The increase in adolescent sport participation over the past years and the specialization in a single sport with intensive repetitive activity at younger ages increase the risk of sustaining an overuse injury.1–3 This risk is nearly twice higher compared to adolescents with low specialization.1–3 In total, 30% of sports-related injuries among throwing young athletes occur in the shoulder.3 Consensus statements released by healthcare and sports organisations recognize the importance of upper limb screening examination as part of the periodic young athlete's health evaluation.4–6

Physical performance tests are a part of this screening examination, provide a more complete picture of the functional status of the athlete's upper extremity, and are routinely used for performance enhancement or post-rehabilitation outcome measures.7–10 Furthermore, physical performance tests are, most of the time, easily performed in many different environments and contexts with minimal material,5,11 and are thus, very popular. The Closed Kinetic Chain Upper Extremity Stability Test (CKCUEST) is one of the most popular physical performance tests to assess upper body function.11 Despite reliable and valid results within the adult population,9,12–16 data about the CKCUEST on adolescents is scarce.17 Moreover, results on adults should not be generalized to young athletes because of the differences in anthropometric characteristics. In previous studies, the CKCUEST was performed in accordance with the guideline described by Tucci et al.9 and Goldbeck and Davies14 with participants adopting a push-up position with a fixed 91.4cm distance between hands. However, this standardized distance may not be appropriate for adolescents by being too wide for some young individuals or too narrow for others.12,17,18 Other studies performed the test with participants adopting a push-up position with their hands located directly under their shoulders12,19 or with hands placed at a width 50% of the participant's height.12

Overuse injuries have been associated with strength impairments affecting the balance between shoulder internal and external rotator muscles.20–22 To the best of our knowledge, only a few studies have examined the relationship between the CKCUEST and shoulder rotation strength.23,24 Lee and Kim23 reported a high correlation between the CKCUEST mean touches and isokinetic shoulder rotation strength while Decleve et al.24 showed moderate correlations between the CKCUEST mean touches and power score and shoulder rotation strength. However, normalizing the hand spacing at shoulder width may influence the relationship found in these previous studies.

The primary purpose of our study was to determine the relative and absolute reliability of the Modified CKCUEST (MCKCUEST)12 in a population of healthy adolescent volleyball and basketball players. We hypothesized that normalizing the hand spacing at shoulder width may be an alternative to improve the consistency in the CKCUEST on adolescent athletes. The secondary purpose of our study was to examine if performance on the MCKCUEST was correlated with the isometric shoulder rotation strength of adolescent volleyball and basketball players.

Seventy-three adolescent healthy basketball and volleyball players completed the study. Demographic characteristics of all participants are presented in Table 1. Results of the MCKCUEST and isometric shoulder rotation strength stratified by sex and sports for day 1 and day 2 are summarized in Table 2.

Intra-session reliability between trials within day 1 and day 2 showed good reliability with ICC2,1 values ranging from 0.86 (95% CI=0.80, 0.90) on day 2 to 0.89 (95% CI=0.81, 0.93) on day 1. The test–retest reliability between day 1 and day 2 showed excellent reliability with ICC3,1 value of 0.93 (95% CI=0.63, 0.97). The SEM was 1.1 touches and the MDC95 was 3.04 touches.

Results of the correlation analysis are summarized in Table 3. Weak non-significant correlations were found for all variables except for the MCKCUEST power score that showed significant weak correlation with shoulder strength variables (r ranging from 0.3 to 0.4).

The primary purpose of our study was to evaluate the intra-session and test–retest reliability of the MCKCUEST on adolescent athletes. This test was modified from the original testing procedure proposed by Goldbeck and Davies14 by spacing hands at shoulder width using the inter-acromial distance to adapt the starting position to the anthropometric characteristics of the individuals.12,18

