- Email: [email protected]
The effect of disc thickness and trauma on disc surface friction in the porcine temporomandibular joint
Archives of Oral Biology 46 (2001) 155 – 162 www.elsevier.com/locate/archoralbio
The effect of disc thickness and trauma on disc surface friction in the porcine temporomandibular joint J.C. Nickel a,b,*, L.R. Iwasaki a, D.E. Feely b, K.D. Stormberg c, M.W. Beatty d a
Department of Growth and De6elopment, Uni6ersity of Nebraska Medical Center, College of Dentistry, P.O. Box 830740, Lincoln, NE 68583 -0755, USA b Department of Oral Biology, Uni6ersity of Nebraska Medical Center, College of Dentistry, P.O. Box 830740, Lincoln, NE 68583 -0755, USA c Pri6ate Practice, 4333 Palm A6enue, Suite C, La Mesa, CA 91941, USA d Department of Adult Restorati6e Dentistry, Uni6ersity of Nebraska Medical Center, College of Dentistry, P.O. Box 830740, Lincoln, NE 68583 -0755, USA Accepted 8 August 2000
Abstract The pathomechanics of osteoarthritis in the human temporomandibular joint (TMJ) are unknown. Compromised lubrication is a potential factor, but, lubrication within even the normal TMJ is not understood completely. Weeping lubrication is a concept that may be applicable to the TMJ. A characteristic of weeping lubrication is a slow increase in friction during static loading. The rate of increase in friction is related to the rate of lateral movement of synovial fluid away from the loading area. The TMJ disc is expected to be the main source of TMJ lubrication. This study tested two variables, disc thickness and magnitude of trauma to the disc, as factors that can affect the rate of flow of synovial fluid and thus alter lubrication of the disc surfaces. To test these variables, TMJ disc surface friction was measured before and after an impulse load. Before the impulse load, all discs demonstrated a gradual increase in friction during light static loading. The rate of increase in friction was inversely related to the disc thickness (R 2 =0.75). After an impulse load of known magnitude and peak force, disc surface friction was higher. The magnitude of this surface friction was correlated with the magnitude of the impulsive blow (R 2 = 0.89) and the area of surface damage (R 2 =0.85). Disc thickness was a significant factor in determining the minimal impulse needed to produce higher surface friction (R 2 =0.99). These results confirm that disc thickness and trauma to the disc affect surface friction in the TMJ, and therefore may be important factors in compromised lubrication and the development of osteoarthritis. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: TMJ disc; Trauma; Impulse load; Surface friction; Pathomechanics; Osteoarthritis
1. Introduction It has been suggested that the development of osteoarthritis in the temporomandibular joint (TMJ) may
* Corresponding author. Tel.: +1-402-472-1307; fax: + 1402-472-5290. E-mail address: [email protected] (J.C. Nickel).
be traced to a single traumatic event that occurred many years before the manifestation of symptoms of the disease (Solberg et al., 1985). As the disc is the primary mechanism of stress distribution and lubrication in the joint, it has been speculated that trauma has its effect by altering the mechanical properties of the disc, reducing stress distribution and lubrication and increasing the likelihood of osteoarthritis (Westessen and Rohlin, 1984; Castelli et al., 1985; Solberg et al.,
0003-9969/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 3 - 9 9 6 9 ( 0 0 ) 0 0 1 0 1 - 1
156
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
1985; Nickel and McLachlan, 1994a,b; Iwasaki et al., 1997). The long-term health of all synovial joints, including the TMJ, depends on the effectiveness of mechanisms to control stresses in the articular tissues. One obvious source of shear and tensile stress on the surface of articular joints is friction at the start of movement. The remarkable lubrication mechanism in the TMJ results in very low friction at the start of movement (Nickel and McLachlan, 1994a). The disc appears to be the principal means of reducing surface friction. Friction on the surfaces of the condyle, without the aid of the disc, is at least three times greater than when the disc is in place (Nickel and McLachlan, 1994d). The lubrication capabilities of the articular eminence are expected to be similar to those of the condyle, but have not been quantified. The mechanism that produces the low friction in the TMJ has yet to be defined. One likely model of synovial joint lubrication is a combination of weeping and boundary lubrication. Weeping lubrication has been described as a form of self-pressurized hydrostatic lubrication (McCutchen, 1962). In engineered hydrostatic bearings, a pressurized fluid prevents contact between the surfaces. In natural bearings, the model of weeping lubrication assumes that irregularities between the loaded surfaces trap liquid. This helps to form an incomplete ‘squeeze film’ (McCutchen, 1962, 1983). The synovial fluid is squeezed between the loaded surfaces, and when it leaks out from between them, it is replenished by fluid from within the cartilage matrix (Fig. 1). As long as a load is maintained on the cartilage, the tissue continues to lose fluid, and the load is transferred to the phospholipid/glycoprotein boundary lubricant adhering to the cartilage surface (Swann et al., 1981, 1985; Hills and Butler, 1984; Hatton and Swann, 1986; Hills, 1989). However, at very low magnitudes of stress,
Fig. 1. Weeping lubrication (see Section 1 for explanation) (modified from Nickel and McLachlan, 1994a).
the phospholipid/glycoprotein boundary layer fails (McCutchen, 1983), causing an increase in dry contact of the solid skeleton with the cartilage, which itself causes an increase in friction. Thus, a feature of weeping lubrication, the predominant component of the model, is low initial friction which increases with duration of loading. The rate of increase in surface friction during continuous loading of cartilage may be related to the rate of loss of lubricant due to the lateral flow of the fluid through the matrix and over the surface of the disc (Fig. 1). The faster the lateral flow, the less fluid there is available to maintain the ‘hydrostatic bearing’. The velocity of lateral fluid flow has been estimated by the equation (Hou et al., 1992): 6t = ff
z 2
'
k (P ma (r
where: 6t is the tangential or lateral fluid velocity, ff is the volume fraction of the matrix occupied by the fluid, z is the thickness of the fluid film on the surface of the cartilage, k is the apparent porosity of the cartilage matrix, ma is the apparent viscosity of the fluid in the matrix, and #P/#r is the lateral pressure gradient within the matrix. Lateral pressure gradients produced by statically loading the discs of the mandibular joints were increased with thinness of the disc (Nickel and McLachlan, 1994b). Therefore according to the equation, with all other factors being the same, the velocity of lateral fluid flow (6t) within a thinner disc should be faster. Thus, the surface friction of thinner discs may increase more quickly under static load than that of thicker discs. To date, to the best of our knowledge, data demonstrating the effect of disc thickness on surface friction have not been reported. The equation also predicts that for increased lateral permeability, reflected by an increase in apparent porosity of the cartilage matrix (k), lateral fluid velocity (6t) also increases. Impulse loads to the disc surface produced a linearly related increase in surface damage and surface friction in preliminary experiments (Tait et al., 1997). The ability of an impulse load to produce damage, and thus, increase permeability, may be affected by disc thickness. Data describing this phenomenon have not, we believe, been published. Our specific aims now were to determine: (1) if the change in disc surface friction with duration of static load in the TMJ is related to the thickness of the disc; (2) if impulse loading to the disc alters the surface friction; (3) if the thickness of the disc determines the minimal impulse magnitude necessary to induce the changes in surface friction; and (4) if the increase in friction is related to the area of surface damage. The specific aims were chosen to test the hypothesis that weeping lubrication is a useful model to describe how the disc minimizes surface friction.
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
2. Materials and methods
2.1. Description of the sample In order to address our specific aims, it was necessary to test the frictional properties of pristine, freshly dissected discs. Study of the mechanics of the human disc ideally requires healthy human material. Due to the difficulty in procuring and maintaining unpreserved healthy human discs, the pig was used to study disc mechanics. The choice of the porcine model was based on similarities between pig and human TMJ (Herring, 1976; Stro¨m et al., 1986; Nickel and McLachlan, 1994a,b). Discs were obtained from a local abattoir (Farmland Foods Corp., Crete, NE). Each was dissected free of the joint shortly after the animal was killed, and stored in warmed 0.3 M phosphate buffer. Within 1 h, the samples were transported to the laboratory where the experiments were conducted on the same day. A total of 67 discs from 34 animals were used in this study. Disc pairs showed good right–left symmetry visually and with respect to minimum thickness ( 90.4 mm) and static coefficient of surface friction (ms) at 300 s of static load ( 910%). Five pristine disc pairs were used as control samples (n=10). The remaining 28 disc pairs and one single disc were used in experiments to demonstrate the effects of disc thickness and impulse loading on surface friction.
