SPINE Volume 19, Number 5, pp 495-501 (019%, J.B. Lippincott Company MH = Mechanoreceptor Endings in Human Cervical Facet Joints Robert F. McLain, MD ‘Twenty-one cervical facet capsules, taken from three normal human subjects, were examined to determine the type, density, and distribution of mechanoreceptive nerve endings in these tissues. Clearly identifiable mechanoreceptors were found in 17 of 21 specimens and ware classified according to the scheme for encap- sulated nerve endings established by Freeman and Wyke. Eleven Type |, 20 Type ll, and 5 Type Ill receptors were identified, as wall as a number of small, uneneap- sulated nerve endings. Type | receptors were small globular structures measuring 25-60 um in diameter. ‘Type Il receptors varied in size and contour, but were characterized by their oblong shape and broad, lamel- lated capsule. Type ll receptors were relatively large oblong structures with an amorphous capsule, within which a reticular meshwork of fine n bedded. Free (nociceptive) nerve endings were found in subsynovial loose areolar and dense capsular tis- sues, The presence of mechanoreceptive and nocicep- tive nerve endings in cervical facet capaules proves that these tissues are monitored by the central nervous sys tem and implies that neural input from the facets is im- ception and pain sensation in the cer- vical spine. Previous studies have suggested that protective muscular reflexes modulated by these types of mechanoreceptors are important in proventing joint instability and degeneration. It is suggested that the surgeon take steps to avoid inadvertently damaging ‘these tissues when exposing the cervical spine. [Key ‘words: cervical spine, mechanoreceptors, nerve end: ings, proprioception] Encapsulated mechanoreceptors and free nerve endings have been identified in the periarticular tissues of all major joints of the body and in every articular tissue, except for cartilage.142527.35 Polacek, in 1966, summa- rized the voluminous literature on articular innervation and his own studies of mechanoreceptor morphology, and produced a simple classification for receptor end- ings found in joints.° Freeman and Wyke, building on the work of Polacek and others, published their classi- fication scheme in 1967. This system was, at the same time, more comprehensive and simpler to use than pre- vious schemes and is currently the most commonly used in studies of receptor morphology.'' Studies of pe- ripheral joints suggest that these receptors monitor joint From the Department of Orthopaedie Surgery, University of Califor: » Davis, Sacramento, California cepted for publication January 13, 1993. excursion and capsular tension and may initiate protec- tive muscular reflexes important in preventing joint de- generation and instability.64617213: Although mecha- noreceptors have been identified in some spinal tissues, receptor type and distribution have not been specifically analyzed. Wyke and Molina reported finding Type | and I mechanoreceptor endings in cervical facets taken from cats, but found no Type Ill receptors and provided no information about the distribution and density of the endings they did find.263° Recent electrophysiologic studies have demonstrated the presence of propriocep- tive endings in facet and paraspinous tissues and have documented their response to pressure and facet mo- tion.1037 The anatomic location and type of the recep- tors were implied, but have not been confirmed histo- logically. ‘In the present study, our purpose was to analyze tissues from normal human cervical facet joints to char- acterize the types and distribution of capsular mecha- noreceptors. m Materials and Methods ‘Twenty-one cervical facet capsules were taken from three nor- mal human subjects. Thirteen capsular specimens were te- moved from one donor at necropsy (C1-2 to C7~T1). Four capsules each were removed from two other normal subjects at the time of cervical stabilization for traumatic injuries (C4—5 to C5~6 in one, and C34 and C56 in the other). Only intact capsules at levels adjacent to injury were included. The cervi- cal joints were examined at each sampled level to rule out pre-existing degenerative changes. All facets were judged to be free of joint degeneration by x-ray and direct visual examina- tion. None of these patients had any systemic disease that might be associated with neuropathy or cervical arthropathy. Facet capsules were excised en-bloc with a sharp scalpel and forceps. Care was taken to avoid crushing or distorting, the tissue. The electro-cautery knife was not used. Tissue specimens were frozen in saline before processing, The