Teoria tradicional de la terminologia

Results of the fixed effect meta-analysis (Relph et al., 2014) indicated there are statistically significant differences in the proprioception, in terms of JPS acuity and threshold to detection of movement, of patients with ACL injury in that they have poorer proprioception than people without such injuries and poorer proprioception in the injured than uninjured leg. The proprioception of people whose ACL was reconstructed was statistically significantly better than those whose ligament is left unreconstructed (ACL deficient). These differences are seen whether the comparator group is a patient’s uninjured leg, or a control group of people with no injuries; suggesting that either can be used as a control group in future research.

However, results of an additional random effect analysis revealed ACL patients may have worse JPS than their contra-lateral leg, but not compared to an external control group. Therefore, results are inconsistent and no clear conclusions can be made regarding ACL injury and proprioception. This is probably due to the large variation in approaches to JPS measurement, indicated in the analysis by the high I2 _{scores (these ranged from 42% to} 78%). Therefore, a consistent knee proprioception measurement protocol must be developed in order to distinguish whether ACL injury does indeed decline proprioceptive ability. However, the significant differences that were reported in the meta-analysis were seen most clearly when joint position sense was measured but were less apparent when threshold to detect passive motion measurement techniques were used; the meta-analysis revealed greater differences in joint position sense (JPS) than studies using TTDPM. Furthermore, comparison of the fixed and random effects model results produced opposite findings, again suggesting the result of an ACL injury on dynamic proprioception is not clear. Techniques may be insufficiently sensitive to detect the responses of rapid receptors such as the pacinian corpuscles in the ACL (Barrack and Munn, 2000) as measurements incorporate the participants’ reaction time, which is unrelated to their injury. JPS methods may be more sensitive as these measurements also incorporate the slower responses of the ruffini nerve endings and Golgi tendon organs (Schultz et al., 1984) and allow the conscious perception of joint motion and position. Therefore, joint position sense should be used to measure knee proprioception.

These findings were supported by the majority of literature excluded from the meta-analysis (Relph et al., 2014). As stated previously, it is thought mechanoreceptors in the ACL provide

afferent information on the relative position and movement of the knee joint (Riemann and Lephart 2002a, Johansson et al., 2000, Schultz et al., 1984). Therefore, ACL injury may well impair proprioception through disruption to the transmission of this sensory information (Barrack and Munn, 2000). Marks et al., (2007) suggest other articular structures in the knee joint may attempt to compensate for the loss of ACL afferent signals. However, these compensatory signals may be ‘nonphysiologically disorganised’ (Marks et al., 2007 p.42), and hence the central nervous system and consequently joint position are disturbed and the knee become more unstable. The differences in directional (i.e. flexion or extension) proprioception may be due to the location of the injury. The anteromedial bundles of the ACL are most taut in flexion, the posterolateral bundle tautest in extension. Therefore the area of deficiency may determine which direction the deficits in proprioception lie. However, there is significant research to suggest injury may not reduce proprioceptive ability. In a similar study design to Relph et al., (2014), Fyhr et al., (2014) reported results of their meta-analysis on shoulder injuries and proprioception; it was suggested there is only limited to moderate evidence for a proprioceptive deficit following injury. Ambiguity in previous research can be attributed to differences in methods, for example research design, participant injury type, rehabilitation completed, equipment used, proprioception methods and outcome measures. Furthermore, very few studies on ACL injury and knee proprioception include information on the reliability, sensitivity and measurement error of the measurement techniques used. Generally the statistical analysis does not provide appropriate detail. For example only two studies in the meta-analysis (Mir et al., 2008, Angoules et al., 2011) reported whether the data was normally distributed and hence justified the use of parametric statistics. Many studies used ‘home-made’ measurement devices prepared specifically for data collection but the reliability and sensitivity were infrequently reported. Again, only two of the studies included in the meta-analysis reported reliability statistics. Mir et al., (2008) stated test-retest reliability using a correlation coefficient (0.99); however this was from a previous study which was not referenced. Angoules et al., (2011) did comprehensively report the accuracy of their data collection methods, reporting the standard error of measurement (SEM), coefficient of variation (CV), smallest detectable differences (SDD) and intraclass correlation coefficients (ICCs) for each of their seven measures of knee proprioception. Hence, as reliability and validity is lacking in the majority of studies it is possible that the differences in proprioception found after an

ACL injury are due to measurement error and/or the measurement techniques were insufficiently sensitive to detect clinically significant differences (Relph et al., 2014). Another explanation is that the comparisons included in the meta-analysis could be under- powered because the sample was too small, (again, very few of the studies discussed calculated sample size using power estimations). However the pooled data from the meta- analysis (Relph et al., 2014) involved nearly 200 patients and the 95% confidence intervals of the comparisons made were small, indicating that a lack of power was not an issue. Further research is needed to evaluate the sensitivity and reliability of techniques to measure proprioception at the knee, before they can meaningfully be used as an evaluation tool.