2.2. Measurements To address our specific aims, measurements were made of disc thickness, disc surface friction before an impulse load, impulse magnitude and peak force imposed on the disc surface, disc surface friction after an impulse load, and area of disc surface damage. The minimum thickness of each disc was measured using a linear voltage differential transformer, the displacement and subsequent electrical output of which were calibrated to facilitate the measurement of disc thickness to within 90.05 mm. The static coefficient of friction (ms) on the surface of each disc while under light static load was measured using a previously described friction pendulum (Nickel and McLachlan, 1994a). Measurements were made before imposing an impulse load on the surface of the experimental discs. In brief, the friction pendulum was designed to produce a known static load on the disc whilst measuring the frictional forces on the cartilage surface during the static-loading period. This period was 10–12 min in duration and timed with a digital stopwatch and tape-recorder counter. The apparatus measured the peak frictional forces associated with the start of movement. To make the measurements, the disc was removed from the warmed buffer solution and
157
placed on a sandpaper surface fixed to an aluminium platform. The sandpaper stopped the disc from slipping off the platform during the experiments. The friction pendulum was lowered on to the disc until its full weight was supported by the disc. Right and left angular deflections were imposed on the pendulum, its resting positions were recorded, and ms was calculated based on the angular deflection in static equilibrium. The basis for the calculations was described by Nickel and McLachlan (1994a). Measurement accuracy of ms was to within 9 0.0006. Errors in timing the period of static load were within 915 s. Once the pre-impulse friction measurements were complete, the disc was placed in warmed buffer for 30 min to recover fluid lost during these measurements. To test the repeatability of the method, the ms for each control disc was measured, then measured again after the disc had been placed in warmed buffer for 30 min, and measured a third time after the disc had been placed in warmed buffer for another 30 min. Following the recovery phase, an impulse load was imposed on each experimental disc using a hinged hammer (Tait et al., 1997). The hammer struck an acrylic indenter which rested on the disc surface. Two strain gauges, arranged in series and attached to the indenter, measured the time course of the impulse. The linear load – strain characteristics of the indentor/strain gauge apparatus were calibrated using static loads. During impulse loading, a trigger mechanism initiated a display of the strain-gauge electrical signal on a digital oscilloscope. A tracing was made of the impulse signal for analysis to determine impulse magnitude and peak force. The error associated with the measurement of the physical variables of the impulse was 9 3.8% for impulse magnitude and 9 2.4% for peak force. To investigate a range of impulse magnitudes and peak forces, the height and mass of the hammer were changed. The relation between the drop height of the hammer, its mass, and the magnitude of the impulse is described in the following equation: Impulse =
&
t2
t1
F (t=m 2g(ht1 − ht2)
where: m is the mass of the hammer (in kg), g is the gravitational constant (in m/s2), h is the height of the hammer (in m), F is the force (in N), and t is time (in s) associated with the impulse (in N/s). Each experimental disc was loaded at a different impulse magnitude and peak force, made possible by either increasing or decreasing the mass of the hammer with a commensurate raising or lowering the height that it dropped before striking the instrumented indenter. After the impulse load, each experimental disc was placed in warmed phosphate-buffered solution for 30 min, and then positioned under the friction pendulum for a second set of measurements of surface friction
158
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
during static loading. The post-impulse load friction measurements were made in the same manner as the pre-impulse load friction measurements. Following measurements of friction, the anterior and posterior boundaries of the impulse loading area were identified using a dissecting needle covered with India ink. Each disc was placed in balanced salt solution for 30 min to allow for rehydration. The discs were then fixed in 2% glutaraldehyde/5% sucrose/2% tannic acid in 0.1 M phosphate buffer (pH =7.3) for 24 h, followed by 2% osmium tetroxide in phosphate buffer and 2% lysine (Hopwood, 1985). The discs were trimmed to a suitable size for scanning in an electron microscope, while preserving the India ink-stained punctures in the specimen. The trimmed discs were processed through graded alcohols, chemically dried with hexamethyldisilazine, and mounted on aluminium stubs using colloidal graphite. Specimens were stored in a dessicator under vacuum and in the presence of anhydrous calcium sulphate. In order to reduce the costs associated with producing photoelectron micrographs, 21 experimental discs were selected to investigate the relation between changes in surface friction and surface damage. Those discs subjected to impulse loads in the ranges of 51.2 and ]1.6 N/s were included. These two impulse load ranges represent the extremes of low and high impulse magnitudes used in the experiments, which optimizes the possibility of detecting whether or not there is a correlation between the friction effects of impulse loading and the area of damage. Each of the discs was examined with an electron microscope (JSM-61W; JEOL, Tokyo, Japan), using accelerating voltages between 10 and 25 kV. Each specimen was photographed with a reference of known size in the field of view. Using the stained disc-surface punctures as guides, the area of cracking was identified on a clear acetate overlay, which was then digitized using a commercial software package (Bioquant II; R and M Biometrics Inc., Nashville, TN). The area of damage was measured in units of mm2. Error studies determined that the repeatability of area measurements was within 9 5%. All friction measurements were plotted with respect to time of static load. Trends in the data were demonstrated using a best-fit least-squares polynomial regression. To demonstrate the effect of an impulse load on disc surface friction, a friction ratio was calculated according to the following equation: ms(post-impulse) Friction ratio = ms(pre-impulse) where ms was the coefficient of static friction measured at 300 s on the plots for friction versus time of static load. The ms for this time-point was used for two reasons. First, preliminary data had shown that the maximum differences between pre- and post-impulse
Fig. 2. Time-dependent changes in friction and disc thickness. The change () in the friction coefficient per second of static loading (ms(pre-impulse)/s) for pristine discs is plotted on the vertical axis. This value was calculated as the change over a period of 300 s of static loading with the friction pendulum averaged for each disc pair (n =33 disc pairs). Disc thickness (in mm), measured by the linear voltage differential transformer prior to measuring surface friction and also averaged for each disc pair, is plotted on the horizontal axis.
friction occurred between 200 and 400 s after the start of static-loading friction measurements. Second, cumulative periods of nocturnal bruxism have been shown to last for 300 s (Rugh and Harlan, 1988). Therefore, the friction measurements were made after static loading of duration similar to that in bruxing patients. The importance of disc thickness was evaluated using plots of friction ratio versus impulse magnitude for experimental discs grouped according to thickness. A linear regression coefficient was calculated to determine the variance in the relationship for each disc-thickness group. The equation derived from each of these plots was used to determine the x-axis intercept, which represented the minimum threshold impulse magnitude required to affect the coefficient of friction for each disc-thickness group. Similarly, the results for the friction ratio versus the peak force of the impulse load were plotted and linear regression coefficients were calculated. The importance of disc surface damage was evaluated using plots of friction ratio versus area of surface damage. A linear regression equation was calculated to determine the variance in the relationship. 3. Results
3.1. Minimum disc thickness The minimum thickness of each disc was measured and the 57 experimental discs were then divided into four groups according to that thickness. The range of disc thicknesses and the number of discs (n) per group are shown in Fig. 2.
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
3.2. Surface friction associated with pristine discs All pristine discs demonstrated a gradual increase in coefficient of static friction (ms) with duration of static load. The rate of increase in ms per second was correlated (R 2 =0.74) with the thickness of the disc (Fig. 2). The rates of increase in friction for the thinnest discs were approximately twice the rates measured for discs four times thicker. The coefficient of static friction for the control discs was first measured once, repeated after placing the discs in warmed buffer for 30 min, and repeated again after placing them in warmed buffer for a second 30 min. The mean variation for all of the measurements combined was 11%.