tissue ‘was processed and stained with a modified gold chloride tech- nnique.? After thawing, each specimen was placed, whole, in a solution of three parts commercial lemon juice and one part 88% formic acid. The specimen was held in solution for 15 ‘minutes at room temperature, then removed and dried and placed in a 1% solution of gold chloride for 15 minutes in darkness. The gold chloride solution was decanted, and the specimens were dried and placed in a 25% formic acid solu- tion for 4—6 hours. They then were rinsed twice in ethanol and placed in glycerol overnight. After they were removed 495496 Spine * Volume 19 + Number 5 + 1994 Table 1, Classification of Articular Receptor Systems Probable Functional ——_Other Eponymous or ‘Tye Morphological Features Average Dimensions Location Characteristics Dascriptive Designations ‘ype Globular receptors; round, 40-100 pm Jong Found in fibrous capsule of | Mechanoreceptor Rufin's ending, Gogi-Mazzoni ‘oval, or "bean sheped,” joint and in perarticular (slowly adapting, ending, Missner’s corpusce, thinly encapsulated, usully ligaments and tendons; usu- low threshold" basket or sprey-type ending found in clustors aly in suparicil layers of aflerent ending) dense connective tissue ‘ype ll Gyiindrical corpuscles with 250-200 um long. Found in deeper layers of - Mechanoreceptor Pecinian corpusce, Vater-Pacin- dencepsule- and 100 um wide —brous cepsulo, at junctions (rapidly adapting, ian corpuscle; modified Pacin- al axon cof fibrous tissue and fa, low threshold" ian corpusce; Paciniform cor- on may be and in fat pads; often ac- efferent ending) ‘Moissnor’s corpse! companiad by vascl fazzoni body; bulbous leash; oriented with con- nective tissue fibers in dense capsule or ligament ‘Type Il Fusiform corpuscles with @ Upto 600 ym long; Found in ligaments and ten- Mechanorecaptor —Golg's ending, Golg-Mazzoni capsule surrounding @ 100 wm long dons, es well as in dense (vary slowiy corpuscle orang nerve fibrous connective tissue of adapting, high fine neuritis Joint threshold afer rmashwork visible at high ent) ‘magnification Tyre Found inal perianioular and Nociceptor (non- for carta intraarticular tissues, except adapting) ‘adapted Tom roforenes 17 from the glycerol, the specimens were dehydrated and embed- ded in paraffin, Twenty-fve-micron serial sections were cut and mounted on glass slides. After deparaffnization, the spec- mens were cover-slipped in Eukitt Mounting Reagent (Cali- brated Instruments, Inc., Hawthorne, NY). No post-staining was necessary. No sampling techniques were applied in this study, Because of uncertainty about the expected density of receptors in these tissues, each specimen was entiely sectioned and every section was reviewed. Hence, the likelihood of a sampling error in this study is small All sections were analyzed under light microscopy using an Olympus BH-2 microscope with photographic head (Olym- pus, New Hyde Park, NY). No section was excluded because of inadequate staining or artifact. Each section was examined under X140 final magnification for the presence of large and small nerves, fine myelinated and unmyelinated axons, free nerve endings, and encapsulated nerve endings. Encapsulated endings were classified according to Freeman and Wyket! (Table 1). For an identified ending to be confirmed and clas- sified as a mechanoreceptive nerve ending, three criteria had to bbe met: 1) it had to have evidence of encapsulation and a ‘morphology consistent with previously described receptors; 2) it had to be identifiable on consecutive sections without evi- dence of artifacts; and 3) in most cases, a parent neuron also could be identified leading to the structure, again helping to verify that the structure was a nerve ending and not an artifact for vessel. Oil-emersion (X700) magnification was used to verify encapsulation and to distinguish between axon fila- ‘ments, the central axon core, and the darkly stained vessel lumen. The presence of free nerve endings also was noted. The nature of the tissue in which the receptive ending was located (dense connective, loose arcolar, synovial) was recorded. im Results ‘Mechanoreceptor endings were found in 17 out of 21 capsule specimens. Six capsules contained only one re- ceptor, five capsules contained two receptors, and four capsules contained three; one each contained four and five receptors, respectively. When capsules contained more than one receptor, the receptor types usually were different. The majority of endings resembled the Type Il mechanoreceptors identified in cat joint studies.!12627 Fewer Type I receptors also were identified, and only a handful of