A more likely, but controversial, explanation of such ambiguous findings is that ACL injuries do not have a major impact on proprioception at the knee. This might support the view that muscle, rather than ligaments, provide the primary afferent information in the sensorimotor system (Beard and Refshauge 2000) which is not a surprise given that only 1- 2.5% of the ACL total area is made up of proprioceptive receptors (Barrack and Munn, 2000)and that receptors are often still deficient six months after reconstructive surgery (Barrack and Munn, 2000). It may, to some degree, also explain the inconclusive evidence for reconstructive surgery and conservative (non-surgical) rehabilitation (Beard and Refshauge 2000, Friden et al., 2001, Tagesson et al., 2008), while some patients ‘cope’ with an ACL-deficiency and have an apparently stable knee even after complete rupture, others do not ‘cope’ despite reconstructive surgery and apparent passive stability (Barrack and Munn, 2000, Beard and Refshauge, 2000, Herrington and Fowler, 2006, Friden et al., 2001). Given that joint stability relies on synergy between muscles and ligaments (Ryder et al., 1997, Lephart et al., 2000, Huston et al., 2000, Smith et al., 2010), once the ligament is damaged, patients may adapt by using proprioceptive information from the muscles to a greater extent to compensate for the lack of information from the ligament. This may explain why some patients cope better with ACL injury (however it is managed) than others (Herrington and Fowler, 2006); some may be more able to make that adaption more than others. The surgical treatment used to reconstruct the ACL may also influence proprioceptive rehabilitation. Although evidence shows auto-graft techniques can produce structural improvements such as stiffness and ultimate load that exceed uninjured ACLs (Woo et al., 2005), it is unclear whether the tissues used in this surgery allow mechanoreceptor regeneration or optimisation of the remaining mechanoreceptors. Therefore, surgical techniques may also hinder proprioceptive ability.

A limitation of research into ACL injury and proprioception is that all data collection is retrospective, which inevitably means pre-injury proprioception is unknown. It is possible that patients who suffered injuries had poorer proprioception which predisposed them to injury. Large scale normative studies are needed to give insight into the distribution of proprioception abilities across the population and whether this predisposes people to ACL injury. Such studies should consider a measurement technique that explores the full range of knee motion and direction using large sample sizes that represent the complete ACL patient population and normative data on proprioception ability.

It must also be noted that heterogeneity of variance in the referenced meta-analysis (Relph

et al., 2104) and the fixed effect model was greater than the recommended level of 50%

(Deeks et al., 2008) in all but one comparison; this may be due to variability in the recruitment strategies across studies. The time since injury when proprioception was measured and the use of rehabilitation programmes was not consistent. Highly varied measurement techniques were also evident, which is a limitation that hampers further analysis. Different pieces of measuring equipment and varied knee movements, in terms of direction and speed of motion, were employed (see appendix 1c). Proprioception increases towards the extremes of range of movement in order to protect the joint from injury (Barrack and Munn, 2000, Borsa et al., 1997), thus studies that do not include measurements across the whole range of movement may either be under- or over- estimating knee proprioception. These inconsistent methods of measuring proprioception could have contributed to the high levels of heterogeneity in the current analysis. However as there is no gold standard method of measuring knee proprioception, this variation was unavoidable.

This section of the thesis examined the effect of an ACL injury on proprioception, in terms of joint position sense and threshold to detect passive motion. The results indicate that patients with ACL injury may have poorer proprioception than people without such

injuries and poorer proprioception in the injured than uninjured leg. The proprioception of people whose ACL is reconstructed may be better than those whose ligament is left

unreconstructed (ACL- deficient). This may be due to an increase in knee stability. These differences are seen whether the comparator group is a patient’s uninjured leg, or a control group of people with no injuries; suggesting that either can be used as a control group in future research. However there is variability in proprioceptive measurement techniques and a lack of reliability and validity statistics. There is also inconsistency in findings when

a fixed effect model is compared to a random effects model, again providing evidence that a standardised protocol for collecting knee proprioception is needed. There is also a need for large scale normative data to make appropriate comparisons to injured populations, indeed Stillman (2002) concludes “clearly there is a need for more normative data derived from reliable instrumentation…” (p.559). The following sections consider the topics of proprioception, and the current measurement tools of this “sixth sense” (Berthoz, 2002).