Editor,—We would like to air three points that arise from the paper Sonographic incidence of tendon microtears in athletes with chronic Achilles tendinosis.¹

Firstly, unfortunately the word “partial” has been omitted from the conclusion (abstract). The conclusion currently reads “There appears to be an association between microtear formation and Achilles tendon rupture”. The distinction needs to be made, as the paper's Take home message concludes with the words “There may be a role for ultrasound in identifying athletes at high risk of tendon rupture”, and in this case it appears that the authors mean complete tendon rupture. It is important that this ambiguity is clarified.

Secondly, although it is well accepted that in many cases tendinosis precedes “spontaneous” tendon rupture,² it does not necessarily follow that sonographic abnormality will lead to (complete) tendon rupture. The tissue based pathologies found by Kannus and Jozsa may be more subtle than can be detected by sonography. We feel that, without longitudinal data in a large cohort with an appropriate control group, Gibbon and colleagues are not in a position to draw a conclusion on sonographic screenings of athletes.

Thirdly, given the cross sectional study design, the authors should provide caveats emphasising that their theory of tendon remodelling paralleling that of bone is purely speculative. Cross sectional athlete-control comparisons (such as those provided to sustain the argument on tendon remodelling) do not provide evidence of causality—such differences can result merely from selection bias. If athlete-control difference were causally related, then one could take up basketball in order to grow tall.

We emphasise the need for prospectively collected data to predict future outcome. Researchers and clinicians should desist from speculating as to the longitudinal clinical significance of imaging data collected at one point in time only. To our knowledge, there is only one published prospective ultrasonography and clinical correlation study in the tendon literature at present,³ and it shows rather poor predictive value of ultrasonography in terms of development of symptoms and clinical findings of patellar tendinopathy in female basketball players. Until researchers and reviewers acknowledge the importance of longitudinally collected data for predicting clinical outcome, our imaging field will continue to be mired in a debate based on speculation rather than science.