3.3. Surface friction associated with discs post-impulse The range of impulse loads delivered to the experimental discs was 0.19–2.55 N/s, with a median value of 1.34 N/s. The range of peak forces delivered to the experimental discs was 206–970 N, with a median value of 534 N. Impulse loading affected disc surface friction. For a given disc and a given time of static loading in the range 50–400 s, the coefficient of static surface friction was higher after an impulse load. The magnitude of the
159
change in surface friction was expressed using the ratio of ms(post-impulse)/ms(pre-impulse) for each disc at 300 s of static load. The results for friction ratio versus impulse magnitude (N/s) showed a strong correlation when disc thickness was taken into account (Fig. 3A – D; R 2 = 0.89 for each of the four disc-thickness groups). With increased disc thickness, the critical threshold of the impulse magnitude to affect surface friction also increased. This was demonstrated by the higher value for the x-axis intercept point for discs of greater thickness. The correlation between disc thickness and the critical threshold of the impulse magnitude to affect surface friction was strongly linear (R 2 = 0.99) (Fig. 4). The results for friction ratio versus peak force did not show a strong correlation overall (R 2 = 0.45). When this correlation was grouped according to disc thickness, the coefficients were low (R 2 5 0.46) except for group C (R 2 = 0.89). Of the 21 experimental discs subjected to impulse loads 51.2 N/s and ]1.6 N/s and evaluated for surface damage, five discs were from each of groups A, B, and D, and six were from group C. The magnitude of the change in static surface friction as a result of impulse loading was related (R 2 = 0.85) to the area of damage (Fig. 5). The nature of the relation was logarithmic and indicated that surface damage must increase by 104 for the surface friction to double. The
Fig. 3. Friction ratio versus impulse magnitude. A set of data grouped according to disc thickness is presented. The disc-thickness ranges and number (n) for each group are (A) 0.6–1.2 mm (n = 11); (B) 1.3 – 1.6 mm (n =19); (C) 1.7 – 1.9 mm (n =11); and (D) 2.0–3.0 mm (n = 16). The friction ratio (ms(post-impulse)/ms(pre-impulse)) is plotted on the vertical axis, and impulse magnitude (in N/s) is plotted along the horizontal axis. A best-fit straight-line regression demonstrates the strength of each relationship (R 2 values). The x-axis intercept is the critical threshold of impulse magnitude required to affect an increase in surface friction for the given range of disc thicknesses. Note that as disc thickness increased (A– D), so did the critical threshold.
160
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
Fig. 4. Average disc thickness versus critical threshold of impulse magnitude. The relation between the average disc thickness (in mm) for each disc group (A–D) and critical threshold of impulse magnitude (x-axis intercept of friction ratio versus impulse magnitude plot for each disc group, in N/s) is shown. Critical threshold of impulse magnitude increased with disc thickness (R 2 = 0.99). An impulse load to the disc above the thickness-dependent threshold would be expected to induce increased surface friction.
minimum threshold of surface damage to affect surface friction predicted by the plot (Fig. 5) was 2.6×103 mm2.
4. Discussion
4.1. Weeping lubrication theory and disc thickness Surface friction increased gradually with the duration of static loading of the TMJ disc by the friction pendulum. This finding is consistent with the theory of weep-
ing lubrication, in which the fluid reservoir within the cartilage matrix is depleted gradually and more areas of opposing cartilage surfaces come into direct contact. Under these circumstances there is a time-dependent increase in friction. Here, our thickest pristine discs were four times thicker than our thinnest. The thickest discs showed a rate of change in friction with static load that was one-half that of the thinnest (Fig. 5). The lateral flow of fluid in the cartilage matrix may have contributed to this effect. The faster this lateral flow, the less fluid there is available to move to the surface where it can be used to recharge the ‘hydrostatic bearing’. A previous report showed a strong correlation between disc thickness and lateral pressure gradient (#P/#r), where #P/#r was 3.5-fold higher for the disc that was 0.8 mm thick that for one that was 2.8 mm thick (Nickel and McLachlan, 1994b). As the present range of disc thicknesses was greater (0.68 – 2.85 mm) than in the previous report, lateral pressure gradients under the thinnest of our discs were expected to approach values four times those found under the thickest. If the difference between the thinnest and the thickest discs subject to static load was reflected in the pressure gradients (#P/#r) in this way, then according to equation (1), the lateral velocity of fluid flow (6t) for the thinnest discs should be double that for the thickest. That is, the thinnest discs should have higher lateral pressure gradients, which would cause faster lateral fluid movement through the cartilage matrix. This high 6t would deplete the fluid that would otherwise be available to recharge the surface and provide hydrostatic lubrication. As a result, and in keeping with our findings, surface friction would increase at a faster rate in thinner discs. The clinical significance of these phenomena lies in the potential for increased surface scuffing in the intact TMJ as a consequence of chronic high loads, such as those experienced in bruxism. In addition, thinner discs are likely to lose the weeping lubricant more quickly, leading to increased friction within the joint. Mechanical fatigue of the joint surfaces and debris in the synovial fluid might result. Debris causes an increase in the metabolic activity of macrophages in the synovial cavity, which in turn release various catabolic agents typically associated with the processes of osteoarthritis (Kubota et al., 1998).
4.2. Impulse loading and surface friction Fig. 5. Friction ratio versus area of damage. The friction ratio (ms(post-impulse)/ms(pre-impulse)) is plotted against the area of damage measured on the surface of the discs subjected to the highest (1.6 – 2.0 N/s) and lowest (0.8–1.2 N/s) impulse magnitudes. Five discs from groups A, B, and D, and six from group C were evaluated for surface damage and the plot of friction ratio versus area of damage (in mm2) is shown. The best-fit straight-line regression had an R 2 of 0.85.
The impulse magnitude, which was the amount of energy introduced on to the experimental disc surface, determined whether there were post-impulse changes in friction. Impulse-induced changes in lateral permeability could account for the increase in friction. The impulse magnitude determined the amount of disc surface damage, which in turn, increased permeability to
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
lateral fluid flow. However, thicker discs required more energy introduced to the surface before any changes were seen. The data suggest that disc thickness is positively correlated with the critical threshold of impulse magnitude which must be exceeded to cause an increase in friction (Fig. 4). Numerical modelling studies of peak force on cartilage have suggested that a thicker layer of hyaline cartilage, which is bound to bone, is more susceptible to damage than a thinner layer (Armstrong, 1986). Others report that the magnitude of surface shear stress was the factor most important to the occurrence of surface damage, and that peak force was strongly correlated with surface yielding (Atkinson et al., 1998). We show that larger impulse loads are required to affect surface friction in thicker discs and that peak impulse force does not correlate strongly with changes in surface friction (R 2 =0.45 overall). The apparent contradiction between previous reports and the current findings may be due to the physical constraints of bound hyaline cartilage, which are quite different from the boundary conditions associated with the TMJ disc, where friction is very low. In addition, the material properties of that disc are anisotropic (Bruno et al., 1999) and the previous studies applied isotropic methods of analysis.
4.3. The pathomechanics of TMJ osteoarthritis Osteoarthritis of the TMJ is common according to reports, but the mean age of onset differs widely depending on the methods of examination and the distinguishing criteria employed. Mean age of onset ranges between 25 and 35 years (Helo¨e and Helo¨e, 1975; Solberg et al., 1979; Nilner, 1981; Pullinger et al., 1988; Stegenga et al., 1989; Ong and Franklin, 1996; Israel et al., 1998; Murakami et al., 1998), approximately a decade earlier than the mean age of onset reported for osteoarthritis of the hip (Lawrence et al., 1989; Felson et al., 1997; Vingard et al., 1997). The earlier appearance of symptoms in the mandibular joint may relate in part to its disc being its principal stress-reducing mechanism. The surface congruency in the mandibular joint is poor over the crest of the eminence due to the eminence’s mechanism of growth (Nickel and McLachlan, 1994c). The disc distributes loads over the surfaces and maintains relatively frictionless sliding between them. In other joints such as the hip, the high degree of surface congruency between the hyaline cartilage on both rubbing surfaces serves to maintain minimal stresses. The knee, like the TMJ, has a fibrocartilaginous meniscus, but it is incomplete. The congruency and lubricating properties of the hyaline cartilage covering the femoral condyles and tibial plateau can minimize stresses to a certain extent, but like the TMJ, loss of the meniscus results in osteoarthritis. This probably accounts for the earlier expression of osteoarthritis in
161
knees than hips. In the mandibular joint, however, the fibrous linings of the condyle and temporal bone are not well-suited to distributing loads or lubrication (Nickel and McLachlan, 1994b,d), which is why the mechanical integrity of the disc is vital for the health of this joint over decades of function. We tested two variables, disc thickness and magnitude of trauma to the disc, as factors important to the mechanical integrity of the disc. The results confirm that disc thickness and trauma, specifically impulse load magnitude, affect disc static surface friction and, therefore, may be important factors in compromised lubrication and the development of osteoarthritis in the TMJ.
