higher loading of their quadriceps extensor mechanism resulted in better performance in the jump-testing programme. In their population the right knee was more affected than the left knee. These authors subsequently hypothesised that the observed higher loading, in combination with the generally preferred take-off technique during the offensive spike jump, provides greater knee flexion angle, knee abduction moments and tibial external rotation for the right leg.3 This possibly predisposes the volleyball player to the development of patellar tendinopathy.

During landing after the spike jump, the generated kinetic energy has to be properly absorbed to avoid injury. The importance of appropriate joint flexion to dissipate this kinetic energy during the landing and its association with increased knee injury risk has been extensively studied.6^–9 Additionally, in our previous cross-sectional study, we compared the landing technique after drop jumps of healthy volleyball players and asymptomatic volleyball players with a previous patellar tendinopathy. Overall data analysis indicated that a stiffer landing technique might be a risk factor for the development of patellar tendinopathy. Furthermore, we found indications that volleyball players with current injuries performed a pain-avoiding landing strategy.10

The “hazardous” eccentric loading of the quadriceps extensor mechanism occurs during the first part of the take-off phase and during the landing in volleyball spike jumping. The purpose of this study was to compare the ankle and knee joint dynamics during the performance of the volleyball spike jump between healthy volleyball players and asymptomatic volleyball players with previous patellar tendinopathy to find possible biomechanical risk factors for the development of patellar tendinopathy. When healthy subjects are compared with symptomatic subjects, it remains unclear if the outcomes can be related to the injury or the causal mechanism. Therefore, we compared the spike jump dynamics between healthy volleyball players and asymptomatic volleyball players with a previous patellar tendinopathy.

METHODS

Subjects

In March 2004, an inventory of possible participants for our investigation was made among 89 male volleyball players from the northern part of The Netherlands. All invited players gave their informed consent for the interview, the clinical examination by an experienced sports physician, and the biomechanical testing. Players completed a questionnaire assessing type, history, prevalence and severity of knee injuries in volleyball.

The following inclusion procedure and measurements took place in September 2004, which was the beginning of the volleyball season. Our volleyball population was divided into two groups: an asymptomatic group of volleyball players with previous patellar tendinopathy and a non-symptomatic group of healthy volleyball players. The diagnostic criteria for these two groups has been published previously.10 Further player characteristics are summarised in table 1.

Data acquisition

The spike-jump movement was measured with an Optotrak motion analysis system (Northern Digital Inc., Waterloo, Ontario, Canada) with two cameras containing three sensors each, which captured the three-dimensional coordinates at a sampling rate of 200 Hz. Moulded rigid frames (3.2 mm Aquaplastic) were tightly attached to the thigh and shank segments with wide neoprene bandages and Velcro fasteners. On each frame, four light emitting markers were fitted, and another four were attached to the right shoe at the lateral stiff side of the calcaneus to measure foot segment position. Ground reaction force data were collected by a Bertec force plate (type 4060–08) at a sampling rate of 1000 Hz. The position of the centre of pressure was computed afterwards. After amplifying, all force plate signals were converted to digital by the 16 bit A/D converter of the Optotrak system.

Procedures

Before testing, participants followed a warming-up period, including 10 minutes cycling on a bicycle ergometer and stretching. All participants were right-handed and reportedly preferred a right–left, step–close take-off technique, which they were asked by the investigator to perform during the measurements. Spike-jump take-off was measured for the right leg, with only the right foot hitting the force plate. The same procedure was applied for the spike-jump landing, where the player landed with the right foot on the force plate. For both series, data acquisition was continued until five successful trials (ie adequate landing on the force plate) were available for further analysis. Because of the limited amount of height in our biomechanics laboratory, the players were instructed by the investigator to imagine a ball coming from the right which they had to spike maximally. Before measurement, players needed some practice trials before they consistently performed the movement in a natural way with the right foot on the force plate. During the measurements, participants wore their own indoor sport shoes. To minimise an interfering role of fatigue on the spike-jump performance, players had 5 minutes of rest between each series. Video registration was used after the trials to verify adequate foot contact with the force plate. To verify group classification, players were asked after each trial to report on a scale of 1 to 5 (1, no pain; 5, intense pain) pain in the patellar tendon region when jumping.

