Re: “Physical Management of Scar Tissue: A Systematic Review and Meta-Analysis” by Deflorin et al.

We read with great interest the article by Deflorin et al. ¹ reviewing the reported effects of physical treatments on scar tissue. This kind of study is welcome, because little is known about the different approaches available to clinicians to prevent or treat scar dysfunctions.

In general, this meta-analysis shows that physical scar management has a significant positive effect compared with control or no treatment. However, to rely on data as an accurate indicator of the treatment outcome and a basis on which to improve clinical decision making, it is necessary that the tools used be valid, reliable, and responsive. In our opinion, some important issues regarding the assessment methods used in the selected randomized controlled trials (RCTs) need to be addressed, because they could possibly mitigate the results.

First, a prerequisite of psychometric scales is that they have a minimum of three items, for only multiple items enable the assessment of internal reliability and reliability. ² Apart from a few RCTs, which employed instrumental evaluation methods, this prerequisite was not met. Mostly single-item ordinal scales (e.g., visual analogue scale, numeric rating scale, single-item Likert scale) or separate items of the Vancouver Scar Scale (VSS) were used to evaluate variables. These measures do not have regular score intervals, and the absence of a clinical anchor makes it difficult to correctly interpret the statistical variations observed.

Second, the VSS is probably the most widely used scar scale, but it has several pitfalls ^3,4 : (1) poor inter-rater reliability (intraclass correlation coefficient range: 0.03–0.64); (2) the pigmentation is a categorical item—that is, normal color, hypopigmentation, or hyperpigmentation—and cannot be used numerically; (3) different and nonvalidated versions of the original scale have been proposed in the literature, so authors should specify which one they used (at least one RCT measuring pain and pruritus used one such modified version).

Third, the healing process of a scar depends on the time elapsed since its formation, and the same effect size can have different meanings depending on whether the treatment was prompt or delayed. Unfortunately, this information is lacking.

Finally, physiotherapists focus on the dysfunctional aspect—this includes commonly measured physical variables such as pliability and thickness—but scar adhesions can also induce major dysfunctions when they prevent mobility between skin and underlying soft tissues. As becomes apparent from this meta-analysis, assessment of impairment and disability caused by scar adhesions was not included in any of the RCTs examined. A valid, reliable (ICC intrarater reliability: 0.96; inter-rater reliability: 0.87), and responsive (effect size: 1.4/1.2; minimum detectable change: 0.20) tool—namely the Adheremeter—already exists to measure this, ⁵ and we hope it will soon be used not only in clinical rehabilitation but also in research.

In conclusion, although the results of this meta-analysis are encouraging, future systematic reviews need to pay more attention to the measuring instruments used in the clinical studies.