Type Ill receptors were found in any of the tissues. In all, there were 20 Type Il receptors, 11 Type T receptors, and five Type Ill receptors identified in this study. Parent axons were identified for most of these receptors and usually measured 3-5 yum in diameter. ‘The Type Il mechanoreceptive endings varied consid- ‘erably in size and shape (Figure 1). The largest of the endings measured 500 im in its greatest length and was characterized by a broad capsule with numerous lamel- Jae, Cross-sections through the capsule of Type Il recep- tors revealed concentric layers of faintly pink material alternating with light gray lamellar lines stained with gold chloride (Figure 2). These receptors were found in dense, fibrous capsular tissue, usually along a thin cleft of fatty tissue through which the parent axon and mi- crovasculacure pass. Type I mechanoreceptive endings, which are small globular endings measuring 25-50 um in diameter, usu- ally were found in loose areolar tissue and always were solitary receptors. No clusters of nerve endings were found in these specimens (Figure 3). ‘Type Ill mechanoreceptors were found at the junc- tion between the dense fibrous capsule and the loose areolar subsynovial tissue. Inspection at higher magni- fication revealed a reticular meshwork of fine neutites contained within an oblong capsule (Figure 4).Maechanoreceptor Endings in Human Cervical Facet Joints * McLain 497 Y Fine, unmyelinated nerves (less than 5,0 um in diam- eter) were identified in the dense capsular tissues and in the synovial and areolar tissues; these fine filaments may represent nociceptive nerve endings (Figure 5). Such f- bers also accompanied many of the vessels within the dense fibrous tissue of the capsule, ‘There was no apparent difference in the distribution of receptors within the facet capsules of the upper cet- vical versus lower cervical spine. No specific level ap- peared to be more richly innervated than any others. Likewise, there was no apparent difference from indi- vidual to individual regarding the density of receptors found in their facet capsules. Figure 1. Type I! mechanorecep- tors in human cervical facets. Type Il receptors vary consi ably in size and shape. These re- ceptors are the most common type found in cervical facets, and ‘are characterized by a thick cep- sule (10-12 lamina thick) and 2 fusiform or cylindrical shape. (A) Large, encapsulated mechan ceptor from dense connecti tissue of C2-3 cervical facet cap- ‘ule. A small strip of fat accom- panies the parent axon and cap- illary supply through the capsule, nt 100 jm). {B) This tong, fusiform Type Il receptor from a C3 ticulation is folded back on at its proximal end; seri tions demonstrated an overall length of 350 pa. A parent axon can be seen following a thin cleft af ftthrough the dense capsular tissue (arrows; bar = 100 um) {C) Smaller, oblong receptor from 1-2 level, with parent axon (ar- rou), This section shows multiple lamellae typical of Type Il cap- sule (bar = 100 um). Synovial joints are composite tissues, specialized to bear loads and allow motion through specific, well-defined ares, Their separate components have evolved to pro- vide painless range of motion while maintaining joint stability. Numerous studies document the presence of mechanoreceptive endings in these articular tissues and periarticular ligaments. Except for the articular carti- lage, all of the tissues of the synovial joint receive some form of innervation. Disturbance of this innervation ‘may play a role in the development of degenerative and inflammatory diseases and joint dysfunction. Synovial498. Spine * Volume 19 + Number 5 + 1994 Figure 2. High-power view of the capsule of Type Il receptor. arrowhead) alternate between layers of lighter staining material. These lamellae are arranged concentri- cally around the central axon core, An outermost lamina (arrow) separates the structure from the surrounding connective tissue and a small aggregation of fat (bar = 25 ym), joints enjoy a dual pattern of innervation—primary ar- ticular nerves are independent branches from larger pe- ripheral nerves that specifically supply the joint capsule and ligaments; accessory articular nerves reach the joint after passing through muscular or cutaneous tissues to receptor. Small, lobular receptor found issue of the capsule of a C2-3 facet joint. re just visible at this magnifi- ; bar = 50 jam Fine fi cation, and a parent axon is present (aro Figure 4, Type Ill mechanoreceptor. The thin, amorphous capsule (smell arrow) that separates the complox web of fine neural elements from the surrounding connective tissue