Shoulder muscles in the archery draw: Reddy 2015

This is a simulation, not a measurement on archers. The authors filmed the draw of a single adult male archer with an optoelectronic system, using markers on the upper body only, and fed that movement into a musculoskeletal model of the human body, applying a draw weight of 48 pounds. The model works backwards from movement to forces and computes how hard the shoulder muscles work during the roughly four seconds of the draw.

The result the authors state is a ratio: in the model the infraspinatus worked about twice as hard as the supraspinatus as a percentage of maximal voluntary contraction, while the teres minor was activated at the same level as the supraspinatus. From this they conclude that the three rotator cuff muscles considered are loaded more than the accessory muscles, that is the rhomboids, trapezius and deltoid, and therefore carry a higher chance of injury.

The actual numbers are absent. The paper contains no table of values: muscle effort can only be read off six time plots, whose vertical scales reach 8% for the supraspinatus, 20% for the infraspinatus, 10% for the teres minor and the trapezius, 15% for the deltoid and 3% for the rhomboids. Those are the ceilings of the axes rather than stated peak values, and they give an order of magnitude: in the model none of these muscles exceeds a fifth of its maximal capacity during the draw. The joint reaction forces, announced among the aims in the paper's own abstract, are never reported.

The toolchain is described in five steps. An adult male archer with markers on the upper body performs the movement; the motion is captured with a Vicon system and saved in C3D format, with markers limited to the upper limbs to cut computation time; the file is inspected in the MOKKA viewer, which displays markers and segments over time; the C3D and a solid model of the bow in STL format are fed into the AnyBody software, which optimises motion and parameters and then runs an inverse dynamic analysis on the musculoskeletal model.

The muscles considered are split into primary, that is the supraspinatus, infraspinatus and teres minor of the rotator cuff, and accessory, that is the rhomboids, trapezius, deltoid and latissimus dorsi. Effort is expressed as a percentage of maximal voluntary contraction and reported as a curve over time, across an interval that in the plots runs from about 0.03 to 4.15 seconds. The draw weight applied in the model is 48 pounds, given as a figure for an adult male.

What the paper does not describe is almost everything else. Nothing is known about the archer beyond sex and adulthood, not his level, discipline, draw length or dominant hand; it is not stated how many repetitions were captured, how the bow's force-draw curve was represented in the model, or how the movement was divided into the phases of the shot. There is no comparison with an independent measurement, for instance electromyography on the same archer, and no sensitivity analysis on the model's parameters. The authors announce as future work the calculation of stresses on the shoulder blade and collarbone using finite elements.

This is a short conference paper, and this record is short because the material is thin. The movement comes from one person, so there is no between-archer variability, and the archer is not described: without knowing his level there is no way to know whether the captured movement is that of a skilled archer. The model was not validated against anything, and in an inverse dynamic analysis the way load is shared between muscles depends on the optimisation criterion chosen by the software, which the paper does not discuss.

The results can only be read off plots. There is no table, no peak values and no intervals: the single quantitative statement is the two-to-one ratio between the infraspinatus and the supraspinatus. The joint reaction forces and the maximum muscle stress, both named among the aims and in the conclusion, never appear in the results.

The largest leap is in the conclusions. From the fact that in the model the rotator cuff works at a higher percentage of its capacity, the authors derive a higher chance of injury and a set of recommendations about stretching and strengthening: none of that was measured, since no injuries were observed, no shooting volume was recorded and there is no follow-up. The text also contains descriptive slips, such as the passage where locking the shoulder is attributed to the bow arm while the muscles being modelled belong to the drawing arm.

For an archer this paper offers one pointer and no usable numbers. The pointer is a hierarchy: in a simulated draw at 48 pounds, the rotator cuff muscles of the drawing shoulder carry a larger share of their capacity than the trapezius, rhomboids and deltoid, and among the three the infraspinatus is the most loaded. An archer with a nagging string shoulder finds a hypothesis here about where to look, not evidence.

Everything else needs caution. The percentages readable from the plots are low in absolute terms, under a fifth of maximal contraction, and in a simulation of one person they say nothing about the risk run by someone shooting thousands of arrows: shooting volume, which is the factor people mean when they talk about overload in archery, does not exist in this study. The exercises suggested in the conclusions are the authors' recommendation resting on the general literature, not a result of the model. In this registry the source counts as a model rather than a measurement on archers, and its place is beside the studies that measure with electromyography what is only computed here.