This study included 20 participants in the EG, which comprised 10 men and 10 women with chronic achilles tendinopathy (onset period: 7.67±1.64 month) and running experience of more than 1 year, as well as 20 participants in the CG, which comprised individuals of similar age and physical condition compared with those in the EG. Subjects were recruited from the K Rehabilitation Hospital Research Center located in Seoul (Table 1). The inclusion criteria for EG patients were a diagnosis of achilles tendinopathy based on structural abnormalities found on only one side by sonography, achilles tendinopathy for at least 6 months, availability for outpatient follow–up, and ability to achieve independent gait without assistant devices.

The exclusion criteria were ankle range limitations (ankle dorsiflexion 30 degrees or less, plantar flexion 50 degrees or less), previous orthopedic surgery on the lower limb, use of foot orthoses, neurologic injuries or lesions, and other disorders such as osteoarthritis, rheumatoid arthritis, or osteoporosis in the feet in addition to achilles tendinopathy. Ample explanation was provided to the test subjects by the researcher regarding the purpose and the protocol of the study before the study was conducted, and all the subjects understood the explanation and agreed to participate. The project was approved by the University of Sahmyook Research Ethics Review Committee (SYUIRB2010-007, 27 February 2010), and the study protocol was conducted in strict accordance with the Declaration of Helsinki. Written informed consent was obtained from each subject.

The general features of the subjects, such as sex, age, weight, height, and body fat percentage, were recorded using bio-impedance body composition analyzer (Inbody 570, Biospace, Korea) before the test. For the EG, the clinical characteristics of achilles tendinopathy were also recorded, such as duration. Sixteen motion analysis markers (8 makers each side) were then attached to the subjects, who then were asked to walk in a manner suggested in a previous study, namely, repeating a distance of 13 m at their normal walking speed 5 times (Ryan et al., 2009).

For the gait analysis test in the EG and CG, three-dimensional movement analysis equipment was used, which comprised 9 infrared cameras, a signal control box, a computer, and the necessary software (VICON v8i motion analysis system; Vicon, Los Angeles, USA). The cameras received reflected infrared light from each marker, and location movement data were collected at 120 frames per second and an average accuracy of 0.85 mm.

Kinematic data were collected as primary data for body modeling by using Nexus 1.7 software (Vicon). From the collected data, the anatomic position was reanalyzed and used for bone structure modeling. Using the basic principles of physics, simple location data were converted to bone movement data with Polygon software (Oxford Metrics Ltd., Oxford, UK), whereby gait parameters such as cadence, single-limb support, double-limb support, limb index, step length, step time, step width, stride length, stride time, and walking speed were calculated and used for analysis.

To measure the joint moments of the EG and CG, a 60 cm × 90 cm-sized dual AMTI force platform (Advanced Mechanical Technology Inc., MA, USA) was used. After the current runs through the amplifier and is changed to a digital value by Nexus 1.7 software, it can be collected as primary data. Using basic principles of physics, hip, knee, and ankle moment data were then derived by Polygon software and used for analysis. The leg side used for comparisons between the EG and CG remains controversial. This study used the method described in a previous study by Choi et al., namely, use of the affected side in EG subjects and the dominant side in CG subjects (Choi et al., 2011).

Subjects were asked to walk a distance of 13 m on bare feet at their own walking speed and performed the test after a training period to fully understand the movements required. All subjects walked back and forth over the whole distance 5 times during the test, with a 1-minute rest between gait routes. All subjects who participated in the test wore black pants, which minimize the interference of clothing with joint movement and have excellent elasticity and a comfortable feeling on the skin. The location of the markers attached for three-dimensional movement analysis is shown in Figure 1.

The markers were plotted on the X-axis (coronal plane), Y-axis (sagittal plane), and Z-axis (transverse plane) on the basis of gait time and force platform data. The joint moment was calculated using the moment equation reported by Flanagan and Salem (2005), and the limb index was calculated by dividing the total limb support (single- + double-limb support) by the total limb support of the contralateral side of the lower body. All procedures were conducted directly by the researcher.

SPSS ver. 12.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. The mean and standard deviation were calculated using descriptive statistics to compare the general features, gait parameters, and hip, knee, and ankle moments between the EG and CG. Independent t tests were used to compare subject characteristics, gait parameters, and hip, knee, and ankle moments between the 2 groups,. The significance level (α) of all statistical tests was set below 0.05.

The results of this study show that, among the gait parameters assessed in runners with achilles tendinopathy, double-limb support, step length, step width, stride length, stride time, and walking speed showed significant differences compared with the CG. Walking speed has been identified as an indicator of functional activity and quality of life after diagnosis of the condition, and walking speed has been shown to affect activities of daily living (Jonsdottir et al., 2009). Moreover, walking speed is closely associated with other gait parameters. At a gait speed of 6 km per hour, which can generally be said to constitute fast walking, the increase in speed also increases cadence, step length, and stride length and produces a decrease in stance phase and double-limb support. This change has been shown not only in normal individuals but also in patients (Gaudreault et al., 2010; Stoquart et al., 2008).