Data analysis

To determine the take-off and landing dynamics, a Matlab V.6.5-based motion analyis program (BodyMech; Free University Amsterdam, The Netherlands; http://www.bodymech.nl) processed both kinematic and force plate data. Additional protocols completed the inverse dynamics analysis.

Position data were smoothed through a second order, low-pass, zero phase Butterworth filter with a cutoff frequency of 20 Hz before the first and second derivatives were calculated. From these filtered marker trajectories, joint angular position11 and joint angular velocities were calculated. In addition to to the maximum joint angle, joint range of motion (ROM) and joint angle at the time of touch-down, we determined the clinically relevant joint angles when the vertical ground reaction force reached its peak12 and the joint ROM from the time of touchdown until the time of peak vertical ground reaction force. Force plate data were smoothed with a cut-off frequency of 100 Hz through a second order, low-pass, zero phase Butterworth filter. Three dimensional joint moment values for ankle and knee joint were calculated by combining anthropometric,13 kinematic and force plate data via inverse dynamics analysis, using a three-segment rigid-body model.14 Solely for the assessment of joint moment, both position and force plate data were equally smoothed with a cut-off frequency of 20 Hz to minimise impact peak errors in the moment calculation.15 16 The loading rate of knee extensor moment was calculated for the eccentric phase preceding the push-off phase, and for the landing. These loading rates were defined as the peak value of the first derivative of the regarding moment curve. The loading rate of the vertical ground reaction force was defined as the peak vertical ground reaction force value divided by the time from touch-down until this peak value. Finally, to reduce intersubject variability, biomechanical variables were presented as dimensionless measures, normalised and expressed according to Hof.17

Statistical analysis

Statistical analysis was performed using SPSS V,11.01. For continuous biomechanical variables, acting as possible risk factors for patellar tendinopathy, univariate logistic regression analyses were performed to estimate the probability of the dichotomous variable (healthy players or asymptomatic players with previous patellar tendinopathy). The log likelihood (with improved χ2) was used to assess if the specific biomechanical variable would have a significant effect on the predictive ability of the model. Statistical significance was set at p⩽0.05.

RESULTS

From the population who were asked to participate in this study, seven subjects with previous patellar tendinopathy were included. These players had been symptom-free for a period of 5–12 months before measurements took place. In the previously injured group, three players had a history of bilateral patellar tendinopathy: two of these reported more palpation tenderness in the right knee and one in the left knee. Of the four remaining players with unilateral previous injuries, all had a history of patellar tendinopathy in the right knee. During measurements, none of the players reported pain in the patellar tendon region or physical problems when asked after each trial. Figure 1 represents a graphical representation of the determined biomechanical variables of one participant while performing the volleyball spike jump.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Take-off and landing phase dynamics of one participant performing the volleyball spike jump. Ma, ankle moment; Mk, knee moment; VGRF, vertical ground reaction force. The directions of the analysed joint moments were: the sagittal plane ankle plantar flexion(–)/dorsal flexion(+) moment; and knee flexion(–)/extension(+) moment; the frontal plane knee abduction(–)/adduction(+) moment; and the transversal plane knee external(–)/internal(+) rotation moment (tibial rotational moment). For joint angles the ankle dorsal flexion and knee extension had a positive direction.

For all biomechanical variables describing the spike-jump take-off and landing, a univariate logistic regression was performed. Three individual variables showed a significant effect on the ability to predict the presence or absence of previous patellar tendinopathy in our population. The predicted group classification for all players is presented in table 2.

The first biomechanical variable that significantly improved this prediction was the ankle plantar flexion angle at the time of touchdown of landing (p = 0.05), which correctly predicted group classification in 12 of 15 players. The healthy players showed an mean of 41.8 (SD 6.5) degrees of plantar flexion at the time of touchdown, whereas players with previous patellar tendinopathy showed less plantar flexion at the time of touch-down (34.9 (SD 7.0) degrees (table 3). Hence, the smaller the plantar flexion angle, the greater the likelihood that the volleyball player suffered from a previous patellar tendinopathy.