Acknowledgements We acknowledge financial support of this project from a University of Nebraska Medical Center, College of Dentistry Seed Grant. We are grateful to Farmland Foods Corporation of Crete, Nebraska, for providing access to fresh TMJ discs. We thank C. Anderson and M. Leiker, who undertook the task of microscopic evaluation of the disc surfaces, and K.R. Theesen, Illustrator, University of Nebraska Medical Center, College of Dentistry Learning Resources, who helped produce the figures.
References Armstrong, C.G., 1986. An analysis of the stresses in a thin layer of articular cartilage in a synovial joint. Eng. Med. 15, 55 – 61. Atkinson, T.S., Haut, R.C., Altiero, N.J., 1998. Impact-induced fissuring of articular cartilage: an investigation of failure criterion. J. Biomech. Eng. 120, 181 – 187. Bruno, M.J., Beatty, M.W., Nickel, J.C., Iwasaki, L.R., 1999. Strain rate dependent anisotropic properties of the porcine TMJ disc. J. Dent. Res. (SI) 78, 125. Castelli, W.A., Nasjleti, C.E., Diaz-Perez, R., Caffesse, R.G., 1985. Histopathologic findings in temporomandibular joints of aged individuals. J. Prosthet. Dent. 53, 415 – 419. Felson, D.T., Zhang, Y., Hannam, M.T., Naimark, A., Weissman, B., Aliabadi, P., Levy, D., 1997. Risk factors for the incident of radiographic knee osteoarthritis in the elderly: the Framingham study. Arthritis Rheum. 40, 728 – 733. Hatton, M.N., Swann, D.A., 1986. Studies on bovine temporomandibular joint synovial fluid. J. Prosthet. Dent. 56, 635 – 638. Helo¨e, B., Helo¨e, L.A., 1975. Characteristics of a group of patients with temporomandibular joint disorders. Community Dent. Oral Epidemiol. 3, 72 – 79. Herring, S.W., 1976. The dynamics of mastication in pigs. Arch. Oral Biol. 21, 473 – 480. Hills, B.A., 1989. Oligolamellar lubrication of joints by surface active phospholipid. J. Rheumatol. 16, 82 – 91.
162
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
Hills, B.A., Butler, B.D., 1984. Surfactants identified in synovial fluid and their ability to act as boundary lubricants. Ann. Rheum. Dis. 43, 641–648. Hopwood, D., 1985. Cell and tissue fixation, 1972–1982. Histochem. J. 17, 389–442. Hou, J.S., Mow, V.C., Lai, W.M., Holmes, M.H., 1992. An analysis of the squeeze film lubrication mechanism for articular cartilage. J. Biomech. 25, 247–259. Israel, H.A., Ramamurthy, N.S., Greenwald, R., Golub, L., 1998. The potential role of doxycycline in the treatment of osteoarthritis of the temporomandibular joint. Adv. Dent. Res. 12, 51 – 55. Iwasaki, L.R., Nickel, J.C., McLachlan, K.R., 1997. Relationship between growth, function, and stress in the temporomandibular joint. In: McNeill, C. (Ed.), Science and Practice of Occlusion. Quintessence Publishing, Chicago, pp. 125 – 136. Kubota, E., Kubota, T., Mastumoto, J., Shibata, T., Murakami, K.I., 1998. Synovial fluid cytokines and proteinases as markers of temporomandibular joint disease. J. Oral Maxillofac. Surg. 56, 192–198. Lawrence, R.C., Hochberg, M.C., Kelsey, J.L., McDuffie, F.C., Medsger, T.A., Felts, W.R., Shulman, L.E., 1989. Estimates of the prevalence of selected arthritic and musculoskeletal diseases in the United States. J. Rheumatol. 16, 427 – 441. McCutchen, C.W., 1962. The frictional properties of animal joints. Wear 5, 1 – 17. McCutchen, C.W., 1983. Lubrication of and by articular cartilage. In: Hall, B.K. (Ed.), Cartilage: Biomedical Aspects. Academic Press, New York, pp. 87–107. Murakami, K.I., Shibata, T., Kubota, E., Maeda, H., 1998. Intra-articular levels of prostaglandin E2, hyaluronic acid, and chondroitin-4 and -6 sulfates in the temporomandibular joint synovial fluids of patients with internal derangement. J. Oral Maxillofac. Surg. 56, 199–203. Nickel, J.C., McLachlan, K.R., 1994a. In vitro measurement of the frictional properties of the temporomandibular joint disc. Arch. Oral Biol. 39, 323–331. Nickel, J.C., McLachlan, K.R., 1994b. In vitro measurement of the stress-distribution properties of the pig temporomandibular joint disc. Arch. Oral Biol. 39, 439–448. Nickel, J.C., McLachlan, K.R., 1994c. An analysis of surface congruity in the growing human temporomandibular joint. Arch. Oral Biol. 39, 315–321.
.
Nickel, J.C., McLachlan, K.R., 1994d. A comparison of the lubrication properties of the mandibular condyle and TMJ disc. J. Dent. Res. (SI) 73, 304. Nilner, M., 1981. Prevalence of functional disturbances and diseases of the stomatognathic system in 15 – 18 year olds. Swed. Dent. J. 5, 189 – 197. Ong, T.K., Franklin, C.D., 1996. A clinical and histopathological study of osteoarthrosis of the temporomandibular joint. Br. J. Oral Maxillofac. Surg. 34, 186 – 192. Pullinger, A.G., Seligman, D.A., Solberg, W.K., 1988. Temporomandibular disorders. Part I: functional status, dentomorphologic features and sex differences in a non-patient population. J. Prosthet. Dent. 59, 228 – 235. Rugh, J.D., Harlan, J., 1988. Nocturnal bruxism and temporomandibular disorders. Adv. Neurol. 49, 329 – 341. Solberg, W.K., Woo, M.W., Houston, J.B., 1979. Prevalence of mandibular dysfunction in young adults. J. Am. Dent. Assoc. 98, 25 – 34. Solberg, W.K., Hansson, T.L., Nordstro¨m, B., 1985. The temporomandibular joint in young adults at autopsy: a morphologic classification and evaluation. J. Oral Rehabil. 12, 303 – 321. Stegenga, B., de Bont, L.G.M., Boering, G., 1989. Osteoarthritis as the cause of craniomandibular pain and dysfunction: a unifying concept. J. Oral Maxillofac. Surg. 47, 249 – 256. Stro¨m, D., Holm, S., Clemensson, E., Haraldson, T., Carlsson, G.E., 1986. Gross anatomy of the mandibular joint and masticatory muscles in the domestic pigs (Sus scrofa). Arch. Oral Biol. 31, 763 – 768. Swann, D.A., Hendren, R.B., Radin, E.L., Sotman, S.L., Duda, E.A., 1981. The lubricating activity of synovial fluid glycoproteins. Arthritis Rheum. 24, 22 – 30. Swann, D.A., Silver, F.H., Slayter, H.S., Stafford, W., Shore, E., 1985. The molecular structure and lubricating activity of lubricin isolated from bovine and human synovial fluids. Biochem. J. 225, 195 – 201. Tait, D.C., Nickel, J.C., McLachlan, K.R., 1997. Impulse loading and weeping lubrication on the surface of the TMJ disc. J. Dent. Res. (SI) 76, 447. Vingard, E., Alfredsson, L., Malchau, H., 1997. Osteoarthrosis of the hip in women and its relation to physical load at work and in the home. Ann. Rheum. Dis. 56, 293 – 298. Westessen, P.L., Rohlin, M., 1984. Internal derangement related to osteoarthrosis in temporomandibular joint autopsy specimens. Oral Surg. Oral Med. Oral Pathol. 57, 17 – 22.