does not have separate lamellae, This receptor, taken from C3-4 facet joint, is located at the junction of the capsular connective tissue and the attachment of the paraspinous musculature. Skeletal muscle fi- bers (large arrow) are seen investing the fibro-fatty connective tissue, The parent axon is seen at lower left (bar = 50 yum). which they provide primary innervation.1235 These pri- mary and accessory articular nerves terminate in a va- riety of encapsulated and uncapsulated nerve endings. Freeman and Wyke described four basic types of afferent nerve endings in articular tissues, and docu- mented the presence of those endings in a wide variety of joints." Polacek found numerous variations in the morphology of encapsulated receptors, even within the definitions of specific receptor types, and demonstrated ‘marked variations in morphology and distribution be- ‘owen tissue types and between subject species.%° Histo- logic studies have demonstrated these receptors in a variety of tissues, including ligaments,”#92¢ ligamentous Figure 5. Free nerve ending in dense connective tissue. A spray of fine neural elements is seen coursing through the capsule of a 5-6 facet. Nociceptive free nerve endings (arrow) are unencap- sulated and are found throughout the capsule, synovium, and loose areolar tissue (bar = 20 wm).‘Mechanoreceptor Endings in Human Cervical Facet Joints * McLain 499 joint capsule,"%7 and meniscal tissues,27 as well as in periarticular fat and muscle.*1! Types I through Ill are encapsulated receptors thought to be primarily involved in proprioception. These endings respond primarily to the extremes rather than to the mid-range of joint mo- tion.*H6 Ie also is thought that these mechanoreceptors may influence the activity of muscles crossing the joint, modulating protective muscular reflexes important to joint function and stability.42392923 ‘Wyke and others have presented a series of studies demonstrating that stimulation of articular nerves can generate reflex contraction of periarticular muscles, but have not specifically identified the receptor endings in- volved.2#t Type IV receptors are unencapsulated free nerve endings found in a variety of soft tissues. With recent advances in neurohistochemical techniques, in- vestigators have demonstrated greater numbers of these small diameter nerve fibers in the facet capsule and synovium than were previously reported using standard histologic methods.1#22 Although the Type IV receptors are the only ones thought to be exclusively nociceptive, itis known that the encapsulated mechanoreceptors are capable of responding to excessive joint excursion as a noxious stimulus. Focusing primarily on the knee joint, electro-physio- logic studies have demonstrated that altered mechano- reception has a direct effect on reflex activity of muscle groups crossing the joint7#° DeAndrade and Kennedy both found that knee joint effusion produces significant reflex inhibition of the quadriceps muscles, probably because of distension of the posterior capsule.62! Using a dog knee model, O’Connor and others demonstrated that peripheral joint deafferentation may accelerate de- generative articular changes in joints that have been surgically destabilized.2® Although the articular innervation of the knee and the large joints of the extremities have been studied extensively, the facet joints of the spinal column have only recently begun to attract the attention of investi- gators. Freeman and Wyke previously documented the presence of receptors in the posterior elements of the spinal column, but did not comment on the density of the receptor population or on the distribution of the different receptor types.tt# Giles and Harvey demon- strated nociceptive free nerve endings in capsular tissue of human facets, and reported similar endings in the facet synovium.'415 Gronblad et al identified numerous fine nerves traveling with the vessels of the synovial plica, and occasional free nerve endings within the s novium.20 These investigators have not reported the presence of encapsulated nerve endings in facet capsular tissue. Also, these studies have focused on the lumbar spine, and have not addressed the innervation of other segments of the spine. ‘The present study has documented the presence of nerve endings in the capsule of normal human cervical facets. These endings were consistent with the descrip- tions advanced by Freeman and Wyke and others for articular tissues in the cat, dog, and human,t1252632 ‘Axons and small vessels usually were easily distin- guished in well-stained tissues. The authenticity of “pu- tative receptors” reported in