Another predictive kinematic landing characteristic was the knee ROM from the time of touchdown until the time of peak vertical ground reaction force. Average knee flexion trajectory during this first part of impact was 27.5 (SD 8.0) degrees for healthy players, and 19.1(SD 7.5) degrees for players with previous patellar tendinopathys. This variable significantly improved the prediction of the presence or absence of previous patellar tendinopathy (p = 0.04) in 11 of 15 players. Hence, the likelihood of previous patellar tendinopathy increased with smaller knee ROM during the first part of impact. From the take-off phase, the loading rate of knee extensor moment during the eccentric movement preliminary to the push-off phase significantly improved the prediction of the division of the two groups (p = 0.04); 9 of 15 players were correctly predicted, and a higher loading rate was associated with players with previous patellar tendinopathy (an average of 1.57(SD 0.62) and 2.32(SD 0.76) BM×g^1½×l₀^½ for healthy players and players with previous patellar tendinopathy, respectively (table 4)).

From all univariate tested biomechanical variables that had p⩽0.25, interaction effects were determined (see table 2 for the predicted group classification for each player). From these interaction effects, three couples of variables significantly improved prediction of a history of patellar tendinopathy. Ankle and knee ROM (from the time of touch-down until the time of peak vertical ground reaction force for both) predicted the presence or absence of previous patellar tendinopathy correctly in 11 of 15 players (p = 0.05). The smaller the ankle and knee flexion trajectory during the first part of impact, the greater the likelihood of previous patellar tendinopathy (table 3). Another significant contributor to better prediction was the interaction between the loading rate of knee extensor moment during the eccentric phase of take-off and landing (p = 0.02). This interaction effect correctly predicted the presence or absence of previous patellar tendinopathy in 12 of 15 players. Players with previous patellar tendinopathy showed higher values of the rate of knee moment development (table 4). In addition, presence or absence of previous patellar tendinopathy was correctly predicted in 12 of 15 players by the interaction between knee angular velocity during the eccentric phase of take-off and landing (p = 0.04). Players with previous patellar tendinopathy showed higher ankle and knee angular velocities than did healthy players (table 3).

DISCUSSION

In this study we investigated the take-off and landing dynamics of the volleyball spike jump of healthy volleyball players and players with previous patellar tendinopathy to find possible biomechanical risk factors for the development of patellar tendinopathy.

An important finding in this study was that the volleyball players with previous patellar tendinopathy landed from the spike jump with less ankle and knee flexion, especially during the first part of impact. Univariate logistic regression analysis revealed that ankle plantar flexion at touch-down and the critical knee ROM during the first part of impact (from touch-down until the time of peak vertical ground reaction force) both significantly improved the prediction of the presence or absence of previous patellar tendinopathy. Furthermore, the ankle and knee ROM during the early part of impact (both from touch-down until peak vertical ground reaction force) as interaction effects together improved prediction of presence or absence of previous patellar tendinopathy. In this study, asymptomatic volleyball players with previous patellar tendinopathy absorbed the kinetic energy during the first part of impact from the jump with less ankle and knee flexion than did their healthy counterparts. An adequate landing technique with proper joint flexion is essential to accommodate the excessive impact forces efficiently, and is for that reason thought to be related to the risk of injury.18

Together with the reduced ankle and knee flexion of the previously injured group, knee angular velocity during landing showed a trend (p = 0.052) towards improving prediction of the group classification. However, when l joint kinetics and energetics of the performed spike-jump landings were examined, no single variable during the eccentric phase of landing provided a significant contribution to improving prediction of the presence or absence of previous patellar tendinopathy among the players.

When performing a volleyball spike jump, the quadriceps extensors are eccentrically loaded during the landing to absorb the gained kinetic energy of the jump, and, during the countermovement phase of take-off, to develop a high level of active state and muscle force before the start of pushoff.19 During this eccentric countermovement phase, volleyball players with previous patellar tendinopathy showed a higher loading rate of knee extensor moment. This was a significant predictor for presence or absence of previous patellar tendinopathy. In particular, when the loading rates of the knee extensor moment during the eccentric countermovement of the take-off phase and landing were considered together as interaction effects, they correctly predicted a substantial 80% of the group classification. This finding was accompanied by a significant interaction effect of knee angular velocity during both the eccentric phases of take-off and landing in the improvement of predicting the presence or absence of previous patellar tendinopathy. Previous reports had already pointed to a higher rate of knee extensor moment development and higher knee angular velocities during the eccentric contraction of landing as a possible risk factor for the development of patellar tendinopathy.10 20