The MCKCUEST displayed good to excellent intra-sessions and test–retest reliability on the adolescent athletes which is in contrast to the study of Oliveira et al.17 that found weak ICC value (ICC=0.68) on adolescents using the initial push-up position with hands placed 91.4cm apart. The initial standardized hands distance may be a factor that contributes negatively to their results.17,18 Spacing hands at shoulder width using the inter-acromial distance placed adolescent athletes at an advantage when performing the MCKCUEST compared to the original CKCUEST. Anthropometric characteristics of individuals such as narrower shoulder width and/or shorter arm length may interfere with the results by increasing effort needed to stabilize the upper body when the initial standardized hands distance is applied.34 In our study, both sexes used a push-up position while Oliveira et al.17 adopted a knee push-up position for adolescent females. If we compare our results to other reliability studies performed on adults (ICC range: 0.77–0.96; SEM range: 0.9–2.8; MDC95 range: 2.6–7.8),9,12–16,23,35 we show similar results. However, a variety of testing procedures used in the literature, makes comparisons between these studies difficult regarding the average number of touches, SEM, and MDC.7,9,12–17,19,23,35 The CKCUEST as originally described by Goldbeck and Davies14 was performed by athletic male students in a push-up position with their hands touching parallel pieces of tape placed 91.4cm apart. They reported an ICC of 0.92 and number of touches of 27.814 while Lee and Kim23 reported on adults an ICC of 0.97 and an average number of touches of 13. In the study of Sciascia and Uhl13 on symptomatic and asymptomatic participants they reported an ICC ranging from 0.85 to 0.86, an average of touches of 22, a SEM of 2, and MDC of 4.

Some other studies adopted a modified knee push-up position for female participants.9,17 Tucci et al.9 evaluated the test–retest reliability of the CKCUEST in sedentary and active individuals with or without shoulder injury. They demonstrated an ICC ranging from 0.82 to 0.96, an average number of touches ranging from 11 to 30, a SEM of 1.4 to 2.8, and an MDC of 2 to 4.9 De Oliveira et al.17 studied the reliability of the CKCUEST on male and female adolescents using a knee push-up position for female participants. They reported an ICC of 0.68, an average number of touches of 26.8, a SEM of 2.2, and MDC of 6.

Other studies adapted the starting positions of the hands.12,19,34,36 Tarara et al.36 instructed the participants to assume a push-up position with hands located directly under the shoulders to begin the test. The parallel pieces of tapes remained 91.4cm apart. They reported on active college students an ICC ranging from 0.73 to 0.78 and SEM from 7 to 8. Hollstadt et al.19 reported on college basketball players an average of touches ranging from 25 to 32 and Taylor et al.34 reported a range from 23 to 25 on collegiate athletes. Callaway et al.12 evaluated the reliability of 4 variations of hand starting position on adult males. The first variation was the standard test described by Golbeck and Davies.14 The second adaptation spaced hands at shoulder width using the inter-acromial distance similar to our testing procedure. The third variation started with hands at shoulder width and reached the 91.4cm distance, and the last variation placed the hands at 50% of the participant's height. They determined an ICC from 0.84 to 0.93, a SEM ranging from 1.5 to 2.8, and MDC ranging from 4.1 to 7.8 touches. The second variation from Callaway et al.12 is of interest because it is similar to our testing position and comparisons can be drawn. Even though our study shows similar good ICC values compared to our mean scores, the SEM and MDC95 are slightly lower. Our mean score is 30.8 touches on day 1 and 32.2 touches on day 2 while they found 34.2 touches on adult male population. We obtained a SEM of 1 and MDC95 of 3 compared to 2.8 and 7.8, respectively. These SEM and MDC95 values enhance sensitivity to change. Our MDC95 value suggests that a change of 3.04 touches (approximately 10% of the mean score) is indicative of a true change while their study12 shows a %MDC95 of 23%.

From a clinical perspective, the CKCUEST as originally described14 may not be appropriate for assessing all athletes interchangeably. Our study demonstrates that the MCKCUEST is a reliable alternative to CKCUEST that enables to standardize the test to any adolescent population regardless of individual anthropometric difference or sex.

Our results showed only weak correlation between the MCKCUEST power score and shoulder strength variables. Previous studies23,24 have examined the relationship between the CKCUEST and isokinetic or isometric shoulder rotation strength. Lee and Kim23 found a high correlation with the isokinetic shoulder external and internal rotation strength at angular speed of 60°/s and 180°/s (r range=0.87–0.94) while Decleve et al.24 showed moderate correlation with isometric shoulder external and internal rotation strength (r range=0.49–0.65). In their studies23,24 participants performed the initial CKCUEST with the standardized between hands distance of 91.4cm in a push-up position23 or a knee push-up position for female.24 The body position when performing the initial CKCUEST requires a substantial amount of coordination between scapular, glenohumeral, elbow, and forearm muscles and also trunk strength and stability.16 In our study, the modification of hands spaced at shoulder width using the inter-acromial distance might influence the results with adolescents relying more on other upper limb or trunk muscles to perform the MCKCUEST.

Relative reliability of the test was good to excellent and absolute reliability was clinically acceptable in adolescent basketball and volleyball players. Finally, a weak correlation was found between the MCKCUEST power score and isometric shoulder internal and external rotation strength.

The authors declare no conflict of interest.