The effect of disc thickness and trauma on disc surface friction in the porcine temporomandibular joint J.C. Nickel a,b,*, L.R. Iwasaki a, D.E. Feely b, K.D. Stormberg c, M.W. Beatty d a
Department of Growth and De6elopment, Uni6ersity of Nebraska Medical Center, College of Dentistry, P.O. Box 830740, Lincoln, NE 68583 -0755, USA b Department of Oral Biology, Uni6ersity of Nebraska Medical Center, College of Dentistry, P.O. Box 830740, Lincoln, NE 68583 -0755, USA c Pri6ate Practice, 4333 Palm A6enue, Suite C, La Mesa, CA 91941, USA d Department of Adult Restorati6e Dentistry, Uni6ersity of Nebraska Medical Center, College of Dentistry, P.O. Box 830740, Lincoln, NE 68583 -0755, USA Accepted 8 August 2000
Abstract The pathomechanics of osteoarthritis in the human temporomandibular joint (TMJ) are unknown. Compromised lubrication is a potential factor, but, lubrication within even the normal TMJ is not understood completely. Weeping lubrication is a concept that may be applicable to the TMJ. A characteristic of weeping lubrication is a slow increase in friction during static loading. The rate of increase in friction is related to the rate of lateral movement of synovial fluid away from the loading area. The TMJ disc is expected to be the main source of TMJ lubrication. This study tested two variables, disc thickness and magnitude of trauma to the disc, as factors that can affect the rate of flow of synovial fluid and thus alter lubrication of the disc surfaces. To test these variables, TMJ disc surface friction was measured before and after an impulse load. Before the impulse load, all discs demonstrated a gradual increase in friction during light static loading. The rate of increase in friction was inversely related to the disc thickness (R 2 =0.75). After an impulse load of known magnitude and peak force, disc surface friction was higher. The magnitude of this surface friction was correlated with the magnitude of the impulsive blow (R 2 = 0.89) and the area of surface damage (R 2 =0.85). Disc thickness was a significant factor in determining the minimal impulse needed to produce higher surface friction (R 2 =0.99). These results confirm that disc thickness and trauma to the disc affect surface friction in the TMJ, and therefore may be important factors in compromised lubrication and the development of osteoarthritis. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: TMJ disc; Trauma; Impulse load; Surface friction; Pathomechanics; Osteoarthritis
1. Introduction It has been suggested that the development of osteoarthritis in the temporomandibular joint (TMJ) may
* Corresponding author. Tel.: +1-402-472-1307; fax: + 1402-472-5290. E-mail address: [email protected] (J.C. Nickel).
be traced to a single traumatic event that occurred many years before the manifestation of symptoms of the disease (Solberg et al., 1985). As the disc is the primary mechanism of stress distribution and lubrication in the joint, it has been speculated that trauma has its effect by altering the mechanical properties of the disc, reducing stress distribution and lubrication and increasing the likelihood of osteoarthritis (Westessen and Rohlin, 1984; Castelli et al., 1985; Solberg et al.,
0003-9969/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 3 - 9 9 6 9 ( 0 0 ) 0 0 1 0 1 - 1
156
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
1985; Nickel and McLachlan, 1994a,b; Iwasaki et al., 1997). The long-term health of all synovial joints, including the TMJ, depends on the effectiveness of mechanisms to control stresses in the articular tissues. One obvious source of shear and tensile stress on the surface of articular joints is friction at the start of movement. The remarkable lubrication mechanism in the TMJ results in very low friction at the start of movement (Nickel and McLachlan, 1994a). The disc appears to be the principal means of reducing surface friction. Friction on the surfaces of the condyle, without the aid of the disc, is at least three times greater than when the disc is in place (Nickel and McLachlan, 1994d). The lubrication capabilities of the articular eminence are expected to be similar to those of the condyle, but have not been quantified. The mechanism that produces the low friction in the TMJ has yet to be defined. One likely model of synovial joint lubrication is a combination of weeping and boundary lubrication. Weeping lubrication has been described as a form of self-pressurized hydrostatic lubrication (McCutchen, 1962). In engineered hydrostatic bearings, a pressurized fluid prevents contact between the surfaces. In natural bearings, the model of weeping lubrication assumes that irregularities between the loaded surfaces trap liquid. This helps to form an incomplete ‘squeeze film’ (McCutchen, 1962, 1983). The synovial fluid is squeezed between the loaded surfaces, and when it leaks out from between them, it is replenished by fluid from within the cartilage matrix (Fig. 1). As long as a load is maintained on the cartilage, the tissue continues to lose fluid, and the load is transferred to the phospholipid/glycoprotein boundary lubricant adhering to the cartilage surface (Swann et al., 1981, 1985; Hills and Butler, 1984; Hatton and Swann, 1986; Hills, 1989). However, at very low magnitudes of stress,
Fig. 1. Weeping lubrication (see Section 1 for explanation) (modified from Nickel and McLachlan, 1994a).
the phospholipid/glycoprotein boundary layer fails (McCutchen, 1983), causing an increase in dry contact of the solid skeleton with the cartilage, which itself causes an increase in friction. Thus, a feature of weeping lubrication, the predominant component of the model, is low initial friction which increases with duration of loading. The rate of increase in surface friction during continuous loading of cartilage may be related to the rate of loss of lubricant due to the lateral flow of the fluid through the matrix and over the surface of the disc (Fig. 1). The faster the lateral flow, the less fluid there is available to maintain the ‘hydrostatic bearing’. The velocity of lateral fluid flow has been estimated by the equation (Hou et al., 1992): 6t = ff
z 2
'
k (P ma (r
where: 6t is the tangential or lateral fluid velocity, ff is the volume fraction of the matrix occupied by the fluid, z is the thickness of the fluid film on the surface of the cartilage, k is the apparent porosity of the cartilage matrix, ma is the apparent viscosity of the fluid in the matrix, and #P/#r is the lateral pressure gradient within the matrix. Lateral pressure gradients produced by statically loading the discs of the mandibular joints were increased with thinness of the disc (Nickel and McLachlan, 1994b). Therefore according to the equation, with all other factors being the same, the velocity of lateral fluid flow (6t) within a thinner disc should be faster. Thus, the surface friction of thinner discs may increase more quickly under static load than that of thicker discs. To date, to the best of our knowledge, data demonstrating the effect of disc thickness on surface friction have not been reported. The equation also predicts that for increased lateral permeability, reflected by an increase in apparent porosity of the cartilage matrix (k), lateral fluid velocity (6t) also increases. Impulse loads to the disc surface produced a linearly related increase in surface damage and surface friction in preliminary experiments (Tait et al., 1997). The ability of an impulse load to produce damage, and thus, increase permeability, may be affected by disc thickness. Data describing this phenomenon have not, we believe, been published. Our specific aims now were to determine: (1) if the change in disc surface friction with duration of static load in the TMJ is related to the thickness of the disc; (2) if impulse loading to the disc alters the surface friction; (3) if the thickness of the disc determines the minimal impulse magnitude necessary to induce the changes in surface friction; and (4) if the increase in friction is related to the area of surface damage. The specific aims were chosen to test the hypothesis that weeping lubrication is a useful model to describe how the disc minimizes surface friction.
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
2. Materials and methods
2.1. Description of the sample In order to address our specific aims, it was necessary to test the frictional properties of pristine, freshly dissected discs. Study of the mechanics of the human disc ideally requires healthy human material. Due to the difficulty in procuring and maintaining unpreserved healthy human discs, the pig was used to study disc mechanics. The choice of the porcine model was based on similarities between pig and human TMJ (Herring, 1976; Stro¨m et al., 1986; Nickel and McLachlan, 1994a,b). Discs were obtained from a local abattoir (Farmland Foods Corp., Crete, NE). Each was dissected free of the joint shortly after the animal was killed, and stored in warmed 0.3 M phosphate buffer. Within 1 h, the samples were transported to the laboratory where the experiments were conducted on the same day. A total of 67 discs from 34 animals were used in this study. Disc pairs showed good right–left symmetry visually and with respect to minimum thickness ( 90.4 mm) and static coefficient of surface friction (ms) at 300 s of static load ( 910%). Five pristine disc pairs were used as control samples (n=10). The remaining 28 disc pairs and one single disc were used in experiments to demonstrate the effects of disc thickness and impulse loading on surface friction.