In our cross-sectional study design, use of an asymptomatic group of volleyball players with a previous patellar tendinopathy instead of current patellar tendinopathy was a key feature. Risk factors from a case–control study comparing healthy players with current injured players are subject to debate of whether or not these risk factors are an effect of the injury.21 Owing our inclusion of asymptomatic volleyball players with previous patellar tendinopathy instead of symptomatic players with current patellar tendinopathy, we can make the assumption that the performed ankle and knee joint dynamics by the asymptomatic players may be associated with tdevelopment of patellar tendinopathy. It should be noted that the motor behaviour of the previously injured players could hypothetically be a residual effect of a previous patellar tendon injury, used to avoid high force production of the quadriceps extensor mechanism (eg patellar tendon) with the knees flexed. However, this view is inconsistent with the findings found in our previous study, which found that recently injured players landed with a technique to avoid high patellar tendon loading, whereas the previously injured group did not show these load avoidance strategies, and in fact, even showed a stiffer landing technique with a higher rate of moment development during landing.10 These differences in landing technique supports our belief that it is very unlikely that asymptomatic volleyball players with previous patellar tendinopathy have adapted their jump and landing kinematics and kinetics in the past to avoid high loads on their quadriceps mechanism.

In our previous data set we showed the possible importance of the eccentric phase in drop-jump landings in the development of patellar tendinopathy.10 Additionally, biomechanical analysis of players performing functional spike-jump landing sequences confirms this issue of the eccentric phases in relation to its possible role in the developmental mechanism of the chronic sports injury patellar tendinopathy. Insufficient landing absorption techniques characterised by reduced ankle and knee flexion during the first part of spike-jump landing, together with a higher rate of load development of the knee extensor structures might be associated with the development of patellar tendinopathy.

Although data analysis revealed some general movement characteristics that improved prediction of the presence or absence of patellar tendinopathy among our population, it still could not predict correct group classification for all players. It seems that for some players other characteristics play a role to determine whether or not they are at risk. This is supported by some individual cases in the group of players with previous patellar tendinopathy (see table 4 for the predicted group classification for each individual player). The study by Lian et al3 had already shown a relationship between the amount of power generation during take-off and patellar tendinopathy among elite volleyball players. When we examined the spike-jump dynamics of players 1 and 2, it appeared that these two players showed the highest power generation during take-off and the highest jump height of our entire study population. These movement characteristics are in line with the findings of Lian et al,3 and might therefore be possible risk factors for these specific individual players. As already mentioned by Reeser et al,18 changes in training intensity and load might make the player more susceptible to patellar tendinopathy. They stated that talented young players who are abruptly moved from a safe training environment to intensive practice run the risk of developing patellar tendinopathy. In our previously injured group, players 2 and 3 were the only two players who had intensively increased their training load compared with the preceding season. Thus, based on these findings, we can state that besides a general movement pattern that can be associated with the development of patellar tendinopathy, other characteristics such as training intensity and power generation during take-off might play a role in the development of patellar tendinopathy for some players.

This strengthens the idea that the search for the aetiology of sports injuries requires a dynamic model that accounts for the multifactorial nature of sports injuries, where the interaction and sum of the intrinsic and extrinsic factors makes an athlete susceptible to injury.21^–23

The practical implementations of these results are that trainers should be encouraged to be cautious with players who undergo an abrupt change in training regimen by making adjustments in the training programme to avoid the player having to perform too many jump-landing sequences. Furthermore, volleyball trainers should ensure appropriate strength training of the leg extensor musculature and be aware of the need of appropriate ankle and knee joint flexion during the first part of impact.

CONCLUSION

Comparison of volleyball spike jump and landing dynamics between non-symptomatic healthy volleyball players and asymptomatic players with previous patellar tendinopathy revealed that inappropriate joint flexion during the first part the spike-jump landing impact and the rate of knee moment development during the eccentric phases of the spike-jump landing sequence may be risk factors in the development of patellar tendinopathy in volleyball. In addition, for some individual players, other mechanisms might play a role in the development of patellar tendinopathy. These results should enable more goal-orientated dynamic research model with multiple risk factors to be carried out to find the causal mechanisms of patellar tendinopathy. Furthermore, trainers should be aware of the importance of eccentric loading of the leg extensor mechanism during both the countermovement phase of take-off and landing of their players.