Shoulder strength in archers: Kim 2018 study

Fifteen high-level Korean archers (7 men and 8 women) were compared with 30 healthy non-athletes matched for sex and age, using an isokinetic dynamometer to measure peak torque in six shoulder movements on both sides: flexion, extension, abduction, adduction, external rotation and internal rotation. All participants were right-handed, so the left shoulder is the bow shoulder and the right the drawing shoulder.

The firmest result concerns the bow shoulder. In the archers, peak internal rotation torque was higher at all three speeds (39.8 against 29.9 newton metres at 60 degrees per second, 38.0 against 28.9 at 120, 36.2 against 30.2 at 180) and adduction was higher at the two slower speeds (66.7 against 53.2 and 64.2 against 51.6). On the right none of these differences reaches significance.

The second result is a ratio rather than a force. The quotient of flexion over extension was lower in the archers on both shoulders and at every speed, for instance 49.4 against 61.2 per cent on the left shoulder at 60 degrees per second and 51.8 against 59.6 on the right: the extensors are relatively stronger than the flexors. The abduction to adduction ratio also separates the groups, but only on the left and only at the two faster speeds: 48.6 against 59.4 per cent at 120 degrees per second and 47.2 against 58.3 at 180, so lower in the archers, which indicates relatively stronger adduction. The ratio of external to internal rotation, used clinically as an indicator of shoulder risk, does not separate the groups at the slow speeds: at 60 degrees per second it is 58.3 and 51.0 per cent in the archers against 62.3 and 58.5 in the controls, close to the general-population reference values the authors cite. In absolute terms flexion tended to be lower in the archers and extension higher, with one isolated significant difference (right flexion at 180 degrees per second, 30.9 against 40.3).

The archers were a national squad aged 18 to 35; the controls, 15 men and 15 women aged 19 to 31, were healthy volunteers who did not play sport. Both groups excluded anyone with pain, tenderness or restricted range of motion in either shoulder and anyone with a history of injury, trauma or surgery. The study was approved by the ethics committee of Korea University Anam Hospital.

The measurements come from a Biodex isokinetic dynamometer. Participants sat strapped across the chest, with the axis of the humerus aligned to the machine's axis of rotation, the elbow supported at 90 degrees and the forearm in neutral rotation. Each pair of movements had its own position: for flexion and extension the shoulder in forward flexion with the elbow extended; for abduction and adduction the shoulder at 90 degrees of abduction; for the two rotations, 30 degrees of shoulder abduction with 30 degrees of scapular plane elevation. Testing always began on the right shoulder and always followed the same order of movements. Each movement called for 15 repetitions, with five minutes of rest between trials, and the same protocol for everyone.

Between-group comparisons, for both peak torques and the three ratios, used the Mann-Whitney U test in SPSS 12 with a threshold of 0.05; sex matching was checked with a chi-square test and the anthropometric characteristics with the same U test. No shooting variable was measured: no score, no distance and no electromyography.

The sample is small and the authors list that first among their limitations: with fifteen archers many differences stay below the threshold even when the means point the same way, as with extension, higher in the archers on both shoulders but not significantly so. Their remark that a larger sample would make those differences significant is a prediction, not a finding.

The more serious problem is the comparison itself. The two groups were matched for age and sex but not for build: the archers weighed 76.5 kilograms on average against 64.0 and had a body mass index of 25.2 against 22.0, both significant differences. Peak torques are reported in absolute terms, with no normalisation for body mass or muscle mass, and the authors acknowledge that this interferes with the analysis. Part of what reads as an adaptation to shooting could simply be size.

There is also an inconsistency in the methods: the abstract says testing was done at 30, 60 and 180 degrees per second, while the text and tables report 60, 120 and 180. This record follows the tables. In the peak torque table an asterisk appears on the left external rotation value at 180 degrees per second, whose p value is 0.763, which is a printing slip. More awkward is a contradiction about the abduction to adduction ratio: the abstract calls it greater in the archers on the left shoulder, while the table and the discussion make it smaller (48.6 against 59.4 per cent at 120 degrees per second). This record follows the table.

The paper does not say whether the assessors knew which group each participant belonged to, and the test order was fixed, always starting on the right shoulder, so any fatigue or learning effect falls on the same side every time. Muscular endurance, which matters in archery, and periscapular muscle activity, which the authors could not assess because the required attachments were unavailable, are both missing by their own admission. The archers were recruited symptom-free, so the study describes shoulders in good order and says nothing about injury risk. The reading that extensors and adductors serve to stabilise the bow shoulder during aiming and release is the authors' biomechanical argument, built on the geometry of the bow arm, not something measured on a shot. The closing conclusion comes from the same chain: that a programme selectively strengthening adductors and extensors might improve performance. No training was given and no score was collected.

The study describes what the shoulder of a pain-free high-level archer looks like next to that of someone who does not play sport, and the picture is asymmetrical. Almost all the measurable differences sit on the bow shoulder, in internal rotation and adduction, the movements that pull the arm towards the body's axis and rotate it inwards: the muscles that hold steady a shoulder being pushed sideways by the bow. On the drawing shoulder, where many would expect the bulk of the adaptation, peak torques do not stand apart from the controls.

For anyone training in the gym the interesting figure is the flexion to extension ratio, lower in the archers on both sides. It is a profile shifted towards the extensors on both shoulders, though it licenses no conclusion about scores: nobody shot an arrow in this study, and the suggestion of selectively strengthening adductors and extensors remains the authors' hypothesis.

Just as useful is the ratio that does not change. The quotient of external over internal rotation, used clinically to monitor the shoulder, stays close to general-population reference values at the slow speeds in these symptom-free archers. It serves as a snapshot of healthy shoulders, not as an assessment criterion: shoulder pain needs to be looked at by a doctor or a physiotherapist, and the mean isokinetic profile of fifteen athletes says nothing about any one person.