past studies has been ques- tioned, because the structures illustrated resembled vessels and fixation artifacts rather than genuine mech- anoreceptors.”3? In this study, ifthe distinction between myelinated axon and a vessel was found to be difficult, review of serial sections usually brought the observer to a branch point in the vessel or a lumen that would make identification possible. Following the axon through se- rial sections frequently brought the observer to a node of Ranvier, which again confirmed the identification, Although large nerves containing multiple axons could be identifed at the musculo-tendinous junction, most of the neural structures within the dense fibrous capsule were single axons, The small number of mechanoreceptor endings en- ‘countered in the facet capsules suggests that these recep- tors have a relatively large receptive field or that their individual functional contribution is of little impor- tance. Because the receptors are relatively large, it is likely that receptive fields are large and that one of two nerve endings may be sufficient to monitor the atea of each individual facet capsule. If ths is the case, damage to even a small part of a capsule might denervate that articular structure. This could have important implica- tions for long-term joint function. O'Connor et al have shown that when joint instability (accomplished by transection of the anterior cruciate ligament) was super- imposed upon a deafferented hind-limb, the develop- ment of arthropathy was greatly accelerated compared to dogs with unstable joints, but intact nerves.2# If the same relationship between joint afferent innervation and protective muscular reflexes pertains in the cervical spine, it is most important that surgeons protect the cervical facets during operative procedures, such as lam- inectomy or laminoplasty. In the present study, free nerve endings also were identified in the loose areolar tissue, dense connective tissue, and synovial lining of the cervical facets. These fine endings probably have nociceptive function and the facet capsule and its lining probably play a role in neck pain in otherwise normal patients, Other authors have identified Substance P immunoreactivity in encapsulated receptors of the posterior longitudinal ligament and the periosteum of long bones, suggesting that encapsulated mechanoreceptors also play a role in nociception.%®23 ‘Whether these receptors are active in pain production in the cervical spine at the extremes of joint motion or whether they respond to other stimuli is not known, m Conclusions Encapsulated mechanoreceptors are a consistent finding in normal human cervical facets. The presence of these receptors in the facet capsule indicate that the mechan-500 Spine * Volume 19 + Number 5 + 1994 ical state of the capsule (position, tension, pressure, etc.) is under the surveillance of the central nervous system. ‘Whether the information collected concerns joint mo- tion during routine daily activities or relates to poten- tially damaging stresses at the extremes of motion is unknown. It is likely that different receptors respond to different states of excursion and that these receptors play a role in protective muscular reflexes and in pain caused by excessive motion, Although muscular reflexes initiated by these receptors may not be important in routine activities, they may be crucial for protecting the injured or unstable joint from early degenerative dis- ease. Therefore, itis important the surgeon respects the margins of the facet capsule and takes care to see that these articular nerves and nerve endings are carefully protected during surgical procedures. m Acknowledgments This research was funded in part by a grant from the Orthopaedic Research and Education Foundation, Rosemont, Illinois. References 1. Baxendale RH, Ferrell WR. Modulation of transmission in flexion reflex pathways by knee joint afferent discharge in decerebrate cat. Brain Res 1980;202:497-500. 2, Baxendale RH, Ferrell WR. Modulation of transmission in forelimb flexion reflex pathways by elbow joint afferent discharge in decerebrate cats. Brain Res 1981;221:393-6. 3. Baxendale RH, Ferrell WR. Facilitation of joint afferent mediated reflex effects by stretch-related muscle afferent charge in decerebrate cats. J Physiol (Lond) 1982;329:60-1. 4, Burgess PR, Clark FJ. Characteristics of knee joint recep- tors in cats. J Physiol (Lond) 1969;203:317-35. 'S. Clark FJ, Burgess PR. Slowly adapting receptors in cat kknee joint, Can they signal joint angle? J of Neurophysiol 1975;38:1448-1463. 6. DeAndrade JR, Grant C, Dixon A. Joint distension and reflex muscle inhibition in the knee. J Bone Joint Sucg 19653 47A3313-32, 7. DeAvila GA, O'Connor BL, Visco DM, Sisk TD. The mechanoreceptor innervation of the human fibular collateral ligament, J Anat 1989;162:1~7, 8, Dee RM. The innervation of joints. In: Sokoloff L, ed. Joints and Synovial Fluid. New York, Academic Press, 1978. 9. Eckholm J, Eklund G, Skoglund S. On the reflex effects from the knee joint of the cat. Acta Physiol Scand 1960350: 167-74, 10, ElBohy A, Cavanaugh JM, Getchell ML, Bulas T, Getch- cell TV, King AI. Localization of substance P and neurofilament immunoreactive fibers in the lumbar facet joint capsule and ‘supraspinous ligament of the rabbit, Brain Res 1988;460: 370-82. 11, Freeman MAR, Wyke BD. The innervation of the knee joint: An anatomical and histological study in the cat. J Anat 1967;101:505-32. 12, Gardner E, The distribution and termination of nerves in the knee joint of the cat. J Comp Neurol 1944;80:11-32. 13, Gardner E. The innervation of the knee joint. Anat Rec 1948;101:109-30, 14, Giles LGF, Harvey AR. Immunobistochemical demon- stration of nociceptors in the capsule and synovial folds of human 2ygapophyseal joints. Br J Rheumatol 1987;26:362—4, 15. Giles LGF, Taylor JR. Innervation of lumbar zygapophy- seal joint synovial folds. Acta Orthop Scand 1987558:43-6. 16. Grigg P. Mechanical factors influencing response of joint afferent neurons from cat knee. J Neurophysiol 1975;38: 1473-83. 17. Grigg P, Hoffman AH, Fogarty KE. Properties of Golgi- ‘Mazzoni afferents in cat knee joint capsule as revealed by mechanical studies in isolated joint capsule, J Neurophysiol 1982547:31-40. 18. Gronblad M, Liesi P, Korkala O, Karaharju E, Polak J. Innervation of human bone periosteum by peptidergic nerves. Anat Rec 1984;209:297-9, 19. Gronblad M, Konttinen Y, Korkala O, Liesi P, Hukkanen M, Polak J. Neuropeptides in synovium of patients with sheu- matoid arthritis and osteoarthritis. J Rheumatol 1988;15: 1807-10. 20. Gronblad M, Korkala O, Konttinen Y, et al. Silver im- ppregnation and immunohistochemical study of nerves in lum- bar facet joint plical tissue. Spine 1991516:34-8. 21. Kennedy JC, Alexander IJ, Hayes KC, Nerve supply ofthe hhuman knee and its functional importance. Am J Sports Med 1982510:329-35. 22. Kidd BL, Mapp Pl, Blake DR, Gibson SJ, Polak JM. Neurogenic influences in arthritis. Ann Rheum Dis 1990} 649-52. 23. Liesi P, Gronblad M, Korkala O, Karaharju E, Rusanen 1M. Substance P: A neuropeptide involved in low back pain? Lancet 1983;1:1328-9. 24. Molina F, Ramcharan JE, Wyke BD. Structure and fune- tion of articular receptor systems in the cervical spine. J Bone Joint Surg 1976;58B:255-256, 25. O'Connor BL. The mechanoreceptor innervation of the posterior attachments of the lateral meniscus of the dog knee capsule. J Neurophysiol 1982;47:41-54. 26. O'Connor BL, Gonzales J. Mechanoreceptors of the me- dial collateral ligament of the cat knee joint. J Anat 1979;129: 719-29. 27. O'Connor BL, McConaughey JS. The structure and in- netvation of cat knee menisci, and ther relation to a ‘sensory hypothesis’ of meniscal function, Am J Anat 19785153:431— 42, 28. O'Connor BL, Palmoski MJD, Brandt KD. Neurogenic ‘acceleration of degenerative joint lesions. J Bone Joint Surg 1985;67A:562-72. 29, Palmer I, Pathophysiology of the medial ligament of the knee joint. Acta Chirurgica Scandinavica 1958;115:312-8. 30, Polacek P. Receptors of the joints: Their structure, vai ability, and classification. Acta Facultatis Medicae Universita- tis Brunensis 1966;23:1-107. 31. Ramcharan JE, Wyke BD. Articular reflexes in the knee joint; an electromyographic study. Am J Physiol 1972;223: 1276-80. 432, Schutte MJ, Dabezies EJ, Zimny ML, Happel LT. Neural anatomy of the human anterior cruciate ligament. J Bone Joint Surg 1987;69A:243-7. 33. Solomonow WM, Baratta R, Zhou BH, et al. The syner- gistic action of the anterior cruciate ligament and thigh mus- cles in maintaining joint stability. Am J Sports Med 1987;15: 207-13.‘Mechanoreceptor Endings in Human Cervical Facet Joints * McLain $01 34, Wyke B, The neurology of joints. Ann R Coll Surg Engl 1967541:25-50. 35. Wyke B. Articular neurology—a review. Physiotherapy 1972;58:94-9, 36. Wyke B, Molina F, Articular reflexology of the cervical spine. Proc 6th Int Cong Phys Med, Barcelona, 1972;61-66. 37. Yamashita T, Cavanaugh JM, El-Bohy AA, Getchell TV, King Al. Mechanosensitive afferent units in the lumbar facet joint. J Bone Joint Surg 1990;72A:865~70, Address reprint requests to Robert F. McLain, MD Department of Orthopaedic Surgery University of California, Davis 2230 Stockton Boulevard Sacramento, CA 95817