Comments on ‘Experimental versus computational analysis of micromotions at the implant - bone interface’

We read the paper by Tarala et al., which was recently published in this Journal. We believe the authors addressed a very important issue. In fact, we agree that artifacts are induced by bone deformation when implant–bone micromotion is measured at points that are far from each other, and more, in general we agree on the importance of assessing the precision of experimental measurements (as well as of numerical simulations). We would like to comment on some points. First of all, when quoting some of our previous studies they state ‘a single LVDT implant–bone relative motion measurement system is mounted transcortically by means of an anchorage set-up; the motion is measured between the pin connected to the stem and the linear variable differential transformers (LVDTs) support attached to the bone surface’. This is partly true for one of the two quoted papers (even in this paper three linear variable differential transformers (LVDTs) and one extensometer were used, not just ‘a single LVDT’). This statement is inaccurate for the other paper: in fact, as shown in that paper, the reference frame of each of the four LVDTs was connected to the bone close (500micron) to the stem–bone interface (as opposed to the bone surface) by means of a tiny sleeve inserted onto the transcortical holes (Figure 1). We used the same type of fixation for the LVDTs in a number of later studies on cemented and cementless stems, many of which are published in this Journal, which Tarala et al. did not mention. Therefore, the statement that ‘None of these experimental methods allow for micromotion measurement at the actual implant–bone interface’ seems unjustified. Second, referring to the errors induced by elastic deformation of bone, Tarala et al. state ‘It is not possible to assess the magnitude of these errors with the currently available experimental methods’. This is incorrect. In fact, in a study that they quoted we measured experimentally the relative motion of different points across the thickness of cortical wall. The anchorage of the LVDTs was fixed to the external surface of the bone, while the pin attached to the LVDT probe sensed the shear motions corresponding to the elastic strains at three controlled depths across the cortex (500micron below the external surface, 500micron above the bone–stem interface, and midway between these two levels). The LVDTs we used had an intrinsic precision of 1micron, and the entire measurement system had an overall precision of better than 2.3micron. The relative motions between a point on the external bone surface and the point close (500micron) to the stem–bone interface did not exceed 5micron in any of the bone areas where such artifact was assessed. Furthermore, based on the micromotions measured at three depths, it can be seen that such relative motion varied, more or less linearly, across the thickness of the cortical wall. Similarly, in a latter study a different group investigated this type of error combining in vitro measurements with a validated FE model. Third, Tarala et al. aim at assessing the error of experimental measurements. As a general rule, in order to measure the uncertainty of a measurement system it is necessary to use something that is more precise than the system under investigation. However, the precision of their finite element (FE) models is questionable: