Proprioception in the ACL-ruptured knee: The contribution of the medial collateral ligament and patellar ligament. An in vivo experimental study in the cat

1. Introduction 

In the absence of anterior cruciate ligament (ACL), secondary restraints such as menisci, ligaments, and tendons restrict anterior knee laxity. The most significant secondary restrictors to the ATT are the medial meniscus, the medial collateral ligament (MCL), the patellar tendon (PT) acting as a passive ligament, and the posterolateral corner structures [1]. These not only contribute to mechanical control of anterior knee laxity, but the remaining knee proprioception after ACL injury is linked to the remaining articular and periarticular structures that intervene in the joint stabilisation during knee motion and activities. It has been hypothesized that these secondary restraints may be the proprioception agents in the ACL deficient knee [2], but their role is unclear. In the long-term, their inefficiency to properly signal the anterior tibial translation (ATT) that elicits a control response, may originate giving-way episodes. The assessment of in vivo instability in ACL-deficient knees would require a new strategy of stress–strain measurements in overloaded structures after ACL injury, like the MCL [3] or the PT [4]. Strain detection at these sites could define the contribution of this alternative signalling system to knee proprioception after ACL injury. However, accurate detection of ATT is a difficult task.

An excessive ATT at certain knee flexion angles is the standard evaluation of anterior knee laxity after ACL injury. Anterior translation values have been measured in humans, with the help of devices such as KT-1000 or KT-2000 [5], and even in animals, with more or less sophisticated methods [6], [7]. EMG has been used as an indirect technique of evaluating ATT during time, as muscular activity changes (amplitude and latency) may inform about the quadriceps and hamstrings response to knee motion. However, these data are non-specific and highly variable among individuals [8]. Different devices have been developed to directly measure stress and strain in soft tissues. From stress measurement techniques to pressure detection, or even magnetic field properties to deduce strain, different authors explored sensor capabilities to detect knee instability [9], [10], [11], [12], [13], [14], [15], [16], [17]. Surface strain gauges based on electric variations of a metallic resistance [18], [19] may be a simple way to detect surface tensile variations at ligament insertions, previously used in biomechanical experiments dealing with ligament stress in human cadaver knees [20]. Insertion location of the gauge would avoid confusion due to viscoelastic hysteresis at the ligament body during knee motion. Its sensitivity could be defined as the relationship of the output voltage per time or the output voltage per applied load [11].

We hypothesized that data from surface strain gauges on the insertions of secondary restraints to anterior tibial displacement due to ACL injury may be sufficiently sensitive and specific to differentiate stable from unstable knees after ACL injury, during anterior tibial displacement manoeuvres. Firstly, we aimed to detect if significant differences exist between gauge measurements at the MCL and the PT before and after ACL section. Secondly, we aimed to differentiate gauge measurements during ATT from those obtained during knee flexion and extension.

3. Results 

3.2. Changes in ligament strain with anterior tibial traction 

After ACL section, the first peak amplitude increased when recorded both at the MCL (from 0.38 to 2.61 mV; p<0.001) and at the PT gauge (from 0.36 to 3.26 mV; p<0.001) during ATT with a 90° flexed knee. At the 30° flexed knee, voltage variations were similar (0.43 to 3.36 mV, p<0.001, for MCL; 0.64 to 3.33 mV, p<0.001 for PT). The first peak latency showed a statistically significant decrease during ATT 90° (MCL: 795.57 to 409.44 ms, p<0.001; PT: 723.71 to 398.57 ms, p=0.005), but not so clearly at ATT 30° (MCL: 532.32 to 371.79 ms, p=0.03; PT: 495.2 to 360.55 ms, p=0.09). However, the first peak slope was the variable that showed the most definite changes when the ACL was sectioned, and these are summarized in Table 1. The slope increase during ATT both at 90° and 30° flexion showed large increases at the MCL and at the PT.

Table 1. Changes in the first peak slope after ACL section

		MCL			PT
		Stable	Unstable	p values for the difference	Stable	Unstable	p values for the difference
		mean±S.E.	mean±S.E.		mean±S.E.	mean±S.E.
	ATT 90°	0.32±0.06	3.51±0.62	⇑ (<0.001)	0.29±0.06	8.68±2.69	⇑ (<0.001)
	ATT 30°	0.44±0.07	8.12±1.77	⇑ (<0.001)	0.47±0.07	7.46±1.56	⇑ (<0.001)
	Flexion	0.003±0.001	0.19±0.14	0.5	1.67±1.67	0.30±0.21	0.4
	Extension	0.002±0.001	0.12±0.11	0.6	0.08±0.08	0.42±0.41	0.5

MCL = medial collateral ligament, PT = patellar tendon. p values are given for Mann–Whitney's test.

Percentage changes for slope are shown in Fig. 3, and revealed more than a 10-times increase during ATT. Thus, both sensors showed high specificity to detect changes after ACL section (over 85%) during ATT, although sensitivity was not so high (over 70%), as shown in Table 2.

View Large Image Download to PowerPoint

Fig. 3. The first peak slope percentage changes after ACL section. MCL = Medial Collateral Ligament, PT = Patellar tendon. Significant changes with ACL section at the Mann–Whitney test are pointed.

Table 2. Sensitivity and specificity for strain gauges variables to detect knee instability

		Amplitude		Latency		Slope
		Sens	Spec	Sens	Spec	Sens	Spec
	MCL
	ATT 90°	61.54%	86.67%	30.77%	96.67%	74.36%	90.00%
	ATT 30°	76.19%	100.00%	40.48%	93.55%	80.95%	100.00%
	Flexion	20.00%	100.00%	33.33%	86.67%	13.33%	93.33%
	Extension	27.27%	90.91%	36.36%	81.82%	9.09%	90.91%
	PT
	ATT 90°	48.65%	97.06%	5.41%	97.06%	72.97%	85.29%
	ATT 30°	54.76%	96.77%	30.95%	93.33%	78.57%	96.67%
	Flexion	16.67%	94.44%	16.67%	83.33%	0.00%	94.44%
	Extension	38.46%	100.00%	0.00%	100.00%	7.69%	92.31%

Threshold for each variable was set on mean value+2 S.D., for amplitude and slope, and mean−S.D. for latency. MCL = medial collateral ligament, PT = patellar tendon, Sens = sensitivity, Spec = specificity.

4. Discussion 

Biomechanical studies on ACL deficiency are abundant, as this injury is one of the most commonly addressed in orthopaedic biomechanics. Strain gauges have long been used in this research. Our approach, based on the investigation of secondary restraints to anterior tibial displacement when the ACL is absent, provides a different view of the problem that may help to explain changes of instability with time [7]. It also characterises periarticular muscle reaction in the evolution of an ACL deficient knee [2], providing the role of secondary restraints is clear. Defining and comparing knee instability and proprioception experimentally helps interpreting the neuromuscular changes after ACL injury [23], [24], [25].

Neuromuscular studies on proprioception in the ACL deficient knee have been greatly advanced in the cat's knee as an experimental model, from the original description of articular nerve endings [26] to ligamentomuscular reflexes [27] or muscular reaction [17] and neural pathways [28]. Thus, it has allowed significant insight in knee proprioception currently, with a considerable body of literature to support it. This includes the fact that neural endings within articular structures are stimulated by strain. This proprioceptive potential passes to spinal ganglia and spinal cord, where a muscular reaction is mediated via a gamma loop [29]. This muscular reaction includes muscle tone regulation, which is affected following ACL injury due to a proprioceptive deficit [23]. In this context, the remaining proprioception should be located in the remaining mechanically competent structures, the secondary restraints, such as the MCL. Our study evaluated both the MCL and the patellar ligament as a passive restraint [4]; the latter also served as a control. If the muscular control is inadequate due to poor proprioception then instability and giving way will occur. This leads to the critical role of muscular reaction timing.

With the aim of obtaining better time-dependent information about the ACL deficient knee, we planned surface strain detection in secondary restraints to anterior tibial displacement through gauges. Methodological issues include the use and location of strain gauges, the choice of variables providing useful information, and the unavoidable data variability.

Regarding the use and location of gauges, we selected devices of small size and avoided implantation among ligament fibres that could distort the ligament function. These two aspects are important dealing with ACL deficient knees, were the use of DVRT [30], [31], [32] or force probes and Hall effect-based transducers is difficult. Insertion placement skipped tissue disruption at the ligament body, leaving intact the ligament structure [33], and avoiding deformation or hysteresis at the ligament body. Thus, the obtained voltage variations were expected to more precisely reflect the tensional state at the secondary restraints, as signalled by others [4]. These gauges previously used in cadaver knees [20], permit in our study to detect strain on real time by studying a living system.

Regarding the study variables, latency has been previously used [8], [34] to compare a time-related response. First peak latency (the response delay) has been used to synchronize strain changes with other parameters (such as muscle electrical activity, measured ATT, stresses, etc.) [18], [23]. Other authors studied voltage amplitude with time or with load [7], [11], [19]. However, latency and amplitude may be confounding variables as the starting point for latency may be inaccurate, and the voltage amplitude expresses a different reaction if spread along time. We then calculated the slope, a magnitude obtained from amplitude and time of the first peak in the recorded voltage, to address time-depending changes establishing a relationship between strain changes and time [11], [35]. When other factors influencing the gauge output are under control, a higher slope indicates an earlier and more intense deformation under traction. This occurred in our study, especially in the MCL gauge during anterior tibial translation after ACL section.

The final methodological problem was data variability, as voltage changes from the gauge can be affected by many factors. Although sensitivity of a single transducer was independent from room temperature, rotation of the knee, and number of cycles [11], other authors [17] found a 100% inter-specimen variation in strain measurements at the patellar tendon, using an IFT sensor. We decreased variability by following a precise protocol that included technique, position, calibration, and same researcher performing the manoeuvres, but a 30% variability coefficient indicated the same trend found by other investigators [17]. To cope with this variability, we normalized the amplitude values per specimen and performed sequential comparisons in the same knee, before and after ACL section.

This study allowed us to conclude that gauges located on secondary restraints insertions clearly differentiate the knee before and after ACL section. This difference was not only related to intensity of the response, but also to its time distribution: a 10 to 40-fold increase in the response intensity vs. time (slope) was observed after ACL section. With high sensitivity (over 80%) and specificity (100%) during a Lachman-like maneuver, the MCL gauge was a good detector of anterior knee laxity in the absence of the ACL. The application of this result suggests that the MCL is able to detect a forced anterior tibial translation in the ACL-deficient knee, better than knee flexion or extension.

The present experimental work is far from immediate clinical application. However, despite previous clinical experience and biomechanical studies, it was unclear how secondary restrictors respond in the ACL injured knee. Particularly, it was unclear whether if secondary restraints tensioning after ACL injury may stimulate their proprioceptive receptors differently during anterior tibial displacement than during flexion–extension. This is important not only to understand real-time detection of anterior instability that occurs in the ACL-deficient knee, but also to understand how proprioceptive information obtained from those preserved structures can elicit a muscular response in the ACL-deficient knee. The same strain gauges have already detected increased tension in human collateral ligaments in knees submitted to anterior tibial traction at different knee angles [20]. We obtained similar patterns with in vivo strain gauges that could be useful to detect instability. An accurate instability detection can be used to feed a microprocessor that could differentiate stable and unstable knees, as showed by our group in preliminary work [36]. This understanding can help to program specific rehabilitation programs oriented to take advantage of preserved proprioception in secondary restraints such as the MCL.

In conclusion, this is an experimental study obtaining the evidence that particularly the MCL can differentiate the anterior tibial translation from the knee flexion or extension when the ACL is not present, and this response can be quantified. This suggests that secondary restraints may offer a substituting proprioceptive system when the proprioception is damaged due to ACL injury, and there are opportunities to better define clinical rehabilitation programs taking advantage of this preserved capability when the ACL is absent. This could include proprioceptive training programs to learn pattern recognition to elicit an early muscular response in the ACL-deficient knee through guided muscular stimulation. Knee proprioceptive remaining elements should be the aim of therapeutic approaches, as its potential to signal abnormal knee displacements has been verified in the ACL deficient knees.