2.2. Measurements To address our specific aims, measurements were made of disc thickness, disc surface friction before an impulse load, impulse magnitude and peak force imposed on the disc surface, disc surface friction after an impulse load, and area of disc surface damage. The minimum thickness of each disc was measured using a linear voltage differential transformer, the displacement and subsequent electrical output of which were calibrated to facilitate the measurement of disc thickness to within 90.05 mm. The static coefficient of friction (ms) on the surface of each disc while under light static load was measured using a previously described friction pendulum (Nickel and McLachlan, 1994a). Measurements were made before imposing an impulse load on the surface of the experimental discs. In brief, the friction pendulum was designed to produce a known static load on the disc whilst measuring the frictional forces on the cartilage surface during the static-loading period. This period was 10–12 min in duration and timed with a digital stopwatch and tape-recorder counter. The apparatus measured the peak frictional forces associated with the start of movement. To make the measurements, the disc was removed from the warmed buffer solution and
157
placed on a sandpaper surface fixed to an aluminium platform. The sandpaper stopped the disc from slipping off the platform during the experiments. The friction pendulum was lowered on to the disc until its full weight was supported by the disc. Right and left angular deflections were imposed on the pendulum, its resting positions were recorded, and ms was calculated based on the angular deflection in static equilibrium. The basis for the calculations was described by Nickel and McLachlan (1994a). Measurement accuracy of ms was to within 9 0.0006. Errors in timing the period of static load were within 915 s. Once the pre-impulse friction measurements were complete, the disc was placed in warmed buffer for 30 min to recover fluid lost during these measurements. To test the repeatability of the method, the ms for each control disc was measured, then measured again after the disc had been placed in warmed buffer for 30 min, and measured a third time after the disc had been placed in warmed buffer for another 30 min. Following the recovery phase, an impulse load was imposed on each experimental disc using a hinged hammer (Tait et al., 1997). The hammer struck an acrylic indenter which rested on the disc surface. Two strain gauges, arranged in series and attached to the indenter, measured the time course of the impulse. The linear load – strain characteristics of the indentor/strain gauge apparatus were calibrated using static loads. During impulse loading, a trigger mechanism initiated a display of the strain-gauge electrical signal on a digital oscilloscope. A tracing was made of the impulse signal for analysis to determine impulse magnitude and peak force. The error associated with the measurement of the physical variables of the impulse was 9 3.8% for impulse magnitude and 9 2.4% for peak force. To investigate a range of impulse magnitudes and peak forces, the height and mass of the hammer were changed. The relation between the drop height of the hammer, its mass, and the magnitude of the impulse is described in the following equation: Impulse =
&
t2
t1
F (t=m 2g(ht1 − ht2)
where: m is the mass of the hammer (in kg), g is the gravitational constant (in m/s2), h is the height of the hammer (in m), F is the force (in N), and t is time (in s) associated with the impulse (in N/s). Each experimental disc was loaded at a different impulse magnitude and peak force, made possible by either increasing or decreasing the mass of the hammer with a commensurate raising or lowering the height that it dropped before striking the instrumented indenter. After the impulse load, each experimental disc was placed in warmed phosphate-buffered solution for 30 min, and then positioned under the friction pendulum for a second set of measurements of surface friction
158
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
during static loading. The post-impulse load friction measurements were made in the same manner as the pre-impulse load friction measurements. Following measurements of friction, the anterior and posterior boundaries of the impulse loading area were identified using a dissecting needle covered with India ink. Each disc was placed in balanced salt solution for 30 min to allow for rehydration. The discs were then fixed in 2% glutaraldehyde/5% sucrose/2% tannic acid in 0.1 M phosphate buffer (pH =7.3) for 24 h, followed by 2% osmium tetroxide in phosphate buffer and 2% lysine (Hopwood, 1985). The discs were trimmed to a suitable size for scanning in an electron microscope, while preserving the India ink-stained punctures in the specimen. The trimmed discs were processed through graded alcohols, chemically dried with hexamethyldisilazine, and mounted on aluminium stubs using colloidal graphite. Specimens were stored in a dessicator under vacuum and in the presence of anhydrous calcium sulphate. In order to reduce the costs associated with producing photoelectron micrographs, 21 experimental discs were selected to investigate the relation between changes in surface friction and surface damage. Those discs subjected to impulse loads in the ranges of 51.2 and ]1.6 N/s were included. These two impulse load ranges represent the extremes of low and high impulse magnitudes used in the experiments, which optimizes the possibility of detecting whether or not there is a correlation between the friction effects of impulse loading and the area of damage. Each of the discs was examined with an electron microscope (JSM-61W; JEOL, Tokyo, Japan), using accelerating voltages between 10 and 25 kV. Each specimen was photographed with a reference of known size in the field of view. Using the stained disc-surface punctures as guides, the area of cracking was identified on a clear acetate overlay, which was then digitized using a commercial software package (Bioquant II; R and M Biometrics Inc., Nashville, TN). The area of damage was measured in units of mm2. Error studies determined that the repeatability of area measurements was within 9 5%. All friction measurements were plotted with respect to time of static load. Trends in the data were demonstrated using a best-fit least-squares polynomial regression. To demonstrate the effect of an impulse load on disc surface friction, a friction ratio was calculated according to the following equation: ms(post-impulse) Friction ratio = ms(pre-impulse) where ms was the coefficient of static friction measured at 300 s on the plots for friction versus time of static load. The ms for this time-point was used for two reasons. First, preliminary data had shown that the maximum differences between pre- and post-impulse
Fig. 2. Time-dependent changes in friction and disc thickness. The change () in the friction coefficient per second of static loading (ms(pre-impulse)/s) for pristine discs is plotted on the vertical axis. This value was calculated as the change over a period of 300 s of static loading with the friction pendulum averaged for each disc pair (n =33 disc pairs). Disc thickness (in mm), measured by the linear voltage differential transformer prior to measuring surface friction and also averaged for each disc pair, is plotted on the horizontal axis.
friction occurred between 200 and 400 s after the start of static-loading friction measurements. Second, cumulative periods of nocturnal bruxism have been shown to last for 300 s (Rugh and Harlan, 1988). Therefore, the friction measurements were made after static loading of duration similar to that in bruxing patients. The importance of disc thickness was evaluated using plots of friction ratio versus impulse magnitude for experimental discs grouped according to thickness. A linear regression coefficient was calculated to determine the variance in the relationship for each disc-thickness group. The equation derived from each of these plots was used to determine the x-axis intercept, which represented the minimum threshold impulse magnitude required to affect the coefficient of friction for each disc-thickness group. Similarly, the results for the friction ratio versus the peak force of the impulse load were plotted and linear regression coefficients were calculated. The importance of disc surface damage was evaluated using plots of friction ratio versus area of surface damage. A linear regression equation was calculated to determine the variance in the relationship. 3. Results
3.1. Minimum disc thickness The minimum thickness of each disc was measured and the 57 experimental discs were then divided into four groups according to that thickness. The range of disc thicknesses and the number of discs (n) per group are shown in Fig. 2.
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
3.2. Surface friction associated with pristine discs All pristine discs demonstrated a gradual increase in coefficient of static friction (ms) with duration of static load. The rate of increase in ms per second was correlated (R 2 =0.74) with the thickness of the disc (Fig. 2). The rates of increase in friction for the thinnest discs were approximately twice the rates measured for discs four times thicker. The coefficient of static friction for the control discs was first measured once, repeated after placing the discs in warmed buffer for 30 min, and repeated again after placing them in warmed buffer for a second 30 min. The mean variation for all of the measurements combined was 11%.
3.3. Surface friction associated with discs post-impulse The range of impulse loads delivered to the experimental discs was 0.19–2.55 N/s, with a median value of 1.34 N/s. The range of peak forces delivered to the experimental discs was 206–970 N, with a median value of 534 N. Impulse loading affected disc surface friction. For a given disc and a given time of static loading in the range 50–400 s, the coefficient of static surface friction was higher after an impulse load. The magnitude of the
159
change in surface friction was expressed using the ratio of ms(post-impulse)/ms(pre-impulse) for each disc at 300 s of static load. The results for friction ratio versus impulse magnitude (N/s) showed a strong correlation when disc thickness was taken into account (Fig. 3A – D; R 2 = 0.89 for each of the four disc-thickness groups). With increased disc thickness, the critical threshold of the impulse magnitude to affect surface friction also increased. This was demonstrated by the higher value for the x-axis intercept point for discs of greater thickness. The correlation between disc thickness and the critical threshold of the impulse magnitude to affect surface friction was strongly linear (R 2 = 0.99) (Fig. 4). The results for friction ratio versus peak force did not show a strong correlation overall (R 2 = 0.45). When this correlation was grouped according to disc thickness, the coefficients were low (R 2 5 0.46) except for group C (R 2 = 0.89). Of the 21 experimental discs subjected to impulse loads 51.2 N/s and ]1.6 N/s and evaluated for surface damage, five discs were from each of groups A, B, and D, and six were from group C. The magnitude of the change in static surface friction as a result of impulse loading was related (R 2 = 0.85) to the area of damage (Fig. 5). The nature of the relation was logarithmic and indicated that surface damage must increase by 104 for the surface friction to double. The
Fig. 3. Friction ratio versus impulse magnitude. A set of data grouped according to disc thickness is presented. The disc-thickness ranges and number (n) for each group are (A) 0.6–1.2 mm (n = 11); (B) 1.3 – 1.6 mm (n =19); (C) 1.7 – 1.9 mm (n =11); and (D) 2.0–3.0 mm (n = 16). The friction ratio (ms(post-impulse)/ms(pre-impulse)) is plotted on the vertical axis, and impulse magnitude (in N/s) is plotted along the horizontal axis. A best-fit straight-line regression demonstrates the strength of each relationship (R 2 values). The x-axis intercept is the critical threshold of impulse magnitude required to affect an increase in surface friction for the given range of disc thicknesses. Note that as disc thickness increased (A– D), so did the critical threshold.
160
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
Fig. 4. Average disc thickness versus critical threshold of impulse magnitude. The relation between the average disc thickness (in mm) for each disc group (A–D) and critical threshold of impulse magnitude (x-axis intercept of friction ratio versus impulse magnitude plot for each disc group, in N/s) is shown. Critical threshold of impulse magnitude increased with disc thickness (R 2 = 0.99). An impulse load to the disc above the thickness-dependent threshold would be expected to induce increased surface friction.
minimum threshold of surface damage to affect surface friction predicted by the plot (Fig. 5) was 2.6×103 mm2.
4. Discussion
4.1. Weeping lubrication theory and disc thickness Surface friction increased gradually with the duration of static loading of the TMJ disc by the friction pendulum. This finding is consistent with the theory of weep-
ing lubrication, in which the fluid reservoir within the cartilage matrix is depleted gradually and more areas of opposing cartilage surfaces come into direct contact. Under these circumstances there is a time-dependent increase in friction. Here, our thickest pristine discs were four times thicker than our thinnest. The thickest discs showed a rate of change in friction with static load that was one-half that of the thinnest (Fig. 5). The lateral flow of fluid in the cartilage matrix may have contributed to this effect. The faster this lateral flow, the less fluid there is available to move to the surface where it can be used to recharge the ‘hydrostatic bearing’. A previous report showed a strong correlation between disc thickness and lateral pressure gradient (#P/#r), where #P/#r was 3.5-fold higher for the disc that was 0.8 mm thick that for one that was 2.8 mm thick (Nickel and McLachlan, 1994b). As the present range of disc thicknesses was greater (0.68 – 2.85 mm) than in the previous report, lateral pressure gradients under the thinnest of our discs were expected to approach values four times those found under the thickest. If the difference between the thinnest and the thickest discs subject to static load was reflected in the pressure gradients (#P/#r) in this way, then according to equation (1), the lateral velocity of fluid flow (6t) for the thinnest discs should be double that for the thickest. That is, the thinnest discs should have higher lateral pressure gradients, which would cause faster lateral fluid movement through the cartilage matrix. This high 6t would deplete the fluid that would otherwise be available to recharge the surface and provide hydrostatic lubrication. As a result, and in keeping with our findings, surface friction would increase at a faster rate in thinner discs. The clinical significance of these phenomena lies in the potential for increased surface scuffing in the intact TMJ as a consequence of chronic high loads, such as those experienced in bruxism. In addition, thinner discs are likely to lose the weeping lubricant more quickly, leading to increased friction within the joint. Mechanical fatigue of the joint surfaces and debris in the synovial fluid might result. Debris causes an increase in the metabolic activity of macrophages in the synovial cavity, which in turn release various catabolic agents typically associated with the processes of osteoarthritis (Kubota et al., 1998).
4.2. Impulse loading and surface friction Fig. 5. Friction ratio versus area of damage. The friction ratio (ms(post-impulse)/ms(pre-impulse)) is plotted against the area of damage measured on the surface of the discs subjected to the highest (1.6 – 2.0 N/s) and lowest (0.8–1.2 N/s) impulse magnitudes. Five discs from groups A, B, and D, and six from group C were evaluated for surface damage and the plot of friction ratio versus area of damage (in mm2) is shown. The best-fit straight-line regression had an R 2 of 0.85.
The impulse magnitude, which was the amount of energy introduced on to the experimental disc surface, determined whether there were post-impulse changes in friction. Impulse-induced changes in lateral permeability could account for the increase in friction. The impulse magnitude determined the amount of disc surface damage, which in turn, increased permeability to
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
lateral fluid flow. However, thicker discs required more energy introduced to the surface before any changes were seen. The data suggest that disc thickness is positively correlated with the critical threshold of impulse magnitude which must be exceeded to cause an increase in friction (Fig. 4). Numerical modelling studies of peak force on cartilage have suggested that a thicker layer of hyaline cartilage, which is bound to bone, is more susceptible to damage than a thinner layer (Armstrong, 1986). Others report that the magnitude of surface shear stress was the factor most important to the occurrence of surface damage, and that peak force was strongly correlated with surface yielding (Atkinson et al., 1998). We show that larger impulse loads are required to affect surface friction in thicker discs and that peak impulse force does not correlate strongly with changes in surface friction (R 2 =0.45 overall). The apparent contradiction between previous reports and the current findings may be due to the physical constraints of bound hyaline cartilage, which are quite different from the boundary conditions associated with the TMJ disc, where friction is very low. In addition, the material properties of that disc are anisotropic (Bruno et al., 1999) and the previous studies applied isotropic methods of analysis.
4.3. The pathomechanics of TMJ osteoarthritis Osteoarthritis of the TMJ is common according to reports, but the mean age of onset differs widely depending on the methods of examination and the distinguishing criteria employed. Mean age of onset ranges between 25 and 35 years (Helo¨e and Helo¨e, 1975; Solberg et al., 1979; Nilner, 1981; Pullinger et al., 1988; Stegenga et al., 1989; Ong and Franklin, 1996; Israel et al., 1998; Murakami et al., 1998), approximately a decade earlier than the mean age of onset reported for osteoarthritis of the hip (Lawrence et al., 1989; Felson et al., 1997; Vingard et al., 1997). The earlier appearance of symptoms in the mandibular joint may relate in part to its disc being its principal stress-reducing mechanism. The surface congruency in the mandibular joint is poor over the crest of the eminence due to the eminence’s mechanism of growth (Nickel and McLachlan, 1994c). The disc distributes loads over the surfaces and maintains relatively frictionless sliding between them. In other joints such as the hip, the high degree of surface congruency between the hyaline cartilage on both rubbing surfaces serves to maintain minimal stresses. The knee, like the TMJ, has a fibrocartilaginous meniscus, but it is incomplete. The congruency and lubricating properties of the hyaline cartilage covering the femoral condyles and tibial plateau can minimize stresses to a certain extent, but like the TMJ, loss of the meniscus results in osteoarthritis. This probably accounts for the earlier expression of osteoarthritis in
161
knees than hips. In the mandibular joint, however, the fibrous linings of the condyle and temporal bone are not well-suited to distributing loads or lubrication (Nickel and McLachlan, 1994b,d), which is why the mechanical integrity of the disc is vital for the health of this joint over decades of function. We tested two variables, disc thickness and magnitude of trauma to the disc, as factors important to the mechanical integrity of the disc. The results confirm that disc thickness and trauma, specifically impulse load magnitude, affect disc static surface friction and, therefore, may be important factors in compromised lubrication and the development of osteoarthritis in the TMJ.
Acknowledgements We acknowledge financial support of this project from a University of Nebraska Medical Center, College of Dentistry Seed Grant. We are grateful to Farmland Foods Corporation of Crete, Nebraska, for providing access to fresh TMJ discs. We thank C. Anderson and M. Leiker, who undertook the task of microscopic evaluation of the disc surfaces, and K.R. Theesen, Illustrator, University of Nebraska Medical Center, College of Dentistry Learning Resources, who helped produce the figures.
References Armstrong, C.G., 1986. An analysis of the stresses in a thin layer of articular cartilage in a synovial joint. Eng. Med. 15, 55 – 61. Atkinson, T.S., Haut, R.C., Altiero, N.J., 1998. Impact-induced fissuring of articular cartilage: an investigation of failure criterion. J. Biomech. Eng. 120, 181 – 187. Bruno, M.J., Beatty, M.W., Nickel, J.C., Iwasaki, L.R., 1999. Strain rate dependent anisotropic properties of the porcine TMJ disc. J. Dent. Res. (SI) 78, 125. Castelli, W.A., Nasjleti, C.E., Diaz-Perez, R., Caffesse, R.G., 1985. Histopathologic findings in temporomandibular joints of aged individuals. J. Prosthet. Dent. 53, 415 – 419. Felson, D.T., Zhang, Y., Hannam, M.T., Naimark, A., Weissman, B., Aliabadi, P., Levy, D., 1997. Risk factors for the incident of radiographic knee osteoarthritis in the elderly: the Framingham study. Arthritis Rheum. 40, 728 – 733. Hatton, M.N., Swann, D.A., 1986. Studies on bovine temporomandibular joint synovial fluid. J. Prosthet. Dent. 56, 635 – 638. Helo¨e, B., Helo¨e, L.A., 1975. Characteristics of a group of patients with temporomandibular joint disorders. Community Dent. Oral Epidemiol. 3, 72 – 79. Herring, S.W., 1976. The dynamics of mastication in pigs. Arch. Oral Biol. 21, 473 – 480. Hills, B.A., 1989. Oligolamellar lubrication of joints by surface active phospholipid. J. Rheumatol. 16, 82 – 91.
162
J.C. Nickel et al. / Archi6es of Oral Biology 46 (2001) 155–162
Hills, B.A., Butler, B.D., 1984. Surfactants identified in synovial fluid and their ability to act as boundary lubricants. Ann. Rheum. Dis. 43, 641–648. Hopwood, D., 1985. Cell and tissue fixation, 1972–1982. Histochem. J. 17, 389–442. Hou, J.S., Mow, V.C., Lai, W.M., Holmes, M.H., 1992. An analysis of the squeeze film lubrication mechanism for articular cartilage. J. Biomech. 25, 247–259. Israel, H.A., Ramamurthy, N.S., Greenwald, R., Golub, L., 1998. The potential role of doxycycline in the treatment of osteoarthritis of the temporomandibular joint. Adv. Dent. Res. 12, 51 – 55. Iwasaki, L.R., Nickel, J.C., McLachlan, K.R., 1997. Relationship between growth, function, and stress in the temporomandibular joint. In: McNeill, C. (Ed.), Science and Practice of Occlusion. Quintessence Publishing, Chicago, pp. 125 – 136. Kubota, E., Kubota, T., Mastumoto, J., Shibata, T., Murakami, K.I., 1998. Synovial fluid cytokines and proteinases as markers of temporomandibular joint disease. J. Oral Maxillofac. Surg. 56, 192–198. Lawrence, R.C., Hochberg, M.C., Kelsey, J.L., McDuffie, F.C., Medsger, T.A., Felts, W.R., Shulman, L.E., 1989. Estimates of the prevalence of selected arthritic and musculoskeletal diseases in the United States. J. Rheumatol. 16, 427 – 441. McCutchen, C.W., 1962. The frictional properties of animal joints. Wear 5, 1 – 17. McCutchen, C.W., 1983. Lubrication of and by articular cartilage. In: Hall, B.K. (Ed.), Cartilage: Biomedical Aspects. Academic Press, New York, pp. 87–107. Murakami, K.I., Shibata, T., Kubota, E., Maeda, H., 1998. Intra-articular levels of prostaglandin E2, hyaluronic acid, and chondroitin-4 and -6 sulfates in the temporomandibular joint synovial fluids of patients with internal derangement. J. Oral Maxillofac. Surg. 56, 199–203. Nickel, J.C., McLachlan, K.R., 1994a. In vitro measurement of the frictional properties of the temporomandibular joint disc. Arch. Oral Biol. 39, 323–331. Nickel, J.C., McLachlan, K.R., 1994b. In vitro measurement of the stress-distribution properties of the pig temporomandibular joint disc. Arch. Oral Biol. 39, 439–448. Nickel, J.C., McLachlan, K.R., 1994c. An analysis of surface congruity in the growing human temporomandibular joint. Arch. Oral Biol. 39, 315–321.
.
Nickel, J.C., McLachlan, K.R., 1994d. A comparison of the lubrication properties of the mandibular condyle and TMJ disc. J. Dent. Res. (SI) 73, 304. Nilner, M., 1981. Prevalence of functional disturbances and diseases of the stomatognathic system in 15 – 18 year olds. Swed. Dent. J. 5, 189 – 197. Ong, T.K., Franklin, C.D., 1996. A clinical and histopathological study of osteoarthrosis of the temporomandibular joint. Br. J. Oral Maxillofac. Surg. 34, 186 – 192. Pullinger, A.G., Seligman, D.A., Solberg, W.K., 1988. Temporomandibular disorders. Part I: functional status, dentomorphologic features and sex differences in a non-patient population. J. Prosthet. Dent. 59, 228 – 235. Rugh, J.D., Harlan, J., 1988. Nocturnal bruxism and temporomandibular disorders. Adv. Neurol. 49, 329 – 341. Solberg, W.K., Woo, M.W., Houston, J.B., 1979. Prevalence of mandibular dysfunction in young adults. J. Am. Dent. Assoc. 98, 25 – 34. Solberg, W.K., Hansson, T.L., Nordstro¨m, B., 1985. The temporomandibular joint in young adults at autopsy: a morphologic classification and evaluation. J. Oral Rehabil. 12, 303 – 321. Stegenga, B., de Bont, L.G.M., Boering, G., 1989. Osteoarthritis as the cause of craniomandibular pain and dysfunction: a unifying concept. J. Oral Maxillofac. Surg. 47, 249 – 256. Stro¨m, D., Holm, S., Clemensson, E., Haraldson, T., Carlsson, G.E., 1986. Gross anatomy of the mandibular joint and masticatory muscles in the domestic pigs (Sus scrofa). Arch. Oral Biol. 31, 763 – 768. Swann, D.A., Hendren, R.B., Radin, E.L., Sotman, S.L., Duda, E.A., 1981. The lubricating activity of synovial fluid glycoproteins. Arthritis Rheum. 24, 22 – 30. Swann, D.A., Silver, F.H., Slayter, H.S., Stafford, W., Shore, E., 1985. The molecular structure and lubricating activity of lubricin isolated from bovine and human synovial fluids. Biochem. J. 225, 195 – 201. Tait, D.C., Nickel, J.C., McLachlan, K.R., 1997. Impulse loading and weeping lubrication on the surface of the TMJ disc. J. Dent. Res. (SI) 76, 447. Vingard, E., Alfredsson, L., Malchau, H., 1997. Osteoarthrosis of the hip in women and its relation to physical load at work and in the home. Ann. Rheum. Dis. 56, 293 – 298. Westessen, P.L., Rohlin, M., 1984. Internal derangement related to osteoarthrosis in temporomandibular joint autopsy specimens. Oral Surg. Oral Med. Oral Pathol. 57, 17 – 22.