According to gait studies conducted on ankle rheumatoid arthritis patients, a decrease in walking speed is known to comprise a strategic compensatory gait movement (Mejjad et al., 2004; Semple et al., 2007; Turner et al., 2008). In a study by Aronow and Hakim-Zargar (2007), it was reported that, in patients with ankle injury, slower walking speed is associated with shorter stride length, - observed in gait movement pattern. Further, injury, slower walking speed is associated with shorter stride length, - observed in gait movement pattern. Further, Laroche et al. (2007) showed that, compared with normal adults, patients with ankle injuries exhibit slower walking speeds, decreased stride length, and increased double-limb support percentage of gait cycle. Even in the present study, the EG showed a significant reduction in walking speed compared with the CG, and similar results have been observed for step length, stride length, and double-limb support in patients with other conditions. In previous studies, it was found that step width increases as gait time and step length decrease; this finding is similar to the results of the present study, which showed that the EG exhibited a decreased step length and stride length compared with the CG, but an increased step width.

With respect to the moments of lower extremity joints in the sagittal plane under normal gait conditions, the hip generates trunk control, stabilization, and moment for hip extension during the initial contact, whereas in mid-stance, the hip generates flexion moment to reduce hip extension. This flexion moment is generated as a result of the passive force on the anterior capsule of the hip and hip flexor muscle activation. Thereafter, in the swing phase, extension moment is generated for the decrease in hip flexion movement. When the knee moment is observed in the sagittal plane during the initial contact, it can be seen to absorb shock and provide adequate knee array through the flexion moment, which is immediately followed by the occurrence of knee extension that is maintained until mid-stance and is then converted again to flexion moment during the terminal swing phase. The ankle shows a plantar flexion moment until mid-stance, which is decreased thereafter and consistently maintained during the swing phase (Gaudreault et al., 2010). This aspect of normal lower limb joint moment conforms to the average moment changes that occurred during the gait cycle of CG subjects in this study.

The maximal plantar flexion moment of the normal ankle tends to increase in proportion to the increase in walking speed (Stoquart et al., 2008). On the contrary, in patients whose walking speed is reduced because of pain and musculoskeletal problems, the maximal plantar flexion moments decrease (Lelas et al., 2003; Scott and Winter, 1990). The ankle moments recorded in our study show that the plantar flexion moment of EG subjects during the pre-swing phase was decreased compared with CG subjects, which corresponds with the observation that walking speed is decreased in achilles tendinopathy patients. It is believed that the structural changes caused by the damage to the achilles tendon affect plantar flexion (Wagner et al., 2004).

In our study, the hip moment showed significant differences for initial contact, mid-stance, terminal stance, and pre-swing phases. The extension moment, which should normally occur during the initial contact, and the flexion moment, which should appear from mid-stance, were found to be decreased in the EG. Even in the knees, the flexion moment, which occurs during the initial contact and is needed for initial shock absorbance, and the pre-swing extension moment were found to be decreased in the EG compared with the CG. This finding indicates that the turning force that is generated in the hip and knee is decreased in EG subjects. Novak and Brouwer (2011) examined the lower body moments of elderly people, in whom walking speed and stride length are relatively decreased compared with adults in their 20s, and found that the maximal moments in the hip, knee, and ankle during the gait cycle were decreased compared with adults in their 20s. Similar to this previous study, the hip and knee moments in our study were found to be lower during maximal moment generation in the EG compared with the CG. Furthermore, Lewis and Ferris (2008) showed that the change in ankle moment during the stance phase affects the hip moment, and McCrory et al. (1999) found that achilles tendinopathy patients experience kinetic changes in the lower extremities because of the decrease in plantar flexion range of motion in the stance phase, and that the decrease in ankle moment, which is affected by the decrease in walking speed, appears to have an indirect effect on the formation of the hip moment. When looking at existing studies on the co-relationship between lower extremity joints, it can be seen that, in the case of anterior cruciate ligament patients, the knee moment correlates with other lower body joints, resulting in changes in hip and ankle moments during gait (Shimokochi et al., 2009).

Such changes in lower extremity joint moments interrupt the formation of energy-efficient moments of each joint and can cause various changes in these joints (Bulgheroni et al., 2007). Further, a reduction in stance phase moment reduces the overall range of motion in the lower extremities, which may result in claudication (Jonsdottir et al., 2009).

As indicated above, changes in gait parameters and hip, knee, and ankle moments occur in runners with achilles tendinopathy. Therefore, the clinical features of these gait changes should be understood for optimal treatment of achilles tendinopathy, and further research on this topic is required.