Our results demonstrate greater FLX and ER regardless of abduction angle as well as less IR in the dominant compared to non-dominant arm. These findings contrast those of Barnes et al.  that reported no significant bilateral FLX differences in a broad participant population, yet parallel the findings of Johnson  for a college-aged athletic cohort. Factors that may underpin increased FLX for the dominant arm in a related cohort are not currently completely understood, but this phenomenon has been suggested to potentially arise from functional adaptations that permit greater ROM available for the purpose of executing various sports-related tasks, such as overhead throwing . However, similar to Barnes et al. , our results yielded greater active and passive ER regardless of abduction angle as well as less IR for the dominant compared to non-dominant arm. A trend for greater ER and lesser IR in the dominant arm has been previously described for a college-aged athletic population [3, 11, 12] and may be attributed to functional adaptions associated with participation in sports , especially overhead related activities . Furthermore, our results are in agreement with Gunal et al.  who noted greater active and passive CAD in the non-dominant compared to dominant arm. Less CAD in the dominant arm is yet another previously observed trend  in a young athletic population and has been linked to posteroinferior glenohumeral joint capsular stiffness [16, 17], which potentially represents an adaptive phenomenon typically associated with athletic throwing activities . Total active, as well as passive, ARC of IR and ER measured in our cohort were comparable bilaterally, thereby complementing previous findings [7, 18, 19]. Shoulder pain has been identified as a significant factor decreasing total ARC of IR and ER for the dominant compared to non-dominant arm in athletes . None of our participants reported pain, which may explain why no bilateral differences were found.
Contrary to a prior report , our data suggest sex may also influence shoulder complex ROM in an athletic cohort. Greater active and passive FLX and ER0 for the dominant arm as well as greater active and passive FLX and ARC for the non-dominant arm were observed among women compared to men in our investigation, which is comparable to Barnes et al. . Women also exhibited greater bilateral measures of active and passive CAD than men in our study. The influence of sex on shoulder complex ROM measures has been described before [1, 21] and may be associated with muscle morphology differences between men and women . Roy et al.  demonstrated an inverse relation between shoulder complex muscular strength and ROM, and suggested that increased muscle bulk, indicative of greater strength in asymptomatic men compared to women, may yield decreased ROM. Barlow et al.  reported similar observations among healthy, young bodybuilders and non-bodybuilders, and proposed that the observed results were attributed to group differences in muscle mass.
Although we identified similarities and differences for our study and previous reports, comparing our results with prior findings is difficult due to considerable differences in participant populations and methods. Notable sources of variability between our study and preceding investigations are mostly attributed to discrepancies in experimental design and participant demographics. Nonetheless, our study is novel in presenting clinically relevant descriptive data for healthy, young and physically active adults, which represents the population typically treated by sports medicine practitioners. Thus, while our findings are limited in generalizability to other populations, they are well suited for descriptions of young athletes.
Determining descriptive values for an athletic population is essential to providing sports medicine practitioners with a frame of reference with which to interpret clinical measures obtained from respective patients. Such data are also necessary for evaluating patient progression through treatments to determine clinical intervention efficacy. Our investigation indicates that traditional shoulder complex ROM examinations and clinical interpretations, which do not account for arm dominance and sex factors, may be ambiguous in an athletic cohort. These findings suggest shoulder complex ROM differences, both bilaterally as well as among men and women, may be at least partially explained by arm dominance and sex in young, healthy and physically active adults. Therefore, influences of arm dominance and sex should be considered when shoulder complex ROM is clinically evaluated in a similar population.
We observed statistically significant differences for several of the ROM measures examined in our study by arm dominance and sex, yet whether or not these differences are clinically important remains unclear. Due to the large number of observations and associated statistical power, it is likely that even small mean differences reached statistical significance in our study. In fact, the largest mean differences were no greater than 2.5cm for CAD or 5° for any of the goniometric measurements. It has been reported previously that mean differences less than 7°  or 8°  for goniometric measurements may lack clinical relevance, but limited empirical evidence is available to support these values as minimum clinically important differences. The American Medical Association  suggests that differences less than 10° at the shoulder complex may be regarded as clinically insignificant when evaluating permanent impairment. However, much smaller differences may be clinically significant in high level athletes, particularly in sports with considerably upper extremity demands. Thus, further study is needed to empirically quantify minimum clinically important differences for shoulder complex ROM in young athletes.
Our study was associated with certain limitations. One-dimensional goniometric measurements do not concomitantly assess multi-planar joint excursions, which may contribute to the respective resultant net ROM measurement. Thus, more sophisticated, yet non-pragmatic, instrumentation, such as photogrammetry, may be better equipped to comprehensively capture ROM measures. We were also unable to estimate intra- and inter-rater reliability within and between examiners who assessed ROM using standardized procedures, which is an important limitation when interpreting the results of our study. However, we relied on prior reliability estimates [4, 7, 9, 26, 27] for the shoulder complex ROM measures studied. Furthermore, all clinicians completed standardized training prior to data collection to enhance consistency between and within examiners [8, 9]. Although these suggested drawbacks represent important limitations to the internal validity of our study, our measures represent those used in real world settings to assess patients, which considerably contribute to the external validity and generalizability of these results to clinical sports medicine . Continued investigation is necessary to confirm these findings and further contribute to the literature on shoulder ROM